the american journal of medicine - lib.ajaums.ac.irlib.ajaums.ac.ir/booklist/254133.pdf · barbara...
TRANSCRIPT
COMMENTARIES469 Who Will Care for the Frail Elderly?
Joseph S. Alpert and Pamela J. Powers
472 Are Patients with a Patent Foramen Ovale atIncreased Risk of Stroke? A Billion DollarQuestionJames E. Dalen
REVIEWS475 Direct-to-Consumer Advertising of
PharmaceuticalsZiad F. Gellad and Kenneth W. Lyles
The United States and New Zealand are the onlyindustrialized countries to allow direct-to-con-sumer advertising of prescription drugs. Direct-to-consumer advertising in the US has increasedexponentially since 1997 when the FDA releasednew guidelines regulating broadcast advertise-ments. This review examines the effect of direct-to-consumer advertising on patients, physicians,and their interactions.
481 Atrial Fibrillation in the ElderlyMargaret C. Fang, Jane Chen, andMichael W. Rich
Atrial fibrillation is increasingly common in elderlyAmericans. Rate control and anticoagulation withwarfarin has been shown to be as good at preventingembolic strokes as rhythm control with anti-arrhythmic medication.
488 HIV-Associated Nephropathy in the Era ofAntiretroviral TherapyChristina M. Wyatt and Paul E. Klotman
Patients with HIV infection can develop retroviralnephropathy leading to chronic renal failure.Anti-retroviral therapy also can be nephrotoxic,resulting in further injury to the kidney.
UPDATES IN OFFICE MANAGEMENT493 Falls in Older Adults: Risk Assessment,
Management and PreventionKyle C. Moylan and Ellen F. Binder
498 Allergic Respiratory Disease in the ElderlyPaula J. Busse
DIAGNOSTIC DILEMMAS503 High-Speed Heart
Amisha J. Patel, J. Stewart Collins, Jashu R. Patel,and Rajendra H. Mehta
506 Growth by AssociationRita R. Kalyani and Anna R. Hemnes
IMAGES IN DERMATOLOGY509 A Seriously Swollen Lip
Shomeet V. Patel and Joshua E. Lane
ECG IMAGE OF THE MONTH512 Now You See It; Now You Don’t
Norman C. Wang
THE AMERICAN
JOURNAL ofMEDICINE ® CONTENTS
The Green Journal June 2007 Volume 120/Number 6
A5
IMAGES IN RADIOLOGY515 A Wilde Diagnosis
Naseema Gangat and Paul W. Johnson
CLINICAL RESEARCH STUDIES518 Body Mass, Fitness and Survival in Veteran
Patients: Another Obesity Paradox?Paul McAuley, Jonathan Myers, Joshua Abella,and Victor Froelicher
This study noted that obese patients in a VeteransAdministration clinical database had a substan-tially lower mortality risk compared with non-obese individuals. Higher cardiac and respiratoryfitness as well as obesity in later life may accountfor the obesity paradox in this population.
525 Selective Serotonin Reuptake Inhibitor Use byPatients with Acute Coronary SyndromesRoy C. Ziegelstein, Jennifer Meuchel,Thomas J. Kim, Madiha Latif, William Alvarez,Neela Dasgupta, and Brett D. Thombs
Selective serotonin reuptake inhibitor use in patientshospitalized with an acute coronary syndrome wasassociated with reduced rates of recurrent ischemia,heart failure, or cardiac enzyme elevation. However,these patients also experienced increased bleedingcomplications when simultaneous maximal anti-platelet medication and heparin were administered.
531 Improving Risk Assessment with CardiacTesting in Peripheral Arterial DiseaseHarm H. H. Feringa, Abdou Elhendy,Stefanos E. Karagiannis, Peter G. Noordzij,Martin Dunkelgrun, Olaf Schouten,Radosav Vidakovic, Ron T. van Domburg,Jeroen J. Bax, and Don Poldermans
Left ventricular ejection fraction, stress-inducedischemia, ankle-brachial index, and clinical riskfactors all contribute to the assessment of long-term outcome and prognostic risk assessment inpatients with suspected or known peripheral arte-rial disease.
539 The Safety of Warfarin Therapy in the NursingHome SettingJerry H. Gurwitz, Terry S. Field,Martha J. Radford, Leslie R. Harrold,Richard Becker, George Reed, Kristin DeBellis,Jason Moldoff, and Nancy Verzier
Anticoagulation therapy in elderly nursing homeresidents is frequently complicated by excessiveprolongation of the INR with subsequent bleeding
complications. Preventive measures are suggestedto ameliorate this common problem.
545 Maltreatment of Strongyloides Infection:Case Series and Worldwide Physicians-in-Training SurveyDavid R. Boulware, William M. Stauffer,Brett R. Hendel-Paterson,Jaime Luıs Lopes Rocha,Raymond Chee-Seong Seet, Andrea P. Summer,Linda S. Nield, Khuanchai Supparatpinyo,Romanee Chaiwarith, and Patricia F. Walker
Infection with the intestinal parasite Strongyloidesdoes occur in US patients. Strongyloidiasis shouldbe considered when patients present with a historyof appropriate exposure combined with nonspeicficsymptoms including wheezing, and/or eosinophilia(�5% relative or �400 eosinophils/�L).
AJM ONLINECLINICAL RESEARCH STUDY
552 Kidney Disease, Framingham Risk Scores, andCardiac and Mortality OutcomesDaniel E. Weiner, Hocine Tighiouart,John L. Griffith, Essam Elsayed, Andrew S. Levey,Deeb N. Salem, and Mark J. Sarnak
Patients with chronic kidney disease were at highrisk for developing cardiovascular complications.However, the Framingham risk score was able topredict much of this increased risk.
BRIEF OBSERVATION552 High FVIII Level Is Associated with Idiopathic
Portal Vein Thrombosis in South IndiaAbraham Koshy and Mary Jeyakumari
In Indian patients, high blood factor VIII levels areassociated with a markedly increased risk of portalvein thrombosis. Whether this association exists inNorth American patients, remains to be explored.
The following articles are available only inthe online version of the Journal.
CLINICAL COMMUNICATIONS TO THEEDITOR
e1 A Young Woman with Neurofibromatosis andSevere HypertensionRobert Gold, Fredric Adler, Tamara Modilevsky,and Katherine Yu
A7
e3 Thyrotoxic Hypokalemic Periodic Paralysis in aWhite ManKjersti Meyer Kirkeby, David M. Naeger,Allan Pont, and Thomas E. Baudendistel
LETTERSe5 Baclofen Not Comparable to Diazepam for
Alcohol WithdrawalAlec B. O’Connor and Valerie J. Lang
e7 The ReplyGiovanni Addolorato, Lorenzo Leggio,Giovanni Gasbarrini, Roberta Agabio, andGiancarlo Colombo
e9 Baclofen for Alcohol Withdrawal: NotComparable to the Gold Standard(Benzodiazepines)Richard Saitz
e11 The ReplyGiovanni Addolorato, Lorenzo Leggio, andGiovanni Gasbarrini
e13 Indications of Capsule Endoscopy in Henoch-Schonlein Syndrome with GastrointestinalSymptomsJae Il Shin and Jae Seung Lee
e15 Abnormal Valsalva Maneuver Is Not Always aSign of Congestive Heart FailureSatish R. Raj, David Robertson, Italo Biaggioni,and Andre Diedrich
e17 Balancing the Risks and Benefits of CelecoxibBrent Caldwell, Richard Beasley,Mark Weatherall, and Scott Metcalfe
e19 Airway Obstruction in Never SmokersAlbert Miller
e21 Using the Pneumonia Severity Index to GuideHospitalization DecisionsDonald M. Yealy, Thomas E. Auble, andMichael J. Fine
e23 Pneumonia Severity Index (PSI) Lacking inBreadth of ApplicabilityRobert E. Siegel
APM PERSPECTIVES553 Late to the Feast: Primary Care and US
Health PolicyEugene C. Rich and Anna Maio
MEDICAL HUMANITIES PERSPECTIVES560 Ancient Japanese Medicine in The Tale of Genji
Norman A. Desbiens
CLASSIFIED ADSB1 Positions available
A8
Associate EditorsTHOMAS D. BOYER, MDHepatology, Tucson, Arizona
JAMES E. DALEN, MD, MPHCardiovascular Disease & Health Policy,Tucson, Arizona
DAVID B. HELLMANN, MDRheumatology, Baltimore, Maryland
Y. HOWARD LIEN, MD, PHDNephrology, Tucson, Arizona
RICHARD M. MANDEL, MDInfectious Disease, Tucson, Arizona
ESKILD A. PETERSEN, MD, DTM&HInfectious & Tropical Diseases, Tucson,Arizona
BARBARA L. SCHUSTER, MDGeneral Medicine, Dayton, Ohio
ALISON T. STOPECK, MDHematology/Oncology, Tucson, Arizona
Editor, SupplementsWILLIAM H. FRISHMAN, MDValhalla, New York
Specialty EditorsMICHAEL A. BETTMANN, MDRadiology Images Editor, Winston-Salem,North Carolina
STUART R. CHIPKIN, MDDiabetes Editor, Amherst, Massachusetts
RICHARD D. DE SHAZO, MDAllergy/Immunology Editor, Jackson,Mississippi
VIKRAM D. DURAIRAJ, MDOphthalmology Editor, Denver, Colorado
ROBERT J. GOLDBERG, PHDStatistics Editor, Providence, Rhode Island
JULIA H. INDIK, MD, PHDECG Image of the Month Editor, Tucson,Arizona
HELLE MATHIASEN, CAND MAG, PHDMedical Humanities Editor, Tucson, Arizona
PARWATHI ‘UMA’ PANIKER, MDImages in Dermatology Editor, Portland,Oregon
MICHAEL W. RICH, MDGeriatrics Editor, St. Louis, Missouri
CHARLES M. WIENER, MDDiagnostic Dilemma Editor, Baltimore,Maryland
Managing EditorPAMELA J. POWERS, MPH, Tucson, Arizona
Editorial BoardGraciela S. Alarcon, MD, MPHRobert J. Anderson, MDFrank C. Arnett, MDJohn P. Atkinson, MDDavid W. Bates, MD, MScPeter W. Bates, MDCharles L. Bennett, MD, PhD, MPPAl B. Benson, III, MDAndrew B. Bindman, MDWilliam E. Boden, MD, FACCD. Craig Brater, MDGlenn D. Braunstein, MDTroyen A. Brennan, MD, JD, MPHMichael S. Bronze, MDNaga Chalasani, MDGeorge P. Chrousos, MD, FAAP, MACP,
MACEPhilip E. Cryer, MDBurke A. Cunha, MD, MACPPrakash C. Deedwania, MDThomas L. Delbanco, MDGary V. Desir, MDRichard A. Deyo, MD, MPHKim A. Eagle, MDRobert H. Eckel, MDArthur M. Feldman, MD, PhDDavid T. Felson, MD, MPHJohn R. Feussner, MD, MPHStephan D. Fihn, MD, MPHRobert Finberg, MDJoel S. Finkelstein, MDDavid A. Fox, MDVictor F. Froelicher, MDEric P. Gall, MD, MACP, MACRStephen A. Geraci, MD
Bernard J. Gersh, MB, ChB, DPhil, FRCPHenry N. Ginsberg, MDAllan L. Goldman, MDSteve Goldschmid, MDScott M. Grundy, MD, PhDJerry H. Gurwitz, MDMitchell L. Halperin, MDStephen B. Hanauer, MDRobert I. Handin, MDWilliam R. Hazzard, MDJohn E. Heffner, MDMark A. Hlatky, MDF. Allan Hubbell, MD, MSPHRussell D. Hull, MBBS, MScSharon Inouye, MD, MPHDavid Jacoby, MDJohn P. Kane, MD, PhDJohn A. Kastor, MDSanjiv Kaul, MDWilliam A. Knaus, MDBruce C. Kone, MDJoseph A. Kovacs, MDHarlan M. Krumholz, MD, SMMichael A. LaCombe, MD, FACC, MACPLoren Laine, MDC. Seth Landefeld, MDH. Clifford Lane, MDEric B. Larson, MD, PhDMatthew H. Liang, MD, MPHRodger A. Liddle, MDSteven A. Lieberman, MDRichard S. Liebowitz, MDJeffrey R. Lisse, MDJoy Logan, MDDan L. Longo, MDJames D. Marsh, MD
Thomas H. Marwick, MBBS, PhD, FACC, FRACPJoseph R. Masci, MDDavid B. Matchar, MD, FACPJohn R. Michael, MDLarry Moreland, MDPope Moseley, MDRoland W. Moskowitz, MDMaurice A. Mufson, MD, MACPBarbara E. Murray, MDIra S. Nash, MDEric G. Neilson, MDPaul M. O’Byrne, MDG. Richard Olds, MDMark A. Perazella, MD, FACPMichelle A. Petri, MD, MPHRussell S. Phillips, MDHuibert A.P. Pols, MD, PhDEugene Rich, MDDan M. Roden, MDHaya R. Rubin, MD, PhDAnil K. Rustgi, MDSanjay Saint, MD, MPHJonathan M. Samet, MD, MScMerle Sande, MDAndrew I. Schafer, MDW. Michael Scheld, MDRobert W. Schrier, MDBarbara Schuster, MDDavid A. Schwartz, MD, MPHJoseph I. Shapiro, MDMartin Shapiro, MD, PhDSteven Shea, MDEdward H. Shortliffe, MD, PhDMark Siegel, MDAndrew H. Soll, MDFredrick S. Southwick, MD
THE AMERICAN
JOURNAL ofMEDICINE ®
Editor-in-Chief
JOSEPH S. ALPERT, MDRobert S. and Irene P. FlinnProfessor of MedicineSpecial Assistant to the DeanUniversity of Arizona College of MedicineTelephone: 520-207-9415Facsimile: [email protected]
The Green Journal
Meir J. Stampfer, MD, DrPHPaul D. Stein, MDBrian L. Strom, MD, MPHEric J. Topol, MD
Anna N. A. Tosteson, ScDPeter Tugwell, MDStephen I. Wasserman, MDJane C. Weeks, MD, MScWilliam S. Weintraub, MD
William B. White, MDC. Mel Wilcox, MDTimothy J. Wilt, MD, MPHSteven N. Wolff, MDVincent W. Yang, MD, PhD
Manuscript submissions: Submit all manu-scripts to Joseph S. Alpert, MD, Editor-in-Chief, at www.ees.elsevier.com/ajm. New userswill be asked to register and create a user ac-count in the online system. Authors who havesubmitted to the Journal in the past may haveexisting accounts in the system. The online sys-tem includes an author tracking feature thatmonitors the publication process from initialsubmission, through screening, review, revi-sion, and publication. Paper submissions can-not be accepted. Please send editorial questionsto [email protected].
Publisher: The American Journal of Medicine(ISSN: 0002-9343) is published monthly byElsevier Inc., 360 Park Avenue South, NewYork, NY 10010-1710. Business Office: 1600John F. Kennedy Blvd., Suite 1800, Philadel-phia, PA 19103-2899. Editorial Office: 360 ParkAvenue South, New York, NY 10010-1710.Accounting and Circulation Offices: 6277 SeaHarbor Drive, Orlando, FL 32887-4800. Peri-odicals postage paid at New York, NY and ad-ditional mailing offices.
Postmaster: Send address changes to TheAmerican Journal of Medicine, Elsevier Periodi-cals Customer Service, 6277 Sea Harbor Drive,Orlando, FL 32887-4800.
Copyright: © Copyright 2007 by Elsevier Inc.All rights reserved. This journal and the individ-ual contributions contained in it are protected bythe copyright of Elsevier Inc. and the followingterms and conditions apply to their use:
Photocopying: Single photocopies of singlearticles may be made for personal use as al-lowed by national copyright laws. Permissionof the publisher and payment of a fee is re-quired for all other photocopying, includingmultiple or systematic copying, copying for ad-vertising or promotional purposes, resale, andall forms of document delivery. Special ratesare available for educational institutions thatwish to make photocopies for non-profit edu-cational classroom use. In the USA, users mayclear permissions and make payment throughthe Copyright Clearance Center, Inc., 222Rosewood Drive, Danvers, MA 01923; tele-phone (978) 750-8400. In the UK, users mayclear permissions and make payment throughthe Copyright Licensing Agency Rapid Clear-ance Service (CLARCS). In other countrieswhere a local copyright clearance-center exists,please contact it for information on requiredpermissions and payments.
Permission Requests: Please send all inquiriesto Elsevier, Health Sciences Rights Dept., 1600JFK Boulevard, Suite 1800, Philadelphia, PA19103-2899. Phone: 1-215-239-3804, Fax:1-215-239-3805. E-mail: [email protected]. Requests may also be completedonline via the Elsevier homepage (http://www.elsevier.com/locate/permissions).
Derivative Works: Subscribers may repro-duce tables of contents or prepare lists of arti-cles including abstracts for internal circulationwithin their institutions. Permission of thepublisher is required for resale or distributionoutside the institution. Permission of the pub-lisher is required for all other derivative works,including compilations and translations.
Electronic Storage: Permission of the pub-lisher is required to store electronically any ma-terial contained in this journal, including anyarticle or part of an article. Contact the pub-lisher at the address indicated.
Except as outlined above, no part of this publica-tion may be reproduced, stored in a retrieval sys-tem, or transmitted in any form or by any means,electronic, mechanical, photocopying, recordingor otherwise, without prior written permission ofthe publisher.
No responsibility is assumed by either the Pub-lisher or the Association of Professors of Med-icine for any injury and/or damage to personsor property as a matter of products liability,negligence or otherwise, or from any use or op-eration of any methods, products, instructionsor ideas contained in the material herein. Be-cause of rapid advances in the medical sciences,independent verification of diagnoses and drugdosages should be made.
Although all advertising material is expected toconform to ethical (medical) standards, inclu-sion in this publication does not constitute aguarantee or endorsement of the quality orvalue of such product or of the claims made ofit by its manufacturer.
Advertising inquiries should be addressed toStuart Williams, The Braveheart Group LLC, 2Split Rock Road, Mahwah, NJ 07430. Phone:1-201-529-4004, Fax: 1-201-529-4007, E-mail:[email protected] or PeterMurphy, The Braveheart Group LLC, 225Woodside Avenue, Franklin Lakes, NJ 07417.Phone: 1-201-847-1934, Fax: 1-201-847-8166,E-mail: [email protected].
Classified advertising inquiries should be ad-dressed to: Danny Wang, Elsevier Inc., 360
Park Avenue South, New York, NY 10010.Phone: 1-212-633-3158, Fax: 1-212-633-3820.
Reprint inquiries should be addressed to AnneRosenthal, Elsevier Inc., 360 Park AvenueSouth, New York, NY 10010. Phone: 1-212-633-3813, Fax: 1-212-633-3820.
The publication of advertising in the Journaldoes not constitute any guarantee or endorse-ment by the Association of Professors of Med-icine, the Journal, or the Publisher of the adver-tised product or service or of the claims madeby the advertiser. The publication of articlesand other editorial materials in the Journaldoes not represent the policy or recommenda-tions of, or endorsement by, the Association,the Journal, or the Publisher.
Subscriptions: Solicited only from internistsand other physicians. Publisher reserves theright to refuse unqualified subscribers. Pleaseaddress all subscription mail to The AmericanJournal of Medicine®, Elsevier SubscriptionCustomer Service, 6277 Sea Harbor Drive, Or-lando, FL 32887-4800. Single copy price (in-cluding symposia and special issue): $47. 2007Subscription rate for individuals in the field ofpublication: US and Canada: $131.00; all othercountries: $317.00 (air delivery, no surface de-livery rate available). Institutional rate: US andCanada: $336.00; all other countries: $563.00(air delivery, no surface delivery rate available).Publisher is not responsible for replacing miss-ing issues unless the Circulation Department isnotified of nonreceipt within 3 months of issuedate for domestic addresses and 6 months forforeign addresses. Circulation records aremaintained at Excerpta Medica, Inc., MadisonSquare Station, P.O. Box 882, New York, NY10159-0882.
Reprints: For queries about offprints, e-mail:[email protected]. To order 100 ormore reprints for educational, commercial, orpromotional use, contact the Commercial Re-prints Department, Elsevier Inc., 360 Park Ave-nue South, New York, NY 10010-1710. Fax:212-462-1935; e-mail: [email protected] of single article available online may beobtained by purchasing Pay-Per-View accessfor $30 per article on the journal Web site,www.amjmed.com.
Customer Service Numbers: Inside US:1-800-215-4692; Outside US (Includes Can-ada): 1-815-734-1216; Agent Questions: 1-815-734-1207; Customer Service Fax: 1-740-382-5866.
Editorial Board Continued
C
WWlmgpusUtAladt2aa
cgemnr
etctelhd
2ilmpehemm
0d
The American Journal of Medicine (2007) 120, 469-471
OMMENTARY
ho Will Care for the Frail Elderly?snrsfis
Ascltoii
fwaroccapaoaAOila
afafl
lwo
idespread understanding of the principles of a healthyifestyle coupled with remarkable scientific advances inedicine over the last 30 years have enabled the current
eneration of Americans to live considerably longer com-ared with previous generations. This growing elderly pop-lation in the US is changing the social demographics of ourociety rapidly. Between 1950 and 2005, the segment of theS population over the age of 75 years grew more than
wice as fast as the general population.1 While many elderlymericans are remarkably vigorous and active until late in
ife, approximately one third of Americans over 75 years ofge report that they are in fair or poor health.1 As we age,isability eventually descends; many seniors become frail inhe final years of life. Of Americans over 75 years of age,0% have trouble seeing (even with corrective lenses); 15%re deaf or have trouble hearing; 44% have limited mobility;nd 36% have 3 or more chronic conditions (Figure 1).1
In response to the ever increasing number of elderlyitizens in our society, a variety of privately funded andovernment-supported institutions now exist to care for thelderly. Depending upon individual needs, living arrange-ents and care are available at multiple levels, including
ursing homes, assisted living facilities, and continuing careetirement communities.
Nursing homes are designed to house and care for fraillderly people who cannot care for themselves. In 2004,here were 1.8 million nursing home beds and 16,000 Medi-are- and Medicaid-certified facilities in the US.1 The na-ionwide occupancy rate is estimated to be 83%.1 As is to bexpected, nursing home residents have more health prob-ems than the general senior population; 80% of nursingome residents are mobility dependent; 47.3% are eatingependent; and 65.7% are incontinent.1
Seniors who need some daily services but don’t need the4-hour skilled care of a nursing home may choose to liven an assisted living facility. My 93-year-old mother (JSA)ives in such a facility here in Southern Arizona. Likeillions of Americans, I oversee her care and have assumed
ower of attorney because she can no longer see wellnough to write checks and pay her bills. Through her, Iave had a number of insights into the situation of the fraillderly in our affluent society. Assisted living facilities, likey mother’s, offer a number of services, amenities, and a
odest degree of assistance. Although apartments may have t002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.03.006
mall kitchens, meals in a communal dining room are fur-ished. In addition, housekeeping, nursing care, security,eligious services, transportation for appointments andhopping, various forms of entertainment (ie, handicrafts,lms, bridge tournaments, exercise, games), and other re-ources are all available to residents for one monthly fee.
There are many assisted living facilities in Southernrizona and an estimated 20,000 across the US.2 Some are
tand-alone facilities, while others are affiliated continuingare retirement communities, which include independentiving and nursing home care. Active seniors may start inhe independent living section of a facility and move tother sections as they become progressively disabled. Thesenstitutions offer additional care and support, often includ-ng an Alzheimer’s disease unit.
The national average cost to live in an assisted livingacility is $2968 monthly or $35,616 a year.3 Comparedith $206 a day or $75,190 annually for a private room innursing home facility,4 the price for assisted living seems
easonable, given the services and amenities. Unfortunately,nly seniors with additional retirement income beyond so-ial security (ie, individual retirement accounts, long-termare or some other form of insurance, or real estate) canfford to live in these retirement centers. Social securityayments alone would not suffice, and Medicare, Medicaid,nd Veterans benefits do not fund long-term care.5 The costf long-term care must, in most cases, be borne by seniorsnd/or their families. Consequently, only a small number ofmericans needing long-term care can afford to pay for it.ften the elderly and their families are forced to endure
ndignities, crowded living conditions, and low quality ofife because they lack the financial wherewithal to pay forppropriate long-term care.
Millions of frail seniors depend, as does my mother, onssistance from family members and friends. Besides visits,amily members provide support in areas such as financialffairs, home-cooked food, errands, doctor’s visits, and soorth. “Informal” unpaid caregivers provide most of theong-term care services in the US.6
My mother and her friends are the fortunate ones. Mil-ions of elderly Americans are not so fortunate. Everyday,e see these less fortunate individuals admitted to the floorsf our university hospital with acute illnesses or degenera-
ive chronic disease states. When they are discharged from
ois
pdrtgccAc
ibwtcwdsti
dffiWfmtgt
R
470 The American Journal of Medicine, Vol 120, No 6, June 2007
ur acute care hospital, we have improved their health, butt is often impossible for us to improve their livingituations.
Although caring for elderly parents was a widespreadractice in the nuclear families of previous generations, theuration of such care was considerably shorter commensu-ate with the much shorter lifespan. In the very near future,he “baby boomer” generation will begin to arrive at theeriatric threshold. Shortly thereafter, as they age and be-ome frail, this generation will present a major societalhallenge. According to the US government, 15 millionmericans used nursing facilities, alternative residential
are, or home care services in 2000, and this is expected to
Figure 1 Chronic conditions among Americans.1
Table 1 Aging in the US
The Current Population35 million Americans—12% of US population—are aged
65 years or older.8,10
Median age is 74.4 years.8
57% are women.8
85% are white.8
89% are native born.8
40.5% have some type of disability.8
15% are in the labor force.8
80% live in owner-occupied housing.8
31% live alone; 54% live with a spouse; 13% live withrelatives other than a spouse; and 2% live with non-relatives (2001).9
10% live below the poverty level.2
The Trends64% increase in life expectancy—from 47.3 to 77.8 years
of age—(1900-2004).1
108% increase in Medicare enrollment (1970-2005).1
52.8% increase in nursing home residents 65 years of ageand older (1973-1999).1
86.7 million Americans—21% of the US population—willbe 65 years of age or older by 2050.10
ncrease to 27 million by 2050.6 Consequently, as the “babyoomers” progress to old age, the demand for direct careorkers (ie, registered nurses, licensed practical and voca-
ional nurses, nurse aides, and home health and personalare workers) will increase dramatically, as the availableorkforce shrinks.6 In addition to a projected shortfall inirect care workers, medical schools will be hard-pressed toupply the projected 36,000 certified geriatricians necessaryo care for the 70 million Americans who will be 65 or oldern 2030.7
How will the US care for this huge wave of potentiallyisabled elderly? (Table 1, Figures 1 and 2) Who will payor their care? How can we afford the additional societal andnancial burden? These questions must be addressed now.ithout careful planning and investment, the US soon will
ace a rapidly growing group of frail elderly citizens whoseisery will be evident throughout our society. Now is the
ime to begin discussion and debate—in the halls of Con-ress and in the halls of medical and nursing schools acrosshe country—in order to address this impending crisis.
Joseph S. Alpert, MDRobert S. and Irene P. Flinn, Professor of Medicine
Special Assistant to the DeanUniversity of Arizona College of Medicine
Pamela J. Powers, MPHManaging Editor
American Journal of Medicine
eferences1. National Center for Health Statistics. Health, United States, 2006 With
Chartbook on Trends in the Health of Americans. Available at: http://www.cdc.gov/nchs/hus.htm. Accessed March 9, 2007.
2. Assisted Living Federation of America. Frequently Asked QuestionsAbout Assisted Living. Available at: http://www.alfa.org/i4a/pages/index.cfm?pageid�3285. Accessed March 12, 2007.
3. MetLife Mature Market Institute. The MetLife Market Survey ofAssisted Living Costs, October 2006. Available at: http://www.metlife.com/WPSAssets/97714255001163445528V1F2006Assisted
Figure 2 US Population growth, 1950-2050.1
Living.pdf. Accessed March 9, 2007.
1
471Alpert The Frail Elderly
4. MetLife Mature Market Institute. The MetLife Market Survey ofNursing Home and Home Care Costs, September 2006. Available at:http://www.metlife.com/WPSAssets/21052872211163445734V1F2006NHHCMarketSurvey.pdf. Accessed March 9, 2007.
5. US Department of Health and Human Services. Medicare. Paying forLong-Term Care. Updated January 22, 2007. Available at: http://www.medicare.gov/LongTermCare/Static/PayingOverview.asp? Ac-cessed March 12, 2007.
6. US Department of Health and Human Services. Office of the AssistantSecretary for Planning and Evaluation, the Centers for Medicare and MedicadServices, and Health Resource and Services Administration. Department ofLabor’s Office of the Assistant Secretary for Policy, Bureau of Labor Statis-tics and Employment and Training Administration. The Future Supply ofLong-term Care Workers in Relation to the Aging Baby Boom Generation:Report to Congress, May 14, 2003. Available at: http://www.aspe.hhs.gov/dalcp/reports/ltcwk/htm. Accessed March 12, 2007.
7. Butterfield S. American College of Physicians. A New Look at the
Elderly and How to Care for Them. American College of PhysiciansObserver. Available at: http://www.acponline.org/journals/news/jan-feb07/geriatrics.htm. Accessed March 9, 2007.
8. US Census Bureau. SO103. Population 65 Years and Over in theUnited States. 2005 American Community Survey. Available at:http://factfinder.census.gov/servlet/STTable?_bm�y&-geo_id�01000US&-qr_name�ACS_2005_EST_G00_S0103&-ds_name�ACS_2005_EST_G00_&-redoLog�false. Accessed March 9, 2007.
9. US Department of Health and Human Services. Centers for DiseaseControl and Prevention. National Center for Health Statistics. Trends inHealth and Aging. Available at: http://209.217.72.34/aging/TableViewer/tableView.aspx?ReportId�316ttp://209.217.72.34/aging/TableViewer/tableView.aspx?ReportId�316. Accessed March 9, 2007.
0. US Census Bureau, Population Division, Population Projections Branch.Projected Population of the United States, by Age and Sex: 2000 to 2050.Created: March 18, 2004. Available at: http://www.census.gov/ipc/www/
usinterimproj/natprojtab02a.pdf. Accessed March 13, 2007.
C
ARApio
oaacpw2
gw
tctpa
Ps
cAau
dsHsocti
eoe
0d
The American Journal of Medicine (2007) 120, 472-474
OMMENTARY
re Patients with a Patent Foramen Ovale at Increased
isk of Stroke? A Billion Dollar Questioncwp
opsrtwa
WSTP1pp
scpaIippl
to
te
ee
df
pproximately one third of the 600,000 ischemic strokeser year in the US are judged to be cryptogenic; that is, theres no evidence of cerebrovascular disease and no evidencef a cardiac source.1,2
In 1988, Lechat et al3 reported an increased prevalencef patent foramen ovales (PFOs), as detected by transesoph-geal echocardiography (TEE), in 60 patients younger thange 55 years with an ischemic stroke who had a normalardiac examination. Reports since then have shown that therevalence of PFO is approximately 30% to 40% in patientsith cryptogenic strokes1,4 compared with approximately5% in the general population.5,6
These findings indicate that 60,000 to 80,000 crypto-enic ischemic strokes per year in the US occur in patientsith a PFO.The increased prevalence of PFOs in patients with cryp-
ogenic strokes has lead to the assumption that when aryptogenic stroke occurs in a patient with a PFO, it is dueo paradoxical embolism, ie, that a venous thrombus hasassed through the PFO to enter the systemic circulationnd then has embolized the cerebral circulation.
The assumption that cryptogenic strokes in patients withFO are due to paradoxical embolism led to reports ofurgical closure of PFO to prevent recurrent strokes.7
Surgical closure has been replaced by the introduction ofatheter devices that permit transvenous closure of a PFO.t present there are at least 9 different transvenous devices
vailable to close PFOs.5 This procedure can be performednder local anesthesia with minimal serious complications.
The Food and Drug Administration approves this proce-ure in patients who have recurrent unexplained ischemictrokes despite adequate anticoagulation with warfarin.8
owever, the “off-label” use of this procedure is wide-pread and is accelerating. There are reports of the closuref up to 300 to 400 PFOs in patients who have had aerebral ischemic event of unknown cause.9,10 None ofhese reports had controls, and the efficacy of this procedures unproven.
The financial implications of this unproven therapy arenormous. These devices cost $2800 to $5500; the total costf the procedure is approximately $10,000 per patient.2 It is
stimated that, worldwide, 500,000 cryptogenic strokes oc- b002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2006.03.021
ur annually in patients with a PFO; the estimated potentialorld market for this procedure is approximately $2 billioner year.2
Several randomized clinical trials of transvenous closuref PFO are currently underway.11 However, recruitment ofatients to these trials has been slow because of a wide-pread belief that this treatment is effective in preventingecurrent stroke.11 It appears that this procedure is followinghe path of the pulmonary artery catheter; widespreadorldwide use without verification of efficacy by appropri-
tely designed randomized clinical trials.12
HAT IS THE EVIDENCE THAT CRYPTOGENICTROKES IN PATIENTS WITH A PFO ARE DUEO PARADOXICAL EMBOLISM?aradoxical embolism was first described by Cohnheim in877.13 Originally, the diagnosis could only be made atostmortem by finding a venous thrombus straddling aatent foramen ovale or other intracardiac defect.
Johnson, in 1951,14 proposed 3 criteria to make a pre-umptive clinical diagnosis of paradoxical embolism. Theseriteria are: systemic embolism without a cardiac source, theresence of venous thrombosis or pulmonary embolism, andn intracardiac defect that will permit a right-to-left shunt.n nearly all reported cases of paradoxical embolism, thentracardiac defect was a PFO. By 1970, only 8 cases ofaradoxical embolism diagnosed during life had been re-orted in the medical literature.15 The paradoxical embo-ism caused a stroke in 3 of these 8 cases.
Few of the reported cases of cryptogenic stroke in pa-ients with PFO meet these criteria for the clinical diagnosisf paradoxical embolism.14
The first criterion is the occurrence of unexplained sys-emic embolism. It is not certain that cryptogenic strokes arembolic; they could be thrombotic.
The second criterion is the presence of venous thrombo-mbolism. Few of the reported cases have noted the pres-nce of venous thrombosis or pulmonary embolism.
The third criterion is the presence of an intracardiacefect permitting a right-to-left shunt. It is clear that a patentoramen ovale can permit a right-to-left shunt. However,
ecause left atrial (LA) pressure normally exceeds right
atsslbtRar
sdswwd
emRswibavt
hops
ptenoas
AIEtPoosv
Ppat
psnP
aTifssaPhiwti
srgmroraawseasa
idpihTpdiTrP
tr
CTw
473Dalen PFO and Risk of Stroke
trial (RA) pressure, the increased pressure in the LA closeshe flap-like opening in the PFO, preventing right-to-lefthunting across the PFO.15,16 However, if right atrial pres-ure is increased, and exceeds left atrial pressure, right-to-eft shunting can occur.15,16 Most cases of paradoxical em-olism occur in patients with major pulmonary embolismhat has caused right heart failure with a resultant increase inA pressure.17 A valsalva maneuver also can cause righttrial pressure to exceed LA pressure, thereby permittingight-to-left shunting across a PFO.15,16,18
Banas et al18 reported the use of dye curves to detectmall (1 to 12 mm diameter) surgically created atrial septalefects in dogs. Right-to-left shunting across defects asmall as 1.5 mm diameter was demonstrated if the injectionas made during a simulated valsalva maneuver. At rest,ithout a valsalva maneuver, right-to-left shunts could beetected in only 3 of 8 defects.
Right-to-left shunting across a PFO can be detected bychocardiography. When saline is injected intravenously,icro bubbles may be detected crossing the PFO from theA to the LA.19 However, in many cases, right-to-left
hunting can be demonstrated only if the injection is madehile the patient is performing a valsalva maneuver to
ncrease RA pressure. The demonstration by TEE that microubbles may cross a PFO and enter the left atrium in thebsence of a valsalva maneuver does not establish thatenous thrombi could cross a PFO when RA pressure is lesshan LA pressure.
Few of the cryptogenic strokes in patients with a PFOave occurred in patients with major pulmonary embolismr other causes of right heart failure, or while patients wereerforming activities that simulate a valsalva maneuver,uch as straining at the stool.
In summary, the evidence that cryptogenic strokes inatients with a PFO are due to paradoxical embolism is lesshan compelling. The evidence that the strokes are due tombolism is uncertain. In the vast majority of cases, there iso evidence of venous thromboembolism. There is evidencef a PFO, but in most cases there is no evidence that righttrial pressure was elevated, permitting right-to-lefthunting.
RE PATIENTS WITH PFO AT INCREASE RISK OFSCHEMIC STROKE?ven if cryptogenic strokes in patients with PFO are not due
o paradoxical embolism, it is possible that patients with aFO may be at increased risk of stroke by some mechanismther than paradoxical embolism. It has been suggested thatther abnormalities associated with PFO, such as atrialeptal aneurysms,20,21 Chiari networks,1,5 or Eustachianalves,1 may predispose to strokes.
The most accurate way to determine if patients with aFO are at increased risk of ischemic strokes would be toerform an observational study of patients with and withoutPFO as documented by TEE and follow them to determine
he rate of ischemic stroke. If such a study found that l
atients with a PFO were not at increased risk of ischemictrokes, there would be no rationale for closure of PFO, ando reason to perform randomized clinical trials of closure ofFOs in patients with cryptogenic strokes and PFOs.
A small prospective observational study of patients withnd without PFO was recently reported by Meissner et al.6
hey determined the presence or absence of a PFO by TEEn 585 healthy people aged over 45 years. Over a 5-yearollow-up, the incidence of transient ischemic attack, tran-ient ischemic stroke, and cerebrovascular death was theame in the 24% who had PFO as those without a PFO. Inmore recent study, DiTullio et al22 found a prevalence ofFO of 14.9% in 1100 patients older than 39 without aistory of stroke. During follow-up of more than 6 years, thencidence of ischemic stroke was the same in those with orithout a PFO. Additional large studies are needed to de-
ermine if patients with a PFO are at increased risk ofschemic stroke.
Several investigators have reported the recurrence rate oftrokes in patients with cryptogenic strokes. Mas et al20
eported a prospective study of 581 patients with crypto-enic ischemic stroke. They were treated with aspirin, 300g/day, and followed for 4 years. The incidence of recur-
ent stroke was actually higher (4.2%) in 304 patients with-ut a PFO than in 216 patients with a PFO (2.3%). Theyeported an increased incidence of 15.2% in 51 patients withPFO who also had an atrial septal aneurysm.20 Homma et
l4 randomized 265 patients with cryptogenic stroke toarfarin or to aspirin. The recurrence rate of stroke was the
ame in those with or without a PFO in those treated withither warfarin or aspirin. In contrast to the report of Mas etl,20 they found no difference in the incidence of recurrenttroke in patients with PFO with or without an associatedtrial septal aneurysm.4
An additional way to determine if there is an increasedncidence of ischemic strokes in patients with PFO is toetermine if the rate of recurrent stroke is decreased inatients who have had their PFO closed. Several such stud-es have been reported.9,10,23 However, none of these studiesas been randomized. There has been no blinding of theEE readings and no blinding of the determination of endoints.24 In addition, because these studies were not ran-omized, there has been variability in associated conditionsn those undergoing or not undergoing catheter closure.hey are also difficult to assess because they used variable
egimens of antiplatelet or anticoagulant therapy after theFO had been closed.24
A properly designed randomized clinical trial is requiredo determine if closure of PFOs decreases the rate of recur-ent stroke.
ONCLUSIONhe observation that PFO are more prevalent in patientsith cryptogenic strokes than in the general population has
ed to the conclusion that these strokes are due to paradox-
ic
etces
aerpeiiple
R
1
1
1
1
11
1
1
1
1
2
2
2
2
2
2
474 The American Journal of Medicine, Vol 120, No 6, June 2007
cal embolism, and therefore, these patients should undergolosure of their PFO.
The evidence that these strokes are due to paradoxicalmbolism is minimal. It is uncertain that the strokes are dueo embolism, and in most cases there is no evidence ofo-existent venous thrombosis. In the absence of a cause forlevated right atrial pressure, it is uncertain if a right-to-lefthunt has occurred.
It has not been demonstrated that patients with PFO aret increased risk of recurrent stroke. There is currently novidence that transvenous closure of PFOs decreases theisk of recurrent stroke. Given the costs associated with thisrocedure, it would be prudent to await the results of prop-rly designed randomized clinical trials before recommend-ng this unproven procedure. Until the results of a random-zed clinical trial are known, it would be appropriate to treatatients with cryptogenic strokes and a PFO with antiplate-et agents as recommended by the Seventh ACCP Confer-nce on Antithrombotic and Thrombolytic Therapy.25
James E. Dalen, MD, MPHProfessor Emeritus
University of ArizonaTucson
eferences1. Meier B, Lock JE. Contemporary management of patent foramen
ovale. Circulation. 2003;107:5-9.2. Maisel WH, Laskey WK. Patent foramen ovale closure devices.
JAMA. 2005;294:366-369.3. Lechat P, Mas JL, Lascault G, et al. Prevalence of patent foramen
ovale in patients with stroke. N Engl J Med. 1988;318:1148-1152.4. Homma S, Sacco RL, DiTullio MR, et al. Effect of medical treatment
in stroke patients with patent foramen ovale. Circulation. 2002;105:2625-2631.
5. Hara H, Virmani R, Ladich E, et al. Patent foramen ovale: currentpathology, pathophysiology, and clinical status. J Am Coll Cardiol.2005;46:1768-1776.
6. Meissner I, Khandheria BK, Heit JA, et al. Patent foramen ovale:innocent or guilty? J Am Coll Cardiol. 2006;47:440-445.
7. Homma S, DiTullio MR, Sacco RL, et al. Surgical closure of patentforamen ovale in cryptogenic stroke patients. Stroke. 1997;28:2376-
2381.8. Blackshear JL. Closure of patent foramen ovale in cryptogenic stroke.J Am Coll Cardiol. 2004;44:759-761.
9. Wahl A, Krumsdorf U, Meier B, et al. Transcatheter treatment of atrialseptal aneurysm associated with patent foramen ovale for preventionof recurrent paradoxical embolism in high risk patients. J Am CollCardiol. 2005;45:377-380.
0. Krumsdorf U, Ostermayer S, Billinger K, et al. Incidence and clinicalcourse of thrombus formation on atrial septal defect and patent fora-men ovale closure devices in 1,000 consecutive patients. J Am CollCardiol. 2004;43:302-309.
1. Thomson JDR. Percutaneous PFO closure, further data but so manyunanswered questions. Eur Heart J. 2005;27:258-259.
2. Dalen JE, Bone RC. Is it time to pull the pulmonary artery catheter?JAMA. 1996;276:916-918.
3. Cohnheim J. Thrombose und embolie. Vorlesungen uber allgemeinepathologie. 1877;1:134.
4. Johnson BI. Paradoxical embolism. J Clin Pathol. 1951;4:316-332.5. Gazzaniga AB, Dalen JE. Paradoxical embolism: its pathophysiology
and clinical recognition. Ann Surg. 1970;171:137-142.6. Cheng TO. The proper conduct of valsalva maneuver in the detection
of patent foramen ovale. J Am Coll Cardiol. 2005;45:1145-1146.7. Meister SM, Grossman W, Dexter L, Dalen JE. Paradoxical embolism
diagnosis during life. Am J Med. 1972;53:292-298.8. Banas JS, Meister SG, Gazzaniga AB, et al. A simple technique for
detecting small defects of the atrial septum. Am J Cardiol. 1971;28:467-471.
9. De Castro S, Cartoni D, Fiorelli M, et al. Morphological and functionalcharacteristics of patent foramen ovale and their embolic implications.Stroke. 2000;31:2407-2413.
0. Mas JL, Arouizan C, Lamy C, et al. Recurrent cerebrovascular eventsassociated with patent foramen ovale, atrial septal ameurysm or both.N Engl J Med. 2001;345:1740-1746.
1. Overell JR, Bone I, Lees KR. Interatrial septal abnormalities andstroke: a meta-analysis of case-control studies. Neurology. 2000;55:1172-1179.
2. Di Tullio MR, Sacco RL, Sciacca RR, et al. Patent foramen ovale andthe risk of ischemic stroke in a multiethnic population. J Am CollCardiol. 2007;49:797-802.
3. Windecker S, Wahl, A, Nedetchev K, et al. Comparison of medicaltreatment with percutaneous closure of patent foramen ovale in pa-tients with cryptogenic stroke. J Am Coll Cardiol. 2004;44:750-758.
4. Kizer JR, Devereux RB. Patent foramen ovale in young adults withunexplained stroke. N Engl J Med. 2005;353:2361-2372.
5. Albers GW, Amarenco P, Easton JD, et al. Antithrombotic and throm-bolytic therapy for ischemic stroke: the seventh ACCP Conference on
antithrombotic and thrombolytic therapy. Chest. 2004;126:483s-512s.
R
DZa
a
DptppapcOctarptina
DD
0d
The American Journal of Medicine (2007) 120, 475-480
EVIEW
irect-to-Consumer Advertising of Pharmaceuticalsiad F. Gellad, MD, MPH,a Kenneth W. Lyles, MDb
Division of Gastroenterology, Duke University Medical Center, Durham, NC; bDivision of Geriatrics, Duke University Medical Center
nd Geriatric ReScidTmiaiwr
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
search Education and Clinical Center (GRECC), Durham VA Medical Center, Durham, NC.
ABSTRACT
ince the US Food and Drug Administration (FDA) released new guidelines on broadcast direct-to-onsumer advertising in 1997, the prevalence of direct-to-consumer advertising of prescription drugs hasncreased exponentially. The impact on providers, patients, and the health care system is varied andynamic, and the rapid changes in the last several years have markedly altered the health care landscape.o continue providing optimal medical care, physicians and other health care providers must be able toanage this influence on their practice, and a more thorough understanding of this phenomenon is an
ntegral step toward this goal. This review will summarize the history of direct-to-consumer drugdvertisements and the current regulations governing them. It will summarize the evidence concerning thempact of direct-to-consumer advertising on the public, providers, and the health care system, and concludeith observations regarding the future of direct-to-consumer advertising. © 2007 Elsevier Inc. All rights
eserved.
KEYWORDS: Advertising; Direct-to-Consumer; Marketing; Pharmaceutical preparations; United States Food andDrug Administration
rcadowd
BDTubPtlfi(masl
irect-to-consumer advertising of prescription drugs is aowerful force in the health care market. Proponents claimhat good direct-to-consumer advertising educates and em-owers patients in their relationship with their health careroviders.1 Furthermore, they assert that direct-to-consumerdvertising provides an opportunity for patients to talk withroviders about under-diagnosed and under-treated medicalonditions and may lead to improved health outcomes.pponents of the practice argue that there is an inherent
onflict of interest for pharmaceutical companies in adver-ising products directly to patients and that the decision todvertise is not based on concern for the public health butather investment return.2 Physicians and other health careroviders are often caught in the middle of this debate tryingo bridge the divide between consumer perception and med-cal reality. In order to be successful at this task, physicianseed to be optimally informed about direct-to-consumerdvertising and its effects on health care practice. This
Requests for reprints should be addressed to Ziad F. Gellad, MD,ivision of Gastroenterology, Duke University Medical Center, Box 3913,urham, NC 27710.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.09.030
eview will cover the history and prevalence of direct-to-onsumer advertising of prescription drugs, current Foodnd Drug Administration (FDA) guidelines, the evidence toate regarding the effect of direct-to-consumer advertisingn the public, physicians, and the health care system, andill conclude with observations regarding the future ofirect-to-consumer advertising.
ACKGROUNDirect-to-consumer advertising is not a new phenomenon.he first recorded advertisement of a pharmaceutical prod-ct occurred in the early 18th century, and the practiceecame widespread in the 19th and early 20th centuries.3
artially as a response to the proliferation of these adver-isements, the United States Congress undertook a series ofegislative steps to regulate drug advertisements. One of therst steps was the Federal Food, Drug, and Cosmetic ActFDCA) in 1938 that established the FDA and requirededications be proven safe prior to marketing. This act was
mended in 1962 to require that medications be proven bothafe AND efficacious prior to being advertised to the pub-ic.4 The FDCA requires drug companies to provide a sum-
ary of the product label, including all the risk-related
itmcpcpptpenpmhrhtdrcfqmi
hridfiotncfm
snrcm
rb
rcb
.
476 The American Journal of Medicine, Vol 120, No 6, June 2007
nformation in a product’s package labeling, in any promo-ional material. This is known as the brief summary require-
ent. Pharmaceutical companies found these regulationsumbersome and prohibitive to apply to the public, soharmaceutical promotion was directed primarily at healthare providers.5 In the 1980s,harmaceutical companies ex-anded direct-to-consumer adver-ising in magazines and newspa-ers with the intent ofmpowering consumers who wereewly focused on the notion ofatient autonomy.6 After a brieforatorium and review, the FDA
eld that these consumer-di-ected advertisements should beeld to the same standards ashose directed to physicians. Thisecision limited the spread of di-ect-to-consumer advertising be-ause pharmaceutical companieselt that the brief summary re-uirement would negate any pro-otional impact of advertisements
n the minds of consumers.4
In 1997, after a series of publicearings on the topic, the FDAeleased draft guidelines regulat-ng broadcast advertisements andramatically changed the playingeld by allowing the expansionf direct-to-consumer advertising into broadcast and elec-ronic media (Table 1). Furthermore, drug companies wereo longer required to provide a brief summary but insteadould inform consumers of alternative sources for that in-ormation. This is known as the adequate provision require-ent. This requirement could be fulfilled by referring con-
CLINICAL SIGNIF
● The United Statethe only industridirect-to-consumscription drugs.
● Direct-to-consumcreased exponenof new FDA guiding broadcast ad
● Direct-to-consumlates discussionsproviders and mprescription drug
● The future of dirtising will depecians’ ability toon their practice
Table 1 FDA Guidelines on Direct-to-Consumer Drug Advertise
Print Advertisements*
Brief Summary—Advertisements must disclose each side effect,warning, precaution and contraindication from theapproved product professional labeling. FDA-approvedpatient labeling that focuses on the most serious risks andless serious, but most frequently occurring, adversereactions is also acceptable. The latter must include:
(1) All contraindications(2) All warnings(3) Major precautions, including any that describe serious
adverse events(4) The 3-5 most common non-serious adverse reactions
most likely to affect the patient’s quality of life orcompliance with drug therapy
*Amended in 2004 draft guidance from the FDA’s Division of Drug MaResearch, available at www.fda.gov/cder/Guidance/5669dft.pdf.
†Based on 1999 final FDA guidelines on consumer-directed broadcast adver
umers to a physician, a website, a toll-free telephoneumber, and a print advertisement.7 The new guidelinesequired prescription drug broadcast advertisements to in-lude information about the major risks of a drug and theost commonly occurring adverse effects only. This re-
quirement is known as the majorstatement. With the finalization ofthis guideline in 1999, the modernera of direct-to-consumer advertis-ing was born.
PREVALENCE OF DIRECT-TO-CONSUMERADVERTISINGThe FDA divides direct-to-con-sumer advertising into 3 catego-ries (Table 2).5 Reminder adver-tisements contain the name of adrug only and are designed to re-inforce brand recognition; Help-seeking advertisements containinformation about a disease orcondition without mentioning aparticular treatment; and Product-claim advertisements contain spe-cific efficacy and safety informa-tion about a specific drug. Thelatter are regulated by the FDA,which requires that the advertise-ments present a fair balance of the
isks and benefits of a drug. This is referred to as the fairalance requirement.
Since 1997, there has been a significant increase in theate of growth of direct-to-consumer advertising in all 3ategories. In 2000, pharmaceutical companies spent $15.7illion on drug promotion.8 The majority of this money was
CE
New Zealand arecountries to allowvertising of pre-
dvertising has in-since the release
s in 1997 regulat-ements.
dvertising stimu-een patients andimulate increased
o-consumer adver-n part, on physi-age this influence
adcast Advertisements†
or Statement—Advertisements must disclose a product’s majorrisks and most commonly occurring adverse effects in eitherthe audio or audio and visual parts of the presentation
quate Provision—In place of brief summary, advertisementsmay make “adequate provision” for dissemination of packagelabeling with four alternative sources of information:
1) Toll-free telephone number2) Referral to a print advertisement in a concurrently running
print publication, or provision of enough brochures, withrequired product information, in convenient outlets
3) Referral to a health care provider4) Internet web page address
Advertising, and Communications in the Center for Drug Evaluation and
ICAN
s andalizeder ad
er atiallyelinevertis
er abetw
ay stuse.
ect-tnd, iman
ments
Bro
Maj
andAde
((
((
rketing,
tisements, available at www.fda.gov/cder/guidance/1804fnl.pdf.
eddtaatpedepie
da2dtptmdls
DAWdtd
btpottScd
umritpa$t
atced7m4mwctlSte
cCtattd
r4aairsaOb
477Gellad and Lyles Direct-to-Consumer Drug Advertising
armarked for retail samples ($7.9 billion) and physicianetailing ($4 billion), whereas $2.5 billion was spent onirect-to-consumer advertising. Between 1996 and 2003,here was a 400% increase in spending on direct-to-consumerdvertising, from $791 million to $3.2 billion.9 In 2004, themount spent on direct-to-consumer advertising increasedo over $4 billion, another 23% increase from the yearrior.10 Additionally, pharmaceutical companies increasedmphasis on direct-to-consumer advertising as a means ofrug promotion. For example, although total promotionalfforts as a percentage of sales have remained constant, theroportion of sales devoted to direct-to-consumer advertis-ng has increased and has done so at a faster rate thanxpenditures for research and development.11,12
As compared with physician detailing, spending onirect-to-consumer advertising is focused on a more limitedrray of products. In fact, the top 20 advertised drugs in000 accounted for approximately 60% of all spending onirect-to-consumer advertising.12 The most heavily adver-ised classes include antidepressants, antihistamines, antihy-erlipidemics, proton pump inhibitors, and anti-inflamma-ory agents. In 2000, the top spender was Merck. The $161illion they spent advertising Vioxx rivals the marketing
ollars spent by Dell ($160 million), Budweiser ($146 mil-ion), Pepsi ($125 million), or Nike ($78 million) in theame year.12
IRECT-TO-CONSUMER ADVERTISINGND THE PUBLICith such significant expenditure, there is little surprise that
irect-to-consumer advertising reaches a wide audience inhe United States. In 2004, Brownfield et al reported that
Table 2 Categories of Direct-to-Consumer DrugAdvertisements
Category DescriptionFDARegulation
Help-seeking* Contains information about adisease or condition with arecommendation to consulta health care provider ifappropriate; Excludesdiscussions of specifictreatments or drugs
No
Product-claim Includes product’s name anda therapeutic claim,including safety andefficacy
Yes
Reminder Contains product’s nameonly; Designed to reinforcebrand recognition
No
Adapted from: Direct-to-consumer promotion; public hearing. FedRegist. 1995;60:42581-42584.5
*If the only available treatment for a condition is a specific pre-scription drug product, help-seeking advertisements are not allowed.
uring a 1-week period in Atlanta, the 3 major networks h
roadcast 907 advertisements for over-the-counter medica-ions and 428 advertisement for prescription drugs.13 Sixtyercent of all direct-to-consumer advertising in this studyccurred during news programs and soap operas, arguablyargeting older adults and women. The authors concludehat based on average television viewing in the Unitedtates, an adult is exposed to 100 minutes of direct-to-onsumer advertising for each minute they spend with theiroctor each year.13
Furthermore, multiple sources have documented almostniversal awareness of direct-to-consumer advertising. Theost recent FDA survey on the topic reported that 81% of
espondents were exposed to direct-to-consumer advertisingn 2002, a significant increase from 72% in 1999.14 Preven-ion magazine, which annually performs a series of nationalhone surveys on the topic, identified lower recognitionmong minorities and those with household incomes below25,000. Women were more likely to be aware of direct-o-consumer advertising than men.15
Despite widespread recognition of direct-to-consumerdvertising, patient perceptions about the quality of adver-isements is variable. Based on 2002 FDA data, 58% ofonsumers believe direct-to-consumer advertising providesnough information to make a decision about whether toiscuss an advertisement with a doctor. This is down from0% the year prior. Sixty percent of people felt advertise-ents did not provide enough information about risks, and
4% stated that the advertisements lacked sufficient infor-ation about benefit.9 These findings are not surprisinghen studies show that most broadcast advertisements give
onsumers more time to absorb facts about benefits thanhose about risks, and risks are presented at a higher gradeevel (9th grade) as compared with benefits (6th grade).16
imilarly, an analysis of prescription drug websites foundhat the homepage often disproportionately emphasizes ben-fits at the expense of risks.17
Perhaps indicative of these findings, significant miscon-eptions exist on the part of consumers. A 1998 survey inalifornia revealed that 50% of respondents thought direct-
o-consumer advertisements were submitted to the FDA forpproval prior to release. Similarly, 43% of respondentshought that only “completely safe” drugs could be adver-ised, and 21% thought that only “extremely effective”rugs were advertised.18
Another important issue to consider is the effect of di-ect-to-consumer advertising on patient behavior. In 2002,3% of respondents to the FDA survey reported that seeingdrug advertisement led them to look for more information
bout the drug; the majority of respondents (89%) obtainednformation from their doctor, while pharmacists (51%),eference books (40%) and the internet (38%) also wereources.14 Most consumers looked for more informationbout side effects as compared with benefits (61% vs 10%).nly 4% of patients reported seeing their doctors solelyecause of direct-to-consumer advertising. Of those who
ad seen an advertisement, 30%-35% discussed the adver-
tas
DPOctt7tdhfhof2b
etbttmqmfdaipt
DPPttpsirpspswamta
m
HdowtatdmatHaddrs
DPTdttptcecm
ctimrptifmcfelspdt
TTst
478 The American Journal of Medicine, Vol 120, No 6, June 2007
isement with their doctor. This translates into 61.1 milliondditional consumers asking about specific medications in aingle year.11
IRECT-TO-CONSUMER ADVERTISING AND THEHYSICIAN-PATIENT RELATIONSHIPf significant concern to the medical community is the
riticism that direct-to-consumer advertising undermineshe physician-patient relationship.19 In studying this issue,he 2002 FDA survey showed mixed results. For example,3% of patients did not believe direct-to-consumer adver-ising minimized the role of physicians in making productecisions. In fact, 43% felt direct-to-consumer advertisingelped them have better discussions with their doctor (downrom 62% in 1999). However, 10% of respondents wereesitant to talk to their doctor about a drug because of fearf creating the impression of distrust with the doctor (uprom 7% in 1999).14 Prevention magazine in 2001 found7% of respondents felt their visit with their physician wasetter because they discussed an advertisement.15
When polling physicians, a similar mixed picturemerges.14 Forty-one percent of physicians reported direct-o-consumer advertising exposure led to benefits such asetter discussions with patients and greater awareness ofreatments. Eighteen percent felt direct-to-consumer adver-ising led to problems such as increased time to correctisconceptions, requests for unnecessary drugs, and re-
uests for a drug therapy when a nonpharmacologic therapyight be as or more effective. Similarly, 41% of physicians
elt that the patient was confused about the efficacy of arug, and only 40% felt that patients understood the risks ofdrug after seeing a drug advertisement. Overall, advertis-
ng led to tension in a quarter of patient interactions, andrimary care physicians were more likely to report problemshan specialists.14
IRECT-TO-CONSUMER ADVERTISING ANDHYSICIAN BEHAVIORhysicians are more likely to prescribe a medication when
hey believe that the patient has an expectation to receivehat medication.20,21 By stimulating consumer demand,harmaceutical companies hope to increase physician pre-criptions for a particular product. About half of physiciansn the FDA survey reported some pressure to prescribe as aesult of direct-to-consumer advertising, with primary carehysicians again more likely than specialists to report pres-ure. Approximately 73% of primary care physicians feltatients expected a prescription, compared with 63% ofpecialists. In fact, patients who asked for a particular brandere more likely to receive the drug than those who did not
sk.14 The Government Accountability Office (GAO) esti-ates that about 8.5 million Americans received a prescrip-
ion drug in 2000 after viewing a drug advertisement andsking their physician for that drug.11
Evidence of this phenomenon of advertising-induced de-
and also can be found in the medical literature. In 2003, follon et al reported that women familiar with osteoporosisrugs due to direct-to-consumer advertising had 9 times thedds of having bone densitometry performed as comparedith matched controls.22 In 2005, Kravitz et al found that
hat physicians were significantly more likely to prescribentidepressants for standardized patients with depression ifhe patients made a brand-name drug request or a generalrug request as compared with no request at all.23 Further-ore, patients were more likely to receive “minimal accept-
ble initial care” for depression if they requested a medica-ion, illustrating a positive effect of patient empowerment.owever, physicians also were more likely to prescribe
ntidepressants for standardized patients with adjustmentisorder, despite minimal evidence of their benefit in thisisorder. The authors suggest that their findings may indi-ectly support the notion that direct-to-consumer advertisingtimulates prescribing for questionable indications.23
IRECT-TO-CONSUMER ADVERTISING ANDUBLIC HEALTHhe findings by Kravitz et al illustrate the difficulty ineciphering the effect of direct-to-consumer advertising onhe public health system. The effect of increasing prescrip-ions is dependent on the condition for which the drug isrescribed; that is, when an indication is appropriate andhe condition being treated is under-diagnosed, direct-to-onsumer advertising may improve overall health. How-ver, when an indication for therapy is inappropriate and aondition is over-diagnosed, direct-to-consumer advertisingay worsen overall health and strain the health care system.One area that is especially susceptible to strain is health
are costs. Opponents argue that direct-to-consumer adver-ising increases health care utilization and cost by stimulat-ng prescription drug use. In 2002, the GAO reported thatost of the spending increase for heavily advertised drugs
esulted from increased drug utilization.11 In the same re-ort, the GAO estimated that every 10% increase in direct-o-consumer advertising for a drug class results in a 1%ncrease in sales for that class. Furthermore, ecological datarom the Netherlands suggests that help-seeking advertise-ents can lead to increased physician visits for a particular
ondition.24 Prescription drugs costs are already one of theastest growing segments of health care. In 2005, it isstimated that the United States could have saved $20 bil-ion in prescription drug costs if generic drugs had beenubstituted for brand-name equivalents.25 If physicians areressured to prescribe newer, more expensive, prescriptionrugs as a result of direct-to-consumer advertising, the costso the health care system could be significant.
HE FUTUREhe health care market in the United States is a dynamicystem. Direct-to-consumer advertising emerged from rela-ive obscurity in 1997 to become a potent force shaping the
uture of health care. The United States and New Zealand
aSmctitcsrertMpca2tCabp
cakpGctrppctaaau
CPfaupsaroaaeep
ttaAthvaFttsm
R
11
1
1
1
1
1
1
479Gellad and Lyles Direct-to-Consumer Drug Advertising
re the only countries in the world at present to allow it.everal factors may alter the course ahead. First and fore-ost, for direct-to-consumer advertising to be effective in
reating interest in a product, consumers have to be willingo pay attention. FDA data discussed above suggest thatnterest in direct-to-consumer advertising has waned amonghe public.14 Secondly, pharmaceutical companies are con-erned about ever-increasing liability from direct-to-con-umer advertising. In 1999, the New Jersey Supreme Courtuled in Perez v Wyeth Laboratories that drug manufactur-rs have a legal obligation to warn consumers directly of theisks of their products and cannot rely solely on physicianso provide that information to the public.26 The outcome of
erck’s legal battles concerning Vioxx will undoubtedlylay a role in this debate as well. Finally, a collectiveoncern has arisen from the medical community about thesedvertisements. The American Medical Association, in June006, called for a moratorium on direct-to-consumer adver-ising of new prescription drugs.27 Similarly, the Americanollege of Physicians has stated that direct-to-consumerdvertising is inappropriate and in the absence of legislationanning the practice, increased regulation is necessary torotect the public from harm.19
The FDA also is concerned about the effect of direct-to-onsumer advertising on health care but is limited in itsbility to regulate the industry. The Division of Drug Mar-eting, Advertising, and Communications has only 40 em-loyees to review all direct-to-consumer advertising.19 TheAO found that the FDA’s ability is hampered because it
annot ensure that it receives all direct-to-consumer adver-ising for review prior to release, and it cannot releaseegulatory letters to violators in a timely fashion.11 Theharmaceutical industry introduced its own set of guidingrinciples in November 2005 that encouraged its memberompanies to submit all new direct-to-consumer advertisingo the FDA before release, educate health professionalsbout new medications prior to advertising to the public,nd ensure a fair balance of the risks and benefits in alldvertisements.1 It remains unclear what effects these vol-ntary guidelines will have.
ONCLUSIONSharmaceutical companies continue to increase fundingor direct-to-consumer advertising. Evidence discussedbove suggests that direct-to-consumer advertising stim-lates patient demand for pharmaceuticals, may influencehysician prescribing habits, and likely increases drugpending. Consumers are aware of these advertisements,nd discussions about them are becoming part of theoutine provider-patient relationship. As the phenomenonf direct-to-consumer advertising evolves and the prev-lence increases, providers need to continue in their roles advocates for patients. To do this effectively, provid-rs should stay educated about new medications and thevidence for their use. Providers need to understand that
atients may not have adequate information about a par-icular drug from advertisements, especially risks, andhat specific reinforcement of the risks of a drug is key tossuring that the patient makes an informed decision.dditionally, providers may want to prepare printed ma-
erial for patients ahead of time for some of the mosteavily advertised drugs. This may save time during theisit for other concerns and reinforce the provider’s roles advocate rather than create tension during a visit.inally, it is important to understand that patient ques-
ions and inquiries do not necessarily represent expecta-ions for a particular drug, but rather are opportunities totrengthen the provider-patient bond.28 This bond re-ains the foundation of optimal medical care.
eferences1. America’s pharmaceutical industry announces guidelines on direct-to-
consumer advertising. PhRMA Press Release; 2005. Available at:http://www.phrma.org/news_room/ press_releases. Accessed July 11,2006.
2. Hollon MF. Direct-to-consumer advertising: a haphazard approach tohealth promotion. JAMA. 2005;293:2030-2033.
3. Wilkes MS, Bell RA, Kravitz RL. Direct-to-consumer prescriptiondrug advertising: trends, impact, and implications. Health Aff. 2000;19:110-128.
4. Lyles A. Direct marketing of pharmaceuticals to consumers. Annu RevPublic Health. 2002;23:73-91.
5. Direct-to-consumer promotion; public hearing. Fed Regist. 1995;60:42581-42584.
6. Kessler D, Pines W. The federal regulation of prescription drug ad-vertising and promotion. JAMA. 1990;264(18):2409-2415.
7. Draft guidance for industry. Consumer-directed broadcast advertise-ments: availability. Fed Regist. 1997;62:43171-43173.
8. Findlay SD; The National Institute for Health Care Management(NIHCM) Foundation. Prescription Drugs and Mass Media Advertis-ing, 2000. Washington, DC: NIHCM Foundation; 2001.
9. Statement of the American College of Physicians for the Record of thePublic Hearing on Consumer-Directed Promotion of Regulated Med-ical Products; Department of Health and Human Services Food andDrug Administration; November 1, 2005. Available at: http://www.acponline.org/hpp/dtc_fda.pdf. Accessed July 11, 2006.
0. Querna B. The big pill pitch. U.S. News & World Report. 2005;52-53.1. Heinrich J. Prescription Drugs: FDA Oversight of Direct-to-Consumer
Advertising Has Limitations. United States General Accounting Office(GAO); 2002. Available at: www.gao.gov/new.items/d03177.pdf. Ac-cessed July 18, 2006.
2. Rosenthal M, Berndt ER, Donohue JM, et al. Promotion of prescrip-tion drugs to consumers. N Engl J Med. 2002;346:498-505.
3. Brownfield ED, Bernhardt JM, Phan JL, et al. Direct-to-consumer drugadvertisements on network television: an exploration of quantity, fre-quency, and placement. J Health Commun. 2004;9:491-497.
4. Aikin KJ, Swasy JL, Braman AC. Patient and Physician Attitudes andBehaviors Associated with DTC Promotion of Prescription Drugs—Summary of FDA Survey Research Results. Washington, DC: U.S.Department of Health and Human Services, Food and Drug Admin-istration; 2004.
5. International survey on wellness and consumer reactions to DTCadvertising of Rx Drugs: 2000-2001. Prevention. Emmaus, PA:Rodale Press; 2001.
6. Kaphingst KA, Dejong W, Rudd RE, Daltroy LH. A content analysisof direct-to-consumer television prescription drug advertisements.J Health Commun. 2004;9:515-528.
7. Huh J, Cude BJ. Is the information “fair and balanced” in direct-to-consumer prescription drug websites? J Health Commun. 2004;9:529-
540.
1
1
2
2
2
2
2
2
2
2
2
480 The American Journal of Medicine, Vol 120, No 6, June 2007
8. Bell RA, Kravitz RL, Wilkes MS. Direct-to-consumer prescriptiondrug advertising and the public. J Gen Intern Med. 1999;14:651-657.
9. Direct-to-Consumer Prescription Drug Advertising. Philadelphia, PA:American College of Physicians: Position Paper; 2006.
0. Avorn J, Chen M, Hartley R. Scientific versus commercial sources of influ-ence on the prescribing behavior of physicians. Am J Med. 1982;73:4-8.
1. Cockburn J, Pit S. Prescribing behaviour in clinical practice: patients’expectations and doctors’ perceptions of patients’ expectations—aquestionnaire study. BMJ. 1997;315:520-523.
2. Hollon MF, Larson EB, Koepsell TD, Downer A. Direct-to-consumermarketing of osteoporosis drugs and bone densitometry. Ann Pharma-cother. 2003;37:976-981.
3. Kravitz RL, Epstein RM, Feldman MD, et al. Influence of patients’requests for direct-to-consumer advertised antidepressants. JAMA.
2005;293:1995-2002.4. Jong GW, Stricker BH, Sturkenboom M. Marketing in the lay mediaand prescriptions of Terbinafine in primary care: Dutch cohort study.BMJ. 2004;328:931.
5. Agovino T. Generic drugs could have saved U.S. $20 billion. TheWashington Post October 25, 2005.
6. Mello MM, Rosenthal M, Neumann PJ. Direct-to-consumer advertis-ing and shared liability for pharmaceutical manufacturers. JAMA.2003;289:477-481.
7. American Medical Association. AMA calls for temporary moratorium ondirect-to-consumer advertising of new prescription drugs; 2005. Available at:http://www.ama-assn.org/ama/pub/category/print/16467.html. Accessed July11, 2006.
8. Cline RJW, Young HN. Direct-to-consumer print ads for drugs: dothey undermine the physician-patient relationship? J Fam Pract. 2005;
54:1049-1057.
R
AMa
W
ApayiBiwfih
em
COF
M0
0d
The American Journal of Medicine (2007) 120, 481-487
EVIEW
trial Fibrillation in the Elderlyargaret C. Fang, MD, MPH,a Jane Chen, MD,b Michael W. Rich, MDb
Division of General Internal Medicine Hospitalist Group, University of California, San Francisco; bCardiovascular Division,
ashington UnivApbtrmdfipaIma
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ersity School of Medicine, St. Louis, Mo.
ABSTRACT
trial fibrillation is increasingly prevalent among older adults. It causes approximately 24% of strokes inatients aged 80 to 89 years. The management of atrial fibrillation is directed at preventing thromboem-olism and controlling the heart rate and rhythm. Stroke prevention is most effectively accomplishedhrough administering anticoagulants such as warfarin, although older patients have higher hemorrhagicisk. Cognitive dysfunction, functional impairments, and increased fall risk further complicate warfarinanagement in elderly patients. The use of risk stratification schemes can help guide the anticoagulation
ecision, although the benefits of warfarin generally outweigh the risks in most older patients with atrialbrillation. Pharmacologic rate control has been shown to result in similar outcomes compared withharmacologic restoration of sinus rhythm and should be the initial therapy for elderly patients. Anti-rrhythmic medications should be selected based on an individual patient’s coexisting medical conditions.n symptomatic patients who fail pharmacologic therapy, invasive strategies such as AV nodal ablationay help improve quality of life and symptoms, although such strategies do not obviate the need for
ntithrombotic therapy. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Anti-arrhythmic; Anticoagulants; Atrial fibrillation; Elderly; Hemorrhage; Rate control; Rhythmcontrol; Stroke
aisthpowpamohc
SMti
trial fibrillation is a signature disease of aging, with arevalence of about 5% in people aged 65 years and oldernd affecting approximately 10% of those over the age of 80ears (Figure 1).1,2 The prevalence of atrial fibrillation isncreasing, in part because of the aging of the population.3
y year 2050, there will be an estimated 5.6 million peoplen the United States with atrial fibrillation; of these, 50%ill be over the age of 80 years.1 The management of atrialbrillation focuses on preventing thromboembolism andeart rate and rhythm control.
Nonvalvular atrial fibrillation increases the risk of isch-mic stroke by approximately 5-fold and causes an esti-ated 15% of all strokes in the United States. In people
Margaret C. Fang, MD, MPH, is the recipient of a Paul B. Beesonareer Development award in Aging and a Hartford Geriatrics Healthutcomes Research Scholars Award from the American Geriatrics Societyoundation for Health in Aging.
Requests for reprints should be addressed to Margaret C. Fang, MD,PH, University of California, San Francisco, 533 Parnassus Ave., Box
131, San Francisco, CA 94143.
l: [email protected]ront matter © 2007 Elsevier Inc. All rights reserved.ed.2007.01.026
ged 80 to 89 years, this proportion is even higher, approx-mately 24%.3 Strokes occurring from atrial fibrillation re-ult in higher mortality and disability, making the preven-ion of atrial fibrillation-related stroke an important publicealth concern. Oral anticoagulants are highly effective inreventing atrial fibrillation-related stroke,4 but because theldest patients have both the highest risk for stroke offarfarin and the highest risk for hemorrhage on warfarin,rescribing anticoagulation is a central challenge in treatingtrial fibrillation in the elderly. This review addresses theanagement of atrial fibrillation in older patients, focusing
n the risks and benefits of anticoagulation, drug therapy foreart rate and rhythm control, and invasive strategies forontrolling atrial fibrillation.
TROKE PREVENTION IN ATRIAL FIBRILLATIONultiple randomized trials have convincingly demonstrated
he efficacy of oral vitamin K antagonists such as warfarinn reducing the risk of stroke in patients with atrial fibril-
ation.4,5 Warfarin reduces stroke risk by about 68% and
ms
watna3iAlcrih
tsbAarrfrAciatt
deSadw
tttVr
hthtrcc
ac
482 The American Journal of Medicine, Vol 120, No 6, June 2007
ortality by 33%, and it also appears to prevent the mostevere types of strokes.6
Despite its efficacy, warfarin’s narrow therapeuticindow and associated hemorrhagic risks can make antico-
gulation management difficult. Optimal anticoagulation in-ensity, as measured by the Inter-ational Normalized Ratio (INR),ppears to be between 2.0 and.0.7 Low fixed-dose warfarin isneffective in preventing strokes.8
lthough clinicians may targetower INRs in older patients, it islear that INRs �2.0 significantlyaise the risk for stroke while fail-ng to protect against intracranialemorrhage.7,9
If warfarin is contraindicated,herapy with aspirin providesome protection from stroke, al-eit much less than warfarin.10
spirin reduces stroke risk bybout 21% and has fewer hemor-hagic complications than warfa-in; still, gastrointestinal side ef-ects and bleeding with aspirinemain concerns in older patients.
recent small randomized trialomparing warfarin with aspirinn patients between the ages of 80nd 90 years showed that more patients discontinued aspirinherapy compared with warfarin, mostly due to gastrointes-inal side effects.11
Investigations into the efficacy of other antithromboticrugs have thus far failed to identify alternative agentsquivalent to warfarin in terms of efficacy and safety. Thetroke Prevention using an ORal Thrombin Inhibitor intrial Fibrillation (SPORTIF) III and V trials found that theirect thrombin inhibitor ximelagatran was not inferior toarfarin, but the drug was eventually withdrawn from
CLINICAL SIGNIF
● Atrial fibrillatiorected at strokerate and rhythm
● Risk stratificatiguide the anticothough warfarin’weigh its risks.
● Rate and rhythmoutcomes; rate ctial therapy for e
● Invasive strategbrillation can imdo not obviate tbotic therapy.
Figure 1 Prevalence of atrial fibrillation by age and
he market due to an unacceptably high rate of hepaticoxicity.12,13 The warfarin arm of the recent Atrial fibrilla-ion Clopidogrel Trial with Irbesartan for prevention ofascular Events (ACTIVE-W) study found warfarin supe-
ior to combined clopidogrel plus aspirin with similar ratesof hemorrhagic complications.14
Although investigations into otherantithrombotic agents continue,for now warfarin remains the mosteffective drug to prevent stroke inatrial fibrillation.
BLEEDING COMPLICATIONSIN THE ELDERLYThe risk for warfarin-associatedhemorrhage increases withage,9,15,16 and the annual rate of sig-nificant warfarin-associated hemor-rhage ranges from 0.3% to nearly10% in various patient popula-tions.17,18 In one large cohort ofpatients with atrial fibrillation,hospitalization rates for intracra-nial and extracranial hemorrhagewere 0.47% and 0.64% per year,respectively (Figure 2).16 High an-ticoagulation intensity and INRvariability are risk factors for
emorrhage, as is the period following warfarin initia-ion.18,19 Numerous other risk factors for major hemorrhageave been identified, including prior stroke and gastrointes-inal bleeding, hypertension, concomitant aspirin use, anemia,enal insufficiency, neuropsychiatric disease, and malignan-y.9,18-20 Polypharmacy and insufficient education about anti-oagulation also have been linked to hemorrhage.21
Several risk schemes to predict hemorrhage are avail-ble,22-24 although none thus far specifically predict intra-ranial hemorrhage, the most lethal complication of warfa-
CE
nagement is di-ention and heartol.
hemes can helption decision, al-fits generally out-
ol result in similarl is usually the ini-y patients.
control atrial fi-quality of life buted for antithrom-
ICAN
n maprev
contr
on scagula
s bene
controntrolderl
ies toprovehe ne
sex in the ATRIA cohort. Data from Go et al.1
rswpr
POcaepmhtapah
bmtpdtwOafspttt
oeopffifati
cacts
SWAp
WOlhtpyitdnyfaodw
o
Ft
Faf
483Fang et al Atrial Fibrillation in the Elderly
in. One decision analysis examined the relative utilities oftroke versus gastrointestinal hemorrhage and concludedarfarin to be the preferred option for the majority ofatients with atrial fibrillation unless the annual bleedingisk exceeded 10%.25
RESCRIBING BASED ON RISK STRATIFICATIONlder patients are less likely than younger patients to re-
eive anticoagulation and are more likely to be under-nticoagulated.26,27 Although the use of warfarin in thelderly has been increasing, fewer than half of eligibleatients take warfarin.28,29 Surveys of physicians show thatajor deterrents to prescribing warfarin are high fall risk,
istory of bleeding, nonadherence, and dementia, factorshat are prevalent among the elderly.30,31 Physicians gener-lly consider major hemorrhages worse outcomes than doatients32 and may be disproportionately influenced by priordverse experiences,33 making guidelines and decision aidselpful in quantifying the risks and benefits of therapy.
A more sophisticated approach to assessing the risks andenefits of warfarin is through risk factor-guided decision-aking. Clinical risk factors for stroke include prior stroke/
ransient ischemic attack (TIA), older age, female sex, hy-ertension, left ventricular systolic dysfunction, andiabetes mellitus.4,34,35 Although imperfect, risk stratifica-ion schemes can help guide clinicians toward prescribingarfarin for patients with the highest risks for stroke.4,34,36,37
ne widely cited risk scheme is the CHADS2 score, whichssigns points according to the presence of 5 clinical riskactors: 1 point each for congestive heart failure, hyperten-ion, age �75 years, and diabetes mellitus, and 2 points forrior stroke/transient ischemic attack (Figure 3).36 For pa-ients with a CHADS2 score of 0 (annual stroke risk �1%),he absolute benefit of warfarin is not substantially greater
igure 2 Annualized rates of warfarin-associated extracranialnd intracranial hemorrhage by age in the ATRIA cohort. Adaptedrom Fang et al.16
han that for aspirin; for patients with a CHADS2 score of 2 G
r more (annual stroke risk �4%), warfarin is both cost-ffective and associated with a net improvement in qualityf life.38 In patients with a CHADS2 score of 1, it isarticularly important to gauge an individual’s preferencesor care. Because most patients over age 75 years with atrialbrillation will have a CHADS2 score of 2 or higher, itollows that the majority of patients in this age group areppropriate candidates for warfarin. For most older patients,aking warfarin does not appear to significantly reduce qual-ty of life.39,40
The type of atrial fibrillation should not affect the anti-oagulation decision. Because paroxysmal atrial fibrillationnd atrial flutter are associated with similar stroke rates ashronic atrial fibrillation, the anticoagulation decision inhese situations should be guided by the same risk factorchemes.4,41
PECIAL CONSIDERATIONS REGARDINGARFARIN MANAGEMENT IN THE ELDERLYside from the known hemorrhagic complications, warfarinoses a number of other challenges in the elderly.
arfarin Initiation and Dosinglder patients require lower doses of warfarin, are more
ikely to have INR values outside the therapeutic range, andave delayed return to therapeutic range if supratherapeu-ic.42,43 Although studies have not generally found olderatients to have significantly worse INR control thanounger patients, elderly patients are poorly representedn most studies of warfarin initiation and the most effec-ive starting dose and monitoring schedule is not wellelineated.44 One study showed that the weekly mainte-ance warfarin dose was 0.4 mg lower for every additionalear of age, and the average maintenance dose of warfarinor patients older than 70 was �5 mg.45 It is thus appropri-te to exercise greater caution in the initiation of warfarin inlder patients (particularly in women, who require a lowerose of warfarin), by using a starting dose of �5 mg dailyithout a loading dose.Studies also do not provide clear evidence regarding the
ptimal frequency of INR testing during initiation of war-
igure 3 Annual stroke risk by CHADS2 score in 1733 pa-ients with atrial fibrillation not taking warfarin. Data from
age et al.36
fAiwmmptc
CCwaaippof
FHffca3thmcwnhp
plfppa
NMApaoccbpcbd
PRraccbroidflbwbogbm
bfcwhpdrsflwsAhlfiF
484 The American Journal of Medicine, Vol 120, No 6, June 2007
arin in elders. The clinical practice guideline from themerican Geriatrics Society recommends daily INR test-
ng until stable, followed by INR testing 2 to 3 times aeek for 1 to 2 weeks, weekly testing for 1 month, andonthly testing thereafter.46 Such intensive monitoringay produce significant burden on individual patients,
articularly those with mobility impairments or lack ofransportation. In these patients, home INR monitoringould have a positive impact.
ognitive Impairmentognitive and functional impairments have been associatedith supratherapeutic INR values and poorer INR control,
nd psychosocial risk factors for nonadherence have beenssociated with worse outcomes on warfarin.47-50 Interest-ngly, in at least one study older patients were more com-liant with warfarin therapy than younger patients.51 Atresent, however, there are no guidelines addressing the usef warfarin in patients with cognitive dysfunction or severeunctional limitations.
all Riskigh fall risk is an important deterrent to warfarin use,52 but
ew studies have quantified the absolute risk associated withalls. Data are limited in part because most studies of anti-oagulation excluded patients at high fall risk. One decisionnalysis estimated that a patient would have to fall nearly00 times per year for warfarin to no longer be the preferredherapy.53 Another study of Medicare patients found thatospitalized patients with atrial fibrillation who had docu-ented high fall risk had higher subsequent rates of intra-
ranial hemorrhage.54 This study concluded that for patientsith high fall risk but low stroke risk, anticoagulation wasot likely to be beneficial (Table). Conversely, patients withigh fall risk at moderate or high risk for stroke still ap-eared to benefit from warfarin.
Finally, warfarin decreases bone mineral density in ex-erimental studies. One retrospective analysis found thatong-term warfarin use increased the risk of osteoporoticractures by 25%.55 In this study, high fall risk and neuro-sychiatric disease also were risk factors for fractures, em-hasizing the importance of close monitoring in cognitivelynd functionally impaired older adults taking warfarin.
Table Hazard Ratio of Warfarin for Composite Outcome(Out-of-Hospital Death or Hospitalization for Stroke,Myocardial Infarction, or Hemorrhage) in 1245 Patients atHigh Risk for Falls54
CHADS2 scoreHazard ratio(95% CI) P Value
RecommendedAntithromboticTherapy
0-1 0.98 (0.56-1.72) .94 Aspirin or nil2-6 0.75 (0.61-0.91) .004 Anticoagulant
s
EW APPROACHES TO ANTICOAGULATIONANAGEMENTnticoagulation clinics have been shown to improve theroportion of time patients are in therapeutic INR range,lthough data are less compelling on stroke and hemorrhageutcomes.56 Multidisciplinary approaches targeted at anti-oagulation management also may improve anticoagulationontrol and outcomes.57 For patients who have limited mo-ility or highly variable INR control, home INR testing mayrovide anticoagulation control that is at least as effective asonventional care.58,59 At present, however, poor reim-ursement for such services is a barrier to widespreadissemination.
HARMACOLOGIC RATE AND RHYTHM CONTROLandomized trials clearly demonstrate that pharmacologic
hythm control is not superior to rate control in the man-gement of atrial fibrillation, nor do anti-arrhythmic medi-ations prevent ischemic stroke or obviate the need forhronic anticoagulation.60,61 Rhythm control, in fact, haseen associated with higher mortality in older patients60;ate control, therefore, is a reasonable initial therapy forlder patients to reduce symptoms and prevent tachycardia-nduced cardiomyopathy. Beta-blockers, nondihydropyri-ine calcium channel blockers, and digoxin are the mostrequently used medications for rate control in atrial fibril-ation,29 but all may induce excess bradycardia and heartlock in elderly patients. Digoxin especially should be usedith caution due to potential drug toxicity and is generallyest suited for patients with impaired systolic function. Theptimal target heart rate is unclear, but one prominentuideline suggests a target resting heart rate of 60 to 80eats per minute and 90 to 115 beats per minute duringoderate exercise.62
Anti-arrhythmic drugs to maintain sinus rhythm may beeneficial in patients who remain symptomatic despite ef-ective rate control. However, older patients are at in-reased risk for proarrhythmia and drug interactionshen using anti-arrhythmic drugs, in part due to reducedepatic and renal clearance and increased likelihood ofolypharmacy. Although some evidence supports amio-arone as the most effective agent to maintain sinushythm, the choice of drug should depend on patientafety and coexisting medical conditions. Drugs such asecainide and propafenone increase mortality in patientsith structural heart disease,63 and sotalol and dofetilide
hould be avoided in patients with renal impairment.miodarone appears effective and safe in patients witheart failure, although careful monitoring of thyroid,iver, and pulmonary status is required. In addition, war-arin and digoxin dosages should be reduced when start-ng amiodarone due to its effect on drug plasma levels.inally, invasive strategies should be considered in
ymptomatic patients if pharmacologic therapy fails.
IFIfiiaees
pmappmalttfii
fieovtaiytfibota
Ceqmearemsamtpdcqs
robupt
CAoaficoitWnfpgoAtsa
cadbtwcutpccb
R
485Fang et al Atrial Fibrillation in the Elderly
NVASIVE TREATMENT OF ATRIALIBRILLATION IN THE ELDERLYnvasive strategies for rate and rhythm control of atrialbrillation, including AV node ablation with pacemaker
mplantation, catheter-based ablation of atrial fibrillation,nd the Cox-Maze procedure, may eventually prove moreffective than pharmacologic agents. Currently, these strat-gies should be considered for patients who remain highlyymptomatic despite medical management.
AV node ablation provides effective rate control foratients with rapid ventricular response rates refractory toedical management. The procedure involves transvenous
blation of the AV node with radiofrequency energy andlacement of a permanent pacemaker for the resulting com-lete heart block. The procedure is associated with minimalortality and morbidity. In a study of 350 patients with
trial fibrillation (mean age 68 � 11 years), AV node ab-ation with pacemaker implantation did not alter long-erm survival when compared with patients who werereated medically.64 Because it does not eliminate atrialbrillation, chronic anticoagulation continues to be
ndicated.Percutaneous radiofrequency catheter ablation of atrial
brillation is based on the observation that focal sources ofctopic beats, usually arising from the pulmonary veins,ften initiate atrial fibrillation.65 The current technique in-olves circumferential anatomical and electrical isolation ofhe pulmonary veins from the left atrium.66 Success ratespproach 75% to 85%, with complication rates of 1% to 5%n experienced centers67; however, most data are in patientsounger than 65 years without heart disease or comorbidi-ies. In general, radiofrequency ablation for chronic atrialbrillation is less successful than for paroxysmal atrial fi-rillation68 and associated with higher complication rates inlder patients with structural heart disease.69 It is not clearhat warfarin can be safely discontinued after ablation ther-py, especially in patients over 65 years of age.70
The most invasive treatment of atrial fibrillation is theox-Maze procedure, which is performed only by experi-nced cardiothoracic surgeons in selected centers and re-uires cardiopulmonary bypass. The procedure involvesaking a series of incisions in the atria, eliminating the
lectrophysiologic substrate necessary for the developmentnd propagation of atrial fibrillation, followed by atrialeconstruction. Performed since 1988, the procedure has anstablished cure rate of �90% at 10 years, a perioperativeortality of 2% to 3%, and is associated with a decrease in
troke risk to �1%.71 Sinus node dysfunction after surgerypproaches 10% and often requires implantation of a per-anent pacemaker. Because of the increase in complica-
ions that may result from prolonged cardiopulmonary by-ass time, the procedure usually has not been considered aesirable option for elderly patients. However, a more re-ent technique (Cox Maze IV), using bipolar radiofre-uency energy delivered between the jaws of 2 closely-
paced clamps, reduces the time and expertise requiredelative to the original “cut and sew” technique; one studyf 100 patients demonstrated 91% freedom from atrial fi-rillation at 1 year.72 Another study showed that, in patientsndergoing concomitant cardiac surgery and Cox Maze IV,atients �70 years old had pump times and total operationimes similar to younger patients.73
ONCLUSIONtrial fibrillation is a common medical problem affectinglder adults that is associated with significant morbidity anddverse effects on quality of life. The management of atrialbrillation is directed at preventing thromboembolism andontrolling the heart rate and rhythm. Because treatmentptions for atrial fibrillation involve potential risks andnconveniences, a clear determination of the individual pa-ient’s goals of care and preferences should be elicited.
arfarin is highly effective in preventing strokes but cog-itive dysfunction, functional impairments, and increasedall risk can complicate warfarin management in elderlyatients. The use of risk stratification schemes can helpuide the anticoagulation decision, although the net benefitutweighs the harm in most patients with atrial fibrillation.t present there are no risk stratification models that effec-
ively predict warfarin-associated intracranial hemorrhage;uch information is needed to help weigh the risks associ-ted with warfarin.
Heart rate control, generally with beta-blockers or cal-ium channel blockers, is usually considered initial therapylong with stroke prevention. If using anti-arrhythmicrugs, side effects and coexisting medical conditions shoulde considered. AV node ablation with pacemaker implan-ation improves heart rate control and quality of life, albeitithout altering long-term mortality. The safety and effi-
acy of catheter ablation of atrial fibrillation in the elderly isnproven and should be undertaken only in highly symp-omatic patients with due consideration for potential com-lications. In patients who require cardiac surgery, a con-omitant Cox-Maze procedure, particularly in experiencedenters where radiofrequency techniques are available, cane considered.
eferences1. Go AS, Hylek EM, Phillips KA, et al. Prevalence of diagnosed atrial
fibrillation in adults: national implications for rhythm management andstroke prevention: the AnTicoagulation and Risk Factors in AtrialFibrillation (ATRIA) Study. JAMA. 2001;285:2370-2375.
2. Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independentrisk factor for stroke: the Framingham Study. Stroke. 1991;22:983-988.
3. Lakshminarayan K, Solid CA, Collins AJ, Anderson DC, Herzog CA.Atrial fibrillation and stroke in the general Medicare population: a10-year perspective (1992 to 2002). Stroke. 2006;37:1969-1974.
4. Atrial Fibrillation Investigators. Risk factors for stroke and efficacy ofantithrombotic therapy in atrial fibrillation: analysis of pooled datafrom five randomized controlled trials. Arch Intern Med. 1994;154:1449-1457.
5. Hart R, Benavente O, McBride R, Pearce L. Antithrombotic therapy toprevent stroke in patients with atrial fibrillation: a meta-analysis. Ann
Intern Med. 1999;131:492-501.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
486 The American Journal of Medicine, Vol 120, No 6, June 2007
6. Hylek EM, Go AS, Chang Y, et al. Effect of intensity of oral antico-agulation on stroke severity and mortality in atrial fibrillation. N EnglJ Med. 2003;349:1019-1026.
7. Hylek E, Skates S, Sheehan M, Singer D. An analysis of the lowesteffective intensity of prophylactic anticoagulation for patients withnonrheumatic atrial fibrillation. N Engl J Med. 1996;335:540-546.
8. Stroke Prevention in Atrial Fibrillation Investigators. Patients withnonvalvular atrial fibrillation at low risk of stroke during treatmentwith aspirin. JAMA. 1998;279:1273-1277.
9. Fang MC, Chang Y, Hylek EM, et al. Advanced age, anticoagulationintensity, and risk for intracranial hemorrhage among patients takingwarfarin for atrial fibrillation. Ann Intern Med. 2004;141:745-752.
0. van Walraven C, Hart RG, Singer DE, et al. Oral anticoagulants vsaspirin in nonvalvular atrial fibrillation: an individual patient meta-analysis. JAMA. 2002;288:2441-2448.
1. Rash A, Downes T, Portner R, et al. A randomised controlled trial ofwarfarin versus aspirin for stroke prevention in octogenarians withatrial fibrillation (WASPO). Age Ageing. 2006:afl129.
2. Olsson S, Executive Steering Committee on behalf of the SPORTIF IIIInvestigators. Stroke prevention with the oral direct thrombin inhibitorximelagatran compared with warfarin in patients with non-valvularatrial fibrillation (SPORTIF III): randomised controlled trial. Lancet.2003;362:1691-1698.
3. SPORTIF Executive Steering Committee for the SPORTIF V Inves-tigators. Ximelagatran vs warfarin for stroke prevention in patientswith nonvalvular atrial fibrillation: a randomized trial. JAMA. 2005;293:690-698.
4. The ACTIVE Writing Group on behalf of the ACTIVE I. Clopidogrelplus aspirin versus oral anticoagulation for atrial fibrillation in theAtrial fibrillation Clopidogrel Trial with Irbesartan for prevention ofVascular Events (ACTIVE W): a randomised controlled trial. Lancet.2006;367:1903-1912.
5. Torn M, Bollen WL, van der Meer FJ, et al. Risks of oral anticoagulanttherapy with increasing age. Arch Intern Med. 2005;165:1527-1532.
6. Fang MC, Go AS, Hylek EM, et al. Age and the risk of warfarin-associated hemorrhage: the AnTicoagulation and Risk Factors InAtrial Fibrillation (ATRIA) Study. J Am Geriatr Soc. 2006;54:1231-1236.
7. Fanikos J, Grasso-Correnti N, Shah R, et al. Major bleeding compli-cations in a specialized anticoagulation service. Am J Cardiol. 2005;96:595-598.
8. Landefeld CS, Goldman L. Major bleeding in outpatients treated withwarfarin: incidence and prediction by factors known at the start ofoutpatient therapy. Am J Med. 1989;87:144-152.
9. Fihn SD, Callahan CM, Martin DC, et al. The risk for and severity ofbleeding complications in elderly patients treated with warfarin. AnnIntern Med. 1996;124:970-979.
0. Levine MN, Raskob G, Beyth RJ, et al. Hemorrhagic complications ofanticoagulant treatment: the Seventh ACCP Conference on Antithrom-botic and Thrombolytic Therapy. Chest. 2004;126(3 Suppl):287S-310S.
1. Kagansky N, Knobler H, Rimon E, et al. Safety of anticoagulationtherapy in well-informed older patients. Arch Intern Med. 2004;164:2044-2050.
2. Beyth R, Quinn L, Landefeld C. Prospective evaluation of an index forpredicting the risk of major bleeding in outpatients treated with war-farin. Am J Med. 1998;105:91-99.
3. Kuijer P, Hutten B, Prins M, Buller H. Prediction of the risk ofbleeding during anticoagulant treatment for venous thromboembolism.Arch Intern Med. 1999;159:457-460.
4. Gage BF, Yan Y, Milligan PE, et al. Clinical classification schemes forpredicting hemorrhage: results from the National Registry of AtrialFibrillation (NRAF). Am Heart J. 2006;151:713-719.
5. Man-Son-Hing M, Laupacis A. Balancing the risks of stroke and uppergastrointestinal tract bleeding in older patients with atrial fibrillation.
Arch Intern Med. 2002;162:541-550.6. McCrory DC, Matchar DB, Samsa G, et al. Physician attitudes aboutanticoagulation for nonvalvular atrial fibrillation in the elderly. ArchIntern Med. 1995;155:277-281.
7. Quilliam BJ, Lapane KL. Clinical correlates and drug treatment ofresidents with stroke in long-term care. Stroke. 2001;32:1385-1393.
8. McCormick D, Gurwitz JH, Goldberg RJ, et al. Prevalence and qualityof warfarin use for patients with atrial fibrillation in the long-term caresetting. Arch Intern Med. 2001;161:2458-2463.
9. Fang MC, Stafford RS, Ruskin JN, Singer DE. National trends inantiarrhythmic and antithrombotic medication use in atrial fibrillation.Arch Intern Med. 2004;164:55-60.
0. Bungard TJ, Ghali WA, Teo KK, et al. Why do patients with atrialfibrillation not receive warfarin? Arch Intern Med. 2000;160:41-46.
1. Monette J, Gurwitz JH, Rochon PA, Avorn J. Physician attitudesconcerning warfarin for stroke prevention in atrial fibrillation: resultsof a survey of long-term care practitioners. J Am Geriatr Soc. 1997;45:1060-1065.
2. Devereaux PJ, Anderson DR, Gardner MJ, et al. Differences betweenperspectives of physicians and patients on anticoagulation in patientswith atrial fibrillation: observational study. BMJ. 2001;323:1218-1222.
3. Choudhry NK, Anderson GM, Laupacis A, et al. Impact of adverseevents on prescribing warfarin in patients with atrial fibrillation:matched pair analysis. BMJ. 2006;332:141-145.
4. Hart R, Pearce L, McBride R, et al. Factors associated with ischemicstroke during aspirin therapy in atrial fibrillation: analysis of 2012participants in the SPAF I-III clinical trials. Stroke. 1999;30:1223-1229.
5. Fang MC, Singer DE, Chang Y, et al. Gender differences in the risk ofischemic stroke and peripheral embolism in atrial fibrillation: TheAnTicoagulation and Risk Factors In Atrial Fibrillation (ATRIA)Study. Circulation. 2005;112:1687-1691.
6. Gage BF, Waterman AD, Shannon W, et al. Validation of clinicalclassification schemes for predicting stroke: results from the NationalRegistry of Atrial Fibrillation. JAMA. 2001;285:2864-2870.
7. Wang TJ, Massaro JM, Levy D, et al. A risk score for predicting strokeor death in individuals with new-onset atrial fibrillation in the com-munity: the Framingham Heart Study. JAMA. 2003;290:1049-1056.
8. Gage BF, Cardinalli AB, Albers GW, Owens DK. Cost-effectivenessof warfarin and aspirin for prophylaxis of stroke in patients withnonvalvular atrial fibrillation. JAMA. 1995;274:1839-1845.
9. Lancaster T, Singer D, Sheehan M, et al. The impact of long-termwarfarin therapy on quality of life: evidence from a randomized trial.Arch Intern Med. 1991;151:1944-1949.
0. Das AK, Willcoxson PD, Corrado OJ, West RM. The impact oflong-term warfarin on the quality of life of elderly people with atrialfibrillation. Age Ageing. 2007;36:95-97.
1. Ghali WA, Wasil BI, Brant R, et al. Atrial flutter and the risk ofthromboembolism: a systematic review and meta-analysis. Am J Med.2005;118:101-107.
2. Hylek EM, Regan S, Go AS, et al. Clinical predictors of prolongeddelay in return of the international normalized ratio to within thetherapeutic range after excessive anticoagulation with warfarin. AnnIntern Med. 2001;135:393-400.
3. Perez-Gomez F, Iriarte JA, Zumalde J, et al. Antithrombotic therapy inelderly patients with atrial fibrillation: effects and bleeding complica-tions: a stratified analysis of the NASPEAF randomized trial. EurHeart J. 2006 [Epub ahead of print].
4. Copland M, Walker ID, Tait RC. Oral anticoagulation and hemor-rhagic complications in an elderly population with atrial fibrillation.Arch Intern Med. 2001;161:2125-2128.
5. Garcia D, Regan S, Crowther M, et al. Warfarin maintenance dosingpatterns in clinical practice: implications for safer anticoagulation inthe elderly population. Chest. 2005;127:2049-2056.
6. American Geriatrics Society Clinical Practice Committee. The use oforal anticoagulants (warfarin) in older people. American GeriatricsSociety guideline. J Am Geriatr Soc. 2002;50:1439-1445.
7. van Deelen BA, van den Bemt PM, Egberts TC, et al. Cognitive
impairment as determinant for sub-optimal control of oral anticoagu-
4
4
5
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
7
7
7
7
487Fang et al Atrial Fibrillation in the Elderly
lation treatment in elderly patients with atrial fibrillation. Drugs Aging.2005;22:353-360.
8. Schauer DP, Moomaw CJ, Wess M, et al. Psychosocial risk factors foradverse outcomes in patients with nonvalvular atrial fibrillation receiv-ing warfarin. J Gen Intern Med. 2005;20:1114-1119.
9. Lackie CL, Garbarino KA, Pruetz JA. Warfarin therapy for atrialfibrillation in the elderly. Ann Pharmacother. 2002;36:200-204.
0. Palareti G, Legnani C, Guazzaloca G, et al. Risks factors for highlyunstable response to oral anticoagulation: a case-control study. Br JHaematol. 2005;129:72-78.
1. Arnsten JH, Gelfand JM, Singer DE. Determinants of compliance withanticoagulation: a case-control study. Am J Med. 1997;103:11-17.
2. Hylek EM, D’Antonio J, Evans-Molina C, et al. Translating the resultsof randomized trials into clinical practice: the challenge of warfarincandidacy among hospitalized elderly patients with atrial fibrillation.Stroke. 2006;37:1075-1080.
3. Man-Son-Hing M, Nichol G, Lau A, Laupacis A. Choosing antithrom-botic therapy for elderly patients with atrial fibrillation who are at riskfor falls. Arch Intern Med. 1999;159:677-685.
4. Gage BF, Birman-Deych E, Kerzner R, et al. Incidence of intracranialhemorrhage in patients with atrial fibrillation who are prone to fall.Am J Med. 2005;118:612-617.
5. Gage BF, Birman-Deych E, Radford MJ, et al. Risk of osteoporoticfracture in elderly patients taking warfarin: results from the NationalRegistry of Atrial Fibrillation 2. Arch Intern Med. 2006;166:241-246.
6. Matchar DB, Samsa GP, Cohen SJ, et al. Improving the quality ofanticoagulation of patients with atrial fibrillation in managed careorganizations: results of the managing anticoagulation services trial.Am J Med. 2002;113:42-51.
7. Beyth R, Quinn L, Landefeld CS. A multicomponent intervention toprevent major bleeding complications in older patients receiving war-farin. Ann Intern Med. 2000;133:687-695.
8. Cromheecke ME, Levi M, Colly LP, et al. Oral anticoagulation self-management and management by a specialist anticoagulation clinic: arandomised cross-over comparison. Lancet. 2000;356:97-102.
9. Menendez-Jandula B, Souto JC, Oliver A, et al. Comparing self-management of oral anticoagulant therapy with clinic management: arandomized trial. Ann Intern Med. 2005;142:1-10.
0. AFFIRM Investigators. A comparison of rate control and rhythmcontrol in patients with atrial fibrillation. N Engl J Med. 2002;347:1825-1833.
1. Van Gelder IC, Hagens VE, Bosker HA, et al. A comparison of ratecontrol and rhythm control in patients with recurrent persistent atrial
fibrillation. N Engl J Med. 2002;347:1834-1840.2. Fuster V, Ryden LE, Cannom DS, et al. ACC/AHA/ESC 2006 Guide-lines for the Management of Patients with Atrial Fibrillation: a reportof the American College of Cardiology/American Heart AssociationTask Force on Practice Guidelines and the European Society of Car-diology Committee for Practice Guidelines (Writing Committee toRevise the 2001 Guidelines for the Management of Patients WithAtrial Fibrillation): developed in collaboration with the EuropeanHeart Rhythm Association and the Heart Rhythm Society. Circulation.2006;114:e257-e354.
3. Echt DS, Liebson PR, Mitchell LB, et al. Mortality and morbidity inpatients receiving encainide, flecainide, or placebo. The Cardiac Ar-rhythmia Suppression Trial. N Engl J Med. 1991;324:781-788.
4. Ozcan C, Jahangir A, Friedman PA, et al. Long-term survival afterablation of the atrioventricular node and implantation of a permanentpacemaker in patients with atrial fibrillation. N Engl J Med. 2001;344:1043-1051.
5. Haissaguerre M, Jais P, Shah DC, et al. Spontaneous initiation of atrialfibrillation by ectopic beats originating in the pulmonary veins. N EnglJ Med. 1998;339:659-666.
6. Wazni OM, Marrouche NF, Martin DO, et al. Radiofrequency ablationvs antiarrhythmic drugs as first-line treatment of symptomatic atrialfibrillation: a randomized trial. JAMA. 2005;293:2634-2640.
7. Pappone C, Oral H, Santinelli V, et al. Atrio-esophageal fistula as acomplication of percutaneous transcatheter ablation of atrial fibrilla-tion. Circulation. 2004;109:2724-2726.
8. Oral H, Pappone C, Chugh A, et al. Circumferential pulmonary-veinablation for chronic atrial fibrillation. N Engl J Med. 2006;354:934-941.
9. Chen MS, Marrouche NF, Khaykin Y, et al. Pulmonary vein isolationfor the treatment of atrial fibrillation in patients with impaired systolicfunction. J Am Coll Cardiol. 2004;43:1004-1009.
0. Oral H, Chugh A, Ozaydin M, et al. Risk of thromboembolic eventsafter percutaneous left atrial radiofrequency ablation of atrial fibrilla-tion. Circulation. 2006;114:759-765.
1. Ad N, Cox JL. Stroke prevention as an indication for the Mazeprocedure in the treatment of atrial fibrillation. Semin Thorac Cardio-vasc Surg. 2000;12:56-62.
2. Melby SJ, Kaiser SP, Bailey MS, et al. Surgical treatment of atrialfibrillation with bipolar radiofrequency ablation: mid-term results inone hundred consecutive patients. J Cardiovasc Surg (Torino). 2006;47:705-710.
3. Geidel S, Lass M, Ostermeyer J. Permanent atrial fibrillation ablationsurgery in patients with advanced age. Indian Pacing Electophysiol J.
2005;5:254-262.
R
HoCD
Mtcraichriittstp
DG
0d
The American Journal of Medicine (2007) 120, 488-492
EVIEW
IV-Associated Nephropathy in the Eraf Antiretroviral Therapy
hristina M. Wyatt, MD, Paul E. Klotman, MD
ivision of NephWckAsmrthdtA
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
rology, Mount Sinai School of Medicine, New York, NY.
ABSTRACT
ith improved survival in the era of antiretroviral therapy, kidney disease has emerged as an importantomplication of Human Immunodeficiency Virus (HIV) infection and antiretroviral therapy. The classicidney disease of HIV infection, HIV-associated nephropathy, occurs almost exclusively in patients offrican descent. HIV-associated nephropathy is characterized by collapsing focal segmental glomerulo-
clerosis with associated tubular dilatation and interstitial inflammation, although the histology may beore subtle in patients receiving antiretroviral therapy. Renal epithelial cells are infected by HIV-1, which
esults in epithelial cell proliferation and induction of local inflammatory pathways. Even with appropriateherapy, the kidney is a reservoir for HIV-1. Although the widespread introduction of antiretroviral therapyas had a beneficial impact on the epidemiology of HIV-associated nephropathy, the burden of kidneyisease is likely to increase as a result of antiretroviral toxicity, reduction in competing mortality risks, andhe increasing prevalence of HIV-1 infection in patients at risk for kidney disease. © 2007 Elsevier Inc.ll rights reserved.
KEYWORDS: Chronic kidney disease; HIV; HIV-associated nephropathy
c2WmsaactafincHlffic
ore than 25 years after the first published description ofhe acquired immunodeficiency syndrome (AIDS),1 chronicomplications of longstanding human immunodeficiency vi-us (HIV) infection and antiretroviral therapy have emergeds important causes of morbidity and mortality in HIV-nfected patients.2,3 Kidney disease was recognized as aomplication of HIV infection early in the epidemic4-6 butas captured the attention of the HIV community onlyecently.7 With improved survival and aging of the HIV-nfected patient population,8 the spectrum of kidney diseasen the antiretroviral era includes HIV-associated nephropa-hy, immune complex kidney disease, medication nephro-oxicity, and kidney disease related to comorbid conditionsuch as diabetes, hypertension, and hepatitis virus co-infec-ion. The current review will focus on HIV-associated ne-hropathy, the classic kidney disease of HIV infection.
Requests for reprints should be addressed to Christina M. Wyatt, MD,ivision of Nephrology, Box 1243, Mount Sinai School of Medicine, Oneustave L. Levy Place, New York, NY 10029.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2007.01.025
In the United States, there were more than 4200 newases of end-stage renal disease attributed to HIV between000 and 2004, nearly 90% of which occurred in blacks.9
ith the growing burden of HIV infection in susceptibleinority populations, the prevalence of HIV-related end-
tage renal disease is projected to increase.10 Less is knownbout the epidemiology of earlier chronic kidney diseasend acute renal failure associated with HIV infection. In aohort of more than 2000 HIV-infected women, nearly ahird had dipstick proteinuria at baseline, before the avail-bility of combination antiretroviral therapy. Over a 5-yearollow-up period, during which the majority of subjectsnitiated antiretroviral therapy, 2% developed clinically sig-ificant kidney disease as defined by a doubling of serumreatinine.11 In another study, nearly 10% of ambulatoryIV-infected patients experienced at least one episode of
aboratory-defined acute renal failure during 2 years ofollow-up.12 Both chronic kidney disease and acute renalailure are associated with significantly increased mortal-ty in patients with HIV,13,14 and kidney disease is aommon contributing cause of death in the antiretroviral
herapy era.2
CAIiurrptptarcogo
anagtsMcnbdifaHlwdcdsf
PNTessHwnkaclts
fwka
acmitmH
eeemddlsdcna
TTtwtcm
ar ris
489Wyatt and Klotman HIV-Associated Nephropathy
LINICAL COURSE AND HISTOLOGY OF HIV-SSOCIATED NEPHROPATHY
n 1984, three years after the first published report of AIDSn Los Angeles,1 physicians in New York described anique form of collapsing focal segmental glomeruloscle-osis in patients with AIDS.4 Earlyeports of HIV-associated ne-hropathy in the pre-antiretroviralherapy era were characterized byroteinuria and rapid progressiono end-stage renal disease.4,5 HIV-ssociated nephropathy and HIV-elated end-stage renal disease oc-ur almost exclusively in patientsf African descent, suggesting aenetic susceptibility to the devel-pment or progression of disease.9
The glomerular lesion in HIV-ssociated nephropathy is accompa-ied by prominent tubular dilatationnd interstitial inflammation, distin-uishing HIV-associated nephropa-hy from idiopathic collapsing focalegmental glomerulosclerosis.6,15
ore recently, similar histologichanges have been reported in HIV-egative patients receiving theisphosphonate pamidronate.16 Whereas many glomerulariseases are accompanied by tubulointerstitial fibrosis ornflammation, microcystic tubular dilatation is a uniqueeature of the collapsing focal segmental glomerulosclerosisssociated with HIV infection or pamidronate.16 In addition,IV-associated nephropathy and pamidronate-induced col-
apsing focal segmental glomerulosclerosis are associatedith podocyte proliferation, as opposed to the podocyteepletion characteristic of most glomerular diseases. Podo-yte dedifferentiation and proliferation are associated withownregulation of podocyte maturity markers and expres-ion of the cell-cycle protein Ki-67 in kidney biopsy tissuerom patients with HIV-associated nephropathy.17
ATHOGENESIS OF HIV-ASSOCIATEDEPHROPATHYhe pathogenesis of HIV-associated nephropathy has beenxtensively investigated in vitro and in animal models, withupportive evidence from epidemiologic studies and fromtudies using human tissue. The earliest animal model ofIV-associated nephropathy, an HIV-transgenic mouse,as described in 1991.18,19 The Tg26 mouse expresses aoninfectious gag/pol-deleted HIV transgene and developsidney disease indistinguishable from human HIV-associ-ted nephropathy. Although the development of HIV-asso-iated nephropathy in this model could reflect the effects ofocal or systemic transgene expression, reciprocal transplan-ation experiments confirmed that local HIV gene expres-
CLINICAL SIGNIF
● HIV-associatedmost exclusivelydescent.
● Common causesantiretroviral-tremedication toxicsepsis, and liver
● When HIV patiendisease are identprovide a nephdosages of mediexcretion, andand cardiovascul
ion in the kidney is required. In these experiments, kidneys a
rom Tg26 mice developed HIV-associated nephropathyhen transplanted into wild-type hosts, whereas wild-typeidney allografts did not develop HIV-associated nephrop-thy in Tg26 hosts.20
These data are supported by in vitro studies demonstrat-ing dedifferentiation and prolifer-ation of podocytes in response toHIV gene expression, and by thedemonstration of HIV in podo-cytes and tubular epithelial cells inhuman HIV-associated nephropa-thy.21,22 Local HIV infection ofthe kidney also may have implica-tions beyond the development ofHIV-associated nephropathy, withthe kidney serving as a potentialviral reservoir in antiretroviral-treated patients.23,24 Although re-cent data suggest that podocyte-specific expression of viral genesmay be sufficient for the develop-ment of HIV-associated nephrop-athy,25,26 it is likely that HIV in-fection of tubular epithelial cellsand lymphoid tissue also plays animportant, if synergistic, role inthe pathogenesis of HIV-associ-
ted nephropathy. Infection of human tubular epithelialells with HIV, using a pseudotyped virus, induces inflam-atory pathways that may promote the tubulointerstitial
nflammation characteristic of HIV-associated nephropa-hy.27 Tubulointerstitial disease is also the primary renalanifestation in animal models with targeted expression ofIV genes in lymphoid tissues.28
Host factors also play an important role in the pathogen-sis of HIV-associated nephropathy, which occurs almostxclusively in patients of African descent. As a cause ofnd-stage renal disease, HIV-associated nephropathy isore closely associated with black race than any other
isease except for sickle cell.29 Patients with end-stage renalisease due to HIV-associated nephropathy also are moreikely to have a family history of end-stage renal disease,uggesting a genetic predisposition to kidney disease orisease progression.30 A potential genetic susceptibility lo-us has been identified on mouse chromosome 3,31 and aumber of differentially regulated host genes and pathwaysre under investigation.32-34
REATMENT OF HIV-ASSOCIATED NEPHROPATHYreatment recommendations for HIV-associated nephropa-
hy are based on pathogenic insights and observational data,ith no evidence from rigorous clinical trials. In addition to
he intuitive benefits of antiretroviral therapy in a diseaseaused by local HIV infection, there is also strong epide-iologic evidence of a decrease in the incidence of HIV-
CE
opathy occurs al-atients of African
te renal failure inpatients include
volume depletion,se.
ith chronic kidneyphysicians should
y referral, adjusts cleared by renallly manage renalk factors.
ICAN
nephrin p
of acuatedity,disea
ts wified,rologcationcarefu
ssociated nephropathy and HIV-related end-stage renal
drpiudattakt
eislicfssvti
inwcr9etlod
MAIIlcakifnIdtcwma
dt
pcakrinpcewviprinvcvbafce
APDaiamepdaifpinis
CWttpHH
490 The American Journal of Medicine, Vol 120, No 6, June 2007
isease following the introduction of combination antiret-oviral therapy.10,35 Two well-characterized case reportsrovide further support for the use of antiretroviral therapyn the treatment of HIV-associated nephropathy, with doc-mented reversal of histologic changes and recovery ofialysis-dependent renal failure following the initiation ofntiretroviral therapy.23,36 Small cohort studies also suggesthat antiretroviral therapy improves renal survival in pa-ients with biopsy-proved HIV-associated nephropathy,37-39
nd expert guidelines for the management of HIV-relatedidney disease consider chronic kidney disease an indica-ion for antiretroviral therapy.7
Extrapolating from data in other proteinuric kidney dis-ases, the guidelines also recommend the use of ace inhib-tors or angiotensin receptor blockers,7 although evidence inupport of their utility in HIV-associated nephropathy areimited to case reports and cohort studies.40,41 Recent stud-es demonstrating that HIV induces a local inflammatoryascade in renal tubular cells provide a scientific rationaleor the use of corticosteroids,27 which also have demon-trated some benefit in uncontrolled studies.42,43 In the ab-ence of randomized clinical trials and in light of the ad-erse effects of corticosteroids, their use is generally limitedo refractory or rapidly progressive cases with prominentnterstitial inflammation.
Despite significant advances in the treatment of HIVnfection, currently available therapy for HIV-associatedephropathy appears to be only partially protective. Theidespread introduction of antiretroviral therapy was asso-
iated with an initial decrease in the incidence of HIV-elated end-stage renal disease, but there are still more than00 new cases of end-stage renal disease attributed to HIVach year in the US.9 The medical community also shouldake advantage of other opportunities to prevent HIV-re-ated kidney disease, including efforts to prevent the spreadf HIV in minority populations at increased risk for kidneyisease.
ANAGEMENT OF CHRONIC KIDNEY DISEASEND END-STAGE RENAL DISEASE IN HIV-NFECTED PATIENTSn 2005, the Infectious Diseases Society of America pub-ished expert guidelines for the management of HIV-relatedhronic kidney disease, including HIV-associated nephrop-thy.7 Although HIV-associated nephropathy is the classicidney disease of HIV infection, kidney biopsy is increas-ngly important to differentiate HIV-associated nephropathyrom other common glomerular diseases such as diabeticephropathy and hepatitis C-related glomerulonephritis.44
n addition to recommending kidney biopsy for definitiveiagnosis in patients with known chronic kidney disease,he guidelines also emphasize the need to identify earlierhronic kidney disease in patients with HIV. All patientsith HIV infection should be screened for decreased glo-erular filtration rate and proteinuria at the initial visit, with
nnual screening in patients with risk factors for kidney A
isease, such as diabetes, hypertension, hepatitis co-infec-ion, or family history of kidney disease.7
The identification of chronic kidney disease shouldrompt nephrology referral, dose adjustment of medicationsleared by renal excretion, and careful management of renalnd cardiovascular risk factors. The diagnosis of chronicidney disease also should prompt consideration of antiret-oviral therapy,7 although the role of antiretroviral therapyn kidney diseases other than HIV-associated nephropathy isot known.37 In patients with progressive kidney disease,reparation for end-stage renal disease should include dis-ussion of both peritoneal dialysis and hemodialysis,45 witharly referral for arteriovenous fistula creation in patientsho prefer hemodialysis. Kidney transplantation is also aiable alternative in patients with well-controlled HIV,46,47
deally in the context of a clinical trial. In carefully selectedatients with undetectable HIV viral load on stable antiret-oviral therapy, HIV infection has not been associated withncreased infectious complications from transplant immu-osuppression.46,47 Close collaboration between HIV pro-iders and the transplant team is essential to avoid signifi-ant drug interactions with potential implications for bothirologic control and graft function. Similar collaborationetween HIV providers and nephrologists is important forll HIV-infected patients with chronic kidney disease toacilitate definitive diagnosis and treatment, manage theomplications of chronic kidney disease, and prepare fornd-stage renal disease.
CUTE RENAL FAILURE IN HIV-INFECTEDATIENTSespite improvements in overall mortality and morbidity,
cute renal failure remains a common complication of HIVnfection in the antiretroviral era.12,14 Common causes ofcute renal failure in antiretroviral-treated patients includeedication toxicity, volume depletion, sepsis, and liver dis-
ase.12 Predictors of acute renal failure in HIV-infectedatients include traditional risk factors such as older age,iabetes, and chronic kidney disease, as well as factorsssociated with more advanced HIV or hepatitis virus co-nfection.12,14 Consistent with the impact of acute renalailure in the general population, acute renal failure in hos-italized patients with HIV infection is associated withncreased in-hospital mortality.14 In addition to this prog-ostic significance, acute renal failure should prompt med-cation dose adjustment and avoidance of additional insultsuch as hypotension and nephrotoxic agents.
ONCLUSIONSith improved survival and aging of the HIV-infected pa-
ient population in the era of combination antiretroviralherapy, kidney disease is an increasingly important com-lication of prolonged HIV infection and antiviral therapy.IV-associated nephropathy, the classic kidney disease ofIV infection, occurs almost exclusively in patients of
frican descent, creating the potential for an epidemic of
Hppkraiiibpgttapo
R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
491Wyatt and Klotman HIV-Associated Nephropathy
IV-related kidney disease among the estimated 25 millioneople living with HIV in sub-Saharan Africa.48 Early re-orts suggest a high prevalence of proteinuria and reducedidney function among antiretroviral-naïve patients in Af-ica,49-52 with the hopeful suggestion of a beneficial role forntiretroviral therapy in one large study.53 The widespreadntroduction of antiretroviral therapy has had a significantmpact on the epidemiology of HIV-associated nephropathyn industrialized nations, but it is too soon to measure thealance of risk and benefit in resource-poor settings, or toredict the long-term impact of antiretroviral therapy on thelobal burden of HIV-related kidney disease. In the mean-ime, the health care community must work collaborativelyo identify and treat HIV-related kidney disease, improveccess to HIV care and antiretroviral therapy in minorityopulations and resource-poor nations, and limit the spreadf HIV infection in susceptible patient populations.
eferences1. Centers for Disease Control (CDC). Pneumocystis pneumonia—Los
Angeles. MMWR Morb Mortal Wkly Rep. 1981;30:250-252.2. Selik RM, Byers RH Jr, Dworkin MS. Trends in diseases reported on
U.S. death certificates that mentioned HIV infection, 1987-1999. JAcquir Immune Defic Syndr. 2002;29:378-387.
3. Sackoff JE, Hanna DB, Pfeiffer MR, Torian LV. Causes of deathamong persons with AIDS in the era of highly active antiretroviraltherapy: New York City. Ann Intern Med. 2006;145:397-406.
4. Rao TK, Filippone EJ, Nicastri AD, et al. Associated focal and seg-mental glomerulosclerosis in the acquired immunodeficiency syn-drome. N Engl J Med. 1984;310:669-673.
5. Gardenschwartz MH, Lerner CW, Seligson GR, et al. Renal disease inpatients with AIDS: a clinicopathologic study. Clin Nephrol. 1984;21:197-204.
6. Pardo V, Aldana M, Colton RM, et al. Glomerular lesions in theacquired immunodeficiency syndrome. Ann Intern Med. 1984;101:429-434.
7. Gupta SK, Eustace JA, Winston JA, et al. Guidelines for the manage-ment of chronic kidney disease in HIV-infected patients. Recommen-dations of the HIV Medicine Association of the Infectious DiseasesSociety of America. Clin Infect Dis. 2005;40:1559-1585.
8. Palella FJ, Delaney K, Moorman A, et al. Declining morbidity andmortality among patients with advanced HIV infection. HIV Outpa-tient Study Investigators. N Engl J Med. 1998;338:853-860.
9. US Renal Data System (USRDS). USRDS 2004 Annual Data Report:Atlas of End-stage Renal Disease in the United States. Bethesda, MD:The National Institutes of Health, National Institute of Diabetes andDigestive and Kidney Diseases; 2004.
0. Schwartz EJ, Szczech LA, Ross MJ, et al. HAART and the epi-demic of HIV� End Stage Renal Disease. J Am Soc Nephrol.2005;16:2412-2420.
1. Szczech LA, Gange SJ, van der Horst C, et al. Predictors of proteinuriaand renal failure among women with HIV infection. Kidney Int. 2002;61:195-202.
2. Franceschini N, Napravnik S, Eron JJ, Szczech LA, Finn WF. Inci-dence and etiology of acute renal failure among ambulatory HIV-infected patients. Kidney Int. 2005;67:1526-1531.
3. Gardner LI, Klein RS, Szczech LA, et al; HIV Epidemiology ResearchStudy Group. Rates and risk factors for condition-specific hospitaliza-tions in HIV-infected and uninfected women. J Acquir Immune DeficSyndr. 2003;34:320-330.
4. Wyatt CM, Arons RR, Klotman PK, Klotman MK. Acute renal failurein hospitalized patients with HIV: risk factors and impact on in-
hospital mortality. AIDS. 2006;20:561-565.5. Laurinavicius A, Hurwitz S, Rennke HG. Collapsing glomerulopathyin HIV and non-HIV patients: a clinicopathological and follow-upstudy. Kidney Int. 1999;56:2203-2213.
6. Markowitz GS, Appel GB, Fine PL, et al. Collapsing focal segmentalglomerulosclerosis following treatment with high-dose pamidronate.J Am Soc Nephrol. 2001;12:1164-1172.
7. Barisoni L, Kriz W, Mundel P, D’Agati V. The dysregulated podocytephenotype: a novel concept in the pathogenesis of collapsing focalsegmental glomerulosclerosis and HIV-associated nephropathy. J AmSoc Nephrol. 1999;10:51-61.
8. Dickie P, Felser J, Eckhaus M, et al. HIV-associated nephropathy intransgenic mice expressing HIV-1 genes. Virology. 1991;185:109-119.
9. Kopp JB, Klotman ME, Adler SH, et al. Progressive glomeruloscle-rosis and enhanced renal accumulation of basement membrane com-ponents in mice transgenic for human immunodeficiency virus type 1genes. Proc Natl Acad Sci U S A. 1992;89:1577-1581.
0. Bruggeman LA, Dikman S, Meng C, et al. Nephropathy in humanimmunodeficiency virus-1 transgenic mice is due to renal transgeneexpression. J Clin Invest. 1997;100:84-92.
1. Bruggeman LA, Ross MD, Tanji N, et al. Renal epithelium is apreviously unrecognized site of HIV-1 infection. J Am Soc Nephrol.2000;11:2079-2087.
2. Ross MJ, Bruggeman LA, Wilson PA, Klotman PE. Microcyst forma-tion and HIV-1 gene expression occur in multiple nephron segments inHIV-associated nephropathy. J Am Soc Nephrol. 2001;12:2645-2651.
3. Winston JA, Bruggeman LA, Ross MD, et al. Nephropathy and es-tablishment of a renal reservoir of HIV type 1 during primary infec-tion. N Engl J Med. 2001;344:1979-1984.
4. Marras D, Bruggeman LA, Gao F, et al. Replication and compartmen-talization of HIV-1 in kidney epithelium of patients with HIV-associ-ated nephropathy. Nat Med. 2002;8:522-526.
5. Zhong J, Zuo Y, Ma J, et al. Expression of HIV-1 genes in podocytesalone can lead to the full spectrum of HIV-1-associated nephropathy.Kidney Int. 2005;68:1048-1060.
6. Zuo Y, Matsusaka T, Zhong J, et al. HIV-1 genes vpr and nef syner-gistically damage podocytes, leading to glomerulosclerosis. J Am SocNephrol. 2006;17:2832-2843.
7. Ross MJ, Fan C, Ross MD, et al. HIV-1 infection initiates an inflam-matory cascade in human renal tubular epithelial cells. J Acquir Im-mune Defic Syndr. 2006;42:1-11.
8. Hanna Z, Kay DG, Rebai N, Guimond A, Jothy S, Jolicoeur P. Nefharbors a major determinant of pathogenicity for an AIDS-like diseaseinduced by HIV-1 in transgenic mice. Cell. 1998;95:163-175.
9. Abbott KC, Hypolite I, Welch PG, Agodoa LY. Human immunode-ficiency virus/acquired immunodeficiency syndrome-associated ne-phropathy at end-stage renal disease in the United States: patientcharacteristics and survival in the pre highly active antiretroviral ther-apy era. J Nephrol. 2001;14:377-383.
0. Freedman BI, Soucie JM, Stone SM, Pegram S. Familial clustering ofend-stage renal disease in blacks with HIV-associated nephropathy.Am J Kidney Dis. 1999;34:254-258.
1. Gharavi AG, Ahmad T, Wong RD, et al. Mapping a locus for suscep-tibility to HIV-1-associated nephropathy to mouse chromosome 3.Proc Natl Acad Sci U S A. 2004;101:2488-2493.
2. Ross MD, Bruggeman LA, Hanss B, et al. Podocan, a novel smallleucine-rich repeat protein expressed in the sclerotic glomerular lesionof experimental HIV-associated nephropathy. J Biol Chem. 2003;278:33248-33255.
3. Kaufman L, Hayashi K, Ross MJ, Ross MD, Klotman PE. Sidekick-1is upregulated in glomeruli in HIV-associated nephropathy. J Am SocNephrol. 2004;15:1721-1730.
4. He JC, Husain M, Sunamoto M, et al. Nef stimulates proliferation ofglomerular podocytes through activation of Src-dependent Stat3 andMAPK1,2 pathways. J Clin Invest. 2004;114:643-651.
5. Lucas GM, Eustace JA, Sozio S, et al. Highly active antiretroviraltherapy and the incidence of HIV-1-associated nephropathy: a 12-year
cohort study. AIDS. 2004;18:541-516.
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
492 The American Journal of Medicine, Vol 120, No 6, June 2007
6. Wali RK, Drachenberg CI, Papadimitriou JC, Keay S, Ramos E.HIV-1-associated nephropathy and response to highly-active antiret-roviral therapy. Lancet. 1998;352:783-784.
7. Szczech LA, Gupta SK, Habash R, et al. The clinical epidemiologyand course of the spectrum of renal diseases associated with HIVinfection. Kidney Int. 2004;66:1145-1152.
8. Atta MG, Gallant JE, Rahman MH, et al. Antiretroviral therapy in thetreatment of HIV-associated nephropathy. Nephrol Dial Transplant.2006;21:2809-2813.
9. Szczech LA, Edwards LJ, Sanders LL, et al. Protease inhibitors areassociated with a slowed progression of HIV-related renal diseases.Clin Nephrol. 2002;57:336-341.
0. Kimmel PL, Mishkin GJ, Umana WO. Captopril and renal survival inpatients with human immunodeficiency virus nephropathy. Am J Kid-ney Dis. 1996;28:202-208.
1. Wei A, Burns GC, Williams BA, et al. Long-term renal survival inHIV-associated nephropathy with angiotensin-converting enzyme in-hibition. Kidney Int. 2003;64:1462-1471.
2. Smith MC, Austen JL, Carey JT, et al. Prednisone improves renalfunction and proteinuria in human immunodeficiency virus-associatednephropathy. Am J Med. 1996;101:41-48.
3. Eustace JA, Nuermberger E, Choi M, Scheel PJ, Moore R, Briggs WA.Cohort study of the treatment of severe HIV-associated nephropathywith corticosteroids. Kidney Int. 2000;58:1253-1260.
4. Atta MG, Choi MJ, Longenecker JC, et al. Nephrotic range proteinuriaand CD4 count as noninvasive indicators of HIV-associated nephrop-
athy. Am J Med. 2005;118:1288.5. Ahuja TS, Collinge N, Grady J, Khan S. Is dialysis modality a factorin survival of patients with ESRD and HIV-associated nephropathy?Am J Kidney Dis. 2003;41:1060-1064.
6. Stock PG, Roland ME, Carlson L, et al. Kidney and liver transplan-tation in human immunodeficiency virus-infected patients: a pilotsafety and efficacy study. Transplantation. 2003;76:370-375.
7. Kumar MSA, Sierka DR, Damask AM, et al. Safety and success ofkidney transplantation and concomitant immunosuppression in HIV-positive patients. Transplantation. 2005;67:1622-1629.
8. UNAIDS 2006 Global Report. Available at: http://www.unaids.org/en/HIV_data/2006GlobalReport. Accessed October 1, 2006.
9. Han TM, Naicker S, Ramdial PK, Assounga AG. A cross-sectionalstudy of HIV-seropositive patients with varying degrees of proteinuriain South Africa. Kidney Int. 2006;69:2243-2250.
0. Olson D, Thomas E, Allheimen M, et al. Impaired renal function ofpatients initiating HAART in two developing countries [Abstract]. 13th
Conference on Retroviruses and Opportunistic Infections, February 2006.1. Agbaji O, Gwanzhi N, Idoko J. Factors affecting the GFR of ARV-
naïve HIV-1-infected patients at the Jos Teaching Hospital, Nigeria[Abstract]. XVI International AIDS Conference, August 2006.
2. Mugabo JS, Lu J, Binagwaho A, Cohen M, et al. Proteinuria, hematuria,and creatinine clearance in Rwandan women with and without HIVinfection [Abstract]. XVI International AIDS Conference, August 2006.
3. Reid A, Stohr W, Walker S, et al, on behalf of the DART Trial.Glomerular dysfunction and associated risk factors following initiationof ART in adults with HIV infection in Africa [Abstract]. XVI Inter-
national AIDS Conference, August 2006.
OM
FPKa
CS
Nyhiofeonwtiqmi
pl
D1
0d
The American Journal of Medicine (2007) 120, 493-497
FFICE MANAGEMENT: GERIATRICS
ichael W. Rich, MD, Speciality Editoralls in Older Adults: Risk Assessment, Management andrevention
yle C. Moylan, MD,a Ellen F. Binder, MDb
Department of Internal Medicine, Division of General Internal Medicine, University of Missouri–Columbia School of Medicine,olumbia; bDepartment of Medicine, Division of Geriatrics and Nutritional Science, Washington University School of Medicine,
t. Louis, Mo.FeftmpiRa
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ABSTRACT
alls are a common and serious problem for older adults. This article reviews practical aspects of thevaluation and management of this disorder in the ambulatory setting. Older patients should be screenedor falls or changes in mobility as part of their annual health maintenance examination. Most falls are dueo multiple factors, including disorders of gait, balance, strength, and vision. Polypharmacy and certainedications contribute to falls in many patients and can be a remediable factor. Many falls can be
revented through individualized multicomponent interventions. Exercise programs, rehabilitation, med-cation management, and treatment of vitamin D deficiency are the most effective single interventions.eferral to a geriatrician should be considered for patients with other common geriatric syndromes, suchs cognitive impairment, incontinence, or depression. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Balance; Gait; Falls; Mobility; Polypharmacy
ala
sgccMafmpdrmbv
CA
early one third of people aged 65 years or older fall eachear, with consequences ranging from mild to severe. Onealf of these patients fall more than once. The risk of fallsncreases substantially with age, as does the risk of seri-us injury with each fall. Approximately 10% to 15% ofalls result in serious injuries, and unintentional injuries arestimated to be the fifth most common cause of death inlder adults. Many serious injuries are due to fractures, mostotably hip fractures, and up to 75% of hip fracture patientsill not regain their prior level of function. Patients hospi-
alized for a fall also are at high risk for subsequent disabil-ty and nursing home placement. Other serious conse-uences of falls include head injuries, pain, restriction ofobility, fear of recurrent falls, depression, and social
solation.Fear of falling can be a particular problem for older
atients. Patients who fall often limit their activity, whicheads to further functional decline, muscle weakness, dis-
Requests for reprints should be addressed to Kyle C. Moylan, MD,ivision of General Internal Medicine, University of Missouri-Columbia,Hospital Drive, DC043.00, Columbia, MO 65212.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.07.022
bility, and risk of further falls. These patients often fearosing their autonomy and may not volunteer informationbout restricted mobility or falls to their family or physician.
The multifactorial nature of falls cannot be overempha-ized. Successful ambulation depends on the complex inte-ration of cognitive, neuromuscular, sensory, and skeletalomponents. Each of these systems may be affected as aonsequence of degeneration or disease with advancing age.edical conditions and medications associated with falls
re listed in Tables 1 and 2. Many important risk factors foralls can be modified, such as polypharmacy, visual impair-ent, pain, proximal muscle weakness, and gait and balance
roblems. A systematic approach based on available evi-ence can effectively reduce the incidence of falls andesultant injuries, but the first step is to identify patientsost at risk. In this regard, quality indicators have recently
een established for the evaluation, management, and pre-ention of falls.1
LINICAL PRESENTATIONfall is defined as any unintentional positional change that
esults in the patient coming to rest on the ground, floor, or
ooianf
mBtmtseFbhac
frauab
ETfTft(atdmToisaaalclarmd
ppescpvseslrt
494 The American Journal of Medicine, Vol 120, No 6, June 2007
ther lower surface. Those at highest risk are likely to havether “geriatric syndromes,” such as depression, cognitivempairment, polypharmacy, orthostatic hypotension, gaitbnormalities, sensory impairment, and social isolation. Allursing home patients should be considered at high risk foralls.
Ideally, patients at risk for falls should be identified andodifiable risk factors addressed before any serious injury.ecause most falls do not result in serious injuries, practi-
ioners should specifically inquire about falls or changes inobility as part of an annual health maintenance examina-
ion. Two or more falls in 6 months, any fall resulting inerious injury, or fear of falling should trigger a completevaluation and appropriate interventions as outlined in theigure. Postfall evaluations with targeted interventions haveeen shown to reduce the risk of subsequent falls.2,3 Olderospitalized patients also should undergo a fall risk evalu-tion with a fall prevention plan as part of their dischargeare.
Because many older patients may not remember havingallen or may not report falls for other reasons (eg, embar-assment or fear of loss of autonomy), physicians should belert to other signs that may indicate falls. These includenexplained bruises (which also can be a sign of elderbuse), altered cognition, personality changes, or distur-
Table 1 Medical Conditions Associated with Falls
Neurological disorders StrokeParkinsonism and other
movement disordersGait disordersVestibular disordersDementiaDelirium
Musculoskeletal disorders OsteoarthritisJoint deformitiesKyphosisMuscle weaknessPodiatric conditions
Sensory abnormalities Visual impairmentHearing impairmentPeripheral neuropathy
Cardiovascular disease Orthostatic hypotensionSinoatrial dysfunctionArrhythmiasSyncope
Chronic medical conditions AnemiaDiabetes mellitusLung diseaseSleep disordersDepression
Medications See Table 2Polypharmacy (�4 medications)
Miscellaneous Conditions Alcohol UseRecent hospitalizationAcute medical illness
ances in gait or balance.
VALUATIONhe evaluation of a patient who has fallen or is at high risk
or falls should be targeted to identify modifiable factors.he circumstances of any fall should be elicited to help
ocus the evaluation. Associated symptoms may suggesthat the fall was a manifestation of an acute medical illnesseg, pneumonia or heart failure), as the elderly often havetypical and nonspecific clinical presentations. Falls relatedo syncope are a discrete entity that warrants an evaluationirected at potential cardiovascular or other causes. Otheredical conditions that may contribute to falls are listed inable 1. Restricted mobility related to pain, dyspnea, orther treatable conditions also may predispose to falls. Med-cations often play a role (Table 2) and can be one of theimplest risk factors to modify. Many of the medicationsssociated with falls also are identified by the Beers criterias potentially inappropriate for older adults.4 Psychoactivend sedative-hypnotic medications, particularly those with aong half-life, are most often linked with falls. Cardiovas-ular drugs also are frequently implicated, most often re-ated to volume depletion or postural hypotension. Thebsolute number of medications also correlates with theisk of falls. Some commonly used over-the-counteredications also may be implicated, for example,
iphenhydramine.A comprehensive physical examination is an essential
art of the evaluation; office staff may be trained to helperform components of the examination to maximizefficiency. Vital signs, orthostatic blood pressure mea-urement, and body weight should be obtained. A de-rease in height of 2 inches or more may indicate osteo-orosis and should trigger appropriate management. Aision examination should be performed, and referral to apecialist should be considered if vision is impaired or otherye problems are suspected. A neurological examinationhould be performed to assess muscle strength and cerebel-ar function, and to evaluate for a peripheral sensory neu-opathy. A screening test for cognitive impairment, such ashe Mini-Mental State Examination or Short Blessed Test,
Table 2 Medications Frequently Associated with Fallsby Class
Class Specific Agents
Benzodiazepines Chlordiazepoxide, diazepam,alprazolam
Antidepressants Amitriptyline, nortriptyline, fluoxetineAnitpsychotics Fluphenazine, chlorpromazine,
haloperidol, risperidoneAntiepileptics Phenytoin, phenobarbitalAnticholinergics Diphenhydramine, hyoscyamine,
tolterodine, oxybutyninSedative hypnotics All barbiturates, zolpidem, zaleplonMuscle relaxants Cyclobenzaprine, metaxalone,
methocarbamolCardiovascular
medicationsDiuretics, doxazosin, terazosin,
clonidine, digoxin
si
ct5sBscmimA“acopalt
hbtad2mtBaesa
MBvcid
proac
495Moylan and Binder Falls in the Elderly
hould be performed (see resources for health professionalsn the Appendix available online).
Proximal muscle weakness of the lower extremities, aommon and modifiable risk factor for falls, can be ascer-ained by asking the patient to stand from a seated position
times without the use of hands. Taking more than 11econds to complete the task may indicate muscle weakness.alance can be assessed with a progressive Romberg or
ingle-limb stance time. Gait should be assessed, payingareful attention for signs of Parkinsonism or other abnor-alities. More commonly, patients will have subtle changes
n gait and balance that may be evident when the patientakes a turn or is distracted by another simultaneous task.brief standardized method for gait assessment is the
get-up-and-go” test5; patients are instructed to stand fromseated position, walk 10 feet, turn around, return to the
hair, and sit down. In addition to the general informationbtained from observing the patient’s gait and mobility,atients who take more than 30 seconds to complete the testre at higher risk for falls and disability. The web resourcesisted in the Appendix (online) include multimedia websites
Figure Algorithm outlining the ap
hat demonstrate these maneuvers. p
Laboratory evaluation is performed as directed by theistory and physical examination. A comprehensive meta-olic panel (electrolytes, kidney function, and liver functionests) and complete blood count should be considered tossess for metabolic abnormalities, anemia, and nutritionaleficiencies. Vitamin D adequacy can be measured by the5-hydroxyvitamin D level. Serum thyroid-stimulating hor-one (TSH), B12 and folate levels should be obtained if
here is evidence of neuropathy or cognitive impairment.one mineral density assessment should be performed tossess for osteoporosis and increased fracture risk.6 Anlectrocardiogram is usually not necessary in the absence ofyncope or suspected cardiac cause of falls (eg, anrrhythmia).
ANAGEMENTecause most falls are multifactorial in origin, single inter-entions tend to be of limited efficacy. Conversely, usingomplex and multifaceted interventions for every patient isnefficient and wasteful. Therefore, management should beirected to the patient’s unique situation.7,8 The risk/benefit
h to fall prevention in older adults.
rofile of prescribed and over-the-counter medications
sataprmi
pWgdfrcdfaippphpsn
fnphdfcwitbaiosno
gismbstf
r
scpttp
SOpcmrttstoabnihrt
prsafspiwsqlaavb
dfgm
RRppsn
496 The American Journal of Medicine, Vol 120, No 6, June 2007
hould be weighed carefully, with the goal of reducing thebsolute number of medications, utilizing the lowest effec-ive doses, and eliminating medications that are specificallyssociated with falls. Lifestyle and nonpharmacologic ap-roaches to common geriatric problems may be helpful foreducing medication use (such as promoting sleep hygieneethods for insomnia or behavioral interventions for
ncontinence).In general, patients should be encouraged to remain
hysically active to prevent falls and functional decline.hen possible, patients should be directed to specific pro-
rams that suit their needs and preferences. Patients witheficits in strength, balance, or mobility will benefit mostrom physical and occupational therapy services. Specificehabilitation programs are often available for patients witherebrovascular disease, Parkinson’s disease, and vestibularisorders. Tai chi is an effective intervention for reducingalls,9 but its implementation may be limited by a lack ofccess and reimbursement by insurers. Women with signif-cant kyphosis may benefit from postural training; one suchrogram using a postural-training device demonstrated im-roved gait parameters and a reduction in falls and backain.10 Assistive devices, such as a cane or walker, can beelpful but, if inappropriate for the patient’s physical im-airment, they may actually increase the risk of falls. Thepecific device should be selected and fitted to the patient’seeds with the help of a physical therapist or prosthetist.
Environmental factors often interact with patient-relatedactors to induce a fall. Physicians rarely have the opportu-ity to observe the environmental hazards present in theatient’s home, where most falls occur. Many communitiesave occupational therapy services with home programsesigned to assess and modify environmental risk factorsor falls in the home such as cluttered walkways, electricalords, loose rugs, and other identifiable problems. Patientsith impaired vision may benefit from modifications to
mprove lighting and increase visibility of hazards, andhese services may be offered by local programs for thelind or visually impaired if a home fall program is notvailable. Nighttime falls may be reduced by proper light-ng, reducing nocturia, avoiding bedtime sedatives, and usef a bedside commode (if walking to the bathroom is un-afe). Patients should be encouraged to wear supportive,on-slip shoes with a low heel and to avoid wearing slippersr going barefoot.
Correcting vitamin D deficiency with high doses of er-ocalciferol has been shown to be an effective and low-riskntervention for improving balance and reducing falls inome studies. A dose of 50,000 units orally per week toaintain levels above 30 ng/mL is a reasonable strategy
ased on currently available data.11-13 Appropriate diagno-is and management of osteoporosis is recommended,6 al-hough this alone is insufficient for preventing falls andractures.
Hip pads are an attractive treatment option to reduce the
isk of hip fracture resulting from a fall. Although some itudies have shown them to be effective,14 no product isurrently available and Food and Drug Administration ap-roved for this indication. Many nursing homes have insti-uted programs using these simple devices, but many ques-ions remain regarding the optimal device and targetopulation.
PECIFIC CONCERNSlder adults who fall while taking anticoagulants pose aarticularly difficult problem for clinicians. The primaryoncern is the risk of intracranial hemorrhage, most com-only subdural hematoma. Often, older patients are at high
isk for adverse outcomes related to the underlying condi-ion (stroke from atrial fibrillation, or recurrence of venoushromboembolism). The indication for anticoagulationhould be re-assessed, as patients with prior deep venoushrombosis may have completed the recommended durationf therapy. For patients at risk for cardioembolic stroke,nticoagulation should not necessarily be discontinuedased on a single fall. The actual increased risk of intracra-ial hemorrhage while on warfarin is difficult to assess, buts frequently overestimated by clinicians. Recurrent falls oread injury resulting from a fall may unfavorably alter theisk/benefit profile associated with anticoagulation. The pa-ient should play an active role in this discussion.
Concern about falls often triggers consideration of theatient’s safety in their current living situation. Patients mayequire more supervision than is possible in the presentetting. Patients who live alone and are unable to get upfter a fall are at risk for dehydration, rhabdomyolysis, renalailure, decubitus ulcers, and death. Medical alert devicesuch as Lifeline (Philips Electronics North American Cor-oration; New York, NY) may be helpful, but patients withmpaired cognition may not be able to use them properlyhen a fall occurs. Nursing home placement is not neces-
arily the answer to falls prevention, as falls occur fre-uently in the nursing home setting. Assisted living andesser levels of care such as senior apartments often provideservice whereby patients are checked daily. These settingslso may provide access to exercise classes or other inter-entions to maintain physical activity, mobility, andalance.
Of note, the current reimbursement system for physiciansoes not provide a specific ICD-9 code for the diagnosis ofalls. Specific diagnoses related to falls may be used, such asait abnormality, muscle weakness, and adverse effect ofedications.
EFERRALeferral to a subspecialist should be considered for someatients. Geriatricians can be helpful for providing a com-rehensive assessment, particularly when other geriatricyndromes are present that contribute to falls, such as cog-itive impairment, incontinence, or depression. This often
ncludes multidisciplinary team management that integrates
pbgsc
R
1
1
1
1
1
497Moylan and Binder Falls in the Elderly
hysical therapists and social workers, and which can proveeneficial for many patients. Consultation with a neurolo-ist, cardiologist, or physical medicine and rehabilitationpecialist may be indicated based on the patient’somorbidities.
eferences1. Rubenstein LZ, Powers CM, MacLean CH. Quality indicators for the
management and prevention of falls and mobility problems in vulner-able elders. Ann Intern Med. 2001;135:686-693.
2. Becker C, Kron M, Lindermann U, et al. Effectiveness of a multifac-eted intervention on falls in nursing home residents. J Am Geriatr Soc.2003;51:306-313.
3. Tinetti ME, Baker DI, McAvay G, et al. A multifactorial interventionto reduce the risk of falling among elderly people living in the com-munity. N Engl J Med. 1994;331:821-827.
4. Fick DM, Cooper JW, Wade WE, et al. Updating the Beers criteria forpotentially inappropriate medication use in older adults: results of a USconsensus panel of experts. Arch Intern Med. 2003;163:2716-2724.
5. Podsiadlo D, Richardson S. The timed ‘Up & Go’: a test of basicfunctional mobility for frail elderly persons. J Am Geriatr Soc. 1991;
39:142-148.6. Wilkins CH, Birge SJ. Prevention of osteoporotic fractures in theelderly. Am J Med. 2005;118:1190-1195.
7. Tinetti ME. Preventing falls in elderly persons. N Engl J Med. 2003;348:42-49.
8. Chang JT, Morton SC, Rubenstein LZ, et al. Interventions for theprevention of falls in older adults: systematic review and meta-analysisof randomised clinical trials. BMJ. 2004;328:680-683.
9. Wolf SL, Barnhart HX, Kutner NG, et al. Reducing frailty and falls inolder persons: an investigation of Tai chi and computerized balancetraining. J Am Geriatr Soc. 2003;51:1794-1803.
0. Sinaki M, Brey RH, Hughes CA, et al. Significant reduction in risk offalls and back pain in osteoporotic-kyphotic women through a SpinalProprioceptive Extension Exercise Dynamic (SPEED) program. MayoClin Proc. 2005;80:849-855.
1. Bischoff-Ferrari HA, Dawson-Hughes B, Willett WC, et al. Effect ofvitamin D on falls: a meta-analysis. JAMA. 2004;291:1999-2006.
2. Dhesi JK, Moniz C, Close JC, Jackson SH, Allain TJ. A rationale forvitamin D prescribing in a falls clinic population. Age Ageing. 2002;31:267-271.
3. Latham NK, Anderson CS, Reid IR. Effects of vitamin D supplemen-tation on strength, physical performance, and falls in older persons: asystematic review. J Am Geriatr Soc. 2003;51:1219-1226.
4. Kannus P, Parkkari J, Niemi S, et al. Prevention of hip fracture inelderly people with use of a hip protector. N Engl J Med.
2000;343:1506-1513.
497.e1 The American Journal of Medicine, Vol 120, No 6, June 2007
Appendix Resources
For patientsAARP’s website provides useful patient information on fall
http://www.aarp.org/health/staying_healthy/preventionOASIS serves numerous cities across the United States wit
http://www.oasisnet.org/index.htmThe CDC toolkit includes fact sheets and brochures for pat
http://www.cdc.gov/ncipc/pub-res/toolkit/toolkit.htmThe American Geriatrics Society provides patient education
http://www.americangeriatrics.org/staging/education/f
For health professionalsThe American Geriatrics Society website provides numerou
description of the “get up and go” test.http://www.americangeriatrics.org/staging/products/pohttp://www.americangeriatrics.org/staging/education/fhttp://www.americangeriatrics.org/staging/education/0http://www.americangeriatrics.org/staging/education/0
The University of Miami’s GeriU website provides an interaEvaluation.
http://www.geriu.org/angeluploads/applications/tinettiThe University of Utah’s website includes excellent tutoria
prototypical gait abnormalities.http://medlib.med.utah.edu/neurologicexam/html/gait_
The website for the American Medical Director’s Associationursing home, including a CME program.
http://www.amda.com/tools/clinical/falls.cfmThe University of Iowa offers extensive online geriatrics e
MMSE, as well as information on falls.http://www.medicine.uiowa.edu/igec/tools/categoryMen
A copy of the Short Blessed Test can be found on the Washttp://www.strokecenter.org/trials/scales/somct.html
prevention, including additional resources and books./better_balance_prevents_falls.htmlh exercise classes and other activities for older adults.
ients (available in English or Spanish).
materials on their website.alls.shtml
s tools for managing falls, including sample office forms and a
sitionpapers/abstractPF.shtmlalls.shtml1_initial_visit.pdf2_get_up_go_test.pdfctive module that demonstrates the Tinetti Balance and Gait
/tinetti.htmls on the neurological examination, with demonstration of
abnormal.htmln (AMDA) has extensive information on addressing falls in the
ducation material, including cognitive screening tests such as the
u.asp?categoryID�1hington University Internet Stroke Center website.
OR
APM
IBaAdiaDppamtfd
A
EiTnpao
Mo
0d
The American Journal of Medicine (2007) 120, 498-502
FFICE MANAGEMENT: ALLERGY AND IMMUNOLOGY
ichard D. de Shazo, MD, Specialty Editorllergic Respiratory Disease in the Elderlyaula J. Busse, MD
ount Sinai SchAAcdo
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ool of Medicine, Department of Medicine, Division of Clinical Immunology, New York, NY.
ABSTRACT
llergic asthma and allergic rhinosinusitis are typically considered disorders of children and young adults.s a result, these diseases are frequently not considered in the differential diagnosis of nasal obstruction,
ough, wheeze, and dyspnea in older individuals. However, many older adults have long-standing allergicisorders or develop them later in life. This article addresses the complexities of diagnosis and treatment
f allergic respiratory diseases in seniors. © 2007 Elsevier Inc. All rights reserved.tAncrbd“fttrtlilcasppif
siaoiln
NTRODUCTIONy the year 2020, Americans over the age of 65 years willccount for over 20% of the population in the United States.lthough asthma, allergic rhinosinusitis and other atopiciseases are frequently seen in young people, they affectndividuals of all ages. Hospitalization and mortality ratesre highest among patients with asthma over 65 years old.1
iagnosis and treatment of allergic diseases in the olderopulation are complicated by comorbid conditions, as ap-roximately one-half the patients over the age of 75 years ofge have 3 or more diseases and take at least 3 differentedications concomitantly. Additional concerns in the
reatment of seniors are problems in compliance that stemrom physical impairments (eg, visual or motor problems,ifficulty swallowing) or the inability to afford medications.
LLERGIC RHINOSINUSITIS
tiology and diagnosis of allergic rhinosinusitisn the elderlyhe histopathologic changes associated with allergic rhi-itis are almost always present in the mucosa of thearanasal sinuses and are frequently present in the lowerirway as well. In this context, our present understandingf allergic upper airway disease is best expressed by the
Requests for reprints should be addressed to Paula J. Busse, MD, 1425adison Avenue, Division of Clinical Immunology, Mount Sinai School
f Medicine, New York, NY 10129.
s: [email protected].ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.03.024
erm allergic rhinosinusitis rather than allergic rhinitis.llergic rhinosinusitis typically presents with sneezing,asal itching, and congestion, and it is frequently asso-iated with allergic conjunctivitis. “Perennial” allergichinosinusitis is associated with year-round symptomsecause of hypersensitivity to indoor allergens (eg, pets,ust mites, cockroaches, rodents and some molds), whileseasonal” rhinosinusitis is associated with spring and/orall symptoms secondary to outdoor allergens (eg, grass,ree, weed and mold pollens). Allergen specific IgE an-ibodies play a critical role in the pathogenesis of allergichinosinusitis by triggering mast cells in the respiratoryract to produce and release inflammatory mediators,eading to symptoms characteristic of allergic rhinosinus-tis. The mast-cell mediator, histamine, constricts vascu-ar endothelium, resulting in the vascular leak and mu-osal edema that causes nasal and conjunctival swelling,nd triggers irritant receptors that produce itching andneezing. Other mediators activate and recruit eosino-hils and lymphocytes, propagating symptoms. Thus, theresence of nasal itching is the most important symptomn distinguishing allergic rhinosinusitis from otherorms.
The differential diagnosis includes vasomotor rhinitisecondary to nasal vasomotor instability triggered byrritants such as smoke, fumes, and temperature changes,nd continuous positive airway pressure (CPAP) use forbstructive sleep apnea. The condition is frequently seenn seniors who had severe allergic rhinosinusitis early inife. Atrophy of nasal tissue, associated with age, chronicasal inflammation from Wegner’s granulomatosis or
ubsequent to multiple nasal surgeries may produce the
paqcmcd
MeTtNaee1wg
tOdtHsrphpsnc
A
EAcriwoaaeaaaOp“spAaag
DHcphebrAapltefs
499Busse Asthma in the Elderly
urulent nasal discharge and sensation of a foul smellssociated with atrophic rhinitis. Nasal polyps, fre-uently associated with aspirin hypersensitivity, mayause obstructive nasal symptoms or sinusitis. Rhinitisedicamentosa results from rebound mucous membrane
ongestion associated with use of alpha-adrenergic nasalecongestants for extended periods of time.
anagement of allergic rhinosinusitis in thelderlyhe initial approach to treating allergic rhinosinusitis is
he avoidance of the offending antigens, where possible.asal corticosteroids are the first line medical therapy for
llergic rhinitis since these rarely produce serious sideffects and are effective in all forms of rhinitis. Patientducation in their use is essential for effectiveness (Table). Nasal bleeding with topical steroids usually resolvesith switch to a nasal steroid which contains propylenelycol or to a nasal antihistamine (Table 2).
If the nasal steroids are not effective as monotherapy,hey can be combined with oral or nasal antihistamines.ral antihistamines should be used carefully in the el-erly and first generation, histamine-1 (H1) receptor an-agonists should be avoided. They lack specificity for the1 receptor and produce a variety of side effects in
eniors, including anxiety, confusion, sedation, urinaryetention, constipation, disequlibrium, and postural hy-otension. Second-generation H1 receptor antagonistsave reduced ability to cross the blood brain barrier andossess a greater specificity for its receptor, reducingide-effects. We prefer to administer antihistamines byasal spray unless the patient has coexistent allergic
Table 1 Instructions for Use of Topical Nasal Steroids
1) Ask the patient to “prime” the pump unit prior to itsfirst use. (“Priming” is pushing the actuator of the pumpuntil a fine spray is produced).
2) The patient should blow his or her nose gently prior tomedication administration. If there is a large amount ofmucous present, the nose should be cleaned with asaline nose spray prior to use of the nasal steroid.
3) Place the spray tip into one nostril and bend headforward. (Warn patient to aim the tip away from thenasal septum and not to insert the tip far into the nose).
4) Close the other nostril by placing a finger on theoutside, and firmly spray the medication.
5) Repeat this step for the other nostril.6) If the instructions are for more than 1 spray into each
nostril, have the patient administer sprays to alternatingnostrils (ie, do not administer 2 sprays to 1 nostril at atime).
7) The patient should avoid blowing his or her nose for 15minutes after dosing.
onjunctivitis.
STHMA
tiology of asthma in the elderlysthma is an inflammatory airway disease producing bron-
hial hyperresponsiveness and airway obstruction, which iseversible with bronchodilators and corticosteroids. Asthman older adults may reflect persistence of childhood asthma,hich was quiescent in adulthood but returned later in life,r a form of asthma that develops later in life, “late onsetsthma.” Asthma is classified into allergic, with evidence ofllergen specific IgE by either allergy skin testing or serumvaluation and associated symptoms on exposure, or non-llergic asthma, without allergen sensitization. “Late onsetsthma” may be more severe than childhood onset asthmand may develop from different pathologic mechanisms.2
bservational studies suggest that respiratory infectionslay a more important role than atopy in the development oflate onset asthma”.3 “Late onset asthma” also may beecondary to occupational exposures, or intolerance to as-irin or other non-steroidal anti-inflammatory medications.4
lthough most asthma in seniors appears to be nonallergicnd sensitization to allergens appears to decrease with age,subpopulation of elderly patients develop late onset aller-en sensitization and respiratory symptoms.
iagnosis of asthma in the elderlyistory Although the shortness of breath, chest tightness,
ough, and wheezing that characterize asthma in youngeople are present in the elderly, mimicry by congestiveeart failure, chronic obstructive pulmonary disease, isch-mic heart disease, gastroeosphageal reflux, pulmonary em-oli, recurrent aspiration, respiratory track tumors, and la-yngeal dysfunction make the diagnosis a challenge.ngiotensin converting enzyme inhibitor-induced cough is
lso a frequent masquerader. Older individuals often haveoorer perception of airway obstructive symptoms and areess likely to report them. They may falsely attribute symp-oms to “getting older” and avoid activities, including ex-rcise, that trigger asthma symptoms. It is, therefore, criticalor the physician to ask if the patient has modified activitiesecondary to symptoms.
Table 2 Currently Available Topical Nasal Steroids andAntihistamines
Nasal Steroids Dose
Beclomethasonedipropionate
1-2 sprays each nostril BID.
Budesonide 2 sprays each nostril QD.Fluticasone propionate 2 sprays each nostril QD.Flunisolide 2 sprays each nostril BID.Mometasone furoate 2 sprays each nostril QD.Triamcinolone acetonide 2 sprays each nostril QD.
Nasal antihistamine Dose
Azelastine HCl 2 sprays each nostril BID.
Oabppeodittfifieisdcefrct
MMttcotifirh
topmmrTcttmo5ot
os
500 The American Journal of Medicine, Vol 120, No 6, June 2007
bjective measures Office spirometry performed beforend after inhalation of a bronchodilating medication is theest screening tool to make a diagnosis of asthma. Rarely,atients with acute pulmonary edema also may respond, butatients with chronic heart failure do not. Over 90% of 5200lderly study patients in one study were able to performffice spirometry.5 Although, office spirometry allows theiagnosis of asthma when a 15-20% improvement in FEV1
s noted after bronchodilator treatment, it may fail to detecthe presence of coexistent fixed obstructive airway obstruc-ion. Complete pulmonary function testing demonstratingxed airway obstruction with a low diffusion capacity con-rms the presence of chronic obstructive pulmonary dis-ase, whereas a normal or elevated level diffusion capacitys more consistent with asthma. Normal spirometry mayuggest another etiology for the patient’s symptoms, butoes not rule out the diagnosis of asthma. If there is a stronglinical history of asthma in a patient with normal spirom-try, a methacholine airway challenge test should be per-ormed to detect airways inflammation. Complications areare if the baseline FEV1 is greater than 70% predicted. Ahest x-ray, a chest CT, and an ECG are often required in
Table 3 Representative “Controller” Asthma Medicationsfor Seniors
Medications Comments
Inhaled corticosteroidsBudesonide Also available for nebulizer use
(may be easier for someelderly patients to use).
Flunisolide Also available in mint-flavor.Fluticasone propionate Available with long-acting �-
agonist (use carefully inpatients with active cardiacdisease).
Monetasone furoate Available as a twist-inhaler.Triamcinolone acetonide Requires frequent dosing (2
inhalations 3-4 times daily).Long-acting �-agonists
Formoterol fumarate Instruct patients not to use foracute symptoms; potentialcardiac issues. Best used withICS.
Salmeterol As above.Leukotriene receptor
AntagonistsMontelukast May interact with medications
that induce CYP450 andCYP2C8 enzymes; available ingranules (easy to swallow).
Zafirlukast Monitor liver function tests;potentiates warfarin (monitorPT and adjust dose); cautionwith seizure medications andthose metabolized byCYP3A4.
he initial evaluation.
anagement of asthma in elderly patientsanagement of asthma starts with an identification of
riggers and, where possible, avoidance or attention ofhem. If the patient has an allergic family background,oexistent rhinosinusitis, seasonal symptoms, or a historyf exacerbation on exposure to known allergens, referralo an allergist for further evaluation, skin testing, ormmunotherapy may be appropriate. This allows identi-cation of avoidable allergens and detection of those notevealed in the history, for instance cat, cockroach, andouse dust mite.
Patient education is facilitated when asthma medica-ions are classified into “controller” medications, takenn a daily basis to decrease airway inflammation foratients with persistent disease (Table 3), or “rescue”edications with bronchodilatory properties for the treat-ent of acute symptoms. Controllers are indicated if
escue medications are required more than twice a week.he preferred “controller” medications are inhaled corti-osteroids (ICS), which suppress the airway inflamma-ion in patients with persistent asthma, even in asymp-omatic periods. Untreated chronic airway inflammationay produce structural changes in the airways and fixed
bstruction. This is pertinent as a survey of more than00 elderly patients hospitalized for asthma showed thatnly 25% had filled their prescriptions for an ICS duringhe previous year.6
ICS may produce local side effects including hoarsenessr oral candidiasis, which can be prevented by using apacer or by rinsing the mouth after use. If patients receive
Table 4 Recommendations for Minimizing Risks in ElderlyPatients Treated with Corticosteroids
Risk Factor Recommendation
Osteoporosis/fractures Yearly bone density measurement.Supplemental oral calcium with
vitamin D.Encourage exercise.Yearly eye exams to check vision
(prevent falls).Consider bisphonate therapy.
Glaucoma Intraocular pressure testing.Cataract Slit lamp exams.Depression/anxiety Ask patients about changes in
sleeping and eating patterns.Quality of life questions
concerning socialization andhobbies.
Glucose intolerance Follow blood glucose levels forknown diabetics.
Yearly fasting blood glucose.Hypertension Measure blood-pressure with each
visit.Infection Yearly flu-shots.
Pneumovax®.Surgery Consideration of steroid
supplementation.
mhsi
saalpt
sesc�atBu
dnia
d(tdgiavcihc
RA
printed
501Busse Asthma in the Elderly
aximal doses of ICS, (generally over 800 mcg/day), theyave the potential for systemic absorption, with hypophy-eal-pituitary adrenal axis suppression and side effects sim-lar to oral corticosteroids (Table 4).
To reduce the dose of ICS or the need for oral steroids,teroid-sparing medications such as leukotriene receptorntagonists (LTRA) and long-acting beta-agonists may bedded. While currently available LTRAs are generally safe,ong acting beta-agonists should be used with caution inatients with cardiac disease. Theophylline has a narrowherapeutic window and is not recommended for seniors.
Beta 2-receptors may decrease in density with age,uggesting that �2 agonists may be less effective inlderly patients with asthma. Long-acting �2-agonistshould only be used as an add-on therapy in patients whoontinue to properly use ICS without relief. Short-acting2 agonists are relatively safe in the elderly if used on ans-needed basis to prevent exacerbations, although sys-emic absorption can produce tachycardia and tremor.oth short-acting and long-acting �2 agonists must be
Asthma Action Plan for
Doctor’s Phone Number Hospital/Emerg
GREEN ZONE: Doing Well Take These L
• No cough, wheeze, chest tightness, or Medicineshortness of breath during the day or night
• Can do usual activities
And, if a peak flow meter is used,Peak flow: more than(80% or more of my best peak flow)
My best peak flow is:
Before exercise
YELLOW ZONE: Asthma is Getting Worse
• Cough, wheeze, chest tightness, or shortnessof breath, or
• Waking at night due to asthma, or• Can do some, but not all, usual activities
-Or-
Peak flow: to (50% - 80% of my best peak flow)
RED ZONE: Medical Alert!
• Very short of breath, or• Quick-relief medicines have not helped, or• Cannot do usual activities, or• Symptoms are same or get worse after 24 hours
in Yellow Zone-Or-Peak flow: less than(50% of my best peak flow)
DANGER SIGNS• Trouble walking and talking due to shortness of breath• Lips or fingernails are blue
❒
❒
❒
A
Ta
Th✔
✔
If y
❒
❒
❒
❒
❒
First
Second
✔
✔
Figure 1 Example of Asthma Action Plan. Re
sed cautiously in patients with heart disease, as over- c
ose may cause hypokalemia and arrhythmias. An alter-ative to short-acting �2 agonist reliever medication ispratropium bromide. When switching a patient to thisgent, its efficacy should be accessed with spirometry.
Seniors with asthma should be educated about theirisease and receive a written large print “action plan”Figure 1). The “action plan” is a guide to enable patientso recognize when their asthma is worsening and what too next. Patients should be counseled to avoid nonaller-ic factors which may induce their asthma, includingrritants (cigarette smoke, household aerosols) as well asllergens. The patient’s medication list should be re-iewed to detect the addition of any medications whichould exacerbate asthma, such as �-blocker eye drops. Its helpful to explain this plan to family members or theome health professionals assisting with the patient’sare.
elationships of Rhinosinusitis to Asthmapproximately 75% of patients with asthma also have con-
Doctor’s Name Date
m Phone Number
m-Control medicines Each Day (include an anti-inflammatory)
How much to take When to take it
❒ ❒2 or 4 puffs 5 to 60 minutes before exercise
❒❒
❒
❒2 or 4 puffs, every 20 minutes for up to 1 hour
❒2 or 4 puffs or Nebulizer❒
❒
mg. per day For (3-10) days
k-Relief Medicine — and keep taking your GREEN ZONE medicine
edicine:
our doctor NOW. Go to the hospital or call for an ambulance if:till in the red zone after 15 minutes ANDnot reached your doctor.
toms (and peak flow, if used) return to GREEN ZONE after 1 hour of above treatment:
(short-acting beta2-agonist)
(short-acting beta2-agonist)
(oral steroid)
Nebulizer, once
quick-relief medicine every 4 hours for 1 to 2 days.
octor before/ within hours after taking the oral steriod.
he dose of your inhaled steroid for (7-10) days.
mptoms (and peak flow, if used) do not return to GREEN ZONE after 1 hour of above treatment:
❒4 or 6 puffs or Nebulizer❒❒
❒
mg.(short-acting beta2-agonist)
(oral steroid)
4 or 6 puffs of your quick-relief medicine ANDe hospital or call for an ambulance ( ) NOW!
with permission from www.getasthmahelp.org
ency Roo
ong-Ter
dd: Quic
ke this M
en call yYou are sYou have
our symp
Take the
Add:
Call the d
Double t-Or-If your sy
Take:
❒Take Go to th
omitant rhinosinusitis. Studies have suggested that both aller-
gvuas
AAArrtfltacw
SAig
ac
R1
2
3
4
5
6
7
8
502 The American Journal of Medicine, Vol 120, No 6, June 2007
ic and non-allergic rhinosinusitis is a risk factor for the de-elopment of late-onset asthma.7,8 This connection is notnexpected in the context of the pathophysiological discussionbove. Treatment of upper airway disease can improve asthmaymptoms and decrease lower airway hyperresponsiveness.
llergen Immunotherapy in Rhinosinusitis andsthmallergen immunotherapy (IT, “allergy shots”) should be
eserved for patients whose medical management of allergichinosinusitis or asthma fails. Although effective, IT is aime consuming procedure for the patients and is not riskree. Adverse reactions range from local irritation to anaphy-axis and death. �-blockers therapy is a relative contraindica-ion because rescue epinephrine therapy may be ineffective orssociated with acute hypertension. Immunotherapy is alsoontraindicated in asthmatics with an FEV1 less than 60%here adverse respiratory reactions may be difficult to treat.
ummaryllergic diseases including asthma and allergic rhinosinus-
tis are under-diagnosed in the elderly. Avoidance of trig-
ers where possible and a careful selection of newer anti-llergy medications produce welcome relief for theseonditions in seniors.
eferences. Sly RM. Changing asthma mortality. Ann Allergy. 1994;73(3):259-
68.. Miranda C, Busacker A, Balzar S, Trudeau J, Wenzel SE. Distinguish-
ing severe asthma phenotypes: role of age at onset and eosinophilicinflammation. J Allergy Clin Immunol. 2004;113(1):101-8.
. Bauer BA, Reed CE, Yunginger JW, Wollan PC, Silverstein MD.Incidence and outcomes of asthma in the elderly. A population-basedstudy in Rochester, Minnesota. Chest. 1997;111(2):303-10.
. Bottema RW, Reijmerink NE, Koppelman GH, Kerkhof M, Postma DS.Phenotype definition, age, and gender in the genetics of asthma andatopy. Immunol Allergy Clin North Am. 2005;25(4):621-39.
. Enright PL, Kronmal RA, Higgins M, Schenker M, Haponik EF.Spirometry reference values for women and men 65 to 85 years ofage. Cardiovascular health study. Am Rev Respir Dis. 1993;147(1):125-33.
. Hartert TV, Togias A, Mellen BG, Mitchel EF, Snowden MS, GriffinMR. Underutilization of controller and rescue medications among olderadults with asthma requiring hospital care. J Am Geriatr Soc. 2000;48(6):651-7.
. Settipane RJ, Hagy GW, Settipane GA. Long-term risk factors fordeveloping asthma and allergic rhinitis: a 23-year follow-up study ofcollege students. Allergy Proc. 1994;15(1):21-5.
. Guerra S, Sherrill DL, Martinez FD, Barbee RA. Rhinitis as an inde-pendent risk factor for adult-onset asthma. J Allergy Clin Immunol.
2002;109(3):419-25.
DC
HAa
M
PAaesaarifgst
uaemstwaevpmehrp
9
I
C
J9
0d
The American Journal of Medicine (2007) 120, 503-505
IAGNOSTIC DILEMMA: CARDIOLOGY
harles M. Wiener, MD, Section Editorigh-Speed Heartmisha J. Patel, BA,a J. Stewart Collins, MD,a Jashu R. Patel, MD,b Rajendra H. Mehta, MD, MSc
University of Michigan, Ann Arbor; bW. A. Foote Hospital, Jackson, Mich; cDuke Clinical Research Institute and Duke University
HsT
AEoIam1o
huommw
ecmMilmcSk5uws
osa
edical Center, Durham, NC.
RESENTATIONfter 3 hours of increasingly intense chest pain, our patient,50-year-old female smoker, called an ambulance. She had
xperienced 2 hours of intermittent, substernal low-inten-ity, dull aching chest pain, followed by 1 hour of severend continuous pain with radiation to the left shoulder andrm. Prehospital electrocardiogram revealed normal sinushythm and ST elevation in leads II, III, aVF, and V2-V6,ndicating inferior-lateral acute ST-elevation myocardial in-arction (STEMI). (Figure 1) Emergency medical personnelave her 4 aspirin tablets of 81 mg each, to be chewed andwallowed, and sublingual nitroglycerine 0.4 mg on the wayo the hospital.
Upon arrival at the emergency room, the patient contin-ed to complain of substernal chest pain and rated the pains an 8 on a 1- 10 scale, with 10 being the highest. Furthervaluation revealed that the patient had had a previousyocardial infarction in 1990. Given her 60-pack-year
moking habit and medical history, including type 2 diabe-es mellitus, obesity, hypertension, chronic obstructive air-ay disease, and hyperlipidemia, she was at high risk for
nother. Her home medications include fluticasone/salmet-rol inhaler, insulin, clopidogrel, potassium supplement,itamins, lovastatin, lisinopril, omeprazole, metoprolol, as-irin, and furosemide. The patient was allergic to sulfaedications. Additional presenting features included: no
xcessive caffeine intake; no history of substance abuse; noistory of premature coronary artery disease in first-degreeelatives; and an intentional 15-pound weight loss over therevious 5 months due to very restrictive dieting.
The patient’s vital signs were stable with temperature of7° F, heart rate of 70 bpm, blood pressure of 117/63 mm
Source of funding: Dr Mehta is funded by the Duke Clinical Researchnstitute, Durham, North Carolina.
Presented in part at the Michigan Chapter ACC Meeting in Traverseity, Michigan, October 6, 2006.
Requests for reprints should be addressed to Charles Wiener, MD,ohns Hopkins Department of Medicine, 1830 East Monument St, Room030, Baltimore, MD 21205.
oE-mail address: [email protected]
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.02.012
g, and respiratory rate of 19 breaths/minute with oxygenaturations of 99% on 2 liters of oxygen via nasal cannula.he physical exam was otherwise unremarkable.
SSESSMENTlectrocardiogram showed normal sinus rhythm with a ratef 70 bpm and persistent 1-2 mm of ST elevations in leadsI, III, aVF, and V2-V6. A chest x-ray failed to reveal anybnormalities. Laboratory data showed hemoglobin of 13.8g/dL, hematocrit of 41 percent, white blood cell count of
5 000/mL, serum creatinine of 0.7 mg/dL, and troponin Tf 0.36 ng/dL (normal 0.00-0.04).
In the emergency department, she received 5000 units ofeparin bolus followed by continuous infusion at a rate of 1000nits/hour, clopidogrel 600 mg loading dose, a total of 15 mgf intravenous metoprolol in 3 equal doses, 2 mg intravenousorphine sulfate, intravenous nitroglycerin drip at 20 �g/inute, and abciximab bolus and drip in a standard dose. Sheas taken for an emergent cardiac catheterization.Her cardiac catheterization revealed no significant dis-
ase in left main, left anterior descending, or left circumflexoronary arteries, but her right coronary artery had a 95%id lesion, a large thrombus burden, and Thrombolysis inyocardial Infarction (TIMI) flow grade II. After confirm-
ng that her activated clotting time was 255 seconds, theesion was successfully stented with a 3.5 x 28 mm bareetal stent. A TIMI grade III flow was restored in the right
oronary artery with complete resolution of chest pain andT elevation. Left ventriculography revealed mildly hypo-inetic inferior wall with preserved ejection fraction of5%. Her heparin and nitroglycerine drips were discontin-ed at this point, but the abciximab infusion continued. Sheas subsequently transferred to the coronary care unit with
table hemodynamics.On arrival of the intensive care unit she received 50 mg
f metoprolol and 20 mg of atorvastatin orally. The inten-ion was to continue metoprolol at 50 mg twice daily,torvastatin 20 mg at bedtime, enteric coated aspirin 325 mg
nce daily, and clopidogrel 75 mg once daily. Over the
cipWaterwmhvrfb
w2ntpicd
DT
●
●
●
●
●
●
●
●
●
ctmT(mv
MGiwc
inus rh
504 The American Journal of Medicine, Vol 120, No 6, June 2007
ourse of the next 12 hours, the patient’s heart rate graduallyncreased, but she remained comfortable without any com-laints or pain in the chest, back, abdomen, or right groin.ith the exceptions of her elevated heart rate, her temper-
ture, blood pressure, respiratory rate, and oxygen satura-ion remained within normal range. There also was novidence of orthostatic change in blood pressure or heartate. As a result, an additional 50 mg dose of metoprololas given orally; eventually, it was further increased to 100g orally twice daily. Despite the additional medication,
er heart rate rose to 128 bpm, with no changes in her otherital signs parameters, oxygen saturation of 98%, and un-emarkable physical examination. In particular, her rightemoral access site showed no evidence of any hematoma,ruit or pulsatile mass.
A repeat electrocardiogram showed sinus tachycardiaith q waves inferiorly and no other ST-T changes (Figure). An urgent hemoglobin and hematocrit revealed no sig-ificant drop in these values (12 mg/dL and 38%, respec-ively), when viewed in the context of the fact that theatient had just undergone primary percutaneous coronaryntervention. At this point, the etiology of her sinus tachy-ardia was not apparent, and a broad range of differentialiagnosis of her sinus tachycardia was entertained.
IAGNOSIShe differential diagnosis was long but unrevealing.
She had no evidence of hypovolemia or bleeding (basedon orthostatic changes in blood pressure or heart rate andstable hematocrit).With no history of substance abuse, withdrawal was un-likely as the cause of tachycardia.No medications were identified as a potential cause.Although the white blood cell count was elevated, sheremained asymptomatic and afebrile with no clinical ev-idence of an infectious process; this elevation was attrib-
Figure 1 Initial electrocardiogram showing normal s
uted to her STEMI. t
A ventilation/perfusion scan was done to rule out pulmo-nary embolism. This demonstrated low probability of apulmonary embolism.Exacerbation of chronic obstructive airway disease andcongestive heart failure, either spontaneous or secondaryto beta-blocker administration, was considered, but shehad no complains of worsening shortness of breathe; herlung exam was unremarkable without wheezing or rales;and no hypoxia was documented.The patient was given 1 mg intravenous lorazepam forpossible anxiety as a cause of the tachycardia with noimpact on her heart rate.There were no electrocardiographic changes or clinicalsymptoms or signs that suggested recurrent ischemia orpericarditis.Given her STEMI, there was concern about right ventric-ular infarction or a mechanical complication, such as leftventricular free wall rupture, mitral regurgitation, or ven-tricular septal rupture. Although these seemed unlikelybased on her timely reperfusion and physical exam, anechocardiogram was performed that did not reveal thesecomplications.
Finally, attention was directed to other non-cardiacauses of tachycardia. Laboratory results of thyroid functionests showed severely suppressed thyroid stimulating hor-one at �0.01 mIU/L, with a significantly increased free4 of 2.7 �g/dL (normal 0.7-1.8) and total T3 of 357 ng/dL
normal 60-180). Thus, a diagnosis of hyperthyroidism wasade. Her thyroid microsomal autoantibody test revealed
alues �1000 IU/mL (normal �40).
ANAGEMENTiven the iodine load with her recent heart catheterization,
odine-induced thyrotoxicosis (Jod-Basedow phenomenon)as presumed. Her hyperthyroidism and tachycardia were
ontrolled with oral propranolol, a non-specific beta-blocker
ythm and ST elevations in leads II, III, aVF, V2-V6.
hat also reduces peripheral conversion of T4 to more active
Th
mlc
snnlftieeirsCpmbTim
tht2
mpatnkd
R1
2
34
5
ia and
505Patel et al High-Speed Heart
3. She also received methimazole and a short course ofydrocortisone.
The patient’s tachycardia improved with above treat-ent, and she was discharged home 4 days later with fol-
ow-up appointments with an endocrinologist and herardiologist.
Iodine-induced hyperthyroidism is most commonly de-cribed in regions of iodine deficiency. Patients with undiag-osed Graves’ disease or thyroid autonomy due to thyroidodules may be clinically euthyroid, especially if there isimited iodine for hormone production and secretion.1 There-ore, iodine administration could lead to hyperthyroidism inhese patients. However, most areas of the US have a sufficientntake of iodine, making this syndrome extremely rare. Nev-rtheless, cases of iodine-induced thyrotoxicosis can occurven in the US, where iodinization of salt is a norm and iodinentake more than adequate. This is due to a defect in intrathy-oidal iodine metabolism or abnormalities in the body’s ownelf-protective regulation that senses iodine levels (Wolff-haikoff effect).2 In both these scenarios, iodine load canrecipitate hyperthyroidism. Graves’ disease is the most com-on cause of iodine-induced thryotoxicosis3 and is suggested
y the presence of antimicrosomal antibodies, as in this patient.he disease process is usually self limited with avoidance of
odine and can be treated in the short term with anti-thyroid
Figure 2 Subsequent electrocardiogram showing sinus tachycard
edication and beta-blockers.4,5
We believe that the fad diet that she had been on forhe past few months to intentionally lose weight mightave contributed to this phenomenon in her. Notably, herhyroid-stimulating hormone was normal 1 year prior at.98 mIU/L.
In conclusion, our case illustrates that, although not com-only encountered, cardiologists should be aware of the
otential effects of intravenous iodine-based contrast agentsnd therefore iodide load on thyroid function. Awareness ofhe potential to exacerbate or unmask previously undiag-osed hyperthyroidism should be part of a cardiologist’snowledge base where intravenous contrast use is part ofaily routine practice.
eferences. Hurley DL, Gharib H. Evaluation and management of multinodular
goiter. Otolaryngol Clin North Am. 1996;29:527-40.. Wemeau JL. Hypothyroidism related to excess iodine. Presse Med.
2002;31:1670-5.. Cooper DS. Hyperthyroidism. Lancet. 2003;362:459-68.. Bonnema SJ, Bennedbaek FN, Veje A, et al. Continuous methimazole
therapy and its effect on the cure rate of hyperthyroidism using radio-active iodine: an evaluation by a randomized trial. J Clin EndocrinolMetab. 2006;91:2946-51.
. Streetman DD, Khanderia U. Diagnosis and treatment of Graves dis-
resolution of ST segment elevations in leads II, III, aVF, V2-V6.
ease. Ann Pharmacother. 2003;37:1100-9.
DC
GRO
PFow3pilehwcASzs
AOh1awabamehimd
c2
J9
0d
The American Journal of Medicine (2007) 120, 506-508
IAGNOSTIC DILEMMA: ENDOCRINOLOGY
harles M. Wiener, MD, Section Editorrowth by Associationita R. Kalyani, MD, Anna R. Hemnes, MD
uk�ht�tmppamc
b6attTmAmbtio
mmmSpt
DTsd
sler Medical Service, Johns Hopkins Hospital, Baltimore, MD
RESENTATIONrequently, physicians must tease a diagnosis from a jumblef signs and symptoms. A 21-year-old African Americanoman presented to the emergency department with a-month history of chest discomfort. During this period,ain over her left substernal area and neck had progressivelyncreased in frequency, worsening with exertion or whenying flat. Over the preceding month, she lost 20 lbs andxperienced occasional night sweats and chills that soakeder clothes. Additional symptoms in the previous 3 weeksere loose stools, nausea and vomiting, hoarseness, alope-
ia, a mild tremor, palpitations, and shortness of breath.rm weakness made the activities of daily living difficult.he had no odynophagia, dysphagia, double-vision, or diz-iness, and she was not on any medications. She did notmoke, drink alcohol, or use illicit drugs.
SSESSMENTn physical examination, the patient was afebrile with aeart rate of 126 beats per minute, a blood pressure of38/70 mm Hg, a respiratory rate of 22 breaths per minute,nd an oxygen saturation of 97% on room air. She appearedell and in no acute distress. Palpation of her neck revealedslightly enlarged thyroid with no evidence of nodules or
ruits. She had no cervical or supraclavicular lymphadenop-thy. Her head, ears, eyes, nose, throat, lungs, heart, abdo-en, extremities, and skin were all normal. Lid lag was not
vident during a neurologic examination, but the patient didave weakness in her proximal muscle groups, particularlyn her shoulders and hips. Strength was 4 out of 5 in theseuscle groups bilaterally. She had normal reflexes with no
elayed recovery.Laboratory data revealed the following: white blood cell
ount, 7.7 � 103/mm3; hematocrit, 36.3%; platelet count,96 � 103/mm3; mean corpuscular volume, 74.7 fL; blood
Requests for reprints should be addressed to Charles Wiener, MD,ohns Hopkins Department of Medicine, 1830 East Monument St, Room030, Baltimore, MD 21205.
aE-mail address: [email protected]
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.03.015
rea nitrogen, 14 mg/dL; creatinine, 0.2 mg/dL; creatineinase, 50 U/L; creatine kinase-MB, 1 ng/mL; troponin,0.06 ng/mL; and lactate dehydrogenase, 181 U/L. A urine
uman chorionic gonadotropin test was negative. Coagula-ion studies were normal. Thyroid stimulating hormone was
0.02 �IU/mL, free thyroxine was 8.1 ng/dL, and totalriiodothyronine was 469 ng/dL. A calcium level was 11.5g/dL, and an intact parathyroid hormone level was 5
g/mL. Vitamin D 1,25-dihydroxy level was measured at 8g/mL. An anti-acetylcholinesterase antibody test was neg-tive, as were tests for anti-thyroglobulin and anti-microso-al antibodies. A test for thyroid stimulating hormone re-
eptor antibody was positive at 53.7%; normal is �10%.No abnormalities were detected on a chest radiograph,
ut a computed tomography scan with contrast disclosed a.5- � 3.5-cm anterior mediastinal mass of homogeneousttenuation that abutted the heart, a finding consistent withhymic hyperplasia (Figure 1). Also noted were a goiter inhe thyroid gland and an enlarged aortocaval lymph node.hyroid ultrasound showed a heterogeneous thyroid withinimally increased flow and no discrete nodules or cysts.n iodine-123 thyroid-uptake scan demonstrated a mini-ally enlarged thyroid gland with homogeneous uptake in
oth lobes of the thyroid gland. At 23 hours post-ingestion,hyroid uptake was measured at 67.2 %. The free thyroxinendex was high, calculated at 37. Electromyography dem-nstrated a mild nonirritable myopathy.
The differential diagnosis of the anterior mediastinalass included isolated malignant thymoma; malignant thy-oma associated with an autoimmune disease, such asyasthenia gravis, Graves’ disease, ulcerative colitis, orjogren’s syndrome; T-cell lymphoma; isolated thymic hy-erplasia; or thymic hyperplasia related to the aforemen-ioned autoimmune diseases.1,2
IAGNOSIShe patient was diagnosed with thymic hyperplasia as-ociated with Graves’ disease. A link between Graves’isease and thymic hyperplasia was recognized as early
s 1914.3,4 Thymus biopsies from patients undergoing
tppmtdhd
hBamcafimFsbciGfd
tCio
ath
mwatotflifTarfn
chmmiw
pst
Fm
507Kalyani and Hemnes Graves’ Disease-induced Thymic Hyperplasia
hyroidectomy demonstrated thymus medullary lym-hoid-follicle formation—a condition also called lym-hoid hyperplasia—in 32% of samples.5 However, thisorphology was not connected to thymic enlargement. In
rue thymic hyperplasia, x-rays or computed tomographyocument thymic enlargement.6 Since then, case reportsave commented on the relationship between thyroidisease and thymic enlargement.
A 2004 case report described a 24-year-old woman whoad hyperthyroidism complicated by thymic enlargement.7
oth thyroid dysfunction and thymic enlargement resolvedfter administration of methimazole. The regression of thy-us masses with anti-thyroid drugs also was noted in the
ase reports of a 40-year-old woman with Graves’ diseasend a 27-year-old man with hyperthyroidism.8,9 In theormer case, the presence of thyrotropin (thyroid stimulat-ng hormone) receptor was confirmed in the patient’s thy-us using reverse transcription-polymerase chain reaction.urther, sodium iodide symporter channels appear to beimilarly expressed on the thyroid and thymus, as indicatedy accumulation of radioactive iodine in normal thymusells on post-treatment thyroid ablation scans.10 An auto-mmune mechanism for thymic hyperplasia associated withraves’ disease has been suggested by researchers who
ound that immunoglobulins from patients with Graves’isease could stimulate thymocyte mitogenesis in vitro.11
The thymus normally increases in size until puberty, andhen it begins to shrink, a process called age involution.hange in thymus size has been observed following admin-
stration of exogenous hormones. In infants, initiation of
igure 1 Computed tomography of the chest with intravenousass.
ral corticosteroids has been found to cause rapid thymus r
trophy, followed by regrowth after steroids were discon-inued.12 As described in this case, thymic enlargement alsoas been recognized as a complication of hyperthyroidism.
Thyroid hormone administration can directly affect thy-us size. After thyroidectomy, subsequent loss of thymiceight has been demonstrated in rats; this was reversible by
dministration of thyroxine.13,14 A mechanism explaininghis phenomenon has been suggested by studies that dem-nstrate the effect of thyroid hormone on the synthesis ofhymus-derived secretory factor, thymulin. Hypothyroidismollowing thyroidectomy has been found to result in lowerevels of thymulin, while hyperthyroidism has shown toncrease thymulin synthesis.15 Significant correlation isound between the circulating thymulin level and levels of4 and T3. These thymulin changes were reversible withppropriate treatment in both groups.16-18 Conversely, caseeports indicate that thymectomy does not affect thyroidunction, which suggests that true thymic hyperplasia mayot be a causal factor for hyperthyroidism.19
These observations are consistent with histologicalhanges noted in thymocytes after administration of thyroidormone. Significant hyperplasia of the thymus occurs inice after treatment with triiodothyronine during the firstonth of life.18 The absolute number of mononucleate ep-
thelial cells in both the cortex and medulla of the thymusere observed to increase 2-fold in vitro.An understanding of the connection between thymic hyper-
lasia and Graves’ disease can avoid unnecessary surgery,ince treatment of hyperthyroidism can often be the definitivereatment for this reversible condition. Our patient received
t just below the aortic arch demonstrates an anterior mediastinal
contrasadioactive iodine for thyroid ablation. Seven months later, a
cdTthe
R
1
1
1
1
1
1
1
1
1
1
Fm
508 The American Journal of Medicine, Vol 120, No 6, June 2007
omputed tomography scan was repeated. It showed markedecrease in the size of the thyroid and the thymus (Figure 2).he patient’s chest pain was improved, and she was being
reated with levothyroxine for management of post-ablationypothyroidism. Subsequent follow-up has confirmed that novidence of malignancy exists.
eferences1. Tsuchiya M, Asakura H, Yoshimatsu H. Thymic abnormalities and
autoimmune diseases. Keio J Med. 1989;38:383-402.2. Levy Y, Afek A, Sherer Y, et al. Malignant thymoma associated with
autoimmune diseases: a retrospective study and review of the litera-ture. Semin Arthritis Rheum. 1998;28:73-79.
3. Halsted WS. Significance of the thymus gland in Graves’ disease. BullJohns Hopkins Hosp. 1914;25:223-234.
4. Hammar JA. Die thymus bei morbus basedow. Z Mikr Anat Forsh.1929;16:733-771.
5. Gunn A, Michie W. Biopsy of the thymus. Br J Surg. 1965;52:957-963.
6. White SR, Hall JB, Little A. An approach to mediastinal massesassociated with hyperthyroidism. Chest. 1986;90:691-693.
7. Nakamura T, Murakami M, Horiguchi H, et al. A case of thymicenlargement in hyperthyroidism in a young woman. Thyroid. 2004;14:307-310.
8. Brinkane A, Ounadi-Corbille W, Bellamy J, Leroy-Terquem E. Hy-perplasia of the thymus in Graves’ disease: a case report. Rev PneumolClin. 2004;60:239-241.
9. Kawakami M, Tokitsu K, Morita T, et al. Anterior mediastinal mass
igure 2 Seven months later, computed tomography of the chearked decrease in the size of the mass.
with hyperthyroidism; report of a case. Kyobu Geka. 2004;57:513-515.
0. Davidson J, McDougall IR. How frequently is the thymus seen onwhole-body iodine-131 diagnostic and post-treatment scans? EurJ Nucl Med. 2000;27:425-430.
1. Wortsman J, McConnachie P, Baker JR, Jr, Burman KD. Immuno-globulins that cause thymocyte proliferation from a patient withGraves’ disease and an enlarged thymus. Am J Med. 1988;85:117-121.
2. Caffey J, Silbey R. Regrowth and overgrowth of the thymus afteratrophy induced by the oral administration of adrenocorticosteroids tohuman infants. Pediatrics. 1960;26:762-770.
3. Reinhardt WO, Wainman P. Effect of thyroidectomy, castration andreplacement therapy on thymus, lymph nodes, spleen in male rats.Proc Soc Exp Biol (NY). 1942;49:257.
4. Hrinevych IuA, Bendiuh HD, Khranovs’ka NM, et al. Effect of thy-roxin and thymalin on proliferation and apoptosis of thymocytes in ratsafter thyroidectomy. Fiziol Zh. 2004;50:39-43.
5. Hrinevych IuIa, Bendiuh HD, Ostapenko OM. Effect of thymostimulinon endocrine thymus function in thyroidectomized rats during sup-pressive hormone therapy. Fiziol Zh. 2003;49:43-46.
6. Fabris N, Mocchegiani E. Endocrine control of thymic serum factorproduction in young-adult and old mice. Cell Immunol. 1985;91:325-335.
7. Fabri N, Mocchegiani E, Mariotti S, et al. Thyroid function mod-ulates thymic endocrine activity. J Clin Endocrinol Metab. 1986;62:474-478.
8. Scheiff JM, Cordier AC, Haumont S. Epithelial cell proliferation inthymic hyperplasia induced by triiodothyronine. Clin Exp Immunol.1977;27:516-521.
9. Ichiki S, Komatsu C, Ogata H, Mitsudome A. A case of myastheniagravis complicated with hyperthyroidism and thymic hyperplasia in
e level of the aortic arch—this time without contrast—showed a
st at thchildhood. Brain Dev. 1992;14:164-166.
IP
ASa
M
PFm5c(gipaa
APtspHgcn
DGobdecwuebg
C
0d
The American Journal of Medicine (2007) 120, 509-511
MAGES IN DERMATOLOGY
arwathi “Uma” Paniker, MD, Section EditorSeriously Swollen Liphomeet V. Patel, MD,a Joshua E. Lane, MDa,b,c
Division of Dermatology, Department of Internal Medicine, and bDepartment of Surgery, Mercer University School of Medicine,
c ine, TouewatlbRs
pcmwtt
sUtd
utaatietpcPm
ta
acon, Georgia; Division of Dermatology, Department of Medic
RESENTATIONor a variety of reasons, patients often delay medical treat-ent—even as an obvious medical condition progresses. A
2-year-old man presented to our outpatient dermatologylinic for evaluation of chronic swelling of his lower lipFigures 1 and 2). This was present for several years andradually enlarged. He denied any trauma to the lip ornjection of any foreign material. His medical history wasertinent for chronic alcoholism, chronic renal disease, di-betes, pancreatitis, esophageal strictures, poor dentition,nd poor oral hygiene. He denied any prior surgical history.
SSESSMENThysical examination revealed a firm, indurated and hyper-
rophic, vermilion lower lip. His speech was affected by theize of the lesion, but there was no apparent facial nervealsy and no involvement of the tongue or buccal mucosa.istologic examination revealed hyperkeratosis and hyper-ranulosis. A subdermal perivascular infiltrate of lympho-ytes and plasma cells was noted. Congo red staining wasegative for amyloid deposition.
IAGNOSISranulomatous cheilitis is a rare chronic inflammatory dis-rder. It is characterized by non-tender swelling of one oroth lips. Granulomatous cheilitis belongs to a spectrum ofiseases collectively known as orofacial granulomatosis. Thisncompasses diseases that are characterized by lesions withlinical/pathological features resembling Crohn’s disease butithout the gastrointestinal symptoms of Crohn’s.1 Gran-lomatous cheilits, Melkersson-Rosenthal syndrome, Mi-scher’s cheilitis, hypersensitivity reactions, sarcoidosis, tu-erculosis, and Crohn’s disease are included in the orofacialranulomatosis family.
Requests for reprints should be addressed to Joshua E. Lane, MD, 308oliseum Drive, Suite 200, Macon, GA 31201.
gE-mail address: [email protected]
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.01.022
he Medical College of Georgia, Augusta.
Granulomatous cheilitis is characterized by a chronicro-facial swelling involving one or both lips, with thepper lip usually the first affected. Initially the swelling ispisodic, with the first episode of edema usually subsidingithin hours or days. This often leads to a false diagnosis of
ngioedema. Unfortunately recurrent, painless attacks arehe rule with episodes increasing in duration and eventuallyeading to persistent swelling. At first, the swelling tends toe soft, but as episodes recur, the lip(s) become indurated.egression has been reported after a few years, though
pontaneous resolution rarely occurs.2
Diagnosis granulomatous cheilitis often depends uponatient’s history and clinical features, hopefully supported/onfirmed by histology.1-3 The diffuse swelling of granulo-atous cheilitis differentiates it from localized sarcoidosishich typically consists of focal nodular elements.3 Distinc-
ion between the noncaseating granulomas of granuloma-ous cheilitis and sarcoidosis may be difficult.
Granulomatous cheilitis also has been reported repre-enting extraintestinal manifestations of Crohn’s disease.nder these circumstances, the patient also manifests gas-
rointestinal or systemic signs and symptoms of Crohn’sisease or has already been diagnosed with it.2,3
The gold standard for confirming the diagnosis of gran-lomatous cheilitis is histologic analysis, although the his-ologic changes are not always specific or easily identifiable,nd biopsy is not always possible. Histologically, noncase-ting epitheloid granulomatous inflammation is characteris-ic; however, it is not present in all biopsies. Biopsy spec-mens taken during the early stages of the disease show onlydema and perivascular aggregations of lymphocytes.3 Ashe disease progresses, histology shows dilated lymphatics,erivascular lymphocytic infiltration, fibrosis, and non-ne-rotizing/noncaseating granulomas with giant cells. BothAS and Ziehl-Neelsen stain are negative for fungi andycobacteria.1-5
The causative etiology of granulomatous cheilitis andhe exact incidence and prevalence of the condition arell unknown. Many theories of causation exist, including
enetic factors, allergic reactions to food additives, chronic
ffTobutdt
libg
MRo
Fo
510 The American Journal of Medicine, Vol 120, No 6, June 2007
ocal dental infections, autoimmune mechanisms, deepungal infections, and other infectious agents, includingoxoplasma gondii, Treponema pallidum, mycobacteriar herpes simplex virus.1,4,6 Borrelia burgdorferi hadeen included in this list, but has been proven to benrelated to granulomatous cheilitis.6 A relationship be-ween granulomatous cheilitis, sarcoidosis, and Crohn’sisease also has been proposed, but the scientific litera-ure shows this to be unlikely and proposes that granu-
igure 1 A 52 year-old man presented with an enlarged and indf his tongue.
Figure 2 Lateral view of enlarged an
omatous cheilitis is an independent entity.3,4,7 Althought may occur in all ages, patients most often presentetween the second and third decade of life.2 There is noender or racial predilection.2,5
ANAGEMENTecurrent bouts of swelling, as well as an unknown etiol-gy, make granulomatous cheilitis difficult to treat. Since it
lower vermilion lip. He had no facial palsy nor any involvement
uratedd indurated lower vermilion lip.
ii
gnagcdm
macprfmittdc
nascmsgcc
nacs
R
1
1
511Patel and Lane A Seriously Swollen Lip
s an inflammatory disorder, most treatments involve anti-nflammatory regimens.8
A careful history will help identify any allergens causingranulomatous cheilitis; therefore, allergen testing may beecessary in cases with a high index of suspicion. Foodllergens, such as cinnamaldehyde, carnosine, monosodiumlutamate, cocoa, carbone, or sunset yellow, are known toause granulomatous cheilitis.5 Physical examination wouldemonstrate any responsible odontogenic foci. Their re-oval may result in improvement.2,5
Initial treatment with antihistamines is beneficial inany patients. Topical or intralesional corticosteroid ther-
py may be beneficial but only offer temporary relief. Re-alcitrant granulomatous cheilitis has been shown to im-rove with intralesional triamcinolone injections, althoughecurrence was the rule.2,9 It is useful in preventing relapseollowing surgery and should be continued for approxi-ately 2-6 months post operatively.2,10 Surgical procedures
nclude mucosal, submucosal, and tangential muscle resec-ion, crescent shaped commissuroplasty, and facial liposuc-ion are reserved for only patients with severe disfiguringisease.2,10 Systemic corticosteroids are indicated only inases of progressive lip swelling.2,10
Other alternative treatments include clofazamine, metro-idazole, penicillin, erythromycin, sulfasalazine, dapsone,nd ketotifen. These drugs help prevent the excessive use ofteroids, avoiding their long-term side effects. Hydroxy-hloroquine has been used with some success. Similarlyesalamine and sulphasalazine have been used to limited
uccess in granulomatous cheilitis based on reports relatingranulomatous cheilitis and Crohn’s disease.2 There arease reports demonstrating the resolution of granulomatous
heilitis with use of thalidomide and infliximab.8,11 To date,o particular therapy has been completely successful. Man-gement revolves around the reduction of swelling. Topicalorticosteroid has been the primary treatment in our patient–urgical treatment was offered but declined.
eferences1. Mignogna MD, Fedele S, Russo LL, Muzlo LL. The multiform and
variable patterns of onset of orofacial granulomatosis. J Oral PatholMed. 2003;32:200-205.
2. Waal RIF, Schulten EAJM, Scheur MR, et al. Cheilitis Granuloma-tosa. European Academy of Dermatology and Venereology. 2001;15:519-523.
3. Waal RIF, Schulten EAJM, Meij EH, et al. Cheilitis granulomatosa:overview of 13 patients with long-term follow-up results of manage-ment. The International Society of Dermatology. 2002;41:225-229.
4. Kavala M, Sudogan S, Can B, Sarigul S. Geanulomatous cheilitisresulting from a tuberculide. International Journal of Dermatology.2004;43:524-527.
5. Rogers RS. Granulomatous Cheilitis, Melkersson-Rosenthal Syn-drome, and Orofacial Granulomatosis. Arch Dermatol. 2000;136:1557-1558.
6. Muellegger RR, Weger W, Zoechling N, et al. Granulomatous Chei-litis and Borrelia Burgorferi. Arch Dermatol. 2000;136:1502-1506.
7. Dummer W, Lurz C, Jeschke R, et al. Granulomatous Cheilitis andCrohn’s Disease in a 3-year old boy. Pediatric Dermatology. 1999;16(1):39-42.
8. Thomas P, Walchner M, Ghoreschi K, et al. Successful treatment ofGraanulomatous Cheilitis with Tralidomide. Arch Dermatol. 2003;139:136-138.
9. Calderon RP, Gonzalo-Garijo MA, Chaves A, Argila D. Cheilitisgranulomatosa of Melkersson- Rosenthal syndrome: Treatment withintralesional corticosteroid injections. Allergologia et Immunopatho-logia. 2004;32(1):36-38.
0. Tan O, Atik B, Calka O. Plastic Surgical Solutions for Melkersson-Rosenthal Syndrome Facial Liposuction and Cheiloplasty Porcedures.Annals of Plastic Surgery. 2006;56(3):268-273.
1. Barry O, Barry J, Langan S, et al. Treatment of Granulomatous
Cheilitis with Infliximab. Arch Dermatol. 2005;141:1080-1082.
EJ
NND
PTogaitw1axia
nt(flp
AActpedpw
DTb
FE
0d
The American Journal of Medicine (2007) 120, 512-514
CG IMAGE OF THE MONTH
ulia H. Indik, MD, PhD, Section Editorow You See It; Now You Don’torman C. Wang, MD
rn Un
ttm
mrarpovsm
Vaswtspv
MIucTe
baapI1h
ivision of Cardiology, Feinberg School of Medicine, Northweste
RESENTATIONhe following case illustrates the importance of thinkingutside the box. A 92-year-old man arrived in the emer-ency department with fever, cough, shortness of breath,nd hypoxia. He had a history of coronary artery disease,schemic cardiomyopathy, atrial fibrillation and atrial flut-er, and complete heart block; a dual-chamber pacemakeras in place. Significant findings included a temperature of01° F (38.3°C), an oxygen saturation of 88% on room air,white blood cell count of 15.7 x 103 cells/mm3, and a chest-ray revealing pleural effusions and possible pulmonarynfiltrates. The patient was admitted to the telemetry unitnd treated for pneumonia.
On hospital day 2, the managing internal medicine teamoted atrial beats at a fast rate on his telemetry recordinghat, inexplicably, were not seen on a 12-lead surface ECGFigures 1 and 2). The team members, suspecting atrialutter and/or pacemaker malfunction, consulted the electro-hysiology service on hospital day 4.
SSESSMENTphysical examination of the patient uncovered a signifi-
ant tremor of his upper extremities, with a relatively stableorso. While the smaller spikes were seen on telemetry, hisacemaker was interrogated, and a real-time intracardiaclectrogram was recorded (Figures 3 and 4). The pacemakerata clearly demonstrated sinus rhythm with ventricular-aced beats tracking the patient’s native P waves (ie, theyere atrially-sensed and ventricularly paced.)
IAGNOSIShe arrhythmia seen on telemetry was an artifact generatedy the patient’s tremor. Because the right-arm electrode was
Requests for reprints should be addressed to Norman C. Wang, MD,ellow, Cardiac Electrophysiology, Northwestern Memorial Hospital, 251ast Huron Street, Feinberg 8-542, Chicago, IL 60611.
tE-mail address: [email protected].
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.03.004
iversity, Chicago, Ill.
he most distally placed on the upper arm, it was subjectedo the greatest movement. This explains why the artifact wasost pronounced in leads I, II and aVR.A diagnosis of electrocardiographic artifact could be
ade on the basis of the telemetry strip alone. The topecording showed a ventricularly-paced rhythm at a rate ofpproximately 75 beats per minute, with smaller spikes at aate of approximately 250 beats per minute; the ventricularacing spikes and the smaller spikes are dissociated fromne another. At the same time, the bottom recording isirtually free of the smaller spikes, and clear P waves can beeen preceding each ventricular-pacing spike by approxi-ately 120 milliseconds.A review of the 12-lead ECG confirms P waves in lead
1. While the presence of a fine-tremor artifact can beppreciated, the pronounced spikes seen on the telemetrytrip are not present. Pacing spikes can be difficult to seehen the leads are of bipolar configuration, as they were in
his case. Small pacing artifacts were evident precedingeveral QRS complexes. The left bundle branch block mor-hology with a superiorly directed QRS axis suggestedentricular lead placement in the right ventricular apex.1
ANAGEMENTt is essential to recognize an ECG artifact so that the truenderlying rhythm can be determined. Muscle tremorsausing ECG artifact are known to simulate atrial flutter.2
he problem can be avoided by placing the brachiallectrodes on the deltoid areas instead of the arms.
Although cardiologists and electrophysiologists areetter able to discriminate artifact than are internists, theylso can miss a significant number.3 Failing to detect anrtifact can lead to undesired clinical consequences, es-ecially when it is mistaken for ventricular arrhythmia.4
n our case, the patient had his hospital stay prolonged byday. Satisfactory improvement in his pneumonia would
ave allowed for discharge on day 3, had it not been for
he supposed cardiac abnormalities.
Fr
Ft
513Wang ECG Artifact
igure 1 (A) The top recording in this representative telemetry strip obtained on hospital day 4 is from lead II, and (B) the bottom
ecording is from a center chest lead.Figure 2 A 12-lead ECG was obtained simultaneously with the telemetry strip depicted in Figure 1.
igure 3 A standard “full disclosure” telemetry strip depicts the 7 leads recorded with 4 surface electrodes. Seen here, in order from top
o bottom, are leads I, II, III, V, aVR, aVL, and aVF.
R1
2
3
4
Frv
514 The American Journal of Medicine, Vol 120, No 6, June 2007
eferences. Wood MA, Ellenbogen KA. Temporary cardiac pacing. In: Ellenbogen
KA, Wood MA, eds. Cardiac Pacing and ICDs. 4th ed. Malden, Mass:Blackwell Publishing; 2005:163-195.
. Saint-Pierre A. ECG artifacts simulating atrial flutter. JAMA. 1973;224:
igure 4 A real-time intracardiac recording was taken from theecording is from the atrial lead; and (C) the bottom recording is fentricularly-paced beat are marked VP.
1534.
. Knight BP, Pelosi F, Michaud GF, et al. Physician interpretation ofelectrocardiographic artifact that mimics ventricular tachycardia. Am JMed. 2001;110:335-8.
. Knight BP, Pelosi F, Michaud GF, et al. Clinical consequences ofelectrocardiographic artifact mimicking ventricular tachycardia. N Engl
’s pacemaker. (A) At top is the surface recording; (B) the middlee ventricular lead. Atrially-sensed beats are designated AS, while
patientrom th
J Med. 1999;341:1270-1274.
IM
AND
PAo6cssacfittoc
ATlferrol(ttatmtflial
M
0d
The American Journal of Medicine (2007) 120, 515-517
MAGES IN RADIOLOGY
ichael Bettmann, MD, Section EditorWilde Diagnosisaseema Gangat, MD,a Paul W. Johnson, MDb
a b cheste
DOditwdmlsa
Fia
ivisions of Internal Medicine and Radiology, Mayo Clinic, Ro
RESENTATIONpparently, a highly refined demeanor can be perilous tone’s health. An 81-year-old white female presented with a-month history of worsening dyspnea and non-productiveough, which was accompanied by low-grade fevers, nightweats, and a 20-lb weight loss. Her husband mentioned thathe often tried to suppress her cough. She had never smokednd had no preexisting pulmonary disease or immunodefi-iency. Her past medical history was notable for atrialbrillation and macular degeneration. She denied any recent
ravel, tuberculosis exposure, or sick contacts. On presen-ation, she was thin with mild kyphosis, and she had a feverf 101.5° (38.6°). She had decreased breath sounds withrackles in the right, middle, and lower lung zones.
SSESSMENThe patient’s complete blood count, serum chemistries,
iver enzymes, antinuclear antibody titer, and rheumatoidactor titer were unremarkable. Her sedimentation rate waslevated at 52 mm/h, a PPD tuberculin skin test was non-eactive, and an HIV test was negative. An initial chestadiograph revealed increased interstitial markings through-ut both lungs, predominantly in the right middle lobe andingula, secondary to bronchiectasis with mucous pluggingFigures 1 and 2). High-resolution computed tomography ofhe chest demonstrated bronchiectasis and bronchial wallhickening that primarily involved the right middle lobe,long with associated micronodular infiltrates. Some infil-rates had a ”tree-in-bud” configuration, suggesting atypicalycobacterial infection (Figures 3 and 4). Pulmonary func-
ion testing pointed to a mild reduction in forced expiratoryow (FEF25-75) with normal lung volumes, findings that
ndicated obstruction. Three induced sputum samples werecid-fast positive, and Mycobacterium avium-M. intracellu-are complex was identified on cultures by DNA probe.
Requests for reprints should be addressed to Naseema Gangat, MD,ayo Clinic, 200 First Street SW, Rochester, MN 55905.
E-mail address: [email protected] p
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2007.03.014
r, Minn.
IAGNOSISur patient was diagnosed with Lady Windermere syn-rome: middle-lobe bronchiectasis with M. avium complexnfection. It was first described by Reich and Johnson inheir 1992 report of 6 elderly, immunocompetent womenith no prior smoking history or underlying pulmonaryisease who developed Mycobacterium avium complex pul-onary infection limited to the right middle lobe or
ingula.1 They hypothesized that habitual voluntary coughuppression prevented the clearing of secretions from theffected areas, leading to a focus of inflammation that pre-
igure 1 A chest radiograph showed increased interstitial mark-ngs throughout both lungs, predominantly in the right middle lobend lingula. This was secondary to bronchiectasis with mucous
lugging.
dmFs
tOttsdov
mtdamphtophHttLrb
tnoso
Fsit
Ft
516 The American Journal of Medicine, Vol 120, No 6, June 2007
isposed patients to M. avium complex infection. The ail-ent is named for the title character in Lady Windermere’san, a play by Oscar Wilde. Reich and Johnson believed theyndrome was actually the consequence of fastidiousness, a
Figure 2 The lateral view indicated similar findings.
igure 3 High-resolution computed tomography of the chesthowed bronchiectasis and bronchial-wall thickening, primarilynvolving the right middle lobe. Associated micronodular infil-
mrates were evident, and some had a ”tree-in-bud” configuration.
rait they associated with the fictitious Lady Windermere.ther researchers have described M. avium complex infec-
ion of the middle lobe or lingula in thin women withhoracic skeletal abnormalities, and they hypothesized thatkeletal abnormalities could result in ineffective cough andecreased sputum clearance.2,3 Subsequently, case reportsf this syndrome among women who have a history ofoluntary cough suppression have emerged.4-6
Chronic cough, sometimes productive, is the most com-on presenting feature, followed by constitutional symp-
oms, such as fever, fatigue, night sweats, weight loss, andyspnea. Classic radiographic findings are bronchiectasisnd small nodules that are predominantly located within theiddle lobe and lingula.7 Bronchiectasis that exclusively or
rimarily involves the right middle lobe and lingula isighly suggestive of pulmonary M. avium complex infec-ion, even in the face of negative sputum cultures.7 Aboutne-third of patients with thin-section computed tomogra-hy findings of bilateral bronchiectasis and bronchiolitisave nontuberculous mycobacterial pulmonary infection.8
owever, in the absence of specific identifying features inhe history, physical examination, and chest radiography,he isolation of M. avium complex is essential for diagnosis.ady Windermere syndrome has such explicit clinical and
adiographic features, that it can generally be recognizedefore the organism is detected.
A 1997 statement by the American Thoracic Society onhe diagnosis and treatment of M. avium complex pulmo-ary disease recommended that a minimum of 3 sputum orther respiratory specimens be stained and cultured inymptomatic patients presenting with infiltrative, nodular,r cavitary lung disease.9 It is often difficult to achieve a
igure 4 The lingula, viewed with high-resolution computedomography, was similarly involved.
icrobiological diagnosis in patients with Lady Wind-
esmaw
MMitsbactsps5
pllelssa
caca
rcita
R
1
1
517Gangat and Johnson Wilde Diagnosis
rmere syndrome, a dilemma ascribed to the habit of coughuppression. Therefore, a bronchoscopy is often needed toeet the American Thoracic Society’s diagnostic criteria
nd to monitor therapeutic response, particularly in patientsho do not produce M. avium complex in their sputum.10
ANAGEMENTanagement of patients with Lady Windermere syndrome
s problematic, complicated by significant adverse effects ofhe required medication and an unpredictable treatment re-ponse. The American Thoracic Society recommends com-ining a macrolide antibiotic (either clarithromycin orzithromycin), rifampin or rifabutin, and ethambutol, andontinuing treatment with the regimen for 1 year after spu-um has converted to negative.9 To date, there have been notudies evaluating the efficacy of these drugs exclusively inatients with Lady Windermere syndrome. However, atudy reported a 50% failure rate after 1 year of therapy;8% of patients were unable to tolerate their first regimen.11
The role of surgery is unclear. In a study of 13 femaleatients with M. avium complex infection involving theingula and middle lobe, surgical resection of the diseasedobe was associated either with cure or with “far fewer”pisodes when compared with preoperative illness.3 Thoughong-term outcomes of surgery and the necessity for post-urgical antibiotics are unclear, the authors recommendedurgery soon after diagnosis, followed by 12-24 months ofntibiotics.
Since our patient was symptomatic, we treated her withlarithromycin, 750 mg daily; moxifloxacin, 400 mg daily;nd rifampin, 600 mg daily. Ethambutol was avoided be-ause of the patient’s macular degeneration. Two months
fter treatment was initiated, her clinical symptoms hadesolved, the sputum acid-fast stain and mycobacterialultures became negative, and chest radiography showedmprovement in the bronchiectasis and infiltrates. Sheolerated the treatment well, completing 14 months ofntimicrobial therapy.
eferences1. Reich JM, Johnson RE. Mycobacterium avium complex pulmonary
disease presenting as an isolated lingular or middle lobe pattern. TheLady Windermere syndrome. Chest. 1992;101:1605-1609.
2. Iseman MD, Buschman DL, Ackerson LM. Pectus excavatum andscoliosis. Thoracic anomalies associated with pulmonary diseasecaused by Mycobacterium avium complex. Am Rev Respir Dis. 1991;144:914-916.
3. Pomerantz M, Denton JR, Huitt GA, et al. Resection of the rightmiddle lobe and lingula for mycobacterial infection. Ann Thorac Surg.1996;62:990-993.
4. Byrd RP Jr, Mirza S, Roy TM. Lady Windermere’s syndrome: anatypical presentation of an atypical mycobacterium. J Tenn Med Assoc.1995;88:303-305.
5. Dhillon SS, Watanakunakorn C. Lady Windermere syndrome: middlelobe bronchiectasis and Mycobacterium avium complex infection dueto voluntary cough suppression. Clin Infect Dis. 2000;30:572-575.
6. Byrd R Jr, Payne JL, Roy TM. Lingular and middle lobe infiltrates inan elderly woman. Chest. 1995;108:1156-1157.
7. Levin DL. Radiology of pulmonary Mycobacterium avium-intracellu-lare complex. Clin Chest Med. 2002;23:603-612.
8. Koh WJ, Lee KS, Kwon OJ, et al. Bilateral bronchiectasis and bron-chiolitis at thin-section CT: diagnostic implications in nontuberculousmycobacterial pulmonary infection. Radiology. 2005;235:282-288.
9. American Thoracic Society. Diagnosis and treatment of disease causedby nontuberculous mycobacteria. Am J Respir Crit Care Med. 1997;156:S1-25.
0. Chalermskulrat W, Gilbey JG, Donohue JF. Nontuberculous myco-bacteria in women, young and old. Clin Chest Med. 2002;23:675-686.
1. Huang JH, Kao PN, Adi V, Ruoss SJ. Mycobacterium avium-intracel-lulare pulmonary infection in HIV-negative patients without preexist-ing lung disease: diagnostic and management limitations. Chest. 1999;
115:1033-1040.
C
BAPC
Oottspausbc
CM
0d
The American Journal of Medicine (2007) 120, 518-524
LINICAL RESEARCH STUDY
ody Mass, Fitness and Survival in Veteran Patients:nother Obesity Paradox?
aul McAuley, PhD, Jonathan Myers, PhD, Joshua Abella, MD, Victor Froelicher, MD
ardiology DivisPvTvMpnTpRaw(wttCmf
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ion, VA Palo Alto Health Care System/Stanford University, Palo Alto, Calif.
ABSTRACT
URPOSE: The paradox of obesity in patients with heart failure (HF) also has been observed in non-HFeteran patients. Veterans had to have met military fitness requirements at the time of their enlistment.herefore, we assessed the relation of body mass index (BMI) to mortality in a clinical cohort of non-HFeterans, adjusting for fitness.ETHODS: After excluding HF patients (n � 580), the study population comprised 6876 consecutiveatients (mean age 58 [�11] years) referred for exercise testing. Patients were classified by BMI category:ormal weight (BMI 18.5-24.9 kg/m2), overweight (BMI 25.0-29.9 kg/m2), or obese (BMI �30.0 kg/m2).he association between BMI, fitness, other clinical variables, and all-cause mortality was assessed by Coxroportional hazards analysis.ESULTS: During a mean (�SD) follow-up of 7.5 � 4.5 years, a total of 1571 (23%) patients died. Inmultivariate analysis including clinical, risk factor, and exercise test data, higher BMI was associatedith better survival. Expressing the data by BMI category, obese patients were 22% less likely to die
relative risk [RR] � 0.78, 95% confidence interval [CI], 0.69-0.90, P �.001) than patients of normaleight. After further adjustment for cardiorespiratory fitness (CRF), this relationship strengthened such
hat mortality risk for the obese category was 35% lower (RR � 0.65, 95% CI, 0.57-0.76, P �.001), versushe normal weight category.ONCLUSIONS: As has been observed in HF patients, obesity was associated with a substantially lowerortality risk in a clinical population of non-HF veterans. Higher CRF and obesity in later life may account
or an obesity paradox in this population. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Cardiorespiratory fitness; Exercise testing; Mortality; Obesity
sn
psrcstaco
t
besity is associated with an increased risk of adverseutcomes in the general population.1 Paradoxically, in pa-ients with heart failure (HF), an inverse relationship be-ween body mass index (BMI) and mortality has been ob-erved,2-4 including our population of veterans.5 A similarhenomenon has been reported in dialysis patients,6 and tolesser degree in acute myocardial infarction,7 and patientsndergoing coronary artery interventions.8-10 Some havepeculated that such an “obesity paradox” is little more thanasic protection against the wasting from HF and otherachectic disease states.11 Others have suggested that obe-
Requests for reprints should be addressed to Jonathan Myers, PhD,ardiology Division, 111C, VA Palo Alto Health Care System, 3801iranda Ave., Palo Alto CA 94304.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.07.032
ity is protective because it is associated with larger coro-ary arteries10 or increased muscle mass.12
A recent, large-scale epidemiological study in a generalopulation found lower relative risks of mortality with obe-ity than previously observed.13 This resulted from a moreigorous analysis involving more appropriate adjustment ofonfounding factors.14 Because cardiorespiratory fitness isuch a powerful predictor of mortality,15,16 we hypothesizedhat its inclusion as a potential confounder might furtherffect relative risk estimates associated with obesity, espe-ially in veterans, all of whom must meet fitness criteria inrder to qualify for military service.
Although a selected group, veterans referred for exerciseesting for clinical reasons provide objective measurements
f exercise capacity, along with a wealth of other clinical
fiaid
M
PWv5cbpttPBit(stegOk�
BBuam
EAuitwsmwdSlt
trecbal
Pf
ST
fmt
CdpbVtd
soskr2
RBsppoopp
EEp
519McAuley et al Obesity Paradox in Veterans
ndings. The aim of the present study was to determine thessociation between BMI and all-cause mortality in a clin-cal population of veterans without HF, adjusting for car-iorespiratory fitness.
ETHODS
opulatione studied 6876 consecutive
eteran patients (mean age8 � 11 years) referred for exer-ise testing for clinical reasonsetween 1988 and 2004. Studyarticipants were followed fromheir baseline examination untilheir death or until January 2004.atients were excluded if theirMI was �18.5 kg/m2 (n � 58),
f they had HF (n � 580), or ifhere was missing informationn � 145). The presence or ab-ence of HF was coded at theime of the test and was defined by clinical history and anjection fraction �40%. After exclusions, patients wererouped into 1 of 3 BMI categories using World Healthrganization (WHO) criteria:17 normal weight (BMI 18.5-24.9g/m2), overweight (BMI 25.0-29.9 kg/m2), or obese (BMI30.0 kg/m2).
ody Mass Indexefore exercise testing, height and weight were measuredsing standard procedures. Body mass index was calculateds weight in kilograms divided by the square of height ineters.
xercise Testingll patients underwent maximal exercise testing with these of an individualized ramp treadmill protocol.18 This testndividualizes both warm-up and peak walking speeds (onhe basis of a given patient’s height, fitness, and familiarityith treadmill walking) and ramp rate (rate of change in
peed and grade) to yield a test duration between 8 and 12inutes.19-21 A microcomputer automatically increasedorkload after an individualized walking speed and pre-icted values for maximal exercise capacity were entered.tandardized equations were used to determine the calcu-
ated peak metabolic equivalents (METs) on the basis ofreadmill speed and grade.20
Blood pressure was recorded on alternate minuteshroughout the test, and a 12-lead electrocardiogram wasecorded each minute. The patient’s subjective level ofxertion was assessed by the Borg 6-20 scale.22 Standardlinical criteria for terminating the tests (eg, fall in systoliclood pressure, ST-segment depression �2 mm, dangerousrrhythmias) were followed,19,20 but no heart rate or time
CLINICAL SIGNIF
● Although obesitymerous health chigher risk ofobesity has notated with mortal
● In this study, higwith better survAdjustment for fiassociation.
imit was imposed and a maximal effort was encouraged. 8
atients were discouraged from holding onto the handrailsor support.
tatistical Analysisotal (all-cause) mortality was used as the endpoint for
survival analysis with cardiovas-cular disease (CVD) mortality as asecondary endpoint. Survivalanalysis was performed usingKaplan-Meier curves to comparevariables and cut-points, and aCox proportional hazards modelwas used to determine which vari-ables were significantly associatedwith time to death and to developrelative risks for BMI categories(Number Crunching StatisticalSoftware, Kayesville, Utah). Wedeveloped 3 proportional hazardsmodels—first adjusting only forage, sex, and BMI category; sec-ond by adding CVD and CVD risk
actors; and lastly, including METs in the final adjustmentodel. The Armitage test for trends in proportions was used
o test the significance of relative risks for BMI category.Death status was determined as of January 2004. The
alifornia Health Department Service and Social Securityeath indices were used to ascertain the vital status of eachatient. Accuracy of deaths was reviewed by 2 clinicianslinded to exercise test results and confirmed using theeterans Affairs computerized medical records. Deaths due
o CVD included myocardial infarction, sudden cardiaceath (arrhythmic deaths), and stroke.
In order to compare our results with those of previoustudies, we used criteria from the WHO consultation onbesity,17 which classifies all subjects, irrespective of age orex, by BMI. The normal weight category (BMI 18.5-24.9g/m2) was considered the reference group, and relativeisks (RR) were calculated for the other groups (BMI 25.0-9.9 kg/m2; and BMI �30.0 kg/m2).
ESULTSaseline demographic and clinical characteristics of the
tudy group, by BMI category, are listed in Table 1. Obesityrevalence in our population was 30.4%; 44.4% of theopulation was overweight; and the remaining 25.2% weref normal weight. Compared with normal weight subjects,bese subjects were younger, had higher incidences of hy-ercholesterolemia, hypertension, and diabetes and a lowerrevalence of “high” fitness (�10 METs).
xercise Responsesxercise test results are listed in Table 2. The mean (�SD)eak heart rate was 138 � 24 beats/min, corresponding to
CE
ssociated with nu-ions, including avascular disease,uniformly associ-
MI was associated3% per BMI unit).strengthened this
ICAN
is aonditcardiobeenity.
her Bival (tness
5% of the age-predicted maximum heart rate. The mean
(di
w(pfiat(Bo
SDpgg
cm
C(fi(
fiaIlpfs3
odd
sw(
both,
520 The American Journal of Medicine, Vol 120, No 6, June 2007
�SD) peak rating of perceived exertion was 17 � 2, whichid not differ significantly across BMI categories, suggest-ng that a maximal effort was achieved by most patients.
Peak METs were similar for normal weight and over-eight categories (8.8 and 8.6, respectively), but 11% lower
7.8) for the obese category. Additionally, 39% of normalatients and 38% of overweight patients registered “high”tness (�10 METs); whereas only 24% of obese patientschieved this level (Table 1). Nevertheless, a low correla-ion between BMI and peak METs was observedr � �0.12). Similarly, this association persisted acrossMI categories: 0.07, �0.05, and �0.14 for normal weight,verweight, and obese, respectively.
urvivaluring a mean 7.5 � 4.5 years of follow-up, 1571 (23%)atients died. Death occurred in 31% of the normal weightroup, 22% of the overweight group, and 18% of the obeseroup.
Hazard ratios according to clinical variables, BMI, andardiorespiratory fitness are shown in Table 3. After adjust-
Table 1 Baseline Demographic and Clinical Characteristics
Variable
NormalBMI 18.5-(n � 1733
Demographic characteristicsWomen 86 (5%)Ethnicity (non-white) 436 (25%Age (mean [�SD], years) 60 (�12Height (mean [�SD], inches) 69 (�3)Weight (mean [�SD], lbs) 156 (�17BMI (mean [�SD], kg/m2) 23.0 (�1.Low fitness (�5 METs)† 264 (15%High fitness (�10 METs)† 682 (39%
MedicationsCalcium antagonist 375 (22%Beta-blocker 267 (15%Antihypertensive agent 272 (16%
Medical historyHypertension 650 (38%Hypercholesterolemia (�220 mg/dL) 456 (26%Smoking (current) 635 (37%Myocardial infarction 341 (20%Typical angina‡ 270 (16%Diabetes 122 (7%)Pulmonary disease 140 (8%)Claudication‡ 100 (6%)Stroke 48 (3%)Atrial fibrillation‡ 21 (1%)
InterventionsCoronary bypass surgery 168 (10%Percutaneous transluminal coronary angioplasty 119 (7%)
Values are number of participants (%) unless otherwise stated.*Trend for associations with BMI category obtained by modeling the
effect across BMI groups).†METs � metabolic equivalents (calculated from treadmill speed and‡Defined by history, by their occurrence during the exercise test, or
ents for age, sex, and BMI category, presence of CVD and w
VD risk factors, a 13% lower risk of death occurredP �.001). After further adjustment for cardiorespiratorytness, mortality risk was 20% lower per BMI categoryP �.001).
Associations of clinical variables and cardiorespiratorytness to risk of death according to BMI category were thenssessed (Table 4). All models were adjusted for age and sex.n this analysis, including CVD and CVD risk factors, a 22%ower risk of death was observed (P �.001) among obeseatients as compared with patients of normal weight. Afterurther adjustment for cardiorespiratory fitness, this associationtrengthened such that morality risk for the obese category was5% lower (P �.001) than the normal weight category.
Additionally, we entered BMI as a continuous variable inur fully-adjusted multivariate model and found a 3% re-uction in mortality per BMI unit (RR � 0.97, 95% confi-ence interval [CI], 0.96-0.98, P �.001).
Kaplan-Meier survival curves by BMI category demon-trated that the 2 higher BMI categories were associatedith better survival versus the normal weight category
P �.001, Figure 1). Moreover, when the obesity category
/m2OverweightBMI 25.0-29.9 kg/m2
(n � 3050)
ObeseBMI �30.0 kg/m2
(n � 2093) P Value*
72 (2%) 69 (3%) �.001844 (28%) 586 (28%) .3459 (�11) 56 (�11) �.00169 (�3) 69 (�3) .41
186 (�18) 233 (�32) �.00127.3 (�1.4) 34.2 (�4.0) �.001465 (15%) 338 (16%) .62
1145 (38%) 511 (24%) �.001
700 (23%) 546 (26%) .003609 (20%) 490 (23%) �.001600 (20%) 491 (23%) �.001
1437 (47%) 1241 (59%) �.001959 (31%) 728 (35%) �.001894 (29%) 518 (25%) �.001552 (18%) 350 (17%) .06480 (16%) 337 (16%) .89335 (11%) 334 (16%) �.001158 (5%) 152 (7%) �.001124 (4%) 81 (4%) .00792 (3%) 51 (2%) .4648 (2%) 38 (2%) .32
268 (9%) 182 (9%) .49236 (7%) 151 (7%) .52
of each BMI category as a continuous variable (P value represents main
); 1 MET � 3.5 mL/kg/min.using standard clinical criteria.
24.9 kg)
))
)5)))
)))
)))))
)
median
grade
as divided according to WHO obesity class, the trend
ci
DIouifoaseiem7y
sttamshsfricrtnl
PAdeaaAap
irc“on(n
iap
grade
521McAuley et al Obesity Paradox in Veterans
ontinued. Similar patterns were observed for CVD mortal-ty (Figure 2).
ISCUSSIONn contrast to the well-established relationship betweenbesity and mortality observed in numerous general pop-lation studies,1,23,24 we found that mortality was lowern veteran patients with higher BMI. Our findings there-ore confirm that an “obesity paradox” exists in patientsther than those with HF, those undergoing coronaryrtery interventions, or those undergoing dialysis.2-12 Ourtudy has several strengths: all subjects underwent anxtensive physical examination, which provides thoroughnformation on the presence or absence of baseline dis-ase; cardiorespiratory fitness was determined by maxi-al exercise testing; and a large sample size of nearly
000 patients with an average follow-up of more than 7ears (�50,000 person-years).
There are other noteworthy findings from the presenttudy. Importantly, the lower prevalence of high fitness inhe obese category versus the normal weight category iden-ifies a potential confounder often missing from previousnalyses. After we adjusted for cardiorespiratory fitness,ultivariate mortality risk for the obese category fell sub-
tantially (from 22% to 35%). This confirms our originalypothesis that adjusting for fitness would amplify the obe-ity paradox and indicates that among the potential con-ounders of the obesity-mortality relationship, cardiorespi-atory fitness is an independent risk factor at least asmportant as traditional CVD risk factors or pre-existingonditions. The mechanism responsible for such markededuction in mortality risk after adjusting for cardiorespira-ory fitness remains unclear. Perhaps cardiorespiratory fit-ess is a proxy for physical activity15 and other positive
Table 2 Exercise Test Responses
Variable
NormalBMI 18.5-24.9 kg/m2
(n � 1733)
RestHeart rate (beats/min) 75 (�14)Systolic blood pressure (mm Hg) 130 (�21)Diastolic blood pressure (mm Hg) 79 (�11)
Peak exerciseMetabolic equivalents (METs)† 8.8 (�4.0)Heart rate (beats/min) 138 (�25)Systolic blood pressure (mm Hg) 173 (�28)Diastolic blood pressure (mm Hg) 82 (�15)Perceived exertion (Borg scale) 17 (�2)
Values are mean (�SD).*Trend for associations with BMI category obtained by modeling the
effect across BMI groups).†METs � metabolic equivalents (calculated from treadmill speed and
ifestyle practices.
revious Studies Observing an Obesity Paradoxlthough obesity is associated with numerous health con-itions, including a higher risk of cardiovascular dis-ase,1,23,24 in particular populations, obesity has not beenssociated with mortality. For example, obesity was notssociated with mortality risk in Pima Indian women1 orfrican-American women.23 However, in neither case was
ny protective effect of obesity found. Only in certain clinicalopulations has an obesity paradox been observed thus far.
In 1996, Ellis and co-workers25 were among the first todentify higher BMI as a predictor of survival in a clinicallyeferred population. In patients undergoing percutaneousoronary intervention (PCI), the mortality rate in themildly obese” group (BMI 26-34 kg/m2; n � 2566) wasne-fourth that of the “low-normal” group (BMI �25 kg/m2;� 614). Paradoxically, the lowest risk of in-hospital death
0%) was observed in the heaviest patients (�120 kg;� 82).
Horwich and colleagues26 found no significant increasen 5-year mortality in overweight and obese HF patients,nd elevated BMI was an independent predictor of im-roved survival at 1 and 2 years. Gurm et al8 were the first
OverweightBMI 25.0-29.9 kg/m2
(n � 3050)
ObeseBMI �30.0 kg/m2
(n � 2093) P Value*
76 (�14) 78 (�14) �.001133 (�20) 135 (�19) �.00182 (�11) 84 (�12) �.001
8.6 (�3.6) 7.8 (�3.1) �.001139 (�25) 138 (�24) .39179 (�28) 182 (�29) �.00185 (�15) 87 (�15) �.00117 (�2) 17 (�2) .63
of each BMI category as a continuous variable (p value represents main
); 1 MET � 3.5 mL/kg/min.
Table 3 Multivariate Predictors of All-cause Mortality
Adjustments Hazard Ratio 95% CI
Age � sex � BMIcategory
0.88 (0.82-0.94)
Above � CVD* � CVDrisk factors†
0.87 (0.81-0.93)
Above � METs 0.80 (0.75-0.86)
BMI � body mass index; MET � metabolic equivalent (3.5 mL ofoxygen/kg per minute); CVD � cardiovascular diseases.
*History of myocardial infarction, stroke or surgery for CVD.†CVD risk factors include smoking, hypertension and hyper-
median
cholesterolemia.
tctmypi
mypn
FWtsS1llmmo
olemia.
FB
F
522 The American Journal of Medicine, Vol 120, No 6, June 2007
o report a paradoxic protective effect of obesity on out-ome. They followed 11,300 patients after PCI and foundhat obesity was associated with a 20% reduction in 30-dayortality risk with a trend favoring lower mortality at 1
ear. Gruberg et al9 published similar findings in 9633atients undergoing PCI. In the latter study, each 1-unitncrease in BMI conferred a 4% survival benefit.
These reports were followed by several studies docu-enting the existence of an obesity paradox in HF,2-5 dial-
sis patients,6 and a recently published randomized trial inatients with multivessel disease undergoing either coro-ary artery bypass grafting (CABG) or stenting.10
Table 4 Multivariate Relative Risk (RR) of All-cause Mortality
NormalBMI 18.5-24.9 k(n � 1733)
Deaths (all-cause) 530 (31%)Deaths (CVD) 169 (10%)Adjustments
Age � sex 1.0Above � CVD* � CVD risk factors† 1.0Above � METs 1.0
BMI � body mass index; MET � metabolic equivalent (3.5 mL of ox*History of myocardial infarction, stroke or surgery for CVD.†CVD risk factors include smoking, hypertension and hypercholester
igure 1 Kaplan-Meier survival curves of all-cause mortality by
MI category (top) and BMI obesity class (bottom). Bitness and Mortality in Obesityithin the body of obesity-mortality literature, compara-
ively little attention has focused on the role of fitness inurvival. For example, the Aerobics Center Longitudinaltudy24 of over 25,000 men (with a mean follow-up of over0 years) found no increased mortality risk with obesity asong as subjects had a moderate or high age-adjusted fitnessevel. Fitness was therefore a more important marker ofortality risk than obesity. These data suggest that when ainimum age-related criterion fitness level was reached,
besity had no impact on mortality.
I Category
OverweightBMI 25.0-29.9 kg/m2
(n � 3050)
ObeseBMI �30.0 kg/m2
(n � 2093)
671 (22%) 370 (18%)239 (8%) 134 (6%)RR (95% CI) RR (95% CI)
0.73 (0.65-0.82) 0.80 (0.75-0.92)0.73 (0.65-0.82) 0.78 (0.69-0.90)0.70 (0.63-0.79) 0.65 (0.57-0.76)
g per minute); CVD � cardiovascular diseases.
igure 2 Kaplan-Meier survival curves for CVD mortality by
by BM
g/m2
ygen/k
MI category (top) and BMI obesity class (bottom).
doLmisfa
TPrvadricqoscsorgwBriuwlqpa
LAmwwmobcntmmsbevi
SBcfeOobaurobeicbmmf
R
1
1
1
1
523McAuley et al Obesity Paradox in Veterans
Likewise, cardiovascular fitness was a strong indepen-ent predictor of mortality in the present study. However,ur findings differed from those of the Aerobics Centerongitudinal Study in that obesity significantly reducedortality in our population of veteran patients. Differences
n the baseline characteristics of the study populations (eg,ubjects in our study were older, referred to exercise testingor clinical reasons, and had prior military service) mayccount for this.
he Veteran Effectrevalence of obesity (30.4%) in our population mirrors thatecently reported for the general population; however, se-ere (class III) obesity was lower in our population (2.6%),s compared with the general population (4.9%).27 Veteransiffer from other populations of patients in several otherespects. One of the most prominent differences is the meet-ng of selection criteria at the time of enlistment. Theseriteria include, among other things, minimum height re-uirements, maximum weight requirements, and exclusionf recruits having certain preexisting health problems. Con-equently, individuals having obesity in early life are ex-luded from our population. Pediatric obesity has beenhown to have a greater impact on disease outcomes thanbesity that develops in later life. For example, Barker et al28
ecently examined the effects of birth weight and childhoodrowth rates on subsequent disease risk in Finnish men andomen. They reported that the rate of childhood gain inMI between 2 years and 11 years of age was strongly
elated to the risk of coronary events and insulin resistancen later life. It is therefore possible that many such individ-als are excluded from the veteran population by reason ofeight. Finally, the influence of self-selection in our popu-
ation must be considered. Individuals volunteering andualifying for military service may be more likely to beredisposed toward physical fitness or have other healthttributes than those avoiding military service.
imitationslimitation of our study is that it included a predominantlyale population, all of whom had prior military service andere referred for exercise testing for clinical reasons. Thus,e evaluated survival in the context of a clinical population,any of whom were limited by symptoms, medications, and
ther factors related to CVD. Any effort to predict mortalityy using BMI, clinical, or demographic data should beonsidered population-specific. Therefore, our results mayot apply to more general or healthier populations. Al-hough BMI is the most commonly used method to deter-ine obesity status, it is not the optimal measure; otherethods (eg, waist to hip ratio and lean body weight) are
uggested to be superior. Finally, because we only haveaseline data on weight, cardiorespiratory fitness, and otherxposures, we do not know if changes in any of theseariables occurred during follow-up or how this might have
nfluenced the results.ummaryMI was protective for mortality in a heterogeneous clini-ally referred male veteran population. Further adjustmentor cardiorespiratory fitness strengthened this association;ach 1-unit increase in BMI conferred a 3% survival benefit.ur results reveal that the “obesity paradox” previouslybserved in specific clinical populations occurs over aroader spectrum of veteran patients. The interactionsmong obesity, fitness, and mortality are complex and thenderlying mechanisms responsible for this phenomenonequire further elucidation. Therefore, future studies of thebesity paradox phenomenon should be directed towardoth the extent of its occurrence and the nature of itsxistence. The present findings concur with previous studiesn particular clinical populations and suggest that obesityan confer survival benefits in certain conditions. A com-ination of factors such as self-selection, meeting recruit-ent standards for and maintaining physical fitness duringilitary service, obesity in later life, and even psychosocial
actors might explain this phenomenon.
eferences1. National Task Force on the Prevention and Treatment of Obesity. Over-
weight, obesity, and health risk. Arch Intern Med. 2000;160:898-904.2. Curtis JP, Selter JG, Wany Y, et al. The obesity paradox: body mass
index and outcomes in patients with heart failure. Arch Intern Med.2005;165:55-61.
3. Davos CH, Doehner W, Rauchhaus M, et al. Body mass and survivalin patients with chronic heart failure without cachexia: the importanceof obesity. J Card Fail. 2003;9:29-35.
4. Lavie CJ, Osman AF, Milani RV, Mehra MR. Body mass and prog-nosis in chronic systolic heart failure: the obesity paradox. Am JCardiol. 2003;91:891-894.
5. Lissin LW, Gauri AJ, Froelicher VF, Ghayoumi A, Myers J, Giacom-mini J. The prognostic value of body mass index and standard exercisetesting in male veterans with congestive heart failure. J Card Fail.2002;8:206-215.
6. Beddhu S, Ramkumar N, Samore MH. The paradox of the “body massindex paradox” in dialysis patients: associations of adiposity withinflammation. Am J Clin Nutr. 2005;82:909-910.
7. Mehta LS, Skelding KA, Grines LL, et al. Obesity: protective in acutemyocardial infarction patients [abstract]? Circulation. 2002;106(suppl2):402.
8. Gurm HS, Brennan DM, Booth J, et al. Impact of body mass index onoutcome after percutaneous coronary intervention (the obesity para-dox). Am J Cardiol. 2002;90:42-45.
9. Gruberg L, Weissman NJ, Waksman R, et al. The impact of obesity onthe short-term and long-term outcomes after percutaneous coronary inter-vention: the obesity paradox? J Am Coll Cardiol. 2002;39:578-584.
0. Gruberg L, Mercado N, Milo S, et al. Impact of body mass index onthe outcome of patients with multivessel disease randomized to eithercoronary artery bypass grafting or stenting in the ARTS trial: theobestiy paradox II? Am J Cardiol. 2005;95:439-444.
1. Anker SD, Negassa A, Coats AJS, et al. Prognostic importance ofweight loss in chronic heart failure and the effect of treatment withangiotensin-converting-enzyme inhibitors: an observational study.Lancet. 2003;361:1077-1083.
2. Beddhu S, Papas LM, Ramkumar N, Samore MH. Effects of body sizeand composition on survival in hemodialysis patients. J Am Soc Neph-rol. 2003;14:2366-2372.
3. Flegal KM, Graubard BI, Williamson DF, Gail MH. Excess deathsassociated with underweight, overweight, and obesity. JAMA. 2005;
293:1861-1867.
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
524 The American Journal of Medicine, Vol 120, No 6, June 2007
4. Benichou J. A review of adjusted estimators of attributable risk. StatMethods Med Res. 2001;10:195-216.
5. Myers J, Kaykha A, George S, et al. Fitness versus physical activitypatterns in predicting mortality in men. Am J Med. 2004;117:912-918.
6. Myers J, Prakash M, Froelicher V, et al. Exercise capacity and mor-tality among men referred for exercise testing. N Engl J Med. 2002;346:793-801.
7. World Health Organization. Obesity: Preventing and Managing theGlobal Epidemic. Report of a WHO Consultation on Obesity, June 3-5,1997 (WHO Technical Series No. 894). Geneva, Switzerland: WorldHeath Organization; 2000.
8. Myers, J, Buchanan N, Walsh D, et al. Comparison of the rampversus standard exercise protocols. J Am Coll Cardiol. 1991;17:1334-1342.
9. Gibbons RJ, Balady GJ, Beasley JW, et al. ACC/AHA guidelinesfor exercise testing: a report of the American College of Cardiol-ogy/American Heart Association task force on practice guidelines(Committee of Exercise Testing). J Am Coll Cardiol. 1997;30:260-315.
0. American College of Sports Medicine. Guidelines for Exercise Testingand Prescription, 7th ed. Baltimore, MD: Lippincott Williams &
Wilkins; 2005.1. Buchfuhrer MJ, Hansen JE, Robinson TE, et al. Optimizing the exer-cise protocol for cardiopulmonary assessment. J Appl Physiol. 1983;55:1558-1564.
2. Borg GAV. Psychophysical bases of perceived exertion. Med SciSports Exerc. 1982;14:377-381.
3. Calle EE, Thun MJ, Petrelli JM, Rodriquez C, Heath CW Jr. Bodymass index and mortality in a prospective cohort of U.S. adults. N EnglJ Med. 1999;341:1097-1105.
4. Wei M, Kampert JB, Barlow CE, et al. Relationship between lowcardiorespiratory fitness and mortality in normal-weight, overweight,and obese men. JAMA. 1999;282:1547-1553.
5. Ellis SG, Elliott J, Horrigan M, Raymond RE, Howell G. Low-normalor excessive body mass index: newly identified and powerful riskfactors for death and other complications with percutaneous coronaryintervention. Am J Cardiol. 1996;78:642-646.
6. Horwich TB, Fonarow GC, Hamilton MA, et al. The relationshipbetween obesity and mortality in patients with heart failure. J Am CollCardiol. 2001;38:789-795.
7. Ogden CL, Carroll MD, Curtin LR, et al. Prevalence of overweight andobesity in the United States, 1999-2004. JAMA. 2006;295:1549-1555.
8. Barker DJP, Osmond C, Forsen TJ, Kajantie E, Eriksson JG. Trajec-tories of growth among children who have coronary events as adults.
N Engl J Med. 2005;353:1802-1809.
C
SwRWa
PJ
Suc
DC
0d
The American Journal of Medicine (2007) 120, 525-530
LINICAL RESEARCH STUDY
elective Serotonin Reuptake Inhibitor Use by Patientsith Acute Coronary Syndromes
oy C. Ziegelstein, MD,a Jennifer Meuchel, MD,a Thomas J. Kim, MD,a Madiha Latif, MBBS,a
illiam Alvarez, PharmD,b Neela Dasgupta, MD,a Brett D. Thombs, PhDc
Department of Medicine, Johns Hopkins University School of Medicine, Johns Hopkins Bayview Medical Center, bDepartment ofharmacy, Johns Hopkins Hospital, cDepartment of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine,
ohns Hopkins BBarcMcrwRtacdisCsaas
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ayview Medical Center, Baltimore, Md.
ABSTRACT
ACKGROUND: Selective serotonin reuptake inhibitors are commonly used to treat anxiety, depression,nd other conditions that commonly affect patients with coronary artery disease. Selective serotonineuptake inhibitors inhibit platelet activation and may, therefore, affect outcomes in patients with acuteoronary syndromes.ETHODS: A retrospective study was performed of 1254 patients with acute coronary syndromes
omparing in-hospital bleeding and cardiac event rates in 158 patients who received a selective serotonineuptake inhibitor and a propensity score-matched group of patients who did not. All patients were treatedith a glycoprotein IIb/IIIa inhibitor and almost all also received aspirin, clopidogrel, and heparin.ESULTS: Patients who received a selective serotonin reuptake inhibitor were significantly more likely
o experience any bleeding (37.3% vs 26.6%, OR 1.65, 95% confidence interval (CI), 1.02-2.66, P �.04)nd significantly less likely to experience recurrent myocardial ischemia, heart failure, or asymptomaticardiac enzyme elevation while in the hospital (7.0% vs 13.9%, OR 0.46, 95% CI, 0.22-0.99, P �.04). Noifferences were observed in death, myocardial infarction during the hospitalization, urgent revascular-zation, or major bleeding. Bleeding and cardiac events were not affected by antidepressants other thanelective serotonin reuptake inhibitors.ONCLUSIONS: Selective serotonin reuptake inhibitor use during a hospitalization for an acute coronaryyndrome is associated with reduced rates of recurrent ischemia, heart failure, or cardiac enzyme elevationt the expense of increased bleeding in patients receiving maximal conventional antiplatelet medicationsnd heparin. Clinicians should be aware of this association when treating patients with an acute coronaryyndrome. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Acute coronary syndrome; Bleeding; Depression; Myocardial infarction; Selective serotonin reuptakeinhibitor
tfantaoiSs
elective serotonin reuptake inhibitors (SSRIs) are widelysed in the treatment of anxiety, depression, and otheronditions that commonly affect patients with coronary ar-
Requests for reprints should be addressed to Roy C. Ziegelstein, MD,epartment of Medicine, B-1-North, Johns Hopkins Bayview Medicalenter, 4940 Eastern Avenue, Baltimore, MD 21224-2780.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.10.026
ery disease.1 Several lines of evidence suggest that plateletunction might be affected by SSRIs. Neurons and plateletsre the only cells that express serotonin receptors in theonactivated state.2 As part of the hemostatic process, ac-ivated platelets release serotonin, a key step in plateletggregation and vasoconstriction. SSRIs reduce the contentf serotonin in platelets and inhibit adenosine diphosphate-nduced platelet aggregation and dense granule release.3
erotonin-deficient platelets might not be able to releaseufficient serotonin when activated, thereby leading to dys-
unctional hemostasis.4
cetgieut
swotustorfdrttalcerwae
M
PTITAswgtS
raIctdeddd
OBdfr
sGsAdrf
sinmwatmgwpha5wpctcoaPmt
ts.
526 The American Journal of Medicine, Vol 120, No 6, June 2007
A recent review identified more than 120 MEDLINE-ited peer-reviewed articles linking SSRI use to bleedingvents.2 Reported bleeding episodes attributed to SSRIs ashey are used in clinical practice appear to be uncommon andenerally mild.5 An association between SSRI use and gastro-ntestinal bleeding has been report-d,6-8 particularly with concurrentse of nonsteroidal anti-inflamma-ory drugs or aspirin.
Three case-control studies9-11
howed that current use of SSRIsas associated with a lower riskf developing a myocardial infarc-ion (MI) when compared to nose. On the other hand, 2 othertudies showed no relationship be-ween SSRI use and MI,12,13 andne study found an increasedisk.14 The possibility that the ef-ect of SSRIs on platelets mightecrease adverse cardiac eventates in this setting is suggested byhe finding that treatment with ser-raline in depressed patients who had recently experiencedn acute coronary syndrome reduced platelet and endothe-ial activation even though these patients were already re-eiving traditional antiplatelet agents.15 No prior study hasvaluated whether SSRI use by patients who are alreadyeceiving contemporary acute coronary syndrome therapyith aspirin, clopidogrel, a glycoprotein IIb/IIIa inhibitor,
nd heparin, affects the incidence of bleeding or cardiacvents.
ETHODS
atient Sample and Data Collectionhis study was approved by the Johns Hopkins Medicine
nstitutional Review Board. All patients who were admitted tohe Johns Hopkins Hospital in Baltimore, Maryland frompril 1, 2001 through April 30, 2004 with an acute coronary
yndrome (ie, a diagnosis of unstable angina or of MI with orithout ST-segment elevation) and who were treated with anylycoprotein IIb/IIIa receptor inhibitor during the hospitaliza-ion were identified with CView, a Johns Hopkins Healthystem hospital billing software application.16
Both electronic and handwritten medical records wereeviewed. Recorded information included demographics,dmission diagnosis, antidepressant use, glycoprotein IIb/IIa inhibitor agent, medical comorbidities, risk factors forardiovascular disease and bleeding, use of percutaneousransluminal coronary angioplasty, recurrent angina, MIuring the hospitalization, asymptomatic cardiac enzymelevation, bleeding, cardiac death, and total mortality. Theata collection form also recorded the use of aspirin, clopi-ogrel, warfarin, heparin, nonsteroidal anti-inflammatory
CLINICAL SIGNIF
● Acute coronaryreceived an SSRIlikely to experielikely to experienischemia, heart fcardiac enzymehospital.
● Clinicians shouldciation when trsyndrome patien
rugs, and direct thrombin inhibitors. g
utcome Measuresleeding events and adverse cardiac events were monitoreduring the index hospitalization. Information was extractedrom the handwritten and electronic medical records andecorded on an 8-page data collection form. Responses to
these questions were then used tocode whether a bleeding event oradverse cardiac event occurred.Major bleeding was defined asbleeding resulting in death; intracra-nial hemorrhage; retroperitonealbleeding; or bleeding accompaniedby a drop in the hemoglobin con-centration of �3 g/dL.17-19 Minorbleeding was defined as overt bleed-ing not meeting criteria for majorbleeding.19 This included anymention of blood loss in the med-ical record or the need for anyintervention to treat a puncturesite complication that resulted inbleeding. Blood loss resultingfrom coronary artery bypass graft
urgery was not counted under major or minor bleeding.astrointestinal bleeding included any note of blood in the
tool or vomitus (ie, melena, hematemesis, hematochezia).ny need for endoscopy during the hospitalization wasocumented, and then if this was needed, the endoscopyeport was examined to determine whether bleeding wasound at the time of the procedure.
Prior to chart review, adverse cardiac events were clas-ified as “major” or “minor.” Major adverse cardiac eventsncluded death, MI during the hospitalization, and/or theeed for urgent revascularization.20 Death was defined asortality from any cause. An MI during the hospitalizationas defined as an episode of typical ischemic chest pain
ccompanied by new elevation of the MB fraction of crea-ine phosphokinase with at least one value of 8 ng/mL orore or by a new elevation of the cardiac troponin I of
reater than 3 times the upper limit of normal (in a patientith unstable angina as the admission diagnosis). For aatient with an admitting diagnosis of MI, an MI during theospitalization was defined as typical ischemic chest painnd an increase in creatine phosphokinase-MB of at least0% above the previous creatine phosphokinase-MB valueith a characteristic rising and falling pattern of the creatinehosphokinase-MB fraction. In addition to major adverseardiac events, minor adverse cardiac events were moni-ored, including recurrent myocardial ischemia (ischemichest pain with new electrocardiographic changes but with-ut elevation of cardiac biomarkers), new heart failure, andsymptomatic cardiac enzyme elevation (as defined above).atients were classified hierarchically; if criteria for bothajor and minor adverse cardiac events were met, the pa-
ient was included only in the major adverse cardiac event
CE
ome patients whosignificantly morebleeding and lesscurrent myocardial, or asymptomatiction while in the
ware of this asso-g acute coronary
ICAN
syndrwerencece reailureeleva
be aeatin
roup.
SSltppta
aatcSasorrssev
ttcet
cSsppAco
R
SAndcm9aw1hAc
tat
527Ziegelstein et al SSRI Use in Acute Coronary Syndromes
tatistical Analysistatistical significance was defined as a probability value
ess than .05. Chi-square and Fisher’s exact tests were usedo compare categorical variables and 2-tailed t tests to com-are continuous demographic and clinical variables betweenatients who were administered an SSRI at any time duringhe hospitalization and patients who were not administeredn SSRI.
Risk for bleeding and major adverse cardiac events wasssessed by calculating odds ratios for patients who weredministered an SSRI compared to patients not adminis-ered an SSRI. To adjust for differences in demographic andlinical characteristics of patients who were administered anSRI in the hospital and those who were not administeredn SSRI, we used a propensity score approach to generate aet of matched cases.21,22 The propensity score is a measuref the likelihood that a patient received an SSRI based onelevant demographic and clinical variables. This approacheduces many covariates to a single variable, the propensitycore. Since patients with similar propensity scores haveimilar characteristics, once matched, the outcome of inter-st can be compared for patients who received an SSRIersus those who did not receive an SSRI.
A stepwise logistic regression approach with P �.20 ashe limit for variable entry was used to model the likelihoodhat patients received an SSRI based on demographic andlinical variables. A propensity score was generated forach patient, and each patient who received an SSRI during
Table 1 Baseline Demographic and Clinical Characteristics
SSRI (n �
Age (Years), Mean � SD 61.4 �Male Gender, n (%) 90 (57.0Myocardial Infarction*, n (%) 49 (31.0Hypertension†, n (%) 133 (84.2Diabetes Mellitus†, n (%) 66 (41.8Hypercholesterolemia†, n (%) 128 (81.0Smoking†, n (%) 102 (64.6Chronic obstructive pulmonary disease†, n (%) 31 (19.6Prior MI†, n (%) 68 (43.0Congestive heart failure†, n (%) 35 (22.2Left ventricular ejection fraction �35%, n (%) 26 (17.7Liver disease, n (%) 7 (4.4)Renal failure†, n (%) 20 (12.7Gastrointestinal bleed†, n (%) 9 (5.7)Anemia†, n (%) 20 (12.7Cancer†, n (%) 17 (10.8Bleeding diathesis†, n (%) 2 (1.3)Depression†, n (%) 98 (62.0Non-SSRI antidepressant, n (%) 20 (12.7
SSRI � selective serotonin reuptake inhibitor; MI � myocardial inf*Admitting diagnosis.†Any history.
he hospitalization was matched with a patient with the w
losest possible propensity score who did not receive anSRI. The maximum allowable difference between propen-ity scores for matching was set at 0.1. When 2 or moreatients who did not receive an SSRI had an identicalropensity score, the match patient was chosen randomly.23
ll analyses were performed with SPSS version 13.0 (Chi-ago, Ill). Data are reported as means � standard deviationr as odds ratios with 95% confidence intervals (CI).
ESULTS
ample Characteristicstotal of 1254 patients were admitted with an acute coro-
ary syndrome and received a glycoprotein IIb/IIIa inhibitoruring the study period. Baseline demographic and clinicalharacteristics of the patients are shown in Table 1. Theean age of the sample was 63.1 � 12.0 years (range, 26 to
2), and 70.1% were male. Over half of the patients weredmitted with unstable angina pectoris (55.2%), 26.6%ere admitted with non-ST segment elevation MI, and8.2% with ST segment elevation MI. Most patients hadypertension (76.4%) and a history of smoking (57.8%).lmost all patients underwent percutaneous transluminal
oronary angioplasty (98.6%).Approximately 1 of every 7 patients (14.2%) had a his-
ory of depression. A total of 158 patients (12.6%) weredministered an SSRI during the hospitalization. Comparedo patients who were not administered an SSRI, patients
Pre-Propensity ScoreMatching
Post-Propensity ScoreMatching
No SSRI(n � 1096) P Value
No SSRI(n � 158) P Value
63.4 � 12.0 .06 61.2 � 12.1 .84789 (72.0) �.01 91 (57.6) .91513 (46.8) �.01 55 (34.8) .47825 (75.3) .01 133 (84.2) 1.00345 (31.5) .01 65 (41.1) .91866 (79.0) .56 127 (80.4) .89623 (56.8) .07 109 (69.0) .40124 (11.3) �.01 29 (18.4) .77309 (28.2) �.01 64 (40.5) .65143 (13.0) �.01 36 (22.8) .89141 (14.3) .27 27 (19.1) .7518 (1.6) .02 5 (3.2) .5681 (7.4) .02 22 (13.9) .7441 (3.7) .24 5 (3.2) .27
116 (10.6) .43 24 (15.2) .52123 (11.2) .86 10 (6.3) .16
4 (0.4) .17 0 (0.0) .5080 (7.3) �.01 18 (11.4) �.0180 (7.3) .02 19 (12.0) .86
158)
11.8))))))))))
)
))
))
arction.
ho received an SSRI during the hospitalization were sig-
nwhcfttpswa7
ahMruspgstctwwpdtSm
cn9wtw
ttdfpSmans99
s((wbaona
AAmpa(pbAwnc2p(
528 The American Journal of Medicine, Vol 120, No 6, June 2007
ificantly more likely to be female; to have been admittedith unstable angina pectoris; to have had a prior MI; and toave had a history of hypertension, diabetes mellitus,hronic obstructive pulmonary disease, congestive heartailure, renal failure, or liver disease. The majority of pa-ients (62%) administered an SSRI during the hospitaliza-ion had a history of depression, whereas less than 10% ofatients not administered an SSRI had a history of depres-ion. Patients receiving an SSRI during the hospitalizationere more likely to also be receiving another, non-SSRI,
ntidepressant than patients not receiving an SSRI (12.7 vs.3%, P �.02).
The propensity score model included 11 predictor vari-bles (age, gender, admission diagnosis of MI, congestiveeart failure, chronic obstructive pulmonary disease, priorI, history of hypertension, history of smoking, history of
enal failure, liver disease, and non-SSRI anti-depressantse) and 5 interactions (age � prior MI, age � chronic ob-tructive pulmonary disease, gender � chronic obstructiveulmonary disease, gender � history of hypertension,ender � history of smoking). The c-statistic for the propen-ity score model was 0.72, indicating adequate discrimina-ion. Collinearity analysis showed that the variables in-luded in the propensity score were not highly correlated;he highest bivariate correlation between any two variablesas 0.22. All 158 patients who were administered an SSRIere successfully matched: 46 on 5 or more digits of theropensity score, 27 on 4 digits, 47 on 3 digits, 20 on 2igits, and 18 on 1 digit. After propensity score matching,he only significant difference between the SSRI and non-SRI patients was that patients administered an SSRI wereore likely to have a history of depression (Table 1).As shown in Table 2, almost all matched patients re-
eived aspirin (99.1%), clopidogrel (94.6%), and intrave-ous unfractionated heparin (98.3%) as standard therapy;9.2% received either unfractionated or low moleculareight heparin. There were no significant differences be-
ween the patients administered an SSRI and those who
Table 2 Antiplatelet and Anticoagulant Medications during H
Medication SSRI (n�158)
Abciximab, n (%) 8 (5.1)Eptifibatide, n (%) 150 (94.9)Tirofiban, n (%) 1 (0.6)Aspirin, n (%) 156 (98.7)Clopidogrel, n (%) 151 (95.6)Nonsteroidal anti-inflammatory drug, n (%) 4 (2.5)Warfarin, n (%) 20 (12.7)Unfractionated heparin, n (%) 153 (96.8)Low molecular weight heparin, n (%) 23 (14.6)Lepirudin, n (%) 1 (0.6)
SSRI � selective serotonin reuptake inhibitor.
ere not, with respect to medications administered during n
he hospitalization that are known to affect bleeding orhrombosis risk. These medications included aspirin, clopi-ogrel, nonsteroidal anti-inflammatory drugs, warfarin, un-ractionated heparin, low molecular weight heparin, glyco-rotein IIb/IIIa inhibitors, or direct thrombin inhibitors.imilarly, there were no significant differences in the ad-inistration of these medications during the hospitalization
fter propensity score matching. The frequency of percuta-eous transluminal coronary angioplasty was essentially theame among those administered an SSRI (156 of 158,8.7%) and those who did not receive an SSRI (154 of 158,7.5%, P � .68).
Of the 158 patients who were administered an SSRI,ertraline was administered to 75 (47.5%), fluoxetine to 4226.6%), paroxetine to 26 (16.5%), escitalopram to 85.1%), and citalopram to 7 (4.4%). In addition, 100 patientsere administered a non-SSRI antidepressant; 46 receiveduproprion, 22 trazadone, 15 amitriptylline, 9 venlafaxine,nd 15 either doxepine, imipramine, mirtazapine, nefaz-done, or nortriptylline (7 patients were administered twoon-SSRI antidepressants). Twenty patients who receivedn SSRI also received a non-SSRI antidepressant.
dverse Cardiac Events and Bleedingdverse cardiac event rates and bleeding outcomes foratched patients are summarized in Table 3. A total of 15
atients died during the index hospitalization (1.2%). Majornd minor adverse cardiac events occurred in 86 patients6.9%) and 163 patients (13.0%), respectively. A total of 98atients experienced major bleeding (7.8%), 287 minorleeding (22.9%), and 34 gastrointestinal bleeding (2.7%).ny bleeding occurred in 396 patients (31.6%). Patientsho received an SSRI during the hospitalization were sig-ificantly more likely to experience any bleeding episodeompared to those who did not receive an SSRI (37.3% vs6.6%, OR 1.65, 95% CI, 1.02-2.66, P �.04). This was duerincipally to an increase in the risk of minor bleeding27.2% vs 18.4%, OR 1.66, 95% CI, 0.98-2.84, P �.06). Of
ization
Propensity Scoreching
Post-Propensity ScoreMatching
SSRI (n�1096) P Value No SSRI (n � 158) P Value
5 (6.8) .40 14 (8.9) .197 (93.7) .55 144 (91.1) .196 (0.5) .99 2 (1.3) .997 (99.2) .58 157 (99.4) .993 (97.0) .34 148 (93.7) .454 (4.0) .51 6 (3.8) .526 (9.7) .24 15 (9.5) .370 (98.5) .12 157 (99.4) .219 (13.6) .74 17 (10.8) .316 (0.5) .99 0 (0.0) .99
ospital
Pre-Mat
No
7102
108106
410
10814
ote, the bleeding rates observed in this study population are
sca
wcuv0whew(1Ser
ios(apmatacacwgSsa
DTdb
fideeataah
uitratwta
TasdcrSmaDtiistw
sta
529Ziegelstein et al SSRI Use in Acute Coronary Syndromes
imilar to those reported among patients with an acuteoronary syndrome receiving conventional antiplatelet ther-py and heparin.19,24
Patients who received an SSRI during the hospitalizationere significantly less likely to experience a minor adverse
ardiac event (ie, recurrent myocardial ischemia, heart fail-re or asymptomatic cardiac enzyme elevation) than indi-iduals who did not receive an SSRI (7.0% vs 13.9%, OR.46, 95% CI, 0.22-0.99, P �.04). In the SSRI group, thereere 5 patients with recurrent myocardial ischemia, 1 witheart failure, and 5 with asymptomatic cardiac enzymelevation. Among those who did not receive an SSRI, thereere 11 patients with recurrent myocardial ischemia
P �.12 vs SSRI group), 1 with heart failure (P �1.0), and0 with asymptomatic cardiac enzyme elevation (P �.19 vsSRI group). The likelihood of major adverse cardiacvents did not differ between groups that did or did noteceive an SSRI (7.6% vs 6.3%, P �.66).
All analyses were repeated for patients who were admin-stered a non-SSRI antidepressant. The degree of balancingf demographic and clinical characteristics through propen-ity score matching was comparable to the SSRI analysisnot shown). There were 5 major adverse cardiac eventsmong the 100 patients administered a non-SSRI antide-ressant and 5 among matched controls. Compared toatched patients, however, patients who were administered
non-SSRI antidepressant during the hospitalizationended, albeit non-significantly, to experience more minordverse cardiac events (14.0% vs 7.0%, OR 2.16, P �.11)ompared to those patients who did not receive a non-SSRIntidepressant. Non-SSRI antidepressant use was not asso-iated with bleeding (OR 1.04, P �.89) or, individually,ith major (OR 1.09, P �.84), minor (OR 0.95, P �.87), orastrointestinal (OR 0.59, P �.72) bleeding. Results for theSRI and non-SSRI antidepressant analyses did not changeubstantively when the 20 patients who used both types ofntidepressants were excluded from analyses.
ISCUSSIONhis is the first study to examine the association of SSRI useuring hospitalization for an acute coronary syndrome with
Table 3 Adverse Cardiovascular and Bleeding Events During H
EventNo SSRI After Matching(n � 158)
Major Adverse Cardiac Events 10 (6.3%)Minor Adverse Cardiac Events 22 (13.9%)Major Bleeding 13 (8.2%)Minor Bleeding 29 (18.4%)Gastrointestinal Bleeding 2 (1.3%)Any Bleeding 42 (26.6%)
SSRI � selective serotonin reuptake inhibitor.Patients who experienced both a major and a minor adverse cardiac
for definitions). A bleeding event was only counted once in the total if
leeding episodes and adverse cardiac events. The major t
nding of this study was that patients who received an SSRIuring the hospitalization were significantly more likely toxperience a bleeding episode and significantly less likely toxperience recurrent myocardial ischemia, heart failure, orsymptomatic cardiac enzyme elevation while in the hospi-al. The effects of SSRI use were observed in patientslready receiving contemporary standard therapy for ancute coronary syndrome, including aspirin, clopidogrel,eparin, and a glycoprotein IIb/IIIa inhibitor.
SSRI use has been associated with an increased risk ofpper gastrointestinal bleeding in population-based stud-es,6,7,25 especially in elderly patients, in those with a his-ory of gastrointestinal bleeding,25 or in patients concur-ently administered nonsteroidal anti-inflammatory drugs orspirin.6-8 The association between SSRI use and gastrointes-inal bleeding and the increased risk when used concurrentlyith nonsteroidal anti-inflammatory drugs, aspirin, or drugs
hat affect coagulation is noted in the prescribing informationvailable to consumers for all major SSRI antidepressants.26-30
The Sertraline Antidepressant Heart Attack Randomizedrial (SADHART)31 found that the incidence of majordverse cardiac events was lower, although not significantlyo, among patients hospitalized for an acute coronary syn-rome with major depression who were randomized to re-eive sertraline than in those who received placebo (relativeisk [RR] 0.77, 95% CI, 0.51-1.16). Of note, exposure toSRIs in SADHART occurred on average more than aonth after the acute coronary syndrome. A secondary
nalysis of the Enhancing Recovery in Coronary Heartisease Patients (ENRICHD) Randomized Trial32 showed
hat the risk of death or recurrent MI was significantly lowern the 301 patients taking an SSRI to treat depression thann the 1388 depressed patients not receiving an antidepres-ant. In ENRICHD, patients were enrolled within 28 days ofhe index MI, and SSRIs were prescribed a minimum of 5eeks after beginning cognitive behavior therapy.Although our results are consistent with these studies,
everal limitations should be considered in interpretinghese findings. In particular, the results are observationalnd from a retrospective chart review; the number of pa-
lization
SRI (n�158)Odds Ratio(95% Confidence Interval)
PValue
2 (7.6%) 1.22 (0.51-2.90) .661 (7.0%) 0.46 (0.22-0.99) .041 (7.0%) 0.84 (0.36-1.92) .673 (27.2%) 1.66 (0.98-2.84) .065 (3.2%) 2.55 (0.49-13.34) .259 (37.3%) 1.65 (1.02-2.66) .04
ere only counted as having had a major adverse cardiac event (see textn had more than one type (see text for definitions).
ospita
S
1114
5
event wa perso
ients who received an SSRI and the numbers of bleeding
et
daiwesc
ARS
R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
530 The American Journal of Medicine, Vol 120, No 6, June 2007
vents and adverse cardiac events were relatively small; andhe period of exposure to SSRIs is not known.
In summary, this retrospective study shows that SSRI useuring a hospitalization for an acute coronary syndrome isssociated with reduced cardiac events at the expense ofncreased bleeding in patients already receiving therapyith conventional antiplatelet medications and heparin. The
ffect of SSRIs in depressed patients with an acute coronaryyndrome should be examined in a randomized, prospectiveontrolled trial with a larger number of patients.
CKNOWLEDGMENTCZ, JM, and TK were supported by the Miller Familycholar Program.
eferences1. Fava M, Rush AJ, Trivedi MH, et al. Background and rationale for the
sequenced treatment alternatives to relieve depression (STAR*D)study. Psychiatr Clin North Am. 2003;26:457-94.
2. Serebruany VL. Selective serotonin reuptake inhibitors and increasedbleeding risk: are we missing something? Am J Med. 2006;119:113-6.
3. Maurer-Spurej E, Pittendreigh C, Solomons K. The influence of se-lective serotonin reuptake inhibitors on human platelet serotonin.Thromb Haemost. 2004;91:119-28.
4. Maurer-Spurej E. Serotonin reuptake inhibitors and cardiovasculardiseases: a platelet connection. Cell Mol Life Sci. 2005;62:159-170.
5. Nelva A, Guy C, Tardy-Poncet B, et al. Hemorrhagic syndromesrelated to selective serotonin reuptake inhibitor (SSRI) antidepres-sants. Seven case reports and review of the literature. Rev Med Interne.2000;21:152-160.
6. de Abajo FJ, Rodriguez LA, Montero D. Association between selec-tive serotonin reuptake inhibitors and upper gastrointestinal bleeding:population based case-control study. BMJ.1999;319:1106-1109.
7. Dalton SO, Johansen C, Mellemkjaer L, et al. Use of selective serotoninreuptake inhibitors and risk of upper gastrointestinal tract bleeding: apopulation-based cohort study. Arch Intern Med. 2003;163:59-64.
8. Wessinger S, Kaplan M, Choi L, et al. Increased use of selectiveserotonin reuptake inhibitors in patients admitted with gastrointestinalhaemorrhage: a multicentre retrospective analysis. Aliment PharmacolTher. 2006;23:937-944.
9. Schlienger RG, Fischer LJ, Jick H, Meier CR. Current use of selectiveserotonin reuptake inhibitors and risk of acute myocardial infarction.Drug Saf. 2004;27:1157-1165.
0. Sauer WH, Berlin JA, Kimmel SE. Selective serotonin reuptake in-hibitors and myocardial infarction. Circulation. 2001;104:1894-1898.
1. Sauer WH, Berlin JA, Kimmel SE. Effect of antidepressants and theirrelative affinity for the serotonin transporter on the risk of myocardialinfarction. Circulation. 2003;108:32-36.
2. Cohen HW, Gibson G, Alderman MH. Excess risk of myocardialinfarction in patients treated with antidepressant medications: associ-ation with use of tricyclic agents. Am J Med. 2000;108:2-8.
3. Meier CR, Schlienger RG, Jick H. Use of selective serotonin reuptakeinhibitors and risk of developing first-time acute myocardial infarction.Br J Clin Pharmacol. 2001;52:179-184.
4. Tata LJ, West J, Smith C, et al. General population based study of theimpact of tricyclic and selective serotonin reuptake inhibitor antide-pressants on the risk of acute myocardial infarction. Heart. 2005;91:
465-471.5. Serebruany VL, Glassman AH, Malinin AI, et al. Platelet/endothelialbiomarkers in depressed patients treated with the selective serotoninreuptake inhibitor sertraline after acute coronary events: the SertralineAntiDepressant Heart Attack Randomized Trial (SADHART) PlateletSubstudy. Circulation. 2003;108:939-944.
6. Plaut D, CView. Johns Hopkins Health System, Casemix InformationManagement: Baltimore, MD. 2005.
7. Cohen M, Demers C, Gurfinkel EP, et al. A comparison of low-molecular-weight heparin with unfractionated heparin for unstablecoronary artery disease. Efficacy and Safety of Subcutaneous Enox-aparin in Non-Q-Wave Coronary Events Study Group. N Engl J Med.1997;337:447-452.
8. Antman EM, McCabe CH, Gurfinkel EP, et al. Enoxaparin preventsdeath and cardiac ischemic events in unstable angina/non-Q-wavemyocardial infarction. Results of the thrombolysis in myocardial in-farction (TIMI) 11B trial. Circulation. 1999;100:1593-1601.
9. Goodman SG, Fitchett D, Armstrong PW, et al. Randomized evalua-tion of the safety and efficacy of enoxaparin versus unfractionatedheparin in high-risk patients with non-ST-segment elevation acutecoronary syndromes receiving the glycoprotein IIb/IIIa inhibitor epti-fibatide. Circulation. 2003;107:238-244.
0. Wohrle J, Grebe OC, Nusser T, et al. Reduction of major adversecardiac events with intracoronary compared with intravenous bolusapplication of abciximab in patients with acute myocardial infarctionor unstable angina undergoing coronary angioplasty. Circulation.2003;107:1840-1843.
1. D’Agostino RB Jr. Propensity score methods for bias reduction in thecomparison of a treatment to a non-randomized control group. StatMed. 1998;17:2265-2281.
2. Rosenbaum PR, Rubin DB. Reducing bias in observational studiesusing subclassification on the propensity score. Am J Stat Assoc.1984;79:516-524.
3. Parsons LS. Reducing bias in a propensity score matched-pair sampleusing greedy matching techniques. Proceedings of the twenty-sixthannual SAS users group international conference. SAS Institute, Inc.,Cary, NC. 2001.
4. Alexander KP, Chen AY, Roe MT, et al. Excess dosing of antiplateletand antithrombin agents in the treatment of non-ST-segment elevationacute coronary syndromes. JAMA. 2005;294:3108-3116.
5. van Walraven C, Mamdani MM, Wells PS, Williams JI. Inhibition ofserotonin reuptake by antidepressants and upper gastrointestinal bleed-ing in elderly patients: retrospective cohort study. BMJ. 2001;323:655-658.
6. Celexa [package insert]. St. Louis, MO: Forest Pharmaceuticals, Inc.:2006. Available at: http://www.celexa.com/pdf/celexa_pi.pdf. Ac-cessed March 13, 2007.
7. Lexapro. St. Louis, MO: Forest Pharmaceuticals, Inc.:2002. Available at:http://www.fda.gov/cder/foi/label/2003/21323se1-003,se8-007,21365se8-001,se1-004_lexapro_lbl.pdf. Accessed March 13, 2007.
8. Prozac [Medication Guide]. Indianapolis, IN: Eli Lilly and Company:2006. Available at: http://pi.lilly.com/us/prozac.pdf. Accessed March13, 2007.
9. Paxil [Medication Guide]. Research Triangle Park, NC: GlaxoSmithKline:2006. Available at: http://us.gsk.com/products/assets/us_paxil.pdf. AccessedMarch 13, 2007.
0. Zoloft [Medication Guide]. New York, NY: Pfizer Pharmaceuticals:2006. Available at: http://www.pfizer.com/pfizer/download/uspi_zoloft.pdf. Accessed March 13, 2007.
1. Glassman AH, O’Connor CM, Califf RM, et al. Sertraline treatment ofmajor depression in patients with acute MI or unstable angina. JAMA.2002;288:701-709.
2. Taylor CB, Youngblood ME, Catellier D, et al. Effects of antidepres-sant medication on morbidity and mortality in depressed patients after
myocardial infarction. Arch Gen Psychiatry. 2005;62:792-798.
C
IPHMJa
CC
Lta2
ERl
0d
The American Journal of Medicine (2007) 120, 531-538
LINICAL RESEARCH STUDY
mproving Risk Assessment with Cardiac Testing ineripheral Arterial Disease
arm H. H. Feringa, MD,a Abdou Elhendy, MD,b Stefanos E. Karagiannis, MD,a Peter G. Noordzij, MD,c
artin Dunkelgrun, MD,d Olaf Schouten, MD,d Radosav Vidakovic, MD,a Ron T. van Domburg, PhD,a
eroen J. Bax, MD,e Don Poldermans, MDc
Department of Cardiology, Erasmus Medical Center, Rotterdam, the Netherlands; bDepartment of Internal Medicine, Section ofardiology, University of Nebraska, Omaha; cDepartment of Anaesthesiology, and dDepartment of Vascular Surgery, Erasmus Medical
enter, Rotterda e nds.PaidMmp(Rcsca1pLoClrr
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
m, the Netherlands; Department of Cardiology, Leiden University Medical Center, Leiden, the Netherla
ABSTRACT
URPOSE: The study’s objective was to evaluate the prognostic value of left ventricular ejection fractionnd stress-induced ischemia during dobutamine stress echocardiography, in addition to ankle-brachialndex measurements and clinical risk factors in patients with suspected or known peripheral arterialisease.ETHODS: In 852 patients with suspected or known peripheral arterial disease (mean age 63 years, 70%ale), the ankle-brachial index was measured, left ventricular ejection fraction was assessed, and all
atients underwent additional stress testing. Endpoints were all-cause mortality and hard cardiac eventscardiac death or nonfatal myocardial infarction).ESULTS: During a mean follow-up of 7.6 � 4.4 years, death occurred in 288 patients (34%), and hardardiac events occurred in 216 patients (25%). Mean left ventricular ejection fraction was 50% � 17%, andtress-induced ischemia was observed in 352 patients (41%). In multivariate analysis with adjustment forlinical risk factors and ankle-brachial index, each 5% decrease in left ventricular ejection fraction wasssociated with increased all-cause mortality (hazard ratio [HR] 1.05, 95% confidence interval [CI],.02-1.09) and hard events (HR 1.14, 95% CI, 1.08-1.21). Stress-induced ischemia also independentlyredicted all-cause mortality (HR 2.01, 95% CI, 1.38-2.79) and hard events (HR 2.06, 95% CI, 1.39-3.08).eft ventricular ejection fraction and stress-induced ischemia provided incremental prognostic informationver clinical data and ankle-brachial index values (P �.001).ONCLUSIONS: Left ventricular ejection fraction and stress-induced ischemia independently predictong-term outcome and improve prognostic risk assessment, in addition to ankle-brachial index and clinicalisk factors in patients with suspected or known peripheral arterial disease. © 2007 Elsevier Inc. All rightseserved.
KEYWORDS: Ankle-brachial index; Coronary artery disease; Dobutamine stress echocardiography; Left ventriculardysfunction; Peripheral arterial disease; Prognosis
UueotMps
ower extremity peripheral arterial disease is a manifesta-ion of systemic atherosclerosis and has been recognized asgrowing health burden worldwide.1 Prevalence rates up to9% have been reported for peripheral arterial disease in the
Requests for reprints should be addressed to Don Poldermans, MD,rasmus MC, University of Rotterdam, Department of Anaesthesiology,oom H-921, Dr. Molewaterplein 40, 3015 GD Rotterdam, the Nether-
ands.
w: [email protected]ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.06.041
nited States, depending on the age of the study cohort, thenderlying atherosclerosis risk factor profile, and the pres-nce of cardiovascular co-morbidities.2-7 A high prevalencef left ventricular dysfunction among patients with symp-omatic peripheral arterial disease has been observed.8
oreover, coronary artery disease frequently co-exists witheripheral arterial disease because both conditions share theame atherosclerotic risk factors. More than half of patients
ho present with peripheral arterial disease may have evi-
dppt
Atmaiiodtcvtifodca
owiordwfpvtfk
MTro(auea3adcraDaa
omgMP
laPb
ASrrapGaabtdw2spvabea
lues.
532 The American Journal of Medicine, Vol 120, No 6, June 2007
ence of coronary artery disease based on electrocardiogra-hy or medical history.1,9,10 The prognosis of patients witheripheral arterial disease is therefore characterized by a 2-o 6-fold increased risk of cardiovascular death.11-17
The American College of Cardiology/American Heartssociation has emphasized in
heir guidelines for the manage-ent of patients with peripheral
rterial disease the importance ofdentifying and treating underly-ng cardiovascular risk factors andbtaining ankle-brachial indexata for prognostic risk stratifica-ion.1 Although the detection oforonary artery disease and leftentricular dysfunction in this par-icular patient population may bemportant because of the benefitrom subsequent medical therapyr coronary intervention, evi-ence-based recommendations forardiac assessment in literaturere limited.
Dobutamine stress echocardi-graphy is an accurate, safe, andidely used noninvasive imag-
ng technique for the evaluationf coronary artery disease andisk assessment.18,19 However, because ankle-brachial in-ex data can risk stratify this population, it is not knownhether stress testing can further provide prognostic in-
ormation in addition to ankle-brachial index values. Theurpose of this study was to determine the prognosticalue of dobutamine stress echocardiography in additiono ankle-brachial index measurements and clinical riskactors in a large cohort of patients with suspected ornown peripheral arterial disease.
ETHODShe study population consisted of consecutive patients
eferred for evaluation of peripheral arterial disease at theutpatient vascular clinic of the Erasmus Medical CenterMC), Rotterdam, the Netherlands between January 1990nd January 2005. The Erasmus MC is a metropolitanniversity hospital in the southwestern area of the Neth-rlands, serving a population of approximately 3 millionnd acting as a tertiary referral center for approximately0 affiliated hospitals. Patients with suspected peripheralrterial disease had a typical history of intermittent clau-ication or other symptoms of chronic arterial insuffi-iency, including ulceration of the foot, hair loss, oreduced capillary refill. Patients with known peripheralrterial disease had a resting ankle-brachial index �0.90.obutamine stress echocardiography was performed to
ssess the presence and extent of concomitant coronary
CLINICAL SIGNIF
● Patients with peare at increased
● In this cohort stuperipheral arterilence of left vent�35% and corongina pectoris, hfarction, or stwas 12% and 70
● Lower ejectionduced ischemia ioutcome and imassessment overbrachial index va
rtery disease. Specific indications for stress echocardi- s
graphy included typical or atypical chest pain, assess-ent for coronary artery revascularization, and (presur-
ical) prognostic risk stratification. The hospital’sedical Ethical Committee approved the study protocol.
atients who fulfilled inclusion criteria agreed on partic-ipation in the study. Based onhospital records and personal in-terviews at the time of the visit, amedical history was recorded,and data were prospectively en-tered into a computerized data-base. Diabetes mellitus was re-corded if patients presented witha fasting glucose level of �7.0mmol/L or in those who requiredinsulin treatment. Hypertensionwas recorded if patients pre-sented with a blood pressure�140/90 mm Hg or if patientsreceived antihypertensive drugs.Hypercholesterolemia was re-corded if patients presented witha plasma cholesterol level �5.5mmol/L or if patients were tak-ing lipid-lowering agents. Pa-tients were considered to haverenal dysfunction if they pre-sented with a serum creatinine
evel �2.0 mg/dL (177 �mol/L) or if they required di-lysis. Cigarette smoking included only current smoking.atients were assessed for cardiac medication use and aaseline 12-lead electrocardiography was obtained.
nkle-Brachial Index Measurementystolic blood pressures in the right and left brachial artery,ight and left dorsalis pedis artery, and right and left poste-ior tibial artery were measured by trained technicians using
Doppler ultrasonic instrument with an 8-MHz vascularrobe (Imexdop CT� Vascular Doppler, Miami Medical,len Allen, Virginia). The ankle-brachial index in the right
nd left leg was calculated by dividing the right and the leftnkle pressure by the brachial pressure. The higher of the 2rachial blood pressures was used if a discrepancy in sys-olic blood pressure was present. Again, the higher of theorsalis pedis and posterior tibial artery pressure was usedhen a discrepancy in systolic blood pressure between thearteries was measured. The ankle-brachial index was mea-
ured after the participants had been resting in the supineosition for at least 10 minutes. Of the ankle-brachial indexalues obtained in each leg, the lower was used in allnalysis. Inter- and intraobserver agreement for the ankle-rachial index was 97% and 98%, respectively. We consid-red patients with values greater than 1.50 to have calcifiedtherosclerosis. These patients were excluded from the
CE
ral arterial diseaseor late events.
852 patients withsease, the preva-r ejection fractionrtery disease (an-of myocardial in-
nduced ischemia)spectively.
ns and stress-in-endently predictedd prognostic riskal data and ankle-
ICAN
ripherisk f
dy ofal diriculaary a
istoryress-i%, re
fractiondepproveclinic
tudy.
DTa2tttfti�mhhpsmw
tbs�iaorcoswcgbe(w
533Feringa et al Cardiac Testing in Peripheral Arterial Disease
obutamine Stress Echocardiographyhe dobutamine stress echocardiography was performeds previously described.20,21 Patients underwent a resting-dimensional echocardiographic examination. Left ven-ricular end-diastolic and end-systolic volumes were ob-ained from the apical 4- and 2-chamber views by usinghe Simpson’s rule formula, from which the ejectionraction was calculated. Dobutamine hydrochloride washen administered intravenously by infusion pump, start-ng at 10 �g/kg/min for 3 minutes, and increased by 10g/kg/min every 3 minutes to a maximum of 40 �g/kg/in. The dobutamine infusion was stopped if a target
eart rate was achieved (85% of a theoretic maximaleart rate). If the target heart rate was not achieved andatients had no symptoms or signs of ischemia, atropineulphate (starting with 0.25 mg, increased to a cumulativeaximum of 2.0 mg) was given intravenously. Patientsere excluded from the study if the test was prematurely
Table 1 Baseline Characteristics According to Groups DefinedEchocardiography
Characteristic
DSE with NoIschemia, Normal ABI(n � 154)
DSE wAbno(n �
DemographicsAge (years) (�SD) 59 � 12 63Male sex 61.0 74.0Angina 20.8 27.5Previous myocardial infarction 23.4 41.6Previous CABG 13.6 19.4Previous PTCA 1.3 6.1Known CAD (summary variable) 35.7 55.8History of congestive heart
failure8.4 8.1
History of stroke or TIA 3.2 9.0Diabetes mellitus 12.3 17.3Hypercholesterolemia 22.7 20.5Hypertension 27.9 41.3Cigarette smoking 28.6 37.9Renal failure 3.9 4.0Abnormal electrocardiography 31.8 47.7
Q waves 19.5 28.9ST segment changes 1.9 2.3
MedicationsAspirin 13.0 24.9ACE-inhibitor 22.7 28.0Beta-blocker 28.6 28.9Calcium channel blocker 20.1 30.3Digoxin 4.4 7.2Diuretic 8.4 15.0Nitrate 23.4 26.6Statin 23.4 20.5
Ankle-brachial index 1.02 � 0.07 0.62Rest wall motion abnormalities 37.0 47.7Left ventricular ejection fraction 53 � 16 51
DSE � dobutamine stress echocardiography; ABI � ankle-brachial indcoronary angioplasty; CAD � coronary artery disease; TIA � transient isc
Values are expressed in percentages or mean � SD. Abnormal ABI �groups as compared by the chi-squared test for categorical variables an
erminated because of: symptomatic decrease in systoliclood pressure �40 mm Hg from the resting value, or aystolic blood pressure �100 mm Hg; blood pressure240/140 mm Hg; occurrence of cardiac arrhythmias;
ntolerable adverse effects from dobutamine or atropine;nd poor echocardiographic images. Off-line assessmentf echocardiographic images was performed by 2 expe-ienced investigators without knowledge of the patient’slinical data. From 1990 to 1993, a 14-segment 4-pointrdinal scale was used. After 1993, a 16-segment 5-pointcore was used.21-23 Stress-induced myocardial ischemiaas considered if new wall motion abnormalities oc-
urred (ie, if wall motion in any segment worsened by �1rade(s) during the test, with the exception of akinesisecoming dyskinesis). The extent and location of isch-mia were evaluated, and a wall-motion score indextotal score divided by the number of segments scored)as calculated, both at rest and during peak stress. When
kle-Brachial Index Values and Dobutamine Stress
Ischemia,BI
DSE with Ischemia,Normal ABI(n � 77)
DSE with Ischemia,Abnormal ABI(n � 275) P-Value
62 � 12 65. � 11 �.00168.8 70.9 .02729.9 19.3 .04631.2 34.5 .00122.1 14.2 .132.6 2.9 .041
46.8 45.1 �.0013.9 6.5 .53
7.8 5.8 .1118.2 15.3 .5027.3 16.0 .1120.8 37.5 �.00129.9 26.2 .0131.3 6.2 .26
45.5 45.1 .0124.7 26.5 .176.5 5.8 .042
16.9 22.2 .01822.1 24.4 .4827.3 22.9 .3724.7 25.5 .116.5 5.5 .65
14.3 13.5 .2522.1 23.6 .7427.3 16.0 .0941.01 � 0.06 0.60 � 0.20 �.00135.1 37.8 .022
49 � 18 47 � 17 .12
G � coronary artery bypass grafting; PTCA � percutaneous transluminalattack; ACE � angiotensin converting enzyme.brachial index �0.90. The P-values reflect overall differences in the 4sis of variance techniques for continuous variables.
by An
ith Normal A346)
� 11
� 0.17
� 17
ex; CABhemicankle-
d analy
tiw
FDadiCmhnvCoemitCcft
SCaisfotwsaed
asimttfatetAw
RAipawitarcs(0intahiw(ei
534 The American Journal of Medicine, Vol 120, No 6, June 2007
here was disagreement between the 2 assessors, a thirdnvestigator viewed the images and a majority decisionas reached.
ollow-upuring follow-up, study endpoints were all-cause mortality
nd hard cardiac events (cardiac death or nonfatal myocar-ial infarction). Survival status was obtained by approach-ng the referring physician or the municipal civil registries.linical information was obtained by outpatients visits,ailed questionnaires, telephone interviews, and reviewing
ospital records. Nonfatal myocardial infarction was diag-osed when at least 2 of the following were present: ele-ated cardiac enzyme levels (CK level �190 U/L andK-MB �14 U/L, or CK-MB fraction �10% of total CK,r cardiac troponin T �0.1 ng/mL), development of typicallectrocardiographic changes (new Q waves �1 mm or �30s), and typical symptoms of angina pectoris. Death certif-
cates and autopsy reports were reviewed, and general prac-itioners were approached to ascertain the cause of death.ardiac death was defined as death caused by acute myo-ardial infarction, cardiac arrhythmias, or congestive heartailure. Sudden unexpected death in previously stable pa-ients was considered cardiac death.
tatistical Analysisontinuous data were compared using the Student’s t test ornalysis of variance techniques when appropriate. Categor-cal data were compared using the chi-squared test. A finalet of independent predictors of left ventricular ejectionraction �35% and stress-induced myocardial ischemia wasbtained by multivariate analysis with stepwise deletion ofhe least significant variable. The Kaplan-Meier methodith the log-rank test was used to assess differences in
urvival between different groups of patients. Univariatend multivariate Cox hazard regression analysis was used tovaluate the prognostic value of dobutamine stress echocar-iography, ankle-brachial index, and baseline clinical vari-
Table 2 Variables Independently Associated with Left VentricIschemia During Dobutamine Stress Echocardiography Identified
Characteristic
Odds Ratio (95Ejection Fractio(n � 105)*
Male sex 2.08 (1.25-3.49Coronary artery disease 4.07 (2.60-6.36History of heart failure 3.48 (1.75-6.91Diabetes mellitus 2.48 (1.49-4.12Renal failure 3.54 (1.27-9.88Hypercholesterolemia 1.84 (1.14-2.97Ankle-brachial index (per 0.10 decrease)Age (per year increase)Ischemic baseline electrocardiogram
*C-statistic: 0.76.
†C-statistic: 0.60.bles. The incremental value of dobutamine stress test re-ults over clinical variables and ankle-brachial index valuesn the prediction of events was determined according to 3odels. In the first model, clinical variables, baseline elec-
rocardiography, and ankle-brachial index values were en-ered. In the second and third model, left ventricular ejectionraction and stress-induced ischemia, respectively, weredded to the first model. Tests for heterogeneity were usedo reveal a possible interaction between dobutamine stresschocardiography and ankle-brachial index values. For allests, a P-value �.05 (2-sided) was considered significant.nalysis was performed using SPSS-11.0 statistical soft-are (SPSS Inc., Chicago, Ill).
ESULTStotal of 944 patients were referred for ankle-brachial
ndex measurement and dobutamine stress echocardiogra-hy. A total of 30 patients (3%) were excluded because ofnkle-brachial index values �1.50, and 62 patients (7%)ere excluded due to termination of the stress test before an
schemic endpoint (cardiac arrhythmia in 8 patients, hypo-ension in 14, chills and intolerable adverse effects in 11,nd poor echocardiographic images in 29 patients). Theemaining 852 patients were considered for follow-up andonstituted our study population (Table 1). Follow-up wasuccessful in all. Mean age was 63 years, and 598 patients70%) were male. Mean ankle-brachial index was.72 � 0.24, and 621 patients (73%) had an ankle-brachialndex �0.90. During dobutamine stress echocardiography,o fatal complications occurred. Mean left ventricular ejec-ion fraction was 50% � 17%, and 105 patients (12%) hadn ejection fraction �35%. A total of 122 patients (14%)ad an ejection fraction �40%. Rest wall motion abnormal-ties were observed in 469 patients (55%). Ischemia (newall motion abnormalities) was detected in 352 patients
41%). In the patients who presented with left ventricularjection fraction �35%, stress-induced ischemia occurredn 79 patients (75%, 9% of the total study population). In the
ection Fraction �0.35% and Stress-induced Myocardialepwise Multivariate Analysis
Left Ventricular35%
Odds Ratio (95% CI) for Stress-inducedMyocardial Ischemia During DSE(n � 352)†
1.06 (1.01-1.13)1.02 (1.00-1.03)1.48 (1.10-2.01)
ular Ejby St
% CI)n �0.
))))))
cdowitdhwac
PEDitmaeke5mt(r(fsaihpiiptaec1
fsvcsa
DAsrpb
cch7sfhawc7paccsda
Fki
535Feringa et al Cardiac Testing in Peripheral Arterial Disease
urrent patient cohort, the prevalence rate of coronary arteryisease by medical history alone (angina pectoris or historyf myocardial infarction) was 47% and when combinedith dobutamine stress echocardiography, 70%. Variables
ndependently associated with left ventricular ejection frac-ion �0.35% included male sex, history coronary arteryisease, history of heart failure, diabetes, renal failure, andypercholesterolemia. Variables independently associatedith stress-induced myocardial ischemia included age, low
nkle-brachial index values, and ischemic baseline electro-ardiogram (Table 2).
rognostic Value of Dobutamine Stresschocardiographyuring a mean follow-up of 7.6 � 4.4 years, death occurred
n 288 patients (34%) and hard cardiac events in 216 pa-ients (25%; cardiac death in 145 patients and nonfatalyocardial infarction in 71). Kaplan Meier curves stratified
ccording to ankle-brachial index and dobutamine stresschocardiography demonstrated that in patients with an an-le-brachial index �0.90, those with stress-induced isch-mia had a decreased survival (annual mortality rate of.7%), compared with patients without ischemia (annualortality rate of 3.2%, P �.001), and compared with pa-
ients without ischemia and ankle-brachial index �0.90annual mortality rate of 2.5%, P �.001) (Figure). Compa-able results were obtained for the endpoint of hard eventsFigure). In univariate analysis, left ventricular ejectionraction and stress-induced ischemia were significantly as-ociated with adverse outcome (Tables 3, 4). In multivariatenalysis, left ventricular ejection fraction and stress-inducedschemia independently predicted all-cause mortality andard cardiac events, and provided significant incrementalrognostic information over clinical data and ankle-brachialndex values (P �.001) (Tables 3, 4). In a similar modelncluding clinical baseline characteristics, electrocardiogra-hy, ankle-brachial index values, and left ventricular ejec-ion fraction, a higher number of ischemic segments waslso significantly associated with mortality (HR per isch-mic segment: 1.06, 95% CI, 1.01-1.12, P � .01) and hardardiac events (HR per ischemic segment: 1.08, 95% CI,.02-1.14, P � .006).
Tests for heterogeneity revealed no evidence for a dif-erential effect of dobutamine stress echocardiography re-ults among patients with different ankle-brachial indexalues (all interaction terms: P �.05), indicating that de-reased left ventricular ejection fraction and an ischemictress test predicted the risk of death and hard cardiac eventscross the entire spectrum of ankle-brachial index values.
ISCUSSIONtherosclerosis has a common systemic pathogenesis and
imultaneously affects multiple vascular beds.24 Prevalenceates up to 75% of concomitant coronary artery disease ineripheral arterial disease have been published in literature,
ased on coronary angiography or clinical history.9,10 The vurrent results demonstrate that coronary artery disease,erebrovascular disease, and renal disease were remarkablyigh in our study population, with prevalence rates of 70%,%, and 5%, respectively. In addition, 12% of the currenttudy population presented with left ventricular ejectionraction �35%. The high risk of death and the impairedealth-related quality of life in patients with peripheralrterial disease poses a significant health burden world-ide.25 In the current study, death occurred in 34% and hard
ardiac events occurred in 25% during a mean follow-up of.6 years, reflecting the significant adverse consequences oferipheral arterial disease. It is well established that lownkle-brachial index values predict overall and cardiovas-ular mortality, and that ankle-brachial index measurementsan be used for prognostic risk stratification.26 The currenttudy demonstrated supportive results and showed that eachecrease in ankle-brachial index of 0.10 was associated with6% and 7% increased risk of all-cause mortality and hard
igure Kaplan Meier curves in patient with suspected andnown peripheral arterial disease stratified according to stress-nduced myocardial ischemia (DSE�) and ankle-brachial index
alues �0.90 (ABI�).
cf
CTp
piikcc
536 The American Journal of Medicine, Vol 120, No 6, June 2007
ardiac events, respectively, independent of clinical riskactors and baseline electrocardiography.
urrent Recommendationshe prognostic value of dobutamine stress echocardiogra-hy has previously been demonstrated in patients with sus-
Table 3 Univariate and Multivariate Predictors of All-cause M
UnivariateHR (95% CI)Characteristic Mo
Clinical characteristicsAge (per year increase) 1.06 (1.04-1.07) 1.0Male sex 1.38 (1.05-1.80) 1.2Previous myocardial infarction 1.53 (1.21-1.93) 1.3Congestive heart failure 2.35 (1.64-3.37) 1.8History of stroke or TIA 1.52 (1.01-2.29) 1.1Diabetes mellitus 1.40 (1.03-1.90) 1.4Hypercholesterolemia 0.76 (0.55-1.05) 0.9Hypertension 1.09 (0.86-1.39) 1.0Cigarette smoking 1.22 (0.96-1.55) 1.2Renal failure 4.59 (3.12-6.76) 3.1
Ischemic baseline electrocardiogram 1.45 (1.13-1.85) 1.6ABI (per 0.10 decrease) 1.14 (1.09-1.20) 1.0Stress test results —
LVEF (per 5% decrease) 1.07 (1.03-1.11)Angina pectoris during test 0.98 (0.65-1.48)ST changes during test 1.26 (0.89-1.79)New wall motion abnormalities 2.20 (1.57-3.08)
Global �2 78Incremental value
TIA � transient ischemic attack; ABI � ankle-brachial index; LVEF �
Table 4 Univariate and Multivariate Predictors of the Compos
CharacteristicUnivariateHR (95% CI) Mo
Clinical characteristicsAge (per year increase) 1.03 (1.02-1.04) 1.0Male sex 1.46 (1.07-2.01) 1.2Previous myocardial infarction 2.04 (1.56-2.67) 1.5History of heart failure 2.41 (1.58-3.66) 1.9History of stroke or TIA 1.17 (0.68-2.02) 1.0Diabetes mellitus 2.02 (1.46-2.79) 1.6Hypercholesterolemia 1.37 (1.01-1.88) 1.3Hypertension 1.32 (1.00-1.71) 1.0Cigarette smoking 0.97 (0.73-1.30) 1.0Renal failure 5.40 (3.45-8.45) 3.2
Ischemic baseline electrocardiogram 1.71 (1.30-2.25) 1.6ABI (per 0.10 decrease) 1.15 (1.08-1.21) 1.0Stress test results —
LVEF (per 5% decrease) 1.17 (1.13-1.25)Angina pectoris during test 0.87 (0.53-1.41)ST changes during test 1.42 (0.96-2.11)
New wall motion abnormalities 2.34 (1.58-3.47)Global �2 53Incremental value
TIA � transient ischemic attack; ABI � ankle-brachial index; LVEF �
ected or known coronary artery disease.27 A major findingn our study is that dobutamine stress echocardiographymproves risk stratification in patients with suspected ornown peripheral arterial disease, in addition to ankle-bra-hial index values and clinical risk factors. Each 5% de-rease in left ventricular ejection fraction was associated
Multivariate
R (95% CI) Model 2 HR (95% CI) Model 3 HR (95% CI)
3-1.07) 1.04 (1.02-1.06) 1.04 (1.02-1.06)5-1.66) 1.21 (0.84-1.73) 1.25 (0.60-2.57)9-1.69) 1.15 (0.82-1.62) 1.08 (0.57-2.04)2-2.88) 1.60 (0.99-2.41) 1.25 (0.48-3.24)7-1.78) 1.15 (0.72-1.84) 1.48 (0.49-4.52)9-2.05) 1.31 (0.87-1.95) 1.05 (0.50-2.05)7-1.31) 0.90 (0.61-1.34) 0.96 (0.46-2.03)8-1.27) 0.98 (0.72-1.33) 1.04 (0.56-1.94)8-1.60) 1.19 (0.88-1.61) 1.49 (0.81-2.75)0-5.55) 2.86 (1.63-5.01) 2.63 (1.50-4.62)1-2.22) 1.50 (1.10-2.04) 1.48 (1.08-2.01)1-1.12) 1.05 (1.01-1.11) 1.05 (1.01-1.11)
1.06 (1.02-1.11) 1.05 (1.02-1.09)— 1.01 (0.47-1.89)
1.05 (0.67-2.01)2.01 (1.38-2.79)
85 102P � .01 P �.001
tricular ejection fraction.
point of Cardiac Death or Non-fatal Myocardial Infarction
Multivariate
R (95% CI) Model 2 HR (95% CI) Model 3 HR (95% CI)
1-1.05) 1.03 (1.00-1.05) 1.03 (1.00-1.05)9-1.71) 1.09 (0.71-1.68) 0.67 (0.29-1.55)7-2.37) 1.32 (0.79-2.23) 1.37 (0.60-3.10)0-3.30) 1.70 (0.95-3.45) 1.31 (0.46-3.78)5-1.95) 1.00 (0.42-2.01) 1.03 (0.40-2.15)8-2.47) 1.52 (1.03-2.20) 1.73 (0.77-3.92)9-1.94) 1.25 (0.84-1.88) 0.95 (0.43-2.11)1-1.44) 1.11 (0.77-1.59) 1.64 (0.81-3.32)6-1.38) 1.06 (0.73-1.53) 1.03 (0.50-2.13)0-6.46) 2.87 (1.43-5.78) 2.92 (1.45-5.90)6-2.33) 1.45 (0.95-2.57) 1.40 (0.80-2.97)0-1.13) 1.03 (0.98-1.10) 1.03 (0.97-1.11)
1.15 (1.09-1.22) 1.14 (1.08-1.21)— 1.13 (0.46-2.76)
1.76 (0.74-4.19)2.06 (1.39-3.08)
80 93P �.001 P �.001
tricular ejection fraction.
ortality
del 1 H
5 (1.06 (0.92 (0.97 (1.27 (0.79 (1.04 (0.61 (0.75 (0.96 (1.84 (1.26 (1.0
left ven
ite End
del 1 H
3 (1.04 (0.81 (0.99 (1.25 (0.53 (1.08 (0.98 (0.82 (0.71 (1.65 (1.17 (1.0
left ven
whmoceoHgrioddtpackc
WEAawtoticmfwtaccwtpccarcefcpetpie
CTedpsrarapt
R
1
1
1
1
537Feringa et al Cardiac Testing in Peripheral Arterial Disease
ith a 5% and 14% increased risk of all-cause mortality andard cardiac events, respectively. Stress-induced new wallotion abnormalities during dobutamine stress echocardi-
graphy were associated with a 2.0-fold and 2.1-fold in-reased risk, respectively. Furthermore, the extent of isch-mia also was significantly associated with adverseutcome. The American College of Cardiology/Americaneart Association has provided useful evidence-baseduidelines regarding the management of patients with pe-ipheral arterial disease and accentuate the frequent coex-stence of coronary artery disease in these patients. Becausef the paucity of published data, recommendations for car-iac risk assessment are limited. An improvement in theetection of (sub) clinical coronary artery disease and iden-ification of those at increased risk for adverse events ap-ears essential in order to obtain reductions in morbiditynd mortality. The implication of dobutamine stress echo-ardiography in the work-up of patients with suspected ornown peripheral arterial disease should therefore beonsidered.
hen Should Dobutamine Stresschocardiography Be Performed?lthough dobutamine stress echocardiography is a safe and
ccurate noninvasive procedure, the question remainshether all patients should undergo routine screening with
his technique. This will probably depend on the availabilityf dobutamine stress echocardiography in the clinical prac-ice, the presence and local expertise of assessing and scor-ng echocardiographic images, and the consideration ofosts versus benefit. The current results demonstrate thatale patients with coronary artery disease, history of heart
ailure, diabetes, renal failure, and hypercholesterolemiaere more likely to present with left ventricular dysfunc-
ion. A subgroup of patients with advanced age, lowernkle-brachial index values, and ischemic baseline electro-ardiography were more likely to have stress-induced myo-ardial ischemia. Of note, lower ankle-brachial index valuesere directly associated with the severity of new wall mo-
ion abnormalities during dobutamine stress echocardiogra-hy. A limitation that should be addressed is that our studyohort consisted of patients referred to a tertiary referralenter with specific indications for stress echocardiography,nd that it may not reflect “real-world” patients with pe-ipheral arterial disease. Furthermore, it remains to be elu-idated whether other noninvasive modalities, such asxercise electrocardiographic testing and myocardial per-usion imaging are superior to dobutamine stress echo-ardiography for prognostic cardiac risk assessment inatients with suspected or known peripheral arterial dis-ase. Exercise electrocardiography may not be suitable asest of screening for coronary artery disease, becauseatients with peripheral arterial disease often have lim-ted exercise capacity and often present with baseline
lectrocardiographic abnormalities.ONCLUSIONhe current results reveal that decreased left ventricularjection fraction and stress-induced myocardial ischemiauring dobutamine stress echocardiography independentlyredict long-term outcome and improve prognostic risk as-essment in addition to ankle-brachial index and clinicalisk factors in patients with suspected or known peripheralrterial disease. The use of dobutamine stress echocardiog-aphy makes it possible to detect left ventricular dysfunctionnd coronary artery disease in an early stage, to identifyatients at increased risk for future adverse events, andherefore to implement adequate preventive interventions.
eferences1. Hirsch AT, Haskal ZJ, Hertzer NR, et al. ACC/AHA 2005 Practice
Guidelines for the management of patients with peripheral arterialdisease (lower extremity, renal, mesenteric, and abdominal aortic): acollaborative report from the American Association for Vascular Sur-gery/Society for Vascular Surgery, Society for Cardiovascular An-giography and Interventions, Society for Vascular Medicine and Bi-ology, Society of Interventional Radiology, and the ACC/AHA TaskForce on Practice Guidelines (Writing Committee to Develop Guide-lines for the Management of Patients With Peripheral Arterial Dis-ease): endorsed by the American Association of Cardiovascular andPulmonary Rehabilitation; National Heart, Lung, and Blood Institute;Society for Vascular Nursing; TransAtlantic Inter-Society Consensus;and Vascular Disease Foundation. Circulation. 2006;113:e463-e654.
2. Murabito JM, Evans JC, Larson MG, et al. The ankle-brachial index inthe elderly and risk of stroke, coronary disease, and death: the Fra-mingham Study. Arch Intern Med. 2003;163:1939-1942.
3. Criqui MH, Fronek A, Barrett-Connor E, et al. The prevalence ofperipheral arterial disease in a defined population. Circulation. 1985;71:510-515.
4. McDermott MM, Kerwin DR, Liu K, et al. Prevalence and significanceof unrecognized lower extremity peripheral arterial disease in generalmedicine practice. J Gen Intern Med. 2001;16:384-390.
5. Selvin E, Erlinger TP. Prevalence of and risk factors for peripheral arterialdisease in the United States: results from the National Health and Nutri-tion Examination Survey, 1999-2000. Circulation. 2004;110:738-743.
6. Hirsch AT, Criqui MH, Treat-Jacobson D, et al. Peripheral arterialdisease detection, awareness, and treatment in primary care. JAMA.2001;286:1317-1324.
7. Hiatt WR. Medical treatment of peripheral arterial disease and clau-dication. N Engl J Med. 2001;344:1608-1621.
8. Ward RP, Min JK, McDonough KM, Lang RM. High prevalence ofimportant cardiac findings in patients with peripheral arterial diseasereferred for echocardiography. J Am Soc Echocardiogr. 2005;18:844-849.
9. Valentine RJ, Grayburn PA, Eichhorn EJ, et al. Coronary artery dis-ease is highly prevalent among patients with premature peripheralvascular disease. J Vasc Surg. 1994;19:668-674.
0. Park H, Das M, Aronow WS, et al. Relation of decreased ankle-brachial index of prevalence of atherosclerotic risk factors, coronaryartery disease, aortic valve calcium, and mitral annular calcium. Am JCardiol. 2005;95:1005-1006.
1. McKenna M, Wolfson S, Kuller L. The ratio of ankle and arm arterialpressure as an independent predictor of mortality. Atherosclerosis.1991;87:119-128.
2. Newman AB, Siscovick DS, Manolio TA, et al. Ankle-arm index as amarker of atherosclerosis in the Cardiovascular Health Study. Cardio-vascular Heart Study (CHS) Collaborative Research Group. Circula-tion. 1993;88:837-845.
3. Newman AB, Sutton-Tyrrell K, Vogt MT, Kuller LH. Morbidity andmortality in hypertensive adults with a low ankle/arm blood pressure
index. JAMA. 1993;270:487-489.
1
1
1
1
11
2
2
2
2
2
2
2
2
538 The American Journal of Medicine, Vol 120, No 6, June 2007
4. Lee AJ, Price JF, Russell MJ, et al. Improved prediction of fatalmyocardial infarction using the ankle brachial index in addition toconventional risk factors: the Edinburgh Artery Study. Circulation.2004;110:3075-3080.
5. Criqui MH, Langer RD, Fronek A, et al. Mortality over a period of 10years in patients with peripheral arterial disease. N Engl J Med.1992;326:381-386.
6. Vogt MT, Cauley JA, Newman AB, et al. Decreased ankle/arm bloodpressure index and mortality in elderly women. JAMA. 1993;270:465-469.
7. Leng GC, Fowkes FG, Lee AJ, et al. Use of ankle brachial pressureindex to predict cardiovascular events and death: a cohort study. BMJ.1996;313:1440-1444.
8. Marwick TH. Stress echocardiography. Heart. 2003;89:113-118.9. Poldermans D, Fioretti PM, Boersma E, et al. Long-term prognostic
value of dobutamine-atropine stress echocardiography in 1737 patientswith known or suspected coronary artery disease: a single-centerexperience. Circulation. 1999;99:757-762.
0. McNeill AJ, Fioretti PM, el-Said SM, et al. Enhanced sensitivity fordetection of coronary artery disease by addition of atropine to dobut-amine stress echocardiography. Am J Cardiol. 1992;70:41-46.
1. Armstrong WF, Pellikka PA, Ryan T, et al. Stress echocardiography:recommendations for performance and interpretation of stress echo-
cardiography. Stress Echocardiography Task Force of the Nomencla-ture and Standards Committee of the American Society ofEchocardiography. J Am Soc Echocardiogr. 1998;11:97-104.
2. Edwards WD, Tajik AJ, Seward JB. Standardized nomenclature andanatomic basis for regional tomographic analysis of the heart. MayoClin Proc. 1981;56:479-497.
3. Bourdillon PD, Broderick TM, Sawada SG, et al. Regional wall mo-tion index for infarct and noninfarct regions after reperfusion in acutemyocardial infarction: comparison with global wall motion index.J Am Soc Echocardiogr. 1989;2:398-407.
4. Viles-Gonzalez JF, Fuster V, Badimon JJ. Atherothrombosis: a wide-spread disease with unpredictable and life-threatening consequences.Eur Heart J. 2004;25:1197-1207.
5. de Graaff JC, Ubbink DT, Kools EI, et al. The impact of peripheral andcoronary artery disease on health-related quality of life. Ann VascSurg. 2002;16:495-500.
6. Criqui MH, Langer RD, Fronek A, et al. Mortality over a period of 10years in patients with peripheral arterial disease. N Engl J Med.1992;326:381-386.
7. Poldermans D, Fioretti PM, Boersma E, et al. Long-term prognosticvalue of dobutamine-atropine stress echocardiography in 1737 patientswith known or suspected coronary artery disease: a single-center
experience. Circulation. 1999;99:757-762.
C
THJRa
PSe
Q
P
u
0d
The American Journal of Medicine (2007) 120, 539-544
LINICAL RESEARCH STUDY
he Safety of Warfarin Therapy in the Nursingome Setting
erry H. Gurwitz, MD,a,b Terry S. Field, DSc,a Martha J. Radford, MD,c Leslie R. Harrold, MD, MPH,a
ichard Becker, MD,d George Reed, PhD,a Kristin DeBellis, PharmD,a Jason Moldoff, BA,a Nancy Verzier, MSN, RNe
Meyers Primary Care Institute, University of Massachusetts Medical School, Fallon Clinic Foundation, and Fallon Community Healthlan, Worcester, Mass; bDivision of Geriatric Medicine, University of Massachusetts Medical School, Worcester; cNew York Universitychool of Medicine, New York; dDivisions of Hematology and Cardiology, Duke University School of Medicine, Durham, NC;
Qualidigm, Inc,PwMrfPrtoPraRac2ecmtCphm
massmed.edu.
002-9343/$ -see foi:10.1016/j.amjm
Middletown, Conn.
ABSTRACT
URPOSE: We examined the preventability of adverse warfarin-related events and potential adversearfarin-related events (“near misses”) in the nursing home setting.ETHODS: We performed a cohort study of all long-term care residents of 25 nursing homes (bed size
ange, 90-360) in Connecticut during a 12-month observation period. The total number of residents in theseacilities ranged from 2946 to 3212 per quarter. There were 490 residents who received warfarin therapy.ossible warfarin-related incidents were detected by quarterly retrospective review of nursing homeecords by trained nurse abstractors. Each incident was independently classified by 2 physician-reviewerso determine whether it constituted a warfarin-related event, its severity, and its preventability. The primaryutcome was an adverse warfarin-related event, defined as an injury associated with the use of warfarin.otential adverse warfarin-related events were defined as situations in which the international normalizedatio (INR) was noted to be 4.5 or greater, an error in management was noted, but no injury occurred. Welso assessed time in specified INR ranges per nursing home resident day on warfarin.ESULTS: Over the 12-month observation period, 720 adverse warfarin-related events and 253 potentialdverse warfarin-related events were identified. Of the adverse warfarin-related events, 625 (87%) wereharacterized as minor, 82 (11%) were deemed serious, and 13 (2%) were life-threatening or fatal. Overall,9% of the adverse warfarin-related events were judged to be preventable. Serious, life-threatening, or fatalvents occurred at a rate of 2.49 per 100 resident-months; 57% of these more severe events wereonsidered preventable. Errors resulting in preventable events occurred most often at the prescribing andonitoring stages of warfarin management. The percentages of time in the less than 2, 2 to 3, and more
han 3 INR ranges were 36.5%, 49.6%, and 13.9%, respectively.ONCLUSIONS: The use of warfarin in the nursing home setting presents substantial safety concerns foratients. Adverse events associated with warfarin therapy are common and often preventable in the nursingome setting. Prevention strategies should target the prescribing and monitoring stages of warfarinanagement. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Long-term care; Medication safety; Nursing homes; Oral anticoagulants; Warfarin
Cwbmmhw
Supported by a grant from the Agency for Healthcare Research anduality (PO1HS11534), Rockville, Maryland.
Requests for reprints should be addressed to Jerry H. Gurwitz, MD, Meyersrimary Care Institute, 630 Plantation Street, Worcester, MA 01605.
E-mail address: [email protected] or jerry.gurwitz@
sront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.07.045
oncerns relating to the risks of anticoagulation therapyith warfarin are intensified in the long-term care setting,ecause elderly residents of nursing homes are among theost frail patients in the geriatric population. Given thatore than 1.6 million Americans currently reside in nursing
omes and the prevalence of medical conditions for whicharfarin is indicated increases with advancing age, issues
urrounding the management of anticoagulation therapy in
taaitrslttpcahFufmaf
stapstt
M
STtpr(wsU
fccntr
CTrt3frnc
tegsp
rweiwettc
pd
540 The American Journal of Medicine, Vol 120, No 6, June 2007
his population require particular attention.1-3 Factors suchs aging, comorbid conditions, diet, and medications canffect the pharmacodynamics of warfarin, thus making dos-ng and monitoring more challenging with advancing pa-ient age.4 As many as 12% of nursing home residentseceive warfarin for various rea-ons, but the quality of anticoagu-ation care in this setting is subop-imal.1,2 Although oral directhrombin inhibitors hold someromise as less-complicated anti-oagulants, ximelagatran has beenssociated with hepatotoxicity andas not received approval by theood and Drug Administration forse in the United States.5 For theoreseeable future, warfarin re-ains the sole oral anticoagulant
vailable for long-term use in therail elderly.
In prior work examining thecope of adverse drug events inhe long-term care setting, we found that warfarin wasmong the drug types most commonly implicated.6,7 Theurpose of the present study was to carefully assess theafety of anticoagulation therapy with warfarin in the long-erm care setting. A particular focus was placed on assessinghe preventability of adverse warfarin-related events.
ETHODS
tudy Settinghis study was conducted in 25 nursing homes located in
he state of Connecticut. Data were collected during theeriod from April 1, 2003, to March 31, 2004. Patientsesiding in areas of the facilities related to short-term careeg, subacute care, hospital-level care, or rehabilitation)ere not included as potential subjects in the study. The
tudy was approved by the institutional review board of theniversity of Massachusetts Medical School.The bed sizes of the participating nursing homes ranged
rom 90 to 360 beds. Table 1 provides a comparison of theharacteristics of the participating nursing homes with fa-ilities in Connecticut, as well as nationally.8 The totalumber of residents of these 25 facilities, which comprisedhe underlying study population, varied from 2946 to 3212esidents over the 1-year study period.
ase-Finding Approachhe study was limited to warfarin-related incidents occur-
ing in the nursing home setting. Incidents were detectedhrough retrospective review of nursing home records in-month segments performed by trained nurse abstractorsor each eligible resident of the nursing home who waseceiving warfarin at any time during that time period. Theurse abstractors searched for possible warfarin-related in-
CLINICAL SIGNIF
● Adverse events atherapy are comable in the nursi
● Specialized antinursing homes acation protocolsstaff and prescriduce the risk ofevents.
idents using a variety of specified triggers, including in-
ernational normalized ratio (INR) values of 4.5 or greater,vidence of overt bleeding, a hematocrit decrease of 3% orreater from a previous value, an order for vitamin K, atroke or other thrombotic event, and any emergency de-artment visit or hospitalization.
Outcome MeasuresThe primary outcome was an ad-verse warfarin-related event, de-fined as an injury resulting fromthe use of warfarin. This definitionis consistent with definitions usedin previous studies of adversedrug events.6,7,9-12 Adverse warfa-rin-related events may have re-sulted from medication errors (eg,errors in ordering, dispensing, ad-ministration, or monitoring) ormay have occurred in the absenceof any error. We also identifiedpotential adverse warfarin-relatedevents. Potential adverse warfarin-
elated events were defined as situations in which the INRas noted to be 4.5 or greater and no injury occurred, but an
rror in warfarin management was identified. Thus, accord-ng to this study definition, not all INR values 4.5 or moreere considered to be potential adverse warfarin-related
vents. The 4.5 threshold INR level was chosen for use inhis study because various investigations have demonstratedhat the risk of intracranial hemorrhage increases dramati-ally at INRs at or above this point.13,14
Warfarin-related incidents were presented to pairs ofhysician-reviewers (J.H.G., L.R.H., M.J.R., R.B.) who in-ependently classified incidents using structured implicit
CE
ated with warfarinnd often prevent-me setting.
lation services inproved communi-een nursing homephysicians may re-rin-related adverse
Table 1 Comparison of Study Nursing Homes withFacilities in Connecticut and the United States
StudyNursingHomesn � 25 (%)
Connecticutn � 248 (%)
Nationaln � 18,000 (%)
No. of beds 3411 30,504 1,879,600Mean (�SD) 136 (�28.7) 123 (�58.6) 104�50 beds 0 (0) 12 (5) 2100 (12)50-99 beds 1 (4) 79 (32) 7000 (39)100-199 beds 23 (92) 137 (55) 7500 (42)200� beds 1 (4) 20 (8) 1400 (8)
OwnershipProfit 16 (64) 196 (79) 12,000 (67)Nonprofit 8 (32) 50 (20) 4800 (27)Government 1 (4) 2 (1) 1200 (7)
CertificationMedicare and
Medicaid25 (100) 241 (97) 14,700 (82)
AffiliationChain 9 (36) 103 (42) 10,800 (60)Independent 16 (64) 145 (58) 7200 (40)
ICAN
ssocimon ang ho
coagund imbetw
bingwarfa
rvrttuo
sasawSanvabda
wtbpaipa
isardif
eddceqalroptw
fiwl
riilkafsa0
STwlhcemah
2wlpw(vdiwtbD(c
RAr8sTcmna
RPTfsp
541Gurwitz et al Safety of Warfarin Therapy
eview according to the following criteria: whether an ad-erse warfarin-related event or a potential adverse warfarin-elated event was present, the severity of the event, whetherhe event was preventable, and the effects of the event onhe patient. This structured implicit review process has beensed in prior studies of adverse drug events across a numberf clinical settings including the nursing home.6,7,9,10,12,15-17
The severity of adverse warfarin-related events was clas-ified according to modification of the criteria of Landefeldnd colleagues,18 as used by White and colleagues.19 Theeverity of adverse warfarin-related events was categorizeds minor, serious, life-threatening, or fatal. Minor eventsere those with no medical consequence (eg, bruising).erious events were those that required specific treatment ormedical evaluation. Life-threatening events included the
eed for a surgical intervention to stop the bleeding, irre-ersible sequelae (eg, myocardial infarction or stroke), orny 2 of the following: transfusion of 3 or more units oflood, hypotension, critical anemia, or acute bleeding (�3ays). We also sought to identify all thromboembolic eventsssociated with subtherapeutic INR values (INR � 2).
Adverse warfarin-related events and potential adversearfarin-related events were considered to be preventable if
hey were judged to be due to an error and were preventabley any means available. “Preventability” was categorized asreventable, probably preventable, probably not prevent-ble, or definitely not preventable; results were collapsednto the categories of preventable (preventable and probablyreventable) and nonpreventable (probably not preventablend definitely not preventable) in the analysis.
The effects of adverse warfarin-related events on nurs-ng home residents were categorized as up to 1 day ofymptoms, more days of symptoms, nonpermanent dis-bility, permanent disability, and death. The physician-eviewers categorized an event as causing permanentisability based on evidence that a warfarin-associatednjury had caused physical disability or deficits inunctioning.20
The stages of warfarin management during which anrror leading to a preventable event occurred were ordering,ispensing, administration, and monitoring. If an error waseemed to have occurred, the physician reviewers furtherharacterized the error type according to the following cat-gories: wrong dose, wrong drug, missed dose, wrong fre-uency, extra dose, wrong resident, and known drug inter-ction. Monitoring stage errors included inadequateaboratory monitoring of warfarin therapy or a delayedesponse or failure to respond to laboratory evidence of anut-of-range INR value. For a single event, it was oftenossible to identify errors at more than 1 stage of warfarinherapeutic management or to identify more than 1 errorithin a single stage of management.When the 2 physician-reviewers disagreed on the classi-
cation of an incident regarding the presence of an adversearfarin-related event or a potential adverse warfarin-re-
ated event, its severity, or its preventability, they met and r
eached consensus; consensus was reached in all instancesn which there was initial disagreement. We compared thenitial assessments of the physician-reviewers and calcu-ated interrater reliability using the kappa statistic, withappa � 0.59 for judgments regarding the presence of andverse warfarin-related event or a potential adverse war-arin-related event, 0.58 for preventability, and 0.67 foreverity. A kappa score of 0.4 to 0.6 reflects “moderategreement,” 0.6 to 0.8 reflects “substantial agreement,” and.8 to 1.0 is considered “almost perfect.”21
tatistical Analysiso determine crude rates of events, the numbers of adversearfarin-related events and potential adverse warfarin-re-
ated events were divided by the total number of nursingome resident-months, which was estimated (with 95%onfidence interval [CI])22 by obtaining census data for allligible residents on warfarin in the study nursing homes atonthly intervals throughout the course of the study; we
ccounted for absences from the nursing homes (eg, forospitalization) and breaks in warfarin use.
We also assessed the time in specified INR ranges (�2,-3, �3-�4.5, �4.5) per nursing home resident-day onarfarin. INR values for each day were estimated using
inear interpolation between values.23 We excluded anyeriods from our analyses when a nursing home residentas not taking warfarin or not present in the nursing home
eg, during a hospital stay). We stopped interpolating INRalues at the last INR before 1 of these excluded periods andid not restart interpolating until the first INR value follow-ng that period. Overall, only 2.8% of total time for allarfarin-treated nursing home residents was excluded from
he analysis to determine time in specified INR rangesecause of stops and starts of therapy or hospitalizations.ata were available to perform these calculations for 479
97.8%) of the 490 nursing home residents who were in-luded in the study.
ESULTScross the 25 study nursing homes, 490 nursing home
esidents who were receiving warfarin therapy (mean age,2.3 � 10.0 years) yielded 3822.9 resident-months of ob-ervation time; 70% were female and 90% were white.he most common indications for warfarin therapy in-luded stroke prevention in atrial fibrillation (58%), treat-ent/prevention of deep venous thrombosis or pulmo-
ary embolism (26%), and stroke prevention withouttrial fibrillation (12%).
ates of Adverse Warfarin-Related Events andotential Adverse Warfarin-Related Eventshe trained nurse abstractors identified 1501 possible war-
arin-related incidents, of which 720 were judged to repre-ent adverse warfarin-related events and 253 were deemedotential adverse warfarin-related events by the physician-
eviewers. Of the 720 adverse warfarin-related events, 29%
(wd
p(trp2a(v1r
lmme(fdtvt1
TWEtcwElactTiwt
boee
EWAAaoamisatgwrp(
542 The American Journal of Medicine, Vol 120, No 6, June 2007
n � 207) were judged preventable. All potential adversearfarin-related events were deemed preventable, byefinition.
The overall rate of adverse warfarin-related events andotential adverse warfarin-related events combinedn � 973) was 25.5 per 100 resident-months on warfarinherapy (95% CI, 23.9-27.1 per 100 resident-months). Theate of adverse warfarin-related events (n � 720) was 18.8er 100 resident-months on warfarin therapy (95% CI, 17.5-0.3 per 100 resident-months), with a rate of 5.4 preventabledverse warfarin-related events per 100 resident-months95% CI, 4.7-6.2 per 100 resident-months). Potential ad-erse warfarin-related events occurred at a rate of 6.6 per00 resident-months on warfarin (95% CI, 5.8-7.5 per 100esident-months).
Serious, life-threatening, or fatal adverse warfarin-re-ated events occurred at a rate of 2.5 per 100 resident-onths on warfarin (95% CI, 2.0-3.0 per 100 resident-onths). Preventable serious, life-threatening, or fatal
vents occurred at a rate of 1.4 per 100 resident-months95% CI, 1.1-1.8). Of the 95 serious, life-threatening, oratal adverse warfarin-related events, 57% (n � 54) wereeemed preventable, compared with 24% (n � 153) ofhe 625 minor events (Table 2). Overall, more severe ad-erse warfarin-related events were significantly more likelyo be considered preventable (relative risk � 2.3; 95% CI,.9-2.9).
ypes and Effects of Adversearfarin-Related Events
cchymoses, gross hematuria, overt and occult gastrointes-inal bleeding, epistaxis, and microhematuria were the mostommon types of bleeding events (Table 3). Most adversearfarin-related events lasted more than 1 day (Table 4).ight events resulted in permanent disability or death. Of 13
ife-threatening or fatal events, 11 were considered prevent-ble. Four preventable life-threatening or fatal events oc-urred in the setting of supertherapeutic INR levels: 2 gas-rointestinal bleeds, 1 gluteal bleed, and 1 intracranial bleed.wo gastrointestinal bleeds that were deemed life threaten-
ng or fatal were associated with the use of concurrentarfarin, aspirin, and nonsteroidal anti-inflammatory drug
herapy. Preventable life-threatening or fatal thromboem-
Table 2 Severity and Effects of Adverse Warfarin-RelatedEvents
No. (%)
Total(n � 720)
Preventable(n � 207)
Category of severityFatal 5 (1) 3 (1)Life threatening 8 (1) 8 (4)Serious 82 (11) 43 (21)Minor 625 (87) 153 (74)
olic events included 5 events that occurred in the presencef subtherapeutic INR values, including stroke, systemicmbolus, probable mechanical valve thrombosis, pulmonarymbolism, and deep venous thrombosis.
rrors Associated with Preventable Adversearfarin-Related Events and Potentialdverse Warfarin-Related Eventsmong the 207 preventable adverse warfarin-related events
nd the 253 potential adverse warfarin-related events, errorsccurred most commonly at the prescribing (n � 321, 70%)nd monitoring (n � 424, 92%) stages of warfarin manage-ent. Errors accounting for preventable events were rarely
dentified at the dispensing (n � 2) or administration (n � 2)tages. A total of 285 (62%) of the preventable events weressociated with an error at both the prescribing and moni-oring stages of warfarin management. Monitoring errorsenerally referred to inadequate laboratory monitoring ofarfarin therapy or to a delayed response, or a failure to
espond to laboratory results (ie, INR values). Among therescribing errors, the most common were wrong dosen � 259, 81%) and known drug interaction (n � 80, 25%).
Table 3 Types of Bleeding*
Type No. (%)
Ecchymoses 461 (64)Gross hematuria 72 (10)Gastrointestinal (occult) 47 (7)Epistaxis 44 (6)Gastrointestinal (overt) 38 (5)Microhematuria 23 (3)Vaginal 13 (2)Hemoptysis 12 (2)Gingival 9 (1)Subconjunctival 4 (1)Ear 3 (�1)Ocular 2 (�1)Gluteal 1 (�1)Intracranial 1 (�1)
*Adverse warfarin-related events could manifest as more than 1type of bleeding.
Table 4 Effects of Adverse Warfarin-Related Events
No. (%)
Total(n � 720)
Preventable(n � 207)
�1 d of symptoms 251 (35) 65 (31)� 1 d of symptoms 460 (64) 135 (65)Nonpermanent disability 1 (�1) 1 (�1)Permanent disability and death 8 (1) 6 (3)
TRWs31
DRh1mmwavdf
hptrddenicnatbtc
uwctnwtsihoalu
vchrd
nmcsi
pofinlptnbaost
bwawtaeeovafab
tnrh(spfSitrwheat
wi
543Gurwitz et al Safety of Warfarin Therapy
ime in Specified International Normalizedatio Rangeshen all days on warfarin therapy summed across the entire
tudy population were considered, the INR value was �2 in6.5%, 2 to 3 in 49.6%, more than 3 to less than 4.5 in1.9%, and 4.5 or more in 2.0%.
ISCUSSIONesidents of nursing homes who are receiving warfarinave adverse warfarin-related events at high rates (18.8 per00 resident months); approximately 30% of these eventsay be preventable (ie, associated with an error in manage-ent). Of fatal, life-threatening, and serious events, 57%ere considered preventable. Errors in warfarin prescribing
nd monitoring were responsible for nearly all of the pre-entable events. Consistent with previous studies, this studyemonstrates that nursing home residents on warfarin arerequently maintained outside the optimal therapeutic range.
The system of medication management in the nursingome includes the nursing staff within the facility, and thehysicians, laboratories, and pharmacy vendors external tohe nursing home who interact to provide services to theesidents. Although an adverse event in this setting may beirectly linked to a “human error,” the root cause may beefined as the defect in the system that permitted such anrror to occur. In the case of warfarin management forursing home residents, an important root cause is poornformation flow. For example, a frequent occurrence in theare of nursing home residents is a telephone call from theursing home to a covering physician about a resident withurinary tract infection, without noting that the resident is
aking warfarin.24 The result may be an order for an anti-iotic that interacts with warfarin, without adequate moni-oring resulting in a supertherapeutic INR level and in-reased risk of bleeding.
Although our previous work suggested that there may benderuse of warfarin among apparently ideal candidatesho are residents of long-term care facilities,1 warfarin
urrently is commonly prescribed in the nursing home set-ing, with an estimated 200,000 nursing home residentsationwide receiving this therapy at any point in time.6 Fewould argue against the need for a more consistent approach
o managing anticoagulant therapy in the nursing homeetting. Although more widespread use of specialized clin-cs for anticoagulation therapy to provide coordinated careas been promoted to improve the effectiveness and safetyf warfarin in elderly patients,25 to date, the benefits of thispproach relative to usual care have not been firmly estab-ished,26 and specialized anticoagulation services are rarelysed in the nursing home setting.27
Leading-edge, high technology-based strategies to alle-iate problems in prescribing and monitoring warfarin areurrently less amenable to incorporation into most nursingome settings.28 The information technology infrastructureequired to support computerized physician order entry with
ecision support is almost nonexistent in the majority of US sursing homes. Few long-term care facilities have imple-ented such systems owing to cost, complexity, and logistic
hallenges, as well as uncertainty about how effective theystems actually are for reducing drug-related injuries oncemplemented.
We believe that the findings of this study provide com-elling evidence of serious safety concerns around the usef warfarin therapy in the nursing home setting. If ourndings are generalized to residents on warfarin in all USursing homes, there may be approximately 34,000 fatal,ife-threatening, or serious adverse warfarin-related eventser year, of which the majority may be preventable. Fur-hermore, “near misses” are common. Many residents ofursing homes on warfarin are subjected to a high risk ofleeding because of high INR levels that are associated withn error in warfarin management. Nursing home residentsn warfarin also spend considerable amounts of time in theubtherapeutic range, potentially reducing the benefits ofherapy.29
Management of warfarin therapy tends to be substantiallyetter in the context of randomized clinical trials comparedith usual care situations. For example, in SPORTIF V,30
mong patients assigned to receive warfarin, INR valuesere within the 2.0 to 3.0 range 68% of the time and less
han 2.0 only 20% of the time. In our study, these percent-ges were 50% and 37%, respectively. It is important tomphasize that a number of the fatal and life-threateningvents identified in this study were thromboembolic eventsccurring when the INR level was subtherapeutic. The risksersus the benefits of therapy must be carefully assessed inny patient initiated on warfarin therapy, particularly in therail nursing home resident. When such therapy is deemedppropriate, all efforts must be made to provide the fullenefits of therapy, while limiting the risks.
Our study has a number of limitations. Foremost amonghem was our reliance solely on information contained inursing home records to assess the occurrence of warfarin-elated incidents. In randomized controlled trials, compre-ensive ascertainment and careful assessment of end pointsie, thromboembolic events and bleeds) are the priority;ystematic approaches are used to enhance detection (eg,eriodic administration of stroke-symptom questionnairesollowed up by direct clinical assessment of the patient).uch approaches were obviously not possible in our study;
n our study, we relied on information that could be ascer-ained solely through retrospective review of nursing homeecords. Our priority in this study was to assess the safety ofarfarin therapy in the nursing home setting by describingow errors in warfarin management contribute to adversevents and “near misses.” Our study was not designed tossess the effectiveness of warfarin therapy for the preven-ion of thromboembolic events.
Intensified educational efforts concerning the safe use ofarfarin in the nursing home setting are essential. Even if it
s not possible to implement systems-level changes using
pecialized anticoagulation services or computerized physi-
ciisnmbtcammavm
R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
544 The American Journal of Medicine, Vol 120, No 6, June 2007
ian order entry with clinical decision support, efforts tomprove the effectiveness of communication between nurs-ng staff and physicians around the use of warfarin therapyhould be implemented. Physicians spend little time in theursing home setting, and therapeutic decision making com-only occurs over the telephone during brief conversations
etween doctor and nurse. Protocols must be developed andested to provide accurate information to the prescriberoncerning prior INR levels, warfarin dosing information,nd interacting medications. If successful, such an approachay serve as a model for improving the safety of otheredication categories associated with high rates of prevent-
ble adverse drug events and serve to further protect theulnerable nursing home resident who is at special risk foredication-related problems.
eferences1. McCormick D, Gurwitz JH, Goldberg RJ, et al. Prevalence and quality
of warfarin use for patients with atrial fibrillation in the long-term caresetting. Arch Intern Med. 2001;161:2458-2463.
2. Gurwitz JH, Monette J, Rochon PA, Eckler MA, Avorn J. Atrialfibrillation and stroke prevention with warfarin in the long-term caresetting. Arch Intern Med. 1997;157:978-984.
3. McCormick D, Gurwitz JH, Goldberg RJ, Ansell J. Long-term anti-coagulation therapy for atrial fibrillation in elderly patients: efficacy,risk, and current patterns of use. J Thromb Thrombolysis. 1999;7:157-163.
4. Gurwitz JH, Avorn J, Ross-Degnan D, Choodnovskiy I, Ansell J.Aging and the anticoagulant response to warfarin therapy. Ann InternMed. 1992;116:901-904.
5. Gurewich V. Ximelagatran—promises and concerns. JAMA. 2005;293:736-739.
6. Gurwitz JH, Field TS, Avorn J, et al. Incidence and preventability ofadverse drug events in nursing homes. Am J Med. 2000;109:87-94.
7. Gurwitz JH, Field TS, Judge J, et al. The incidence of adverse drugevents in two large academic long-term care facilities. Am J Med.2005;118:251-258.
8. Jones A. The National Nursing Home Survey: 1999 Summary. Na-tional Center for Health Statistics. Vital Health Stat. 2002;13:152.
9. Gurwitz JH, Field TS, Harrold LR, et al. Incidence and preventabilityof adverse drug events among older persons in the ambulatory setting.JAMA. 2003;289:1107-1116.
0. Bates DW, Cullen DJ, Laird N, et al. Incidence of adverse drug eventsand potential adverse drug events implications for prevention. ADEPrevention Study Group. JAMA. 1995;274:29-34.
1. Leape LL, Bates DW, Cullen DJ, et al. Systems analysis of adversedrug events. JAMA. 1995;274:35-43.
2. Leape LL, Cullen DJ, Clapp MD, et al. Pharmacist participation onphysician rounds and adverse drug events in the intensive care unit.JAMA. 1999;282:267-270.
3. Hylek E, Singer DE. Risk factors for intracranial hemorrhage in
outpatients taking warfarin. Ann Intern Med. 1994;120:897-902.4. Fang MC, Chang Y, Hylek EM, et al. Advanced age, anticoagulationintensity, and risk for intracranial hemorrhage among patients takingwarfarin for atrial fibrillation. Ann Intern Med. 2004;141:745-752.
5. Kaushal R, Bates DW, Landrigan C, et al. Medication errors andadverse drug events in pediatric in-patients. JAMA. 2001;285:2114-2120.
6. Bates DW, Leape LL, Cullen DJ, et al. Effect of computerized phy-sician order entry and a team intervention on prevention on seriousmedication errors. JAMA. 1998;280:1311-1316.
7. Bates DW, Spell N, Cullen JJ, et al. The costs of adverse drug eventsin hospitalized patients. JAMA. 1997;277:307-311.
8. Landefeld CS, Anderson PA, Goodnough LT, et al. The bleedingseverity index: Validation and comparison to other methods for clas-sifying bleeding complications of medical therapy. J Clin Epidemiol.1989;42:711-718.
9. White RH, McKittrick T, Takakuwa J, Callahan C, McDonell M, FihnS, and the National Consortium of Anticoagulation Clinics. Manage-ment and prognosis of life-threatening bleeding during warfarin ther-apy. Arch Intern Med. 1996;156:1197-1201.
0. Freedman VA, Martin LG, Schoeni RF. Recent trends in disability andfunctioning among older adults in the United States: a systematicreview. JAMA. 2002;288:3137-3146.
1. Sackett DL, Haynes RB, Guyatt GH, Tugwell P. Clinical Epidemiol-ogy: A Basic Science for Clinical Medicine. 2nd ed. Boston, MA:Little Brown & Co; 1991.
2. Rosner B. Fundamentals of Biostatistics. 3rd ed. Boston, MA: PWS-Kent; 1990.
3. Rosendaal FR, Cannegieter SC, van der Meer FJM, Briet E. A methodto determine the optimal intensity of oral anticoagulant therapy.Thromb Haemost. 1993;69:236-239.
4. Rochon PA, Field TS, Bates DW, et al. Clinical application of acomputerized system for physician order entry with clinical decisionsupport to prevent adverse drug events in long-term care. CMAJ.2006;174:52-54.
5. Knight EL, Avorn J. Quality indicators for appropriate medication usein vulnerable elders. Ann Intern Med. 2001;135:703-710.
6. Matcher DB, Samsa GP, Cohen SJ, Oddone EZ, Gurgelski AE. Im-proving the quality of anticoagulation of patients with atrial fibrillationin managed care organizations: results of the managing anticoagulationservices trial. Am J Med. 2002;113:42-51.
7. Harrold LR, Gurwitz JH, Tate JP, et al. Physician attitudes concerninganticoagulation services in the long-term care setting. J ThrombThromboylsis. 2002;14:59-64.
8. Rochon PA, Field TS, Bates DW, et al. Computerized physician orderentry with clinical decision support in the long-term care setting:Insights from the Baycrest Centre for Geriatric Care. J Am Geriatr Soc.2005;53:1780-1789.
9. Hylek EM, Go AS, Chang Y, et al. Effect of intensity of oral antico-agulation on stroke severity and mortality in atrial fibrillation. N EnglJ Med. 2003;349:1019-1026.
0. SPORTIF Executive Steering Committee for the SPORTIF V Inves-tigators. Ximelagatran vs warfarin for stroke prevention in patientswith nonvalvular atrial fibrillation. A randomized trial. JAMA. 2005;
293:690-698.
C
MWDJKa
b
de
M
H
DM
0d
The American Journal of Medicine (2007) 120, 545-551
LINICAL RESEARCH STUDY
altreatment of Strongyloides Infection: Case Series andorldwide Physicians-in-Training Survey
avid R. Boulware, MD,a William M. Stauffer, MD, MSPH, DTM,a,b Brett R. Hendel-Paterson, MD,a
aime Luís Lopes Rocha, MD,c Raymond Chee-Seong Seet, MRCP,d Andrea P. Summer, MD,e Linda S. Nield, MD,f
huanchai Supparatpinyo, MD,g Romanee Chaiwarith, MD,g Patricia F. Walker, MD, DTMa,b
Division of Infectious Disease and International Medicine, Departments of Medicine and Pediatrics, University of Minnesota, Minneapolis;Center for International Health and International Travel Clinic, Regions Hospital/HealthPartners, St. Paul, Minn; cTravel Clin, Diagnósticosa América, Universidade Federal do Paraná, Curitiba PR, Brazil; dDepartment of Medicine, National University Hospital, Singapore;Department of Pediatrics, Medical University of South Carolina, Charleston; fDepartment of Pediatrics, West Virginia University School of
edicine, Morga gBmtsMiparRprwe9paACaeI©
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
ntown; Department of Medicine, Chiang Mai University, Chiang Mai, Thailand.
ABSTRACT
ACKGROUND: Strongyloidiasis infects hundreds of millions of people worldwide and is an important cause ofortality from intestinal helminth infection in developed countries. The persistence of infection, increasing interna-
ional travel, lack of familiarity by health care providers, and potential for iatrogenic hyperinfection all maketrongyloidiasis an important emerging infection.ETHODS: Two studies were performed. A retrospective chart review of Strongyloides stercoralis cases
dentified through microbiology laboratory records from 1993-2002 was conducted. Subsequently, 363 residenthysicians in 15 training programs worldwide were queried with a case scenario of strongyloidiasis, presentingn immigrant with wheezing and eosinophilia. The evaluation focused on resident recognition and diagnosticecommendations.ESULTS: In 151 strongyloidiasis cases, stool ova and parasite sensitivity is poor (51%), and eosino-hilia (�5% or �400 cells/�L) commonly present (84%). Diagnosis averaged 56 months (intra-quartileange: 4-72 months) after immigration. Presenting complaints were nonspecific, although 10% presentedith wheezing. Hyperinfection occurred in 5 patients prescribed corticosteroids, with 2 deaths. Treatment
rrors occurred more often among providers unfamiliar with immigrant health (relative risk of error: 8.4;5% confidence interval, 3.4-21.0; P �.001). When presented with a hypothetical case scenario, UShysicians-in-training had poor recognition (9%) of the need for parasite screening and frequentlydvocated empiric corticosteroids (23%). International trainees had superior recognition at 56% (P �.001).mong US trainees, 41% were unable to choose any parasite causing pulmonary symptoms.ONCLUSIONS: Strongyloidiasis is present in US patients. Diagnostic consideration should occur withppropriate exposure, nonspecific symptoms including wheezing, or eosinophilia (�5% relative or �400osinophils/�L). US residents’ helminth knowledge is limited and places immigrants in iatrogenic danger.nformation about Strongyloides should be included in US training and continuing medical education programs.
2007 Elsevier Inc. All rights reserved.
KEYWORDS: Eosinophilia; Graduate medical education; Immigrant; Parasites; Refugee; Sensitivity; Strongyloides;Wheezing
ISimA
D.R.B. and W.M.S. received support from the National Institutes ofealth (T32-AI055433).
Requests for reprints should be addressed to David Boulware, MD,ivision of Infectious Disease and International Medicine, University ofinnesota, MMC 250, 420 Delaware St. S.E., Minneapolis, MN 55455.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.05.072
NTRODUCTIONtrongyloidiasis is an extremely common cause of morbid-
ty and mortality worldwide. This disease is endemic inany tropical and subtropical areas, particularly Southeastsia, but also including Latin America and sub-Saharan
frica, as well as temperate areas such as Spain and the
AscgTAi1hSr
tgcucbddoatrmhqgsEctsnsa
dmscskU
MAc3nH1Ca
psRtg
UrdeSptrppcFasqtiammdt
546 The American Journal of Medicine, Vol 120, No 6, June 2007
ppalachian region of the United States.1-3 Intestinal para-ites are common in immigrants and refugees from developingountries, and the prevalence in some Southeast Asian refu-ees approaches 50%.4-6 Refugees from Cambodia, Laos, andhailand are historically at high risk of strongyloidiasis.7-10
mong over 17,000 refugees arriv-ng in Minnesota from 1993 to999, 22% harbored an intestinalelminth and 2.4% had detectabletrongyloides stercoralis (S. sterco-alis) in stool.11
Infection in nonendemic set-ings is predominately in immi-rants or expatriates.12,13 By 2000ensus data, 11% of the US pop-lation is foreign-born. In Appala-hia, the S. stercoralis prevalancey general screening is 4%.14-16 Ineveloped countries, almost alleaths attributable to helminthsccur secondary to strongylodi-isis hyperinfection or dissemina-ion.13 Although once thoughtare, disseminated strongyloidiasisay be relatively common in
igh-risk populations and fre-uently misdiagnosed as isolatedram-negative sepsis or acute re-piratory distress syndrome.17,18
ven though this feared iatrogenicomplication from corticosteroidherapy does occur, strongyloidia-is generally presents with diffuse,onspecific gastrointestinal, dermatologic, or respiratoryymptoms.10,19,20 Most individuals chronically infected aresymptomatic.1,21,22
Minimal data exist regarding S. stercoralis in nonen-emic settings. This study was conducted to review de-ographic and clinical data. Because unrecognized
trongyloidiasis can have devastating iatrogenic compli-ations, a goal of this project was to assess future phy-icians’ knowledge. Therefore, an assessment of basicnowledge of strongyloiodiasis was performed amongS and international physicians-in-training.
ETHODSretrospective clinical descriptive study was conducted of all
ases of S. stercoralis from January 1, 1993 through December1, 2002 at Regions Hospital/HealthPartners in St. Paul, Min-esota, a 427-bed hospital with a Center for Internationalealth serving an immigrant and refugee population. Indeed,3% of St. Paul, Minnesota is of Southeast Asian descent.23
ases were identified through a microbiology database of par-
CLINICAL SIGNIF
● Physicians shoulasis in patientssure history in dgardless of the t
● Clinical clues includistress, and eosor �5% of diffepersists lifelong.
● Strongyloides is uthe United Staterogenic. Steroidschronic, asymptofection can leadperinfection, whdue to gram-neg
● Stool examinatiodetection per exogy is available.
sitology specimens from the hospital laboratory. The charts of b
atients with positive ova and parasite (O&P) specimens for S.tercoralis (rhabditiform or filariform larvae) were reviewed.eviews were conducted by 3 physicians knowledgeable in
ropical medicine. A standardized data form collected: demo-raphic information, US arrival time, clinical manifestations,
complete blood counts with differ-ential (CBC), eosinophil count at di-agnosis and at follow-up, other or-ganisms detected, total number ofstool samples positive, treatmentprescribed, treatment failures de-tected by recurrent positive samplesor persistent eosinophilia, occur-rence of hyperinfection, and ap-propriate management includingfollow-up examination.
Analysis is primarily descriptive.Results are reported as mean�SD.The intra-quartile range (IQR) is re-ported when data have non-normaldistribution. Comparisons are madewith SPSS 13.0 (SPSS Inc., Chi-cago, Ill) presenting relative risk(RR) and 95% confidence intervals(95% CI).
Trainee SurveyAn internet-based survey was con-ducted of 363 Internal Medicine,Medicine-Pediatric, and Pediatricresidents in 15 different residencytraining programs across the
nited States, Brazil, Singapore, and Thailand. Nine USesidency programs participated in regions being en-emic for Strongyloides, with immigrants, and withoutither. The survey consisted of a case presentation of aoutheast Asian immigrant without prior asthma historyresenting with new-onset wheezing and respiratory dis-ress. All values presented were typical of the caseseviewed in this study. For example, the case scenarioresented a 38-year-old patient who immigrated 5 yearsrior. The patient’s peripheral eosinophil count was 900ells/�L (9%) with reportedly normal chest radiograph.or surveys in Brazil and Thailand, the patient was from“rural region” rather than an immigrant. The Brazilian
urvey was administered in Portuguese. Residents wereueried as to their “work up and initial therapy.” Addi-ional queries addressed residents’ helminth knowledgencluding: the “level of absolute eosinophilia consideredbnormal,” “which parasites routinely cause chronic pul-onary symptoms,” and “which helminth causes theost mortality.” Answers were multiple-choice with in-
ividualized computer randomization of answer orderingo eliminate response-order bias. Institutional review
CE
sider strongyloidi-a potential expo-
ping countries, re-ince immigration.
heezing, abdominalilia �400 cells/�Ll count. Infection
e in that deaths innearly always iat-
n to a patient withc Strongyloides in-fe-threatening hy-s �50% mortalitysepsis.
e insensitive (50%ation). IgG serol-
ICAN
d conwitheveloime s
de winophrentia
niqus aregivematito li
ich haative
ns aramin
oard approvals were obtained.
RDsligd2pTymmrnTscvsta
(ssep9cstri
bam
sdp(fp4pBwnmz
asc(
Fm
547Boulware et al Maltreatment of Strongyloides
ESULTSuring the 10-year period, 151 people were diagnosed with
trongyloidiasis. Of 1291 positive stool specimens col-ected, S. stercoralis was the third most common parasitedentified, accounting for 11% of potentially pathogenicastrointestinal parasites (Table 1). Countries of origin wereiverse, but Southeast Asian patients predominated (Table). Stool specimens were positive for S. stercoralis in 149atients, and sputum samples were positive in 3 patients.he mean (�SD) age of infected individuals was 39 � 18ears. The sample consisted of 83 men and 68 women. Theean time of diagnosis from arrival in the US was 45 � 52onths (IQR: 4 to 72 months). If cases of newly arrived
efugees detected at routine screening (n � 28) are elimi-ated, the mean time to diagnosis increases to 55 months.here were 5 cases of documented hyperinfection or dis-emination that occurred as an iatrogenic complication oforticosteroid therapy initiated in chronically infected indi-iduals. Three of the subjects with hyperinfection had re-ided in the US for �5 years. These iatrogenic complica-ions resulted in prolonged stays in an intensive care unit inll 5 patients and led to 2 deaths.
The average absolute eosinophil count was 981 �978IQR: 480 to 1130 cells/�L), and the average percent eo-inophilia was 11.8% � 7.6% (IQR: 6.3% to 16.4%). Eo-inophilia varied substantially with age. Children had higherosinophil counts (Figure 1). Patients harboring multiplearasites (n � 80) had similar levels of eosinophilia (1015 �50; P � 0.7). The sensitivity of an absolute eosinophilutoff of �500 cells/�L for predicting �1 positive stoolample was 73% (94/128). For a cutoff of �400 cells/ �L,he sensitivity improved to 84% (107/128). Using a 5%elative differential as the abnormal cutoff for distinguish-ng eosinophilia had equivalent sensitivity (84%) with good,
Table 1 Number of Patients Infected with IndividualParasite Species 1996-2002
Potentially Pathogenic Species
Positive Specimens
Number Percent (%)
Giardia intestinalis 363 28.1Hookworm species 297 23.0Strongyloides stercoralis 139 10.8Trichuris trichiura 132 10.2Ascaris lumbricoides 115 8.9Entamoeba histolytica/dispar 82 6.4Hymenolepis nana 60 4.6Taenia species 31 2.4Enterobius vermicularis 23 1.8Opisthorchis spp and Clonorchiasis
sinesis17 1.3
Schistosoma mansoni 13 1.0Othera 19 1.5Total 1291 100
aSchistosoma haematobium (n � 6), Diphyllobothrium latum(n � 4), Fascioloidea spp. (n � 4), Paragonimus westermani (n � 3),
4Toxocara (n � 1).
ut not identical, agreement (Kappa � 0.829). Consider-tion of both absolute and relative eosinophilia only mini-ally increased detection sensitivity (86%).Stool ova and parasite examination is a relatively insen-
itive technique for diagnosis of strongyloidiasis with larvaeetected in 51% (262/517) of all collected specimens ofatients with at least one positive sample. Further, 16%24/151) of patients had at least 3 prior negative specimensor Strongyloides (mean 3.6 � 2.1; max 9), confirming theoor sensitivity of stool examination alone. Interestingly,6% (11/24) of those antecedent specimens did harbor otherarasites, indicating a history of fecal-oral contamination.efore the correct diagnosis of strongyloidiasis, patientsith previously negative stool specimens were subjected toumerous invasive procedures and labeled with an array ofisdiagnoses, including irritable bowel syndrome, somati-
ation disorder, and psychogenic pruritis.Presenting symptoms were diverse, with 12% being
symptomatic and identified through routine medicalcreening. Abdominal complaints (40%) and pulmonaryomplaints (22%) were common, including wheezing10%).
igure 1 Absolute eosinophil count versus age. Data representean � SD of absolute eosinophil counts (cells/�L). Intercept is at
Table 2 Country of Origin/Exposure
Country Number Percent (%)
Cambodia 57 30Hmong (ethnicity) 34 18Vietnam 33 17Thailand 28 15Laos 16 8.4Ethiopia 7 3.7Liberia 4 2.1Philippines 3 1.6Eritrea, Guatemala, Mexico, Nigeria,
Somalia, Sudan, and Uganda allcontributed �2 patients each
Note: Totals are �151 due to multiple countries of exposure andoverlapping locales of the Hmong people, who originated in Laos.
00 cells/�L.
9a((awpctesHo(o
pabwaOiyaemt3w
TIhTpwtc(t
pc2cgdOftm
Fr atmen
548 The American Journal of Medicine, Vol 120, No 6, June 2007
Follow-up laboratory investigations were performed in3 subjects (62%) and included: 3 stool samples (n � 53)lone, CBC with differential (n � 3), or both stool and CBCn � 37) (Figure 2). Treatment failures occurred in 7.6%11/145), with 8 documented by positive stool specimensnd 3 by persistent symptoms and eosinophilia. All patientsith primary drug failure by stool examination also hadersistently elevated eosinophil counts (absolute: 975 � 425ells/�L [min 510]; relative: 13.6 � 7.3% [min 5.4%]). Inhe 37 patients with both follow-up tests, none with anosinophil count �400 cells/�L had positive Strongyloidespecimens (0/28; negative predictive value (NPV) 100%).owever, 2 patients did harbor Giardia. Medication failuresccurred with thiabendazole 6% (6/97), albendazole 14%3/22), and mebendazole 50% (2/4). No ivermectin failuresccurred.
Patients with strongyloidiasis frequently harbored otherarasites (56%) (Table 3). Potentially pathogenic parasitesccounted for 63% of positive samples, with hookwormeing the most common co-infection (n � 46). In patientsith hookworm co-infection, no difference was observed in
ge, hemoglobin, or eosinophil count (947 � 660 cells/�L).f hookworm diagnoses, 79% occurred within 3 years of
mmigration and 90% within 6 years from arrival (max 12ears). As hookworm species have a finite 3- to 5-yearverage lifespan, 10%-20% of hookworms either greatlyxceeded the average lifespan or were acquired by postim-igration visiting friends and relatives (VFR) travel. Post-
reatment stool specimens detected persistent parasites in0% (27/90) of patients; 52% (23/44) of detected speciesere potential pathogens.
reatment Errorsnitial diagnostic evaluations were with noninternationalealth or travel clinic providers in 24% (36/151) of cases.hree patients presenting with wheezing to primary carehysicians were treated with corticosteroids for “asthma”ith subsequent development of hyperinfection. Two pa-
ients died. All patients presenting with wheezing to clini-ians familiar with immigrant health or travel medicinen � 12) had an evaluation for strongyloidiasis, appropriate
StrongyloN=
Lost to Follow Up / No treatment
N=12
Treatment FailureN=11
Positive stool O&P N=8 E
igure 2 Patient outcomes. *Eosinophilia defined as �400 ceeceiving therapy occurred (n � 8). These are not counted as a tre
reatment, and confirmatory follow-up evaluation.
Delays of �6 months in treatment from a positive sam-le or highly elevated eosinophilia (�1000 cells/�L) oc-urred in 5% (8/151) of cases, with an average delay of.8 � 3.2 years (IQR: 1 to 3 years, max 10 years). In severalases, multiple referrals to specialists such as dermatolo-ists and gastroenterologists occurred with invasive proce-ures performed without consideration of strongyloidiasis.ne patient underwent multiple procedures for 6 months
ollowing a positive Strongyloides stool sample. Errors inreatment, either prescribing an ineffective medication (eg,ebendazole, metronidazole) or no medication, occurred in
Treatment Success N=120†
Negative stool O&P N=90
No EosinophiliaN=32
ta*
†Delays of �6 months between a positive stool specimen andt success.
Table 3 Other Parasites Discovered in StrongyloidiasisPatients
Parasite
Number ofPatientsn � 151
Positive StoolSpecimens
Percent(%)
Positive/Collected*
Strongyloides stercoralis 151 51 262/517Co-infections of
pathogenic parasitesHookworm 46 76 105/139Giardia intestinalis 14 78 28/36Blastocystis hominis† 12 71 25/35Trichuris trichuria 5 73 11/15Clonorchis sinensis 5 57 10/18Ascaris lumbricoides 4 100 12/12Hymenolepsis nana 4 83 10/12Paragonimus westermani 4 50 4/6Entamoeba histolytica 3 33 3/9
Additional nonpathogenicparasites
Endolimax nana 25 66 53/78Entamoeba coli 12 86 36/42Entamoeba hartmanni 10 78 28/36Iodamoeba butschlii 3 67 6/9
Fascioloidea, Schistosoma japonicum, Schistosoma manconi, Taeniaspecies, and trichostrongylus co-infections were present in 1 patient,each with a detection rate of 67%.
*Collected stool specimens for ova and parasite examinationamong infected individuals.
ides Cases151
s
Persistenosinophili
N=8
lls/�L.
†Controversial as to the pathogenicity of Blastocystis hominis.
6ogI(T
MTsstetsifUCbirlbu9
rpoAca
ttctwdbrFFap6cmrbBaA
DTnlmcibfsrllta3m
3aee5hruusstfs
ppiiSthfhBc
hatrcacsfc
549Boulware et al Maltreatment of Strongyloides
% (9/151). Delays or errors in treatment occurred moreften with health care providers not familiar with immi-rant/travel medicine (RR 8.4; 95% CI, 3.4-21.0; P �.001).n 66% of treatment errors, multiple parasites were presentmean 1.5 � 0.8 other species) perhaps causing confusion.welve patients (8%) were never effectively treated.
edical Resident Trainee Surveyhe survey response rate was 70% (363/520). When pre-ented with a Southeast Asian immigrant (in the case ofurveys in Brazil and Thailand, the individual was reportedo be from a “rural area”) with new-onset wheezing and 9%osinophilia (absolute 900 eosinophils/�L), only 9% of USrainees recommended further evaluation for abnormal eo-inophilia or parasites as compared with 56% (53/94) ofnternational trainees. Empiric corticosteroid use withouturther evaluation was recommended by 23% (61/269) ofS versus 7% (7/94) of international trainees (RR 3.0; 95%I, 1.4-6.4; P � .005). There was no performance differenceetween upper level residents and first or second year res-dents (P � .2). Internal Medicine and Medicine-Pediatricesidents significantly outperformed their pediatric col-eagues in recognition (RR 9.6; 95% CI, 2.1-42; P � .004),ut the recommendation for empiric steroids without eval-ation or treatment for strongyloidiasis was similar (RR 0.7;5% CI, 0.5-1.1; P � .11).
General helminth knowledge was poor, with 51% ofesidents unable to identify the absolute level of eosino-hilia considered abnormal. Most experts agree that eosin-philia above 400-500 cells/�L is considered abnormal.24
mong respondents, 21% had a cutoff too low (250-300ells/�L) and 30% had too high (�750-200 cells/�L) anbnormal threshold.
When presented a list of 16 parasites and asked “whichwo parasites routinely cause chronic pulmonary symp-oms?,” 62% of residents knew Strongyloides stercoralisauses pulmonary symptoms, but 20% also believed Blas-ocystis hominis caused pulmonary symptoms. Ascaris,hich was listed by 43%, can give pulmonary symptomsuring initial larval migration, as can hookworm (12%),ut neither parasite does so chronically. Only 18% cor-ectly identified Paragonimus westermani (Oriental Lungluke) as associated with chronic pulmonary disease.orty-one percent of US trainees were unable to chooseny parasite that caused any pulmonary symptoms. Theerformance of international trainees was superior, with5% correctly naming the 2 main parasitic causes ofhronic pulmonary symptoms. When asked “which hel-inth causes the most mortality in the US?,” 34% cor-
ectly identified strongyloidiasis, with other responseseing: amebiasis (18%), schistosomiasis (11%), andlastocystis hominis (9%). Blastocystis has minimal, ifny, pathogenicity. Resident answers are available in the
ppendix (available online). AISCUSSIONhe principal precept of the practice of medicine is “first, doo harm” (primum non nocere). As 11% of the US popu-ation is foreign-born, and annual immigration exceeds oneillion, basic helminth knowledge is important for physi-
ians. Strongyloides is an important pathogen in refugee andmmigrant populations.10,20 This study demonstrates that,ecause the stool is not always positive, strongyloidiasis isrequently misdiagnosed (eg, psychosomatic), or diagnosisignificantly delayed. In this series, if not detected at newefugee arrival medical screening, the diagnosis was de-ayed by 5 years, on average, from US arrival. Strongy-oides’ unique auto-infective cycle creates indefinite infec-ion. Among World War II British prisoners of war, theverage time from leaving an endemic area to diagnosis was7 � 6 years.25 Thus, potential iatrogenic maleficence re-ains life-long.Strongyloidiasis diagnosis is difficult. The sensitivity of
routine stool specimens by direct examination is reporteds 50%.1,19,26,27 In this study, 3 stool ova and parasitexaminations detected 84% of strongyloidiasis cases; how-ver, 16% had �3 prior false negative specimens. Overall,1% of all stool samples contained S. stercoralis. Thisigher yield may represent the burden of infection in aefugee population but likely may be biased by even morendiscovered cases, as there was no external gold standardtilized, such as agar plate methods of “parasite cultures” orerology.10,19 The Centers for Disease Control (CDC)trongyloides enzyme immunoassay serology is 95% sensi-ive for IgG antibody detection and can be a useful adjuvantor diagnosis.10 Several reference laboratories also haveerology available of unknown validity.
In many, a mild relative (�5%) and absolute eosino-hilia (�400 cells/�L) were the only abnormal finding,redicting a positive stool specimen in 84%. This sensitivitys comparable with the 83% sensitivity of eosinophilia inmmigrants screened by CDC serology.10 Similarly, amongpanish agricultural workers, Román-Sánchez et al found
hat those with eosinophilia (�400 cells/�L) had 73-foldigher odds of strongylodiasis.22 However, when hyperin-ection syndrome occurs, patients without eosinophilia haveigh mortality.3,18 In the largest hyperinfection series fromrazil, Canada, and Minnesota, presenting eosinophilounts �400 cells/�L had 100% mortality.17,18,28
Wheezing was the presenting symptom in 10%. Asthmaas been previously reported,29,30 but this series representshigher occurrence of wheezing. Strongyloides is not causi-
ive for true asthma.30,31 Nevertheless, empiric corticoste-oid therapy is problematic, as empiric corticosteroids mayause life-threatening hyperinfection. Immunosuppressionllows the Strongyloides larvae to penetrate the gut wall,ausing secondary gram-negative sepsis.3,32 In our caseeries, those presenting with wheezing to practitioners un-amiliar with travel or tropical medicine uniformly receivedorticosteroids. In high-risk individuals from Southeast
sia or agricultural workers, we recommend screening (ie,
smciamu
afiaainmopaceWem
tptUtgrbMtcib
CPwnetsigpatseh
AW
R
1
1
1
1
1
1
1
1
1
1
2
550 The American Journal of Medicine, Vol 120, No 6, June 2007
erology) when time allows or empiric treatment with iver-ectin 200 �g/kg before immunosuppressive therapy. Be-
ause the average time to diagnosis was 5 years from arrivaln this series, and hyperinfection has occurred �50 yearsfter immigration, we suggest the aforementioned manage-ent regardless of immigration timing.17,33 This is partic-
larly true with new-onset wheezing in adult immigrants.Treatment failure is common even with appropriate ther-
py. Although not a randomized trial, albendazole’s 14%ailure rate reinforces its inferiority to thiabendazole orvermectin.19,34,35 Ivermectin 200 �g/kg orally and repeatedt 2 weeks has 97% efficacy.34 Ivermectin is inactivegainst hookworm, a common co-infection. Thiabendazoles active against hookworm but has a high rate (�50%) ofausea.35 Post-therapy stool examinations are recom-ended to verify Strongyloides eradication and exclude
ther parasites, as 15% harbored post-treatment pathogenicarasites. Patients with initial eosinophilia (�400 cells/�L)nd subsequently normalized post-treatment eosinophilounts had a 100% NPV for 3 follow-up negative stoolxaminations, corresponding to other observations.19,26
hen repeat stool testing is impractical or unobtainable,osinophil count normalization is a reassuring surrogatearker for treatment success.Although strongyloidiasis may be considered an infec-
ious disease, tropical medicine, or immigrant health issue,atients seek care from primary providers. Physicians’-in-raining basic helminth knowledge is poor, particularly forS resident-physicians. This lack of training was evident in
he 151-subject case series by the poorer quality of careiven by physicians untrained in travel medicine. As the USeceives more immigrants than the rest of the world com-ined and immigrants constitute �25% of the Los Angeles,iami, New York, and San Francisco metropolitan areas,23
his creates a special obligation for US physicians. Althoughomprehensive helminth knowledge among US physicianss unrealistic, basic knowledge of iatrogenic danger posedy corticosteroids and Strongyloides should exist.
ONCLUSIONhysicians should consider strongyloidiasis in patientsith a potential exposure history in developing countries,onspecific symptoms, and �5% eosinophilia (or �400osinophils/�L), regardless of the time since immigra-ion. A significant portion may present with pulmonaryymptoms including wheezing. Screening in high-riskndividuals or empiric anti-helminth treatment in immi-rants with eosinophilia is warranted before immunosup-ression to prevent the severe morbidity and mortalityssociated with hyperinfection syndrome. Knowledge ofhis infection and steps to avoid iatrogenic complicationshould be included in training and continuing medicalducation programs to ensure that physicians first do no
arm.CKNOWLEDGMENTe thank Guilherme Barroso for assistance.
eferences1. Liu LX, Weller PF. Strongyloides and other intestinal nematode in-
fections. Infect Dis Clin North Am. 1993;7:655-682.2. Genta RM. Global prevalence of strongyloidiasis: critical review with
epidemiologic insights into the prevention of disseminated disease.Rev Infect Dis. 1989;11:755-767.
3. Keiser PB, Nutman TB. Strongyloides stercoralis in the immunocom-promised population. Clin Microbiol Rev. 2004;17:208-217.
4. Nutman TB, Ottesen EA, Ieng S, et al. Eosinophilia in Southeast Asianrefugees: evaluation at a referral center. J Infect Dis. 1987;155:309-313.
5. Hoffman SL, Barrett-Connor E, Norcross W, Nguyen D. Intestinalparasites in Indochinese immigrants. Am J Trop Med Hyg. 1981;30:340-343.
6. Lindes C. Intestinal parasites in Laotian refugees. J Fam Pract. 1979;9:819-822.
7. Vannachone B, Kobayashi J, Nambanya S, Manivong K, Inthakone S,Sato Y. An epidemiological survey on intestinal parasite infection inKhammouane Province, Lao PDR, with special reference to Strongy-loides infection. Southeast Asian J Trop Med Public Health. 1998;29:717-722.
8. Wiesenthal AM, Nickels MK, Hashimoto KG, Endo T, Ehrhard HB.Intestinal parasites in Southeast Asian refugees. Prevalence in a com-munity of Laotians. JAMA. 1980;244:2543-2544.
9. Jongsuksuntigul P, Intapan PM, Wongsaroj T, et al. Prevalence ofStrongyloides stercoralis infection in northeastern Thailand (agar plateculture detection). J Med Assoc Thai. 2003;86:737-741.
0. Loutfy MR, Wilson M, Keystone JS, Kain KC. Serology and eosino-phil count in the diagnosis and management of strongyloidiasis in anon-endemic area. Am J Trop Med Hyg. 2002;66:749-752.
1. Swanson S, Lee B, Mamo B, Smith K, Stauffer WM. Changingprevalence of intestinal parasites among newly arrived SoutheastAsian and African refugees after empiric predeparture albendazoletreatment—Minnesota, 1993-2004 [abstract]. 55th Annual EpidemicIntelligence Service Conference, Atlanta, GA. April 24-28, 2006.
2. Hira PR, Al-Ali F, Shweiki HM, et al. Strongyloidiasis: challenges indiagnosis and management in non-endemic Kuwait. Ann Trop MedParasitol. 2004;98:261-270.
3. Muennig P, Pallin D, Sell RL, Chan. The cost effectiveness of strat-egies for the treatment of intestinal parasites in immigrants. N EnglJ Med. 1999;340:773-779.
4. Fulmer HS, Huempfner HR. Intestinal helminths in eastern Kentucky:a survey in three rural counties. Am J Trop Med Hyg. 1965;14:269-275.
5. Walzer PD, Milder JE, Banwell JG, Kilgore G, Klein M, Parker R.Epidemiologic features of Strongyloides stercoralis infection in anendemic area of the United States. Am J Trop Med Hyg. 1982;31:313-319.
6. Berk SL, Verghese A, Alvarez S, Hall K, Smith B. Clinical andepidemiologic features of strongyloidiasis. A prospective study in ruralTennessee. Arch Intern Med. 1987;147:1257-1261.
7. Lim S, Katz K, Krajden S, et al. Complicated and fatal Strongyloidesinfection in Canadians: risk factors, diagnosis and management. CanMed Assoc J. 2004;171:479-484.
8. Newberry AM, Stauffer WM, Hendel-Paterson BR, et al. Strongy-loides hyperinfection presenting as acute respiratory distress and gramnegative sepsis. Chest. 2005;128:3681-3684.
9. Siddiqui AA, Berk SL. Diagnosis of Strongyloides stercoralis infec-tion. Clin Infect Dis. 2001;33:1040-1047.
0. de Silva S, Saykao P, Kelly H, et al. Chronic Strongyloides stercoralisinfection in Laotian immigrants and refugees 7-20 years after reset-
tlement in Australia. Epidemiol Infect. 2002;128:439-444.
2
2
2
22
2
2
2
2
3
3
3
3
3
3
551Boulware et al Maltreatment of Strongyloides
1. Gyorkos TW, Genta RM, Viens P, MacLean JD. Seroepidemiology ofstrongyloides infection in the South East Asian refugee population inCanada. Am J Epidemiol. 1990;132:257-264.
2. Román-Sánchez P, Pastor-Guzmán A, Moreno-Guillén S, Igual-AdellR, Suñer-Generoso S, Tornero-Estébanez C. High prevalence ofStrongyloides stercoralis among farm workers on the Spanish Medi-terranean coast. Analysis of the predictive factors of infection indeveloped countries. Am J Trop Med Hyg. 2003;69:336-340.
3. U.S. Census Bureau, County and City Data Book: 2000. Available at:http://quickfacts.census.gov/qfd/states/00000.html. Accessed Dec 18,2005.
4. Rothenberg ME. Eosinophilia. N Eng J Med. 1998;338:1592-1600.5. Gill GV, Welch E, Bailey JW, Bell DR, Beeching NJ. Chronic
Strongyloides stercoralis infection in former British Far East prisonersof war. QJM. 2004;97:789-795.
6. Grove DI. Strongyloidiasis in Allied ex-prisoners of war in South-EastAsia. BMJ 1980;1:598-601.
7. Pelletier LL. Chronic strongyloidiasis in World War II Far East ex-prisoners of war. Am J Trop Med Hyg. 1984;33:55-61.
8. Ferreira MS, Nishioka Sde A, Borges AS, et al. Strongyloidiasis andinfection due to human immunodeficiency virus: 25 cases at a Brazil-ian teaching hospital, including seven cases of hyperinfection syn-
drome. Clin Infect Dis. 1999;28:154-155.9. Robinson J, Ahmed Z, Siddiqui A, et al. A patient with persistentwheezing, sinusitis, elevated IgE, and eosinophilia. Ann AllergyAsthma Immunol. 1999;82:144-149.
0. Wehner JH, Kirsch CM, Kagawa FT, Jensen WA, Campagna AC,Wilson M. The prevalence and response to therapy of Strongyloidesstercoralis in patients with asthma from endemic areas. Chest. 1994;106:762-766.
1. Leeman BJ, Cabrera MR. No association found between Strongyloidesinfestation and asthma. J Asthma. 1995;32:57-62.
2. Igra-Siegman Y, Kapila R, Sen P, et al. Syndrome of hyperinfectionwith Strongyloides stercoralis. Rev Infect Dis. 1981;3:397-407.
3. Gill V, Beeching NJ, Khoo S, et al. A British Second World Warveteran with disseminated strongyloidiasis. Trans R Soc Trop MedHyg. 2004;98:382-386.
4. Toma H, Sato Y, Shiroma Y, Kobayashi, Shimakukuro I, Takara M.Comparative studies on the efficacy of three antihelminthics on treat-ment of human strongyloidiasis in Okinawa, Japan. Southeast Asian JTrop Med Public Health 2000;31:147-151.
5. Marti H, Haji HJ, Savioli L, et al. A comparative trial of a single doseivermectin versus three days of albendazole for treatment of Strongy-loides stercoralis and other soil transmitted helminth infections in
children. Am J Trop Med Hyg. 1996;55:477-481.
ATh
O
551.e1 The American Journal of Medicine, Vol 120, No 6, June 2007
PPENDIX 1he physicians-in-training survey may be taken at:ttp://www.tropical.umn.edu/Strongyloides.
nline Appendix 2: Resident Survey Answers
Parasite Species*
ChronicPulmonarySymptoms
GreatestMortality
% %
Ancylostoma duodenale (Hookworm) 12 5Ascaris lumbricoides 43 10Blastocystis hominis 20 9Clonorchis sinensis (Oriental liver
fluke)2.3 1
Cryptosporidium parvum 11 —Entamoeba histolytica (Amebiasis) 6 18Paragonimus westermani (Oriental
lung fluke)18 4
Schistosoma species 14 11Strongyloides stercoralis 62 35Taenia species (tapeworm) 3.3 5Trichuris trichuria (whipworm) 5 12
*Endolimax nana, Entamoeba coli, Fasciolides, Giardia, Hymenolep-sis were also listed as possible answers but received �1% each.
Cryptosporidium was not queried as to mortality.
C
KMDAa
sdtmaNsss
DM
0d
The American Journal of Medicine (2007) 120, 552.e1-552.e8
LINICAL RESEARCH STUDY
idney Disease, Framingham Risk Scores, and Cardiac andortality Outcomes
aniel E. Weiner, MD, MS,a Hocine Tighiouart, MS,b John L. Griffith, PhD,b Essam Elsayed, MD,a
ndrew S. Levey, MD,a Deeb N. Salem, MD,c Mark J. Sarnak, MD, MSa
Divisions of Ne b c
BIMerabRdscmA1doCcF
E-mail address
002-9343/$ -see foi:10.1016/j.amjm
phrology, Clinical Care Research, and Cardiology, Tufts-New England Medical Center, Boston, Mass.
ABSTRACT
ACKGROUND: The Framingham equations were developed to predict incident coronary heart disease.t remains unknown how the presence of chronic kidney disease affects their performance.ETHODS: Individuals without preexisting cardiovascular disease aged 45 to 74 years from the Ath-
rosclerosis Risk in Communities and Cardiovascular Health Studies were analyzed. Using sex- andace-specific Cox models, we evaluated the 5-year risk of coronary heart disease and mortality eventsssociated with both chronic kidney disease and Framingham risk score, the absolute risk of events causedy kidney disease, and model discrimination.ESULTS: Among 15,717 subjects, 756 (4.8%) had kidney disease. The Framingham risk score indepen-ently predicted cardiac and mortality events in all subgroups, whereas kidney disease predicted events in allubgroups except cardiac events in white women. After adjustment for traditional risk factors, the increase inardiac and mortality events per 1000 person-years attributable to kidney disease was 4.3 and 13.7 for whiteen, 16.1 and 40.5 for African American men, 1.2 and 5.8 for white women, and 13.6 and 14.2 for Africanmerican women, respectively. This represented an additional 17,000 and 12,000 cardiac events and 63,000 and9,000 deaths per year among whites and African Americans, respectively. Mortality rates attributable to kidneyisease, diabetes, and smoking were comparable. Accounting for kidney disease improved discrimination fornly mortality outcomes in white men and African American women.ONCLUSIONS: Chronic kidney disease in a community-based population is an important predictor ofardiac and mortality events, particularly in African Americans, but it does not improve discrimination oframingham equations. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Cardiovascular disease; Chronic kidney disease; Coronary heart disease; Framingham Risk Score;Renal insufficiency
Tlcfait(tamid
Grant Support: R21 DK068310 and K23 DK71636. Amgen Inc, Thou-and Oaks, Calif, provided partial support for the creation of the pooledatabase. Study sponsors were not involved in data analysis or interpreta-ion of findings. Role of the Sponsors: The Atherosclerosis Risk in Com-unities Study, Cardiovascular Health Studies, and Framingham Heart
nd Framingham Offspring studies are conducted and supported by theational Heart, Lung, and Blood Institute in collaboration with the individual
tudy investigators. This article was not prepared in collaboration with thetudy investigators and does not necessarily reflect the opinions or views of thetudy investigators or the National Heart, Lung, and Blood Institute.
Requests for reprints should be addressed to Daniel E. Weiner, MD, MS,ivision of Nephrology, Box 391, Tufts-New England Medical Center, Boston,A 02111.
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.05.054
he Framingham Coronary Heart Disease Risk Score al-ows clinicians to estimate the individual patient risk oforonary disease by accounting for traditional cardiac riskactors, including sex, age, blood pressure, cholesterol, di-betes, and smoking.1,2 The Framingham Risk Score (FRS)nstrument has been validated in racially diverse popula-ions, including the Atherosclerosis Risk in CommunitiesARIC) and Cardiovascular Health Studies (CHS). Al-hough there was reasonable agreement between predictednd actual event rates in these population-based cohorts,aking recalibration of the equations unnecessary, the abil-
ty of the FRS to discriminate between subjects who did andid not develop coronary disease seemed worse in minority
ohorts.3
kdfdwimwatedatb
M
SWbduieFwA1taCyirch
atiSnttai
ssrc
at
mg
ra
OTtwMaccdaaec
S
FWcFvttapog
score
552.e2 The American Journal of Medicine, Vol 120, No 6, June 2007
Other risk factors for coronary disease include chronicidney disease.4-6 However, the impact of chronic kidneyisease at a public health level remains uncertain. There-ore, we investigated the importance of chronic kidneyisease by assessing sex- and race-specific risk associatedith kidney disease after adjust-
ng for Framingham risk, deter-ining absolute risk associatedith kidney disease in excess of
nd in comparison with tradi-ional cardiac risk factors, andvaluating the impact of kidneyisease on the discriminativebility of the Framingham equa-ions in a biracial, community-ased cohort.
ETHODS
tudy Designe used 2 limited-access data-
ases that evaluate cardiovascularisease in community-based pop-lations: ARIC and CHS. Pool-ng these studies allowed forvaluation of individuals in theramingham study age rangehile increasing the number offrican Americans. From 1987 to 1989, ARIC enrolled5,792 participants aged 45 to 64 years from 4 communi-ies. The Mississippi center is entirely African Americannd comprises 80% of the African Americans in ARIC.7
HS is a population-based study of 5201 subjects aged 65ears or more who were selected from Medicare eligibility filesn 4 communities from 1989 to 1990. An additional 687 Af-ican Americans were recruited from 1992 to 1993.8 We ex-luded subjects aged more than 74 years to match the Framing-am population.
Glomerular filtration rate was estimated with the 4-vari-ble Modification of Diet in Renal Disease study equa-ion.9-11 We calibrated the ARIC and CHS laboratoriesndirectly using National Health and Nutrition Examinationurvey (NHANES) III data.6,12-14 We defined chronic kid-ey disease as an estimated glomerular filtration rate lesshan 60 mL/min/1.73 m2.15 Subjects with a glomerular fil-ration rate less than 15 mL/min/1.73 m2 were excluded tovoid including individuals for whom dialysis was likelymminent.
Baseline characteristics included demographics (age,ex, race); medical history (cardiovascular disease, diabetes,moking); systolic and diastolic blood pressure; and labo-atory variables (total cholesterol, high-density lipoproteinholesterol, creatinine).
Race was defined as white or African American. Cig-rette smoking was dichotomized by current use. Diabe-
CLINICAL SIGNIF
● Kidney disease himpact, particulcans. Among whcans, respectivecounts for 17,00heart disease e19,000 deaths pfor Framingham
● For mortality oukidney disease isdiabetes and sm
● Accounting for tkidney disease geFramingham risk
es was defined by the use of insulin, oral hypoglycemic h
edications, or a fasting glucose level of 140 mg/dL orreater to match the original Framingham definition.16
Baseline cardiovascular disease included a history of bothecognized and silent myocardial infarction, angina, coronaryngioplasty and bypass procedures, stroke, transient ischemic
attack, and intermittent claudicationas defined by consensus committeesfor the studies. Baseline cardiovas-cular disease included heart failurein CHS but was not evaluated inARIC.7,8
The final cohort included15,717 individuals after the fol-lowing exclusions: 402 subjectsmissing data on age, race, sex, orcreatinine, or of non-white/non-African American race; 32 sub-jects with a glomerular filtrationrate less than 15 mL/min/1.73 m2;1915 subjects aged more than 74years; 556 subjects missing base-line cardiovascular disease data;2922 subjects with baseline car-diovascular disease; 125 subjectsmissing baseline blood pressure,cholesterol, diabetes, or smokingdata; and 11 subjects without fol-low-up data.
utcomeshe primary cardiac outcome included myocardial infarc-
ion and fatal coronary heart disease. Consensus committeesithin ARIC and CHS had identical diagnostic criteria.yocardial infarction includes both clinically recognized
nd silent infarctions, diagnosed by electrocardiogramhanges, abnormal cardiac enzymes, and symptoms. Fataloronary disease required chest pain within 72 hours ofeath and/or known ischemic heart disease in the absence ofnother potential noncardiac cause of death.7,8 The second-ry outcome was all-cause mortality, because kidney dis-ase may be an important risk factor for other forms ofardiovascular disease.17,18
TATISTICAL ANALYSIS
ramingham Scorese reproduced the 5-year survival function for cardiac out-
omes by using individual patient data from the limited-accessramingham (11th visit) and Framingham Offspring (baselineisit) datasets, replicating Framingham techniques and defini-ions.2,16,19 With published sex-specific Framingham risk func-ions for predicting “hard coronary heart disease” outcomesnd the values for traditional coronary risk factors (bloodressure and lipid categories, age, diabetes, and smoking) fromur study population, we used Cox proportional hazards re-ression models with coefficients developed by the Framing-
CE
major populationin African Ameri-nd African Ameri-idney disease ac-d 12,000 coronary
and 63,000 andar after adjusting
es, the impact ofar to the impact of.
resence of chronicy does not improvediscrimination.
ICAN
as aarlyites aly, k0 anventser yerisk.
tcomsimil
oking
he pnerall
am investigators to calculate the FRS.2,3
mvTt
s5aati
AAd
mnwhCed
CTwfmadt
552.e3Weiner et al Kidney Disease and Framingham Risk
Next, sex- and race-specific Cox proportional hazardsodels were derived from the pooled cohort using the same
ariables as in the Framingham predictive instrument.hese “best Cox” models yielded different coefficients for
hese risk factors than the original FRS models.All analyses were sex- and race-specific. Proportions of
ubjects with chronic kidney disease were stratified by-year Framingham probability of coronary heart disease,nd these groups were compared using the Cochran-Armit-ge test for trend. The effects of FRS and kidney disease onime to cardiac and mortality outcomes, were analyzed us-ng Cox regression.
bsolute Riskdjusted absolute risk of cardiac and mortality outcomesirectly attributable to chronic kidney disease was deter-
Table 1 Baseline Characteristics and Frequency of Outcomes A
Men
White(n � 5282)
Age (mean � SD) 57.2 � 8.0History of diabetes (%)† 6.3Current smoker (%) 21.5Blood pressure categories (mm Hg)
Optimal (SBP � 120, DBP � 80) 45.2Normal (SBP � 130, DBP � 85) 22.1High normal (SBP � 140, DBP � 90) 15.8Stage I (SBP � 160, DBP � 100) 12.7Stage II-IV (SBP � 160, DBP � 100) 4.3
Total cholesterol categories (mg/dL)�160 9.0160-199 33.8200-239 38.4240-279 15.1�280 3.7
HDL cholesterol categories (mg/dL)�35 23.335-44 35.445-49 14.650-59 16.8�60 9.8
Framingham 5-y probability (%) 3.6% � 2.9%Framingham 10-y probability (%) 8.5% � 6.6%Kidney function
Creatinine (mg/dL) 1.0 � 0.2Estimated GFR (mL/min/1.73 m2) 87.5 � 17.6CKD (%) 4.6
Study of originARIC 4346CHS 936
Cardiac events 231Mortality events 211
SBP � systolic blood pressure; DBP � diastolic blood pressure; HDLfiltration rate; SD � standard deviation.
Continuous variables are presented as mean � SD. For creatinine, to cFor HDL and total cholesterol, to convert milligrams per deciliter to mil
†To duplicate the original FRS criteria, the history of diabetes was defi
mg/dL (7.8 mmol/L).ined to estimate the event reduction if the exposure,amely, kidney disease, were eliminated. Absolute riskas defined as Event Rate*(HRkidney disease�1), where theazard ratio for chronic kidney disease is derived from aox model that includes terms for FRS and kidney dis-ase, and the event rate is for individuals without kidneyisease.13
omparison with Other Risk Factorshe absolute risk associated with chronic kidney diseaseas compared with other dichotomous Framingham risk
actors (diabetes and smoking). We did not evaluateodifiable nondichotomous risk factors (hypertension
nd dyslipidemia) because these analyses would requireropping multiple terms from the FRS, making interpre-ation difficult. We derived an adjusted “best Cox” model
the Pooled Cohort
Women
Black(n � 1513)
White(n � 6548)
Black(n � 2374)
55.3 � 7.6 57.7 � 8.5 55.0 � 7.613.4 5.2 15.736.5 22.2 23.4
25.8 51.5 31.620.5 18.4 21.118.9 13.3 17.122.5 13.0 19.812.3 3.9 10.4
9.5 5.4 8.134.6 27.6 29.634.2 39.3 34.315.0 20.2 19.66.7 7.5 8.3
12.2 4.9 3.927.4 16.5 17.013.9 11.6 12.324.4 25.9 26.122.1 41.0 40.7
3.8% � 3.2% 1.1% � 1.5% 1.3% � 1.8%8.9% � 7.1% 2.8% � 3.8% 3.3% � 4.4%
1.0 � 0.2 0.8 � 0.1 0.8 � 0.299.8 � 21.7 88.3 � 19.2 101.9 � 23.4
2.4 6.1 3.2
1353 4981 2112160 1567 26260 91 54
110 144 89
-density lipoprotein; CKD � chronic kidney disease; GFR � glomerular
milligrams per deciliter to millimolar quantities, multiply values by 88.4.quantities, multiply values by 0.0259.the use of insulin or other diabetes medications or fasting glucose � 140
mong
� high
onvertlimolarned as
wcTmeais
PPAConopamnd
DDreiwgFFdfdsmcfsaC
RAemAacwA
FTm
eItticrw
ARkadympe1wsa
CD
FFwsed
552.e4 The American Journal of Medicine, Vol 120, No 6, June 2007
ith terms for kidney disease and individual Framinghamovariates to obtain the adjusted hazard ratio for diabetes.hen we determined the absolute risk of cardiac andortality outcomes directly attributable to diabetes. This
stimated the event reduction if diabetes were eliminateds an exposure, but kidney disease remained incorporatedn risk models. Identical analyses were performed formoking.
opulation Significanceopulation data were obtained on white and Africanmerican individuals, aged 18 years and older, from USensus Bureau Data (2000), and prevalence data werebtained for individuals with stage 3 and 4 chronic kid-ey disease, aged 20 years and older, from publicationsf NHANES 1999 to 2000 data.20,21 We determined therevalence of kidney disease based on these estimatesnd, in conjunction with race-specific absolute risk fromodels using Framingham coefficients, calculated the
umber of yearly events attributable to chronic kidneyisease.
iscriminationiscrimination is the ability of a prediction model to sepa-
ate those who had events from those who did not havevents and may be quantified by the c-statistic.3 We exam-ned the change in discrimination between models with andithout terms for kidney disease. For each demographicroup, we computed 4 c-statistics. The first applied theramingham function to the cohort. The second applied theramingham function to the cohort and added a kidneyisease term. The third used the “best Cox” model. Theourth used the “best Cox” model and added a kidneyisease term. Models evaluating discrimination are cen-ored at 5 years. C-statistics are compared using a nonpara-etric approach.22 To conceptualize the magnitude of
hanges in c-statistics, we removed variables one at a timerom the Framingham predictive instrument, calculated c-tatistics, and compared them with full models. Data werenalyzed using SAS Version 9.1 (SAS Institute Inc,ary, NC).
ESULTSmong 15,717 individuals without prior cardiovascular dis-
ase, 756 had chronic kidney disease, including 242 whiteen, 36 African American men, 401 white women, and 77frican American women. More than 80% of individuals
re from ARIC. Individuals with chronic kidney diseaseomprised between 2.4% and 6.1% of the study population,ith kidney disease more prevalent in whites than Africanmericans (Table 1).
ramingham Scoreshere was a significant association between the 5-year Fra-
ingham probability of coronary heart disease and the pres- hnce of kidney disease (Ptrend �.001) for all groups (Figure 1).ndividuals with kidney disease had higher event rates, par-icularly for cardiac events in African Americans, and mor-ality in all subgroups (Figure 2). In multivariable analyses,ncreasing FRS posed a highly significant risk for all out-omes, whereas kidney disease was a statistically significantisk factor for all outcomes except cardiac events in whiteomen (Table 2).
bsolute Riskates of cardiac and mortality events attributable to reducedidney function are presented in Figure 3. For white mennd women, the increase in cardiac events caused by kidneyisease was small (4.3 and 1.2 events per 1000 person-ears). The increase in cardiac events in African Americanen and women was more prominent (16.1 and 13.6 events
er 1000 person-years), as was the increase in the mortalityvents in all subgroups (13.7, 40.5, 5.8 and 14.2 deaths per000 person-years for white and African American men andhite and African American women, respectively). Analy-
es including a term for study of origin revealed similarbsolute risk (data not shown).
omparison with Other Risk Factorsiabetic persons and current smokers had significantly
igure 1 Frequency of chronic kidney disease by calculatedramingham probability of developing coronary heart diseaseithin 5 years, stratified by demographic group. Probability is
tratified by sex-specific quartiles. The P value for trend withinach demographic group is less than .001. CKD � chronic kidneyisease.
igher cardiac and mortality event rates than nondiabetic
phicp
PI2adtdderac6a
DTi
FyC
Fo
552.e5Weiner et al Kidney Disease and Framingham Risk
ersons and nonsmokers. Although diabetes and smokingad greater impact than kidney disease on cardiac eventsn whites, kidney disease in African Americans ac-ounted for a marked increase in cardiac events com-ared with diabetes and smoking (Figure 4).
igure 2 Rates of cardiac and mortality events per 1000 person-ears stratified by race, sex, and chronic kidney disease status.KD � chronic kidney disease.
Table 2 Results of Univariate Proportional Hazards Models anOutcomes (Hazard Ratio [95% Confidence Interval])
Men
White Blac
Univariate ModelsCardiac outcomes
FRS 2.08 (1.89-2.29) 1.77CKD 2.35 (1.68-3.28) 3.78
Mortality outcomesFRS 1.92 (1.74-2.10) 1.74CKD 3.75 (2.86-4.91) 4.81
Multivariable modelsCardiac outcomes
FRS 2.04 (1.85-2.25) 1.74CKD 1.48 (1.06-2.08)† 2.93
Mortality outcomesFRS 1.82 (1.65-2.00) 1.70CKD 2.56 (1.94-3.38) 3.76
CKD � chronic kidney disease; FRS � Framingham Risk Score.CKD is defined as eGFR � 60 mL/min/1.73 m2. Hazard ratio for FRS pr
terms for FRS and CKD. The SD for FRS for men was 0.8 for both white an
All P values were �.001, except *P �.01, †P �.05, and ‡nonsignificant.opulation Significancen this predominantly stage 3 chronic kidney disease cohort,.6 additional cardiac events per 1000 person-years in whitesnd 14.6 in African Americans were attributable to kidneyisease after accounting for traditional risk factors comprisinghe FRS. For mortality outcomes, approximately 9.3 additionaleaths per 1000 person-years in whites and 24.4 additionaleaths in African Americans were attributable to kidney dis-ase. When extrapolated to the US adult population, this rep-esents more than 17,000 additional annual cardiac eventsmong whites and approximately 12,000 additional annualardiac events among African Americans, and more than3,000 additional annual deaths among whites and 19,000dditional annual deaths among African Americans.
iscriminationhe addition of kidney disease to predictive models that
ncluded FRS did not improve model discrimination for
ivariable Cox Regression Models for Cardiac and Mortality
Women
White Black
2.15) 2.76 (2.41-3.16) 2.50 (2.05-3.05)8.15)* 1.86 (1.21-2.86)* 7.32 (4.22-12.69)
2.01) 2.22 (2.01-2.46) 1.94 (1.69-2.23)8.02) 2.57 (1.92-3.43) 4.12 (2.53-6.69)
2.12) 2.74 (2.40-3.14) 2.32 (1.90-2.84)6.35)* 1.33 (0.86-2.04)‡ 4.10 (2.32-7.23)
1.97) 2.19 (1.98-2.42) 1.87 (1.62-2.15)6.29) 1.94 (1.45-2.60) 2.60 (1.58-4.27)
isk associated with each 1 SD increase. Multivariable models only includean American men and 1.1 for both white and African American women.
igure 3 The absolute risk of outcomes because of the presencef chronic kidney disease after accounting for the FRS.
d Mult
k
(1.45-(1.76-
(1.50-(2.88-
(1.43-(1.36-
(1.47-(2.24-
esents rd Afric
cf3Aicstas
DIcdehepaokmicaAoi
ivIlcbitaer
pArd1tw
FckC
552.e6 The American Journal of Medicine, Vol 120, No 6, June 2007
ardiac outcomes, although it improved discriminationor mortality in all subgroups except white women (Table). Discrimination was consistently worse in Africanmericans, regardless of kidney disease status. Remov-
ng individual elements of the FRS did not markedlyhange the c-statistic in “best Cox” models. Results wereimilar for removing other risk factors, demonstratinghat discrimination based on multivariable models is rel-tively robust with the removal of a single term (data nothown).
igure 4 The increase in the rate per 1000 person-years ofardiac and mortality events because of the presence of chronicidney disease, diabetes, and smoking derived from adjusted “bestox” models. CKD � chronic kidney disease.
Table 3 Discrimination as Presented by the C-Index (Standard
FRS
Cardiac eventsWhite men 0.738 (0.015)African American men 0.644 (0.032)White women 0.783 (0.022)African American women 0.750 (0.033)
Mortality eventsWhite men 0.650 (0.019)‡African American men 0.641 (0.024)White women 0.707 (0.021)African American women 0.689 (0.028)†
FRS � Framingham Risk Score.“Best Cox” is a fitted Cox regression model using the same covariates
models that include a term for CKD. Comparisons are between c-scores for
†P �.10, and ‡P �.05.ISCUSSIONn the current study, we use a biracial, community-basedohort to assess sex- and race-specific risk of coronary heartisease and mortality associated with chronic kidney dis-ase after adjusting for FRS; determine the excess coronaryeart disease and mortality risk attributable to kidney dis-ase and compare this with the risk associated with com-osite elements of the FRS; and evaluate the impact ofccounting for kidney disease on the discriminative abilityf the Framingham risk equations. We found that bothidney disease and FRS independently predict cardiac andortality events, and that kidney disease was particularly
mportant in African Americans. Although the presence ofhronic kidney disease had a clinically important impact onbsolute risk of cardiac and mortality events, particularly infrican Americans, adding a term for kidney disease tother traditional risk factors did not improve model discrim-nation for cardiac outcomes.
The Framingham predictive instrument helps identifyndividuals at risk for cardiac outcomes, and its externalalidity has been demonstrated in US populations.3,16,23,24
n chronic kidney disease, nontraditional factors are preva-ent and may increase cardiac risk.5 In addition, traditionalardiac risk factors are highly prevalent in kidney diseaseut may be associated with a different magnitude of risk inndividuals with kidney disease than in the general popula-ion.25 Given this uncertainty, our goal was to assess thedditional impact of chronic kidney disease on adversevents beyond that accounted for by traditional Framinghamisk factors.
Our results confirm that chronic kidney disease is an im-ortant risk factor for cardiac and mortality outcomes.4,6,26
fter accounting for FRS, there was a 30% to 50% increasedisk of coronary heart disease and a 100% increased risk ofeath among whites with kidney disease, and an approximately50% to 300% increased risk of coronary disease and a 150%o 300% increased risk of death among African Americansith kidney disease. The mechanism by which kidney disease
) for Study Outcomes Stratified by Demographic Characteristics
CKD Best Cox Best Cox � CKD
(0.015) 0.750 (0.015) 0.751 (0.015)(0.033) 0.675 (0.033) 0.670 (0.033)(0.022) 0.803 (0.021) 0.804 (0.021)(0.034) 0.792 (0.028) 0.790 (0.029)
(0.019) 0.733 (0.016)* 0.737 (0.016)(0.025) 0.732 (0.020)* 0.741 (0.020)(0.021) 0.743 (0.020) 0.745 (0.020)(0.029) 0.716 (0.029)† 0.730 (0.029)
RS models but with coefficients generated for best fit. “� CKD” indicatesnd “FRS � CKD” and between “Best Cox” and “Best Cox � CKD.” *P �.20,
Error
FRS �
0.7400.6420.7840.754
0.6660.6480.7120.706
as the F“FRS” a
iciodim
huorApFreA
atdwcipo
dMcotchds
cesemrbrcodbtAanc
ptcauptHbemfimtakc
dtAappnm
CCvanab
AActyGDpSSlsoS
R
552.e7Weiner et al Kidney Disease and Framingham Risk
s an independent risk factor for cardiovascular disease out-omes is likely multifactorial. Chronic kidney disease maydentify individuals with a longer duration or greater severityf cardiovascular disease risk factors such as hypertension andiabetes. In addition, factors related to kidney disease, includ-ng anemia, left ventricular hypertrophy, and inflammation,ay increase cardiac risk.27-30
Approximately 8 million individuals in the United Statesave chronic kidney disease defined by an estimated glomer-lar filtration rate less than 60 mL/min/1.73 m2.31 On the basisf our results, adverse outcomes of chronic kidney diseaseepresent a sizable public health issue, particularly in Africanmericans, in whom 12,000 cardiac events and 19,000 deathser year are associated with kidney disease after accounting forramingham risk factors. Chronic kidney disease was compa-able to both diabetes and smoking when evaluating cardiacvents in African Americans and mortality events in Africanmerican men and women and white men.Adding chronic kidney disease to analyses that already
ccounted for the FRS only minimally improved discrimina-ion, indicating that there was little change in our ability toetermine who will and will not have an event. This findingas expected because our analyses also showed only minimal
hange in the c-statistic after removing individual risk factors,ncluding diabetes, from prediction models. The relatively lowrevalence of kidney disease in this cohort likely influences theverall lack of significant impact on discrimination.
Although the Framingham equations examined only car-iac outcomes, it is also important to investigate mortality.ortality is not subject to misclassification bias, and, be-
ause individuals with chronic kidney disease have numer-us competing outcomes, one method to account for this iso examine mortality.32-35 Finally, mortality outcomes willapture deaths that, although not considered “hard coronaryeart disease” events, are influenced by the role of kidneyisease on cardiac risk factors, including heart failure andystemic vascular disease.
A major finding of the current study was the marked in-reased risk of adverse outcomes associated with kidney dis-ase in African Americans. This adds to earlier findings bypecifying absolute risk beyond traditional risk factors andvaluating our ability to determine the risk of cardiac andortality outcomes using prediction equations.6 The profound
isk associated with kidney disease in African Americans maye attributable to differences in socioeconomic status or mayepresent geographic variability, because the majority of Afri-an Americans in this study were from the Mississippi cohortf ARIC. Other possibilities include greater severity, longeruration, or worse control of traditional risk factors that cannote accounted for in regression models. Suggestive evidence forhis comes from analysis of NHANES data in which Africanmerican race was an independent risk factor for failure to
chieve blood pressure goals in individuals with reduced kid-ey function, perhaps secondary to decreased awareness of
hronic kidney disease.36,37Our study has several limitations. Because most partici-ants with kidney disease had an estimated glomerular fil-ration rate greater than 40 mL/min/1.73 m2, we cannotomment on the utility of Framingham equations in moredvanced disease. We do not have data on microalbumin-ria, a component of kidney disease that independentlyredicts cardiovascular disease.15,38 Framingham risk equa-ions were derived to calculate cardiac risk, not mortality.owever, given the marked degree of competing eventsetween coronary heart disease outcomes and mortality, welected to analyze predictive ability on both cardiac andortality outcomes as other studies have done.35 Finally,ndings in this article are based on the supposition thatuch of the risk unaccounted for by the Framingham equa-
ions is attributable to chronic kidney disease. Although wecknowledge that some risk may not be directly caused byidney disease, adding kidney disease to statistical modelsaptures important information.
Our study has several strengths. We have a large andiverse population with thorough event ascertainment. Al-hough there may be inherent differences between CHS andRIC, pooling these studies provides a more generalizable
ge range and power to stratify analyses on race. Theseopulations have been pooled in several other studies, andrior work by our group demonstrated no statistically sig-ificant difference by study of origin on incident cardiac andortality outcomes after risk factor adjustment.6,39,40
ONCLUSIONhronic kidney disease is an important risk factor in indi-idual patients, identifying those at increased risk of cardiacnd mortality outcomes. In terms of absolute risk, the mag-itude is similar to that seen for diabetes mellitus. The riskssociated with chronic kidney disease is particularly nota-le in African Americans.
CKNOWLEDGMENTSuthor Contributions: Drs Weiner and Sarnak had full ac-
ess to all of the data in the study and take responsibility forhe integrity of the data and the accuracy of the data anal-sis. Study Concept and Design: Drs Weiner, Tighiouart,riffith, Elsayed, Levey, Salem, and Sarnak. Acquisition ofata: Drs Weiner, Levey, and Sarnak. Analysis and Inter-retation of Data: Drs Weiner, Tighiouart, Griffith, andarnak. Drafting of the Article: Drs Weiner, Tighiouart, andarnak. Critical Revision of the Article for Important Intel-
ectual Content: Drs Weiner, Griffith, Salem, Levey, El-ayed, and Sarnak. Statistical Analysis: Drs Weiner, Tighi-uart, and Griffith. Obtained Funding: Drs Levey andarnak. Study Supervision: Drs Weiner, Salem, and Sarnak.
eferences1. Wilson PW, Castelli WP, Kannel WB. Coronary risk prediction in
adults (the Framingham Heart Study). Am J Cardiol. 1987;59:
91G-94G.
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
552.e8 The American Journal of Medicine, Vol 120, No 6, June 2007
2. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H,Kannel WB. Prediction of coronary heart disease using risk factorcategories. Circulation. 1998;97:1837-1847.
3. D’Agostino RB Sr, Grundy S, Sullivan LM, Wilson P. Validation ofthe Framingham coronary heart disease prediction scores: results of amultiple ethnic groups investigation. JAMA. 2001;286:180-187.
4. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidneydisease and the risks of death, cardiovascular events, and hospitaliza-tion. N Engl J Med. 2004;351:1296-1305.
5. Sarnak MJ, Levey AS, Schoolwerth AC, et al. Kidney disease as a riskfactor for development of cardiovascular disease: a statement from theAmerican Heart Association Councils on Kidney in CardiovascularDisease, High Blood Pressure Research, Clinical Cardiology, andEpidemiology and Prevention. Circulation. 2003;108:2154-2169.
6. Weiner DE, Tighiouart H, Amin MG, et al. Chronic kidney disease as arisk factor for cardiovascular disease and all-cause mortality: a pooledanalysis of community-based studies. J Am Soc Nephrol. 2004;15:1307-1315.
7. The Atherosclerosis Risk in Communities (ARIC) Study: design andobjectives. The ARIC investigators. Am J Epidemiol. 1989;129:687-702.
8. Fried LP, Borhani NO, Enright P, et al. The Cardiovascular HealthStudy: design and rationale. Ann Epidemiol. 1991;1:263-276.
9. Levey AS GT, Kusek JW, Beck GJ. A simplified equation to predictglomerular filtration rate from serum creatinine. J Am Soc Nephrol.2000(Abstract);11:155A.
0. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A moreaccurate method to estimate glomerular filtration rate from serumcreatinine: a new prediction equation. Modification of Diet in RenalDisease Study Group. Ann Intern Med. 1999;130:461-470.
1. Coresh J, Astor BC, McQuillan G, et al. Calibration and randomvariation of the serum creatinine assay as critical elements of usingequations to estimate glomerular filtration rate. Am J Kidney Dis.2002;39:920-929.
2. Weiner DE, Tighiouart H, Stark PC, et al. Kidney disease as a riskfactor for recurrent cardiovascular disease and mortality. Am J KidneyDis. 2004;44:198-206.
3. Shlipak MG, Fried LF, Cushman M, et al. Cardiovascular mortalityrisk in chronic kidney disease: comparison of traditional and novel riskfactors. JAMA. 2005;293:1737-1745.
4. Merkin SS, Coresh J, Roux AV, Taylor HA, Powe NR. Area socio-economic status and progressive CKD: the Atherosclerosis Risk inCommunities (ARIC) Study. Am J Kidney Dis. 2005;46:203-213.
5. Levey AS, Coresh J, Balk E, et al. National Kidney Foundationpractice guidelines for chronic kidney disease: evaluation, classifica-tion, and stratification. Ann Intern Med. 2003;139:137-147.
6. Gordon T, Castelli WP, Hjortland MC, Kannel WB, Dawber TR.Predicting coronary heart disease in middle-aged and older persons.The Framingham study. JAMA. 1977;238:497-499.
7. Fried LF, Katz R, Sarnak MJ, et al. Kidney function as a predictor ofnoncardiovascular mortality. J Am Soc Nephrol. 2005;16:3728-3735.
8. Keith DS, Nichols GA, Gullion CM, Brown JB, Smith DH. Longitudinalfollow-up and outcomes among a population with chronic kidney diseasein a large managed care organization. Arch Intern Med. 2004;164:659-663.
9. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP.An investigation of coronary heart disease in families. The Framing-ham offspring study. Am J Epidemiol. 1979;110:281-290.
0. US Census Bureau. Population by Race and Hispanic or Latino Originfor the United States: 1990 and 2000 Detailed Tables (PHC-T-1). Table1.1 Population by Race and Hispanic or Latino Origin, for All Ages andfor 18 Years and Over, for the United States: 2000. March 12, 2001.Available at: http://www.census.gov/population/cen2000/phc-t1/tab01.xls. Accessed December 7, 2005.
1. Coresh J, Byrd-Holt D, Astor BC, et al. Chronic kidney diseaseawareness, prevalence, and trends among U.S. adults, 1999 to 2000.
J Am Soc Nephrol. 2005;16:180-188.2. DeLong ER, DeLong DM, Clarke-Pearson DL. Comparing the areasunder two or more correlated receiver operating characteristic curves:a nonparametric approach. Biometrics. 1988;44:837-845.
3. Liu J, Hong Y, D’Agostino RB Sr, et al. Predictive value for theChinese population of the Framingham coronary heart disease riskassessment tool compared with the Chinese Multi-Provincial CohortStudy. JAMA. 2004;291:2591-2599.
4. Empana JP, Ducimetiere P, Arveiler D, et al. Are the Framingham andPROCAM coronary heart disease risk functions applicable to differentEuropean populations? The PRIME Study. Eur Heart J. 2003;24:1903-1911.
5. Kalantar-Zadeh K, Block G, Humphreys MH, Kopple JD. Reverseepidemiology of cardiovascular risk factors in maintenance dialysispatients. Kidney Int. 2003;63:793-808.
6. Muntner P, He J, Hamm L, Loria C, Whelton PK. Renal insufficiencyand subsequent death resulting from cardiovascular disease in theUnited States. J Am Soc Nephrol. 2002;13:745-753.
7. Weiner DE, Tighiouart H, Vlagopoulos PT, et al. Effects of anemiaand left ventricular hypertrophy on cardiovascular disease in patientswith chronic kidney disease. J Am Soc Nephrol. 2005;16:1803-1810.
8. McClellan WM, Flanders WD, Langston RD, Jurkovitz C, Presley R.Anemia and renal insufficiency are independent risk factors for deathamong patients with congestive heart failure admitted to communityhospitals: a population-based study. J Am Soc Nephrol. 2002;13:1928-1936.
9. Knight EL, Rimm EB, Pai JK, et al. Kidney dysfunction, inflamma-tion, and coronary events: a prospective study. J Am Soc Nephrol.2004;15:1897-1903.
0. Sarnak MJ. Cardiovascular complications in chronic kidney disease.Am J Kidney Dis. 2003;41(5 Suppl):11-17.
1. Coresh J, Astor BC, Greene T, Eknoyan G, Levey AS. Prevalence ofchronic kidney disease and decreased kidney function in the adult USpopulation: Third National Health and Nutrition Examination Survey.Am J Kidney Dis. 2003;41:1-12.
2. Schwartz GG, Olsson AG, Ezekowitz MD, et al. Effects of atorvastatin onearly recurrent ischemic events in acute coronary syndromes: the MIRACLstudy: a randomized controlled trial. JAMA. 2001;285:1711-1718.
3. Schulgen G, Olschewski M, Krane V, Wanner C, Ruf G, SchumacherM. Sample sizes for clinical trials with time-to-event endpoints andcompeting risks. Contemp Clin Trials. 2005;26:386-396.
4. Wright JT Jr, Bakris G, Greene T, et al. Effect of blood pressure loweringand antihypertensive drug class on progression of hypertensive kidneydisease: results from the AASK trial. JAMA. 2002;288:2421-2431.
5. Aktas MK, Ozduran V, Pothier CE, Lang R, Lauer MS. Global riskscores and exercise testing for predicting all-cause mortality in apreventive medicine program. JAMA. 2004;292:1462-1468.
6. Peralta CA, Hicks LS, Chertow GM, et al. Control of hypertension inadults with chronic kidney disease in the United States. Hypertension.2005;45:1119-1124.
7. Nickolas TL, Frisch GD, Opotowsky AR, Arons R, Radhakrishnan J.Awareness of kidney disease in the US population: findings from theNational Health and Nutrition Examination Survey (NHANES) 1999to 2000. Am J Kidney Dis. 2004;44:185-197.
8. Hillege HL, Fidler V, Diercks GF, et al. Urinary albumin excretionpredicts cardiovascular and noncardiovascular mortality in generalpopulation. Circulation. 2002;106:1777-1782.
9. Howard G, Manolio TA, Burke GL, Wolfson SK, O’Leary DH. Doesthe association of risk factors and atherosclerosis change with age? Ananalysis of the combined ARIC and CHS cohorts. The AtherosclerosisRisk in Communities (ARIC) and Cardiovascular Health Study (CHS)investigators. Stroke. 1997;28:1693-1701.
0. Wong TY, Larsen EK, Klein R, et al. Cardiovascular risk factors forretinal vein occlusion and arteriolar emboli: the Atherosclerosis Riskin Communities & Cardiovascular Health studies. Ophthalmology.
2005;112:540-547.
B
HTAa
M
IAtievEpoI
vi
qb
Kp2
0d
The American Journal of Medicine (2007) 120, 552.e9-552.e11
RIEF OBSERVATION
igh FVIII Level Is Associated with Idiopathic Portal Veinhrombosis in South Indiabraham Koshy, M.D., D.M.,a,b Mary Jeyakumaria,b
Department of Gastroenterology, and bMolecular Biology Unit, Institute of Population Health and Clinical Research, St. John’s
edical CollegeBdMtRctbCt
E-mail: koshya
002-9343/$ -see foi:10.1016/j.amjm
Hospital, Bangalore, India.
ABSTRACT
ACKGROUND: In the majority of the cases of portal vein thrombosis in India, the etiology cannot beetermined. A high level of factor VIII (FVIII) is a risk factor for venous thromboembolism.ETHODS: In this study from South India, FVIII was estimated in 32 patients with idiopathic portal vein
hrombosis (PVT), 12 healthy controls and 12 disease controls with chronic liver disease.ESULTS: Eighty-four percent of the patients with portal vein thrombosis (27/32) vs 17% of healthyontrols (2/12) had high factor VIII levels (P �.0001). FVIII levels were significantly different betweenhe subset of PVT patients with normal liver function compared with healthy controls (P �.001) and alsoetween the subset of PVT patients with abnormal liver function compared with disease controls (P �.05).ONCLUSIONS: A high factor VIII level seems to contribute to the development of portal veinhrombosis in India. © 2007 Elsevier Inc. All rights reserved.
KEYWORDS: Genetics; Human genetic polymorphism; Portal Hypertension
MTdmtwhcabpoof
mgT4
oR
NTRODUCTIONlthough uncommon in Europe and the US, portal vein
hrombosis (PVT) is a frequent cause of portal hypertensionn India.1,2 In the majority of the cases of PVT in India, thetiology cannot be determined. Prothrombin G20210A geneariant, often associated with portal vein thrombosis inurope is absent in India,3 and Factor V Leiden, which alsoredisposes to portal vein thrombosis in Europe, is onlyccasionally associated with portal vein thrombosis inndia.4
Elevated plasma levels of FVIII is associated with deepein thrombosis,5 cerebral venous thrombosis,6 myocardialnfarction7 and bad obstetric history.8-10
The aim of the present study was to determine the fre-uency of elevated FVIII in patients with portal vein throm-osis in South India.
Address for correspondence and reprint requests: Professor Abrahamoshy, Department of Gastroenterology, St. John’s Medical College Hos-ital, Bangalore 560034, India. Tel: �91 80 2206 5239 Fax: �91 (80)553 0070
ront matter © 2007 Elsevier Inc. All rights reserved.ed.2006.02.016
ATERIALS AND METHODShirty-two consecutive patients with portal vein thrombosisiagnosed by Doppler and/or CT examination of the abdo-en between January 2003 and June 2005 were included in
he study. Patients with underlying cirrhosis or pancreatitisere excluded. Cirrhosis was excluded on the basis ofistory, physical findings, and ultrasonography. Acute pan-reatitis was excluded on history, physical findings, serummylase and lipase levels, and chronic pancreatitis on theasis of CT findings. One patient with combined hepatic andortal vein thrombosis who presented with clinical featuresf hepatic vein thrombosis was excluded. Buffy coat wasbtained from blood samples. Genomic DNA was extractedrom buffy coat using QIAamp DNA Midi Blood kits.
Twelve adult healthy controls from the same community,atched for age (27�4, M�SD), sex (6:6, M:F), and blood
roup (5:7, group O:group non-O) served as controls.welve patients with chronic liver disease (bilirubin.0�2.7 mg/dl, INR 1.7�0.5) served as disease controls.
Seventeen healthy first-degree relatives (parents/siblingsr children) of 9 of the patients also were tested for FVIII.esearch guidelines for the protection of human subjects of
he Ethical Committee, St. John’s Medical College, were
fed
acafvp
RTsbpwtdcgpstrSofrhupwsb
lbcP1
DOafet
fltaadtpscp
pafg
ds
wI
ein t
552.e10 The American Journal of Medicine, Vol 120, No 6, June 2007
ollowed. Consent was obtained from the patients after fullxplanation of the purpose, nature, and risks of all proce-ures used.
Factor VIII was measured by APTT based coagulationssay using fresh plasma samples and commercial FVIII defi-ient plasma (Dade Behring, New-rk, Del). All assays were per-ormed in duplicate and the averagealue was calculated for eacherson.
ESULTSwenty-six out of 32 (81%) pre-ented with upper gastrointestinalleeding, 3 with jaundice due toortal hypertensive biliopathy, 1ith abdominal pain, 1 with fa-
igue, and the other was diagnoseduring a routine check-up. Thelinical profile of the patients isiven in table 1. Among the 32atients with portal vein thrombo-is, 9 had associated splenic veinhrombosis and 4 associated supe-ior mesenteric vein thrombosis.even out of 32 (22%) patients were born of consanguine-us parents. None of the patients gave a past history oramily history of thrombosis. No patient gave history ofecent pregnancy, surgery, or trauma. None of the patientsad overt myeloproliferative disorder. One patient who hadpper GI bleeds and transfusions over 18 years was HBsAgositive 10 years after initial presentation. All other patientsere HBsAg, anti-HCV, and anti-HIV negative. Twenty-
even of the 32 patients with idiopathic portal vein throm-osis (84%) had high (�150%) FVIII level. The FVIII
Table 1 Clinical Characteristics of 32 Patients with PortalVein Thrombosis
M � SDa Range
Age (years) 27 � 15 5-60M:F 15:17Duration of disease (years) 3.6 � 4.6 0.1-18.0Liver span (cm) 8.8 � 2.1 5-14Spleen bcmb (cm) 6.1 � 5.6 0-20S. Albumin (g/L) 36 � 5 27-45S. Bilirubin (mg/dL) 2.3 � 5.4 0.2-30.0ALT (xULN)c 0.7 � 0.5 0.3-2.9Platelet count (x109/L) 107 � 59 30-230INR 1.4 � 0.3 0.9-2.8APTT (sec) 34 � 6 26-51Varices grade (0-4) 2 � 1 0-4Factor VIII (%) 272 � 131 88-580Blood group O: non-O 13:19
aMean � Standard deviationbbcm, below costal margin, in 30 patients, 2 had prior splenectomyc
CLINICAL SIGNIF
● The associationlevel, in the maidiopathic portalscribed for the fi
● Variceal bleedingtion of idiopathsis, is usually mcause of variceal
● The presence ofsuggests the poslants may have aiopathic portal v
rALT, alanine aminotransferase; ULN, upper limit of normal
evels were not different between patients with and withoutlood group O (272�134 vs. 266�125). Two of 12 (17%)ontrols had high (�150%) FVIII level (120�49; t-test,�.0001 vs PVT). FVIII level in liver disease controls were69�32. Patients with portal vein thrombosis and abnormal
liver function (defined as serumbilirubin�1.1 mg/dL or INR�1.3,n�15) tended to have higherFVIII level (307�148 vs238�111, P�.16) than those withnormal liver function. However,FVIII levels were significantlydifferent between the subset ofPVT patients with normal liverfunction compared with healthycontrols (t-test, P�.001) and alsobetween the subset of PVT pa-tients with abnormal liver functioncompared with disease controls (t-test, P�.05). Twelve of the 17(71%) family members tested hadhigh FVIII level (225�123, t-test,P�.01 vs controls). The 9 pa-tients, whose family memberswere studied, had FVIII levelsmore than 190% (341�137%).
ISCUSSIONur prospective study shows that high FVIII is significantly
ssociated with portal vein thrombosis. Plasma levels ofactor VIII are higher in those of non-O blood group. How-ver, no difference was observed in the FVIII level betweenhose patients with and without O blood group in our study.
None of the patients had malignancy, renal disease, in-ammatory disease, infection or pregnancy. Patients with
hese conditions were excluded from the study. FVIII is ancute phase reactant, but our patients had chronic rather thancute thrombosis of the portal vein. Although the exacturation of the portal vein thrombosis was difficult to de-ermine as they did not give history of such event, it wasresumably several months to years prior to collection ofamples for FVIII, as all patients had well established portalollaterals, and the average known duration of portal hy-ertension was 3.6 years.
Atherosclerosis is associated with FVIII elevation inlasma. The average age of our patients was only 27 years,nd none had clinical evidence of atherosclerosis. It there-ore appears unlikely that the elevation of FVIII in thisroup of patients with PVT may be due to atherosclerosis.
Familial clustering of high FVIII level may occur.11 Firstegree relatives of our patients with PVT had FVIII levelsimilar to patients and significantly higher than controls.
The present study is the only published series of patientsith PVT in whom details of high FVIII has been reported.
n a recent study from Western India, 6/26 (23%) was
CE
high factor VIIIof patients with
thrombosis is de-me.
major manifesta-tal vein thrombo-ed like any otherd.
othrombotic statety that anticoagu-apeutic role in id-hrombosis.
ICAN
of ajorityvein
rst ti
, theic poranagblee
a prsibilither
eported to have high FVIII level, but further details were
nwe
tGf
CAoe
AWc
R1
1
552.e11Koshy et al Factor VIII & Portal Vein Thrombosis
ot given.4 It appears that high FVIII level is associatedith idiopathic Portal Vein Thrombosis in India, an area
ndemic for portal vein thrombosis.Patients with PVT may have multiple predisposing fac-
ors. None of our patients had associated prothrombin20210A gene variant or factor V Leiden as a predisposing
actor.
ONCLUSIONhigh factor VIII level seems to contribute to the devel-
pment of portal vein thrombosis in South India, an areandemic for portal vein thrombosis.
CKNOWLEDGEMENTe thank Dr. Sitalakshmi for access to some of the healthy
ontrols.
eferences1. Cohen J, Edelman RR, Chopra S. Portal vein thrombosis: a review.
Am J Med. 1992;92:173-182.2. Koshy A, Deshpande JV, Bahl CM. Clinical differentiation between
portal vein thrombosis (PVT), non- cirrhotic portal fibrosis (NCPF)
and cirrhosis by discriminant analysis. Indian J Gastroenterol.1984;3:197-198.
3. Koshy A, Jeyakumari M. Prothrombin G20210A gene variant is notassociated with idiopathic portal vein thrombosis in an area endemicfor portal vein thrombosis. Ann Hematol. 2005;E Pub:13.
4. Shah SR, DasGupta A, Sharma A, et al. Thrombophilic conditions innon-cirrhotic portal vein thrombosis. Indian J Gastroenterol. 2005;24:205-210.
5. Kyrle PA, Minar E, Hirschl M, et al. High plasma levels of factor VIIIand the risk of recurrent venous thromboembolism. N Engl J Med.2000;343:457-462.
6. Cakmak S, Derex L, Berruyer M, et al. Cerebral venous thrombosis:clinical outcome and systematic screening of prothrombotic factors.Neurology. 2003;60:1175-1178.
7. Gorog DA, Rakhit R, Parums D, et al. Raised factor VIII is associatedwith coronary thrombotic events. Heart. 1998;80:415-417.
8. Glueck CJ, Pranikoff J, Aregawi D, et al. The factor V Leiden mutation,high factor VIII, and high plasminogen activator inhibitor activity: etiol-ogies for sporadic miscarriage. Metabolism. 2005;54:1345-1349.
9. senbach-Glaninger A, van TM, Krugluger W, et al. Elevated coagu-lation factor VIII and the risk for recurrent early pregnancy loss.Thromb Haemost. 2004;91:694-699.
0. Marietta M, Facchinetti F, Sgarbi L, et al. Elevated plasma levels offactor VIII in women with early recurrent miscarriage. J ThrombHaemost. 2003;1:2536-2539.
1. Schambeck CM, Hinney K, Haubitz I, et al. Familial clustering of highfactor VIII levels in patients with venous thromboembolism. Arterio-
scler Thromb Vasc Biol. 2001;21:289-292.
C
AS
T
icpaapesNraccwep
CAmpcrmentbcapcfmc
V9
0
The American Journal of Medicine (2007) 120, e1-e2
LINICAL COMMUNICATIONS TO THE EDITOR
twcwrrirkraoc
DIsr
prFwuwsd
ircoit0ha1pppt3
p
Young Woman with Neurofibromatosis andevere Hypertension
o the Editor:
The 2 most common secondary causes of hypertensionn patients with neurofibromatosis type 1 (NF-1) are pheo-hromocytoma and renovascular disease.1,2 The conditionsresent clinically in 2 distinct age patterns. In the pedi-tric population, renovascular pathology predominates assecondary cause and pheochromocytoma is rare. Adult
atients with NF-1 and hypertension have a higher thanxpected incidence of pheochromocytoma.1 A repetitiveearch for a pheochromocytoma in an adult patient withF-1 and severe hypertension may delay the diagnosis of
enovascular disease as an alternative cause. The associ-tion of NF-1 with renovascular disease may be lessommonly appreciated by practitioners of adult medi-ine. We present a case of severe hypertension in a youngoman with NF-1 whose diagnosis of renovascular dis-
ase was delayed because of a prolonged search for aheochromocytoma.
ASE REPORT23-year-old woman presented to our general internal
edicine clinic with a medical history of NF-1 and 3regnancies complicated by midterm fetal demise asso-iated with severe hypertension over a 3-year period. Aeview of her records from an outside hospital docu-ented normal urinary and serum evaluations for cat-
cholamines metabolites on multiple occasions. She de-ied paroxysmal headaches, palpations, sweating, andremulousness. Initial physical examination revealed alood pressure of 190/130 mm Hg with no orthostatichanges, no evidence of congestive heart failure, and nobdominal bruits. Skin examination demonstrated multi-le cutaneous nodules over her entire body and multipleafé au lait spots. Admission laboratory revealed theollowing: Na 138 mmol/L, HCO3 26 mmol/L, Cl 101mol/L, K 3.9 mmol/L, blood urea nitrogen 17 mg/dL,
reatinine 0.8 mg/dL, and a urinalysis negative for pro-
Requests for reprints should be addressed to Robert Gold, MD, Oliveiew-UCLA Medical Center, 14445 Olive View Drive, Sylmar, CA4123.
tE-mail address: [email protected]
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
ein and cells. Two collections of 24-hour metanephrineere 841 and 882 �g (normal range 3-900 �g). Plasma
atecholamines obtained during a hypertensive episodeere as follows: norepinephrine 384 pg/mL (normal
ange 78-520 pg/mL) and epinephrine 10 pg/mL (normalange 10-196 pg/mL). An abdominal magnetic resonancemage was negative for adrenal pathology. A captoprilenal flow study revealed a diminution of flow to the rightidney. A subsequent arteriogram showed a 75% rightenal artery stenosis. The patient underwent a right renalortorenal bypass. Her blood pressure normalized with-ut medication after the procedure. She subsequentlyarried a pregnancy to term without complications.
ISCUSSIONn the pediatric population, the most frequent cause ofecondary hypertension in neurofibromatosis is attributed toenovascular disease.2
The first reports of the association between renovascularathology and NF-1 appeared in Europe in 1945.3 The firsteport in the English medical literature appeared in 1965.3
ossali et al.4 reported on a group of 27 children with NF-1,ith a mean age of 12.8 years, 13 of whom agreed tondergo digital subtraction angiography. Seven childrenere found to have renal artery stenosis (severe hyperten-
ion developed in 3 children, and borderline hypertensioneveloped in 1 child).
The adult medical literature is replete with cases not-ng the association between pheochromocytoma and neu-ofibromatosis. The classic association between the 2onditions was first described in 1910.1 A recent reviewf the association between NF-1 and pheochromocytoma,ncluding more than 118 articles culled from the litera-ure, documents an incidence of pheochromocytoma from.1% to 5.7%, including patients both with and withoutypertension.1 The mean age of the patients (84 womennd 61 men) reviewed in this study was 42 years (range,.5-74 years). Notably, sustained hypertension wasresent in 61% of the patients with NF-1 and documentedheochromocytoma. In comparison, the incidence ofheochromocytoma in the hypertensive general popula-ion is 0.1%.5 Autopsy studies report an incidence of.3% to 13.0% in patients with known NF-1.1
In summary, the most common secondary cause of hy-ertension in adult patients with NF-1 is pheochromocy-
oma, whereas renovascular disease predominates in pedi-
aasm
R1
2
3
4
5
e2 The American Journal of Medicine, Vol 120, No 6, June 2007
tric patients. The overlap age range between childhood anddulthood in hypertensive NF-1 should alert the clinician toearch with equal rigor for both of these otherwise uncom-on disorders.
Robert Gold, MDFredric Adler, MD
Tamara Modilevsky, MDKatherine Yu, MD
Olive-View-University of California at Los Angeles Medical CenterLos Angeles
doi:10.1016/j.amjmed.2006.02.011
eferences. Walther MM, Herring J, Enquist E, et al. Von Recklinghausen’s disease
and pheochromocytoma. J Urol. 1999;162:1582-1586.. Westenend PJ, Smedts F, de Jong MCJW, et al. A 4-year old boy with
neurofibromatosis and severe renovascular hypertension due to renalarterial dysplasia. Am J Surg Pathol. 1994;18:512-516.
. Blum MD, Bahnson RR, Carter MF. Urologic manifestations of vonRecklinghausen neurofibromatosis. Urology. 1985;26:209.
. Fossali E, Signorini E, Inemite RC. Renovascular disease and hyper-tension in children with neurofibromatos. Pediatr Nephrol. 2000;14:806-810.
. Landsberg L, Young B. Pheochromocytoma. In: Kasper DL, Fauci AS,Longo DL, et al. Harrison’s Principles of Internal Medicine, 16th ed.
Columbus, Ohio: McGraw-Hill Companies, Inc.; 2005.
C
Ta
T
opFbAp
ewahwtda1aruadslnestannwe
MC
0
The American Journal of Medicine (2007) 120, e3
LINICAL COMMUNICATIONS TO THE EDITOR
adacprduhcs
st9wbwnprlt
oippb
R1
hyrotoxic Hypokalemic Periodic Paralysis inWhite Man
o the Editor:
Hypokalemic periodic paralysis is a heterogeneous groupf disorders in which a precipitous decrease in the serumotassium level leads to sudden muscle weakness or paralysis.amilial, sporadic, and thyrotoxicosis-associated cases haveeen reported, with the latter almost exclusively reported insian males. We report a rare case of thyrotoxic hypokalemiceriodic paralysis in a white man.
A 29-year-old white man awakened with sudden lower-xtremity paralysis and upper-extremity weakness. Eighteeks prior, he reported weight loss and generalized weakness
nd was diagnosed with hyperthyroidism. Thyroid-stimulatingormone was 0.04 �U/mL (0.34-5.6 �U/mL), and total T4as 21.0 �g/dL (4.5-12.5 �g/dL). He had not yet begun
reatment. The patient reported no family history of thyroidisease or hypokalemic paralysis. On examination, he wasnxious, diaphoretic, and tremulous with a resting heart rate of04 beats/min and blood pressure of 166/54 mm Hg. His motorctivity was limited to moving his digits only. Deep tendoneflexes were absent in his lower extremities and reduced in thepper extremities. His serum potassium was markedly reducedt 1.2 mEq/L (3.5-5.5 mEq/L). The patient’s electrocardiogramemonstrated sinus tachycardia, diffuse ST segment depres-ion, and first-degree arteriovenous block. The patient receivedow-dose potassium (28 mEq orally and 20 mEq intrave-ously) and propranolol (10 mg orally). Within 4 hours, hexperienced dramatic symptomatic improvement, and a repeaterum potassium level was 3.4 mEq/L. His inpatient serumhyroid-stimulating hormone was undetectable, and total T3nd free T4 were markedly elevated at 397 ng/dL (79-149g/dL) and 5.34 ng/dL (0.7-1.8 ng/dL), respectively. A diag-osis of Graves’ disease was made, and he was dischargedith propranolol. He received I-131 ablation and has experi-
nced no further episodes of weakness.
Requests for reprints should be addressed to Kjersti Meyer Kirkeby,D, California Pacific Medical Center, Department of Medicine, 2351lay Street, Suite 380, San Francisco, CA 94115.
E-mail address: [email protected]
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
Thyrotoxic hypokalemic periodic paralysis occurs inpproximately 2% of all thyrotoxic patients of Asianescent, with a male to female ratio of 20:1.1 The prev-lence is unknown in white persons, but only a handful ofase reports document thyrotoxic hypokalemic periodicaralysis in this group. The hypokalemia results fromapid intracellular potassium shift rather than total bodyepletion.1 The inciting mechanism is not completelynderstood; a thyroid hormone-sensitive channelopathyas been proposed, possibly of the Na�/K�ATPase or aalcium channel.1 Hyperinsulinemia and hyperadrenergictates may also play a role in the pathogenesis.1
Management of an acute attack is controversial. Potassiumupplementation can result in rebound hyperkalemia once po-assium redistributes extracellularly, particularly if more than0 mEq is given.1 Doses of 50 mEq or less seem safe, althoughell-substantiated guidelines are lacking. Nonselective beta-lockers have also demonstrated efficacy in the acute setting,ith some authors recommending their use as first-line agents,oting their rapid success in a number of patients in whomotassium supplementation was ineffective.1 To prevent recur-ences during the remainder of the thyrotoxic period, nonse-ective beta-blockers are considered first-line therapy. Resolu-ion of the hyperthyroidism is considered definitive treatment.1
Clinicians should suspect thyrotoxic hypokalemic peri-dic paralysis in anyone presenting with sudden-onset prox-mal paralysis, even in those of non-Asian descent. Appro-riate use of propranolol and low-dose potassium can helprevent morbidity and mortality from this potentially fatal,ut easily treatable, disorder.
Kjersti Meyer Kirkeby, MDDavid M. Naeger, MD
Allan Pont, MD, FACPThomas E. Baudendistel, MD, FACP
California Pacific Medical CenterDepartment of Medicine
San Francisco
doi:10.1016/j.amjmed.2006.02.012
eference. Lin SH. Thyrotoxic periodic paralysis. Mayo Clin Proc. 2005;80(1):
99-105.
L
BA
T
bmtpudAcuaup
tC“cdposwft
ewqvmfbatpccs
0
The American Journal of Medicine (2007) 120, e5
ETTER
rctmae
dr
tpewd
R1
2
3
4
5
6
7
8
9
aclofen Not Comparable to Diazepam forlcohol Withdrawal
o the Editor:
The clinical trial by Addolorato and colleagues1 comparingaclofen with diazepam has several significant limitations thatake interpretation of the results difficult. Unfortunately, as
he authors noted, the 2 randomized groups were not com-arable in withdrawal severity. The authors also used annusual primary end point—the proportional change in Ad-iction Research Foundation Clinical Institute Withdrawalssessment for Alcohol (CIWA-Ar)—which is of unclear
linical significance and, to our knowledge, has not beensed as an end point in alcohol withdrawal trials. Theuthors do not state that these end points and their analysissing analysis of variance and analysis of covariance wererespecified.
The authors report that “baclofen was slightly slower”han diazepam by virtue of having significantly higherIWA-Ar scores on days 2 and 3 (P �.05), but that onsubsequent days the efficacy . . . was comparable.” Inontrast with the authors’ interpretation, this suggests thatiazepam is superior to baclofen. In the majority ofatients, alcohol withdrawal symptoms resolve spontane-usly without treatment in less than 3 days, and CIWAcores decline rapidly even when patients with moderateithdrawal are treated with placebo.2,3 The lack of dif-
erence from days 4 to 10 is necessary, but not sufficient,o conclude equivalence.
The dosing of diazepam is another major concern. Diaz-pam was divided into 6 doses per day, whereas baclofenas given in 3 doses per day. This diazepam dosing fre-uency could have affected treatment compliance in fa-or of baclofen. The long half-life of diazepam and itsetabolites makes this dosing frequency unnecessary. In
act, the diazepam dosing used in the control arm has noteen compared with placebo or another established ther-py; the authors cite 2 published articles, yet neither ofhese articles supports this dosing regimen.4,5 It is im-ossible to establish the efficacy of a study drug (ba-lofen) when the active control (diazepam) has not beenonclusively established to be effective in this dosing
cheme.6002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
The authors report that “the type 2 error was seen to beeasonably low,” but they do not describe their power cal-ulations, what margin of equivalence they used, or whatype 2 error they considered acceptable. These elementsust be reported for readers to conclude that the study was
dequately powered to detect a clinically significant differ-nce between the treatments.7
Finally, the treatment was single-blinded, which intro-uces potential bias into the primary outcome measure, theelatively subjective CIWA-Ar score.
Benzodiazepines have been studied in double-blind, con-rolled trials in a variety of clinical settings and patientopulations, and have reproducibly demonstrated safety andfficacy in a number of hard clinical outcomes.5,8,9 Beforee consider the efficacy of baclofen to be comparable toiazepam, we need considerably more evidence.
Alec B. O’Connor, MDValerie J. Lang, MD
University of Rochester School of Medicine and DentistryRochester, NY
doi:10.1016/j.amjmed.2006.03.041
eferences. Addolorato G, Leggio L, Abenavoli L, et al. Baclofen in the treatment
of alcohol withdrawal syndrome: a comparative study vs diazepam.Am J Med. 2006;119:276.e13-e18.
. Sellers EM, Naranjo CA, Harrison M, et al. Diazepam loading: simpli-fied treatment of alcohol withdrawal. Clin Pharmacol Ther. 1983;34(6):822-826.
. Salloum IM, Cornelius JR, Daley DC, Thase ME. The utility of diaz-epam loading in the treatment of alcohol withdrawal among psychiatricinpatients. Psychopharmacol Bull. 1995;31(2):305-310.
. Saitz R, Mayo-Smith MF, Roberts MS, et al. Individualized treatmentfor alcohol withdrawal. JAMA. 1994;272:519-523.
. Mayo-Smith MF. Pharmacological management of alcohol withdrawal:a meta-analysis and evidence-based practice guideline: American Soci-ety of Addiction Medicine Working Group on Pharmacological Man-agement of Alcohol Withdrawal. JAMA. 1997;278(2):144-151.
. Temple R, Ellenberg SS. Placebo-controlled trials and active-controlledtrials in the evaluation of new treatments. Ann Intern Med. 2000;133:455-463.
. Le Henanff A, Giraudeau B, Baron G, Ravaud P. Quality of reportingof noninferiority and equivalence randomized trials. JAMA. 2006;295:1147-1151.
. Kosten TR, O’Connor PG. Current concepts: management of drug andalcohol withdrawal. N Engl J Med. 2003;348:1786-1795.
. Mayo-Smith MF, Beecher LH, Fischer TL, et al. Management of alco-hol withdrawal delirium—an evidence-based practice guideline. Arch
Intern Med. 2004;164:1405-1412.
L
T
oc
aftst(gComtvtwaptlpsioewqe
ops(
aSAooao
otiadPe(
lsistma
ndbt
R1
2
3
4
0
The American Journal of Medicine (2007) 120, e7
ETTER
he Reply:
We appreciate the interest by O’Connor and Lang inur work,1 although we are not convinced by most of theiromments, specifically as follows.
The randomization was performed as a 1:1 procedure,nd the distribution was not predictable. However, the dif-erent severity of alcohol withdrawal syndrome (AWS) be-ween the 2 groups at baseline constitutes a further elementupporting the baclofen efficacy in AWS. In fact, at baselinehe mean total Clinical Institute Withdrawal AssessmentCIWA-Ar) score was significantly higher in the baclofenroup than in the diazepam group, whereas there were noIWA-Ar score differences between the 2 groups at the endf the study. Moreover, both baclofen and diazepam treat-ents significantly decreased the CIWA-Ar score during
he study, with no significant treatment differences.1 Atariance with the comment made by O’Connor and Lang,he measured variable that underwent statistical analysisas the score assigned to CIWA-Ar or its subscales (ie,
nxiety, agitation, sweating, and tremors) and not the pro-ortion of changes. The latter parameter was used only inhe “Discussion” section when we commented on the col-ected results. Finally, pertaining to the “type 2 error” issue,lease note that according to the statistician at the Univer-ity of Cagliari, no statistical package presently exists thatncludes those functions to directly calculate the occurrencef the type 2 error in analysis of covariance. However, thefficacy of the 2 drugs reached a significant difference onlyhen the factor sample size was multiplied 5-fold. Subse-uently, as written in the article, the occurrence of a type 2rror was thought to be reasonably low.
Contrary to that affirmed by O’Connor and Lang, thebservation that “baclofen was slightly slower” than diaze-am on days 2 and 3 refers only to a single CIWA-Arubscale (ie, agitation) and not to the total CIWA-Ar scoreplease refer to page 276.e16, lines 4-14).
We do not agree with the comment made by O’Connornd Lang on the frequency of diazepam administration.pecifically, Shaw2 reported the efficacy of diazepam onWS when fractioned in 4 administrations per day. On thether hand, Sullivan et al3 reported the efficacy of diazepamn AWS when administered hourly. In agreement with thebove literature data, and as specified in the article,1 previ-
us findings from our laboratory showed that the fractioning002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
f 0.5 to 0.75 mg/kg diazepam in a “medium way” of 6imes per day was both effective and manageable in decreas-ng AWS symptoms in these patients. Accordingly, a seminalrticle by Hall and Zador4 suggested, as a treatment guideline,oses of 5 to 10 mg diazepam given every 4 to 6 hours.ertaining to compliance, in our survey there were no differ-nces between groups of treatment, as stated in the articleplease refer to page 276.e15, “Results” section, lines 1-3).
We agree that the single-blind design could represent aimit of the study. However, it can be reasonably hypothe-ized that the experimental design did not result in anynfluence on the subjective CIWA-Ar score because, aspecified in the article, the investigators who administeredhe CIWA-Ar questionnaire at the different times of treat-ent were always the same individuals and were unaware
s to which drug was being administered to patients.In conclusion, although studies on a larger sample are
eeded, our study suggested that baclofen is comparable toiazepam in the treatment of uncomplicated forms of AWS,ecause CIWA-Ar decreased in baclofen- and diazepam-reated patients without significant treatment differences.1
Giovanni Addolorato, MDLorenzo Leggio, MD
Giovanni Gasbarrini, MDInstitute of Internal MedicineCatholic University of Rome
Rome, Italy
Roberta Agabio, MD“Bernard B. Brodie” Department of Neuroscience
University of CagliariCagliari, Italy
Giancarlo Colombo, PhDC.N.R. Institute of Neuroscience, Section of Cagliari
Cagliari, Italy
doi:10.1016/j.amjmed.2006.05.037
eferences. Addolorato G, Leggio L, Abenavoli L, et al. Baclofen in the treatment
of alcohol withdrawal syndrome: a comparative study vs. diazepam.Am J Med. 2006;119:276e13-276e.18.
. Shaw GK. Detoxification: the use of benzodiazepines. Alcohol Alcohol.1995;30:765-770.
. Sullivan JT, Swift RM, Lewis DC. Benzodiazepine requirements duringalcohol withdrawal syndrome: clinical implications of using a standard-ized withdrawal scale. J Clin Psychopharmacol. 1991;11:291-295.
. Hall W, Zador D. The alcohol withdrawal syndrome. Lancet. 1997;349:
1897-1900.
L
BC(
T
sbegcTcna
bdomurur
0
The American Journal of Medicine (2007) 120, e9
ETTER
a(t
otdc3t
R1
2
3
aclofen for Alcohol Withdrawal: Notomparable to the Gold StandardBenzodiazepines)
o the Editor:
Addolorato et al1 showed that alcohol withdrawal (AW)ymptoms decrease similarly when treated with diazepam oraclofen (albeit more slowly in the baclofen group). How-ver, their conclusions that baclofen is comparable to theold standard benzodiazepines and that baclofen may beonsidered a useful drug for treating AW are unwarranted.he authors do limit their recommendation to “uncompli-ated” AW, but whether seizures or delirium will develop isot known until after a treatment decision has been made,nd the course is completed.
Benzodiazepines are the gold standard because they haveeen proven in clinical trials to prevent AW seizures andelirium tremens.2 Although there are concerns with the usef benzodiazepines, these are not paramount in the treat-ent of AW, particularly when shorter-acting agents are
sed and dosed according to symptoms in patients at highestisk (eg, hepatic synthetic dysfunction and respiratory fail-re). Central nervous system depressant effects are less of a
isk in people tolerant to alcohol. The risk of new benzodi-002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
zepine addiction from short-term management of AWwhich in most cases is �3 days)3 is unknown but not likelyo be high.
Many medications can reduce AW symptoms, but onlyne class has been proven to reduce complications.2 Al-hough a medication for AW that has no side effects isesirable, benzodiazepines should not be replaced by ba-lofen as the gold standard on the basis of a single study of7 patients that did not determine whether it could preventhe morbid and mortal consequences.
Richard Saitz, MD, MPHBoston University School of Medicine
Boston, Mass
doi:10.1016/j.amjmed.2006.03.015
eferences. Addolorato G, Leggio L, Abenavoli L, et al. Baclofen in the treatment
of alcohol withdrawal syndrome: a comparative study vs. diazepam.Am J Med. 2006;119:276.e13-e18.
. Mayo-Smith MF. Pharmacological management of alcohol withdrawal:a meta-analysis and evidence-based practice guideline: American Soci-ety of Addiction Medicine Working Group on Pharmacological Man-agement of Alcohol Withdrawal. JAMA. 1997;278:144-151.
. Sellers EM, Naranjo CA, Harrison M, Devenyi P, Roach C, Sykora K.Oral diazepam loading: simplified treatment of alcohol withdrawal. Clin
Pharmacol Ther. 1983;34:822-826.
L
T
psdoibbf
wwct
pztph
ctpcvsl
0
The American Journal of Medicine (2007) 120, e11
ETTER
aera
R1
2
3
4
5
6
7
he Reply:
We appreciate the interest of Saitz in our work. Althougherformed in a small group of patients, our comparativetudy showed that baclofen is comparable to the gold stan-ard diazepam in the treatment of the uncomplicated formf alcohol withdrawal syndrome (AWS), because the Clin-cal Institute Withdrawal Assessment score decreased inaclofen- and diazepam-treated patients without differencesetween groups.1 A slightly slower effect of baclofen wasound in only 1 of the AWS symptoms (agitation).
Among baclofen-treated patients, no seizures or deliriumas noted during treatment or when the course of therapyas finished and in the subsequent follow-up. This might be
onsistent with the case of suppression of alcohol deliriumremens by baclofen administration recently described.2
Moreover, baclofen presented no relevant side effects inatients treated for alcoholism,1-5 whereas the use of ben-odiazepines is associated with several side effects, such ashe risk of excess sedation and respiratory depression inatients with liver impairment, as is often the case in alco-ol-dependent persons.6
Finally, it should be stressed that baclofen showed its effi-acy in reducing alcohol craving and intake in the long-termreatment of alcohol-dependent patients.3,4 This suggests aotential use of this drug as a promising and unique pharma-otherapy for alcohol dependence, from AWS to relapse pre-ention treatment. These effects should theoretically result in aimplified pharmacotherapy for patients with alcohol prob-ems, increasing the compliance to treatment.
Obviously, further studies on a larger sample of patients
002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
re required to confirm our data and to evaluate the baclofenffect on AWS complication. However, at present baclofenepresents a promising medication for treating alcoholismnd merits further investigation.7
Giovanni Addolorato, MDLorenzo Leggio, MD
Giovanni Gasbarrini, MDInstitute of Internal MedicineCatholic University of Rome
Rome, Italy
doi:10.1016/j.amjmed.2006.03.027
eferences. Addolorato G, Leggio L, Abenavoli L, et al. Baclofen in the treatment
of alcohol withdrawal syndrome: a comparative study vs. diazepam.Am J Med. 2006;119:276.e13-18.
. Addolorato G, Leggio L, Abenavoli L, et al. Suppression of alcoholdelirium tremens by baclofen administration: a case report. Clin Neu-ropharmacol. 2003;26:258-262.
. Addolorato G, Caputo F, Capristo E, et al. Baclofen efficacy in reducingalcohol craving and intake: a preliminary double-blind randomizedcontrolled study. Alcohol Alcohol. 2002;37:504-508.
. Flannery BA, Garbutt JC, Cody MW, et al. Baclofen for alcohol de-pendence: a preliminary open-label study. Alcohol Clin Exp Res. 2004;28:1517-1523.
. Johnson BA, Swift RM, Addolorato G, et al. Safety and efficacy ofGABAergic medications for treating alcoholism. Alcohol Clin Exp Res.2005;29:248-254.
. Mayo-Smith MF. Pharmacological management of alcohol withdrawal:a meta-analysis and evidence-based practice guideline: American Soci-ety of Addiction Medicine Working Group on Pharmacological Man-agement of Alcohol Withdrawal. JAMA. 1997;278:144-151.
. National Institute on Alcohol Abuse and Alcoholism. Neuroscience
research and therapeutic targets. Alcohol Alert. 2004;61:1-6.
L
IHG
T
BaeadHocr(rhjTgo
0
The American Journal of Medicine (2007) 120, e13
ETTER
srnarpcp
R1
2
3
ndications of Capsule Endoscopy inenoch-Schönlein Syndrome withastrointestinal Symptoms
o the Editor:
We read with interest the article, “Bleeding of Smallowel in Henoch-Schönlein Syndrome: The Successful Di-gnostic Role of Video Capsule Endoscopy” by Stancanellit al.1 They used capsule endoscopy in this patient to makedifferential diagnosis between steroid-induced gastroduo-enal bleeding and occult bleeding as a consequence ofenoch-Schönlein purpura (HSP). However, the diagnosisf steroid-induced gastroduodenal bleeding could be ex-luded by normal gastroduodenoscopic findings, and theyegarded the important findings of computed tomographyCT) scan showing edema of the small bowel as negativeesults. Nistala et al reported an HSP patient with jejunalemorrhage who showed a gross mucosal thickening ofejunum on CT scan, suggesting intramural hemorrhage.2
his finding is frequently observed in HSP patients withastrointestinal symptoms and might be an important clue
f the intestinal vasculitis. Also, abdominal pain or even002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
evere mesenteric vasculitis is generally an acute event andesponds well to prednisolone or high-dose methylpred-isolone pulse therapy.3 Therefore, capsule endoscopy indult patients with HSP should be considered in the cases ofecurrent or persistent gastrointestinal bleeding, or sus-ected other chronic inflammatory or neoplastic disordersausing chronic blood loss rather than acute vasculiticrocess.
Jae Il ShinJae Seung Lee
Department of PediatricsYonsei University College of Medicine
Seoul, Korea
doi:10.1016/j.amjmed.2006.02.041
eferences. Stancanelli B, Vita A, Vinci M, Magnano A, Purrello F. Bleeding of
small bowel in Henoch-Schonlein syndrome: the successful diagnosticrole of video capsule endoscopy. Am J Med. 2006;119:82-84.
. Nistala K, Hyer W, Halligan S. Jejunal haemorrhage in Henoch-Schon-lein syndrome. Arch Dis Child. 2003;88:434.
. Wang L, Huang FC, Ko SF, Cheng MT. Successful treatment of mes-enteric vasculitis caused by Henoch-Schonlein purpura with methyl-
prednisolone pulse therapy. Clin Rheumatol. 2003;22:140-142.
L
AS
T
la
N
Fgap
0
The American Journal of Medicine (2007) 120, e15-e16
ETTER
ii
ctpa
ts
bnormal Valsalva Maneuver Is Not Always aign of Congestive Heart Failure
o the Editor:
Felker and colleagues1 are to be commended for high-ighting the clinical utility of the Valsalva maneuver. This isclassic physical examination tool that is easily performed
Supported in part by grants K23 RR020783 and P01 HL56993 from theational Institutes of Health.
igure 1 Absent sympathetically mediated vasoconstriction inenerated Valsalva pressure are shown top to bottom. The blood prend blood pressure after release of Valsalva does not recover back
ressure.002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
n the office or at the bedside. Unfortunately, this simple tests increasingly being forgotten in clinical medicine.
We commonly perform the Valsalva maneuver in ourlinic, because it provides an integrated assessment of au-onomic nervous system function. On the basis of our ex-erience, we would like to add a couple of caveats to theuthors’ excellent review of the Valsalva maneuver.
First, the authors refer to the “absent overshoot” of sys-olic blood pressure after the release of Valsalva as repre-enting mild heart failure. They show a nice example in
mic nervous system failure. The heart rate, blood pressure, andecreases throughout the Valsalva maneuver without late recovery,baseline blood pressure, let alone “overshoot” the baseline blood
autonossure dto the
tsfswF
bufpitptv
tu
R1
2
3
Fs
e16 The American Journal of Medicine, Vol 120, No 6, June 2007
heir Figure 1B. We often see this pattern in patients withympathetic nervous system failure in the absence of heartailure (Figure 1).2 This presumably results from the lack ofympathetically mediated vasoconstriction that, if present,ould lead to the increased peripheral resistance to whichelker et al1 allude.
Second, the authors refer to the “square wave” pattern oflood pressure in response to the Valsalva maneuver (Fig-re 1C in Felker et al1) as being a sign of severe heartailure. This is certainly a common cause of this abnormalattern. We also see this pattern, however, in young, healthyndividuals without congestive heart failure (Figure 2). Al-hough the physiology of this square wave pattern in suchatients is not fully understood, it is important for clinicianso know that this abnormal pattern can be a “normalariant.”3
The “exceptions to the rule” outlined here should beaken as an adjunct to this excellent review of the clinical
igure 2 Square wave pattern Valsalva response in a healthy vtays elevated until the release of Valsalva.
tility of the Valsalva maneuver.
Satish R. Raj, MD, MSCIDavid Robertson, MD
Italo Biaggioni, MDAndré Diedrich, MD, PhD
Autonomic Dysfunction CenterDivision of Clinical Pharmacology
Departments of Medicine and PharmacologyVanderbilt University
Nashville, Tenn
doi:10.1016/j.amjmed.2006.03.022
eferences. Felker GM, Cuculich PS, Gheorghiade M. The Valsalva maneuver: a
bedside “biomarker” for heart failure. Am J Med. 2006;119:117-122.. Mosqueda-Garcia R. Evaluation of autonomic failure. In: Robertson D,
Biaggioni I, eds. Disorders of the Autonomic Nervous System. Luxem-bourg: Harwood Academic Publishers GmbH; 1995.
. Sandroni P, Benarroch EE, Low PA. Pharmacological dissection ofcomponents of the Valsalva maneuver in adrenergic failure. J Appl
r. The blood pressure paradoxically increases with Valsalva and
olunteePhysiol. 1991;71:1563-1567.
L
B
T
ctoprat(k
ttwtiHs
gcdqtdpmmdrsr
t
0
The American Journal of Medicine (2007) 120, e17
ETTER
rtw
R1
2
3
4
5
6
7
8
alancing the Risks and Benefits of Celecoxib
o the Editor:
The 2 major uncertainties regarding the use of celecoxibompared with conventional non-steroidal anti-inflamma-ory drugs (NSAIDs) are whether the initial reduction in riskf serious upper gastrointestinal events is maintained withrolonged treatment,1-3 and the magnitude of the increasedisk of serious cardiovascular thromboembolic events, suchs myocardial infarction.4,5 Regrettably, due to its design,he Successive Celecoxib Efficacy and Safety Study-1SUCCESS) study6 is unable to significantly advancenowledge in either regard.
The relative initial reduction in serious upper gastroin-estinal events identified in the Celecoxib Long-term Arthri-is Safety Study (CLASS)7 was confirmed in SUCCESS,6
hich reported a significant 2-fold greater risk with NSAIDherapy compared with celecoxib. In CLASS the reductionn risk was not significant after 12 months of treatment.1-3
owever, this could not be assessed in the SUCCESStudy,6 as it was of only 12 weeks’ duration.
With regard to serious cardiovascular events, a 5-foldreater risk of myocardial infarction was observed withelecoxib treatment in the SUCCESS study.6 However, thisifference was not statistically significant due to the inade-uate power of the study with only 11 myocardial infarc-ions in total. The lack of power was due to the studyesign—namely the short 12-week duration and, more im-ortantly, the enrollment of individuals at very low risk ofyocardial infarction. Only 0.25% of the subjects with aean age of 62 years had experienced a previous myocar-
ial infarction, at least 20 times lower than the predictedate.8 Thus, although prior cardiovascular disease was nottated as an exclusion criteria, it somehow occurred duringecruitment.
As a result, the findings of the SUCCESS study add little
o current medical knowledge that celecoxib increases the002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
isk of myocardial infarction but does not reduce the long-erm risk of serious gastrointestinal events when comparedith conventional NSAIDs.
Brent Caldwell, MBChBRichard Beasley, MD, DSc
Medical Research Institute of New ZealandWellington, New Zealand
Mark Weatherall, MBChB, BA(Hons)Wellington School of Medicine & Health Sciences
Wellington, New Zealand
Scott Metcalfe, MBChB, FAFPHMContracting public health physician
Wellington, New Zealand
doi:10.1016/j.amjmed.2006.03.018
eferences. Lu HL. Statistical reviewer briefing document for the advisory com-
mittee. Available at: http://www.fda.gov/ohrms/dockets/ac/01/briefing/3577b1_04_stats.doc. Accessed July 25, 2001.
. Hrachovec JB. Reporting of 6-month vs 12-month data in a clinical trialof celecoxib. JAMA. 2001;286:2398.
. Juni P, Rutjes AWS, Dieppe PA. Are selective COX 2 inhibitorssuperior to traditional non steroidal anti-inflammatory drugs? BMJ.2002;324:1287-1288.
. Caldwell B, Aldington S, Weatherall M, Shirtcliffe P, Beasley R. Riskof cardiovascular events and celecoxib: a systematic review and meta-analysis. J R Soc Med. 2006;99:132-140.
. Kearney PM, Baigent C, Godwin J, et al. Do selective cyclo-oxygen-ase-2 inhibitors and traditional non-steroidal anti-inflammatory drugsincrease the risk of atherothrombosis? Meta-analysis of randomisedtrials. BMJ. 2006;332:1302-1308.
. Singh G, Fort JG, Goldstein JL, et al. Celecoxib versus naproxen anddiclofenac in osteoarthritis patients: SUCCESS-I study. Am J Med.2006;119:255-266.
. Silverstein FE, Faich T, Goldstein JL, et al. Gastrointestinal toxicitywith celecoxib vs nonsteroidal anti-inflammatory drugs for osteoarthri-tis and rheumatoid arthritis. the CLASS study: a randomized controlledtrial. JAMA. 2000;284:1247-1255.
. Percentage who report experience of myocardial infarction (ever andrecently), by sex and age, 1998, England. Available at: http://www.
heartstats.org/temp/MSspTabsp1.2aspweb03.xls.
L
A
T
cscf(fiwdatpptMsaa
cFdvO
0
The American Journal of Medicine (2007) 120, e19
ETTER
Ma
wmnmsNw
R1
2
3
irway Obstruction in Never Smokers
o the Editor:
Celli and coworkers1 address a subject of great interest tolinicians and epidemiologists, airway obstruction in nevermokers, that is, in those not exposed to the predominantause of this impairment. Unfortunately, their use of a fixedorced expiratory volume in 1 second/forced vital capacityFEV1/FVC) ratio to define airway obstruction distorts theirndings and conclusions. The FEV1/FVC ratio decreasesith age; a value less than 0.70 is within the 95% confi-ence range of many people aged more than 60 years. Theuthors recognize this: “Limitations [of this study] includehe use of a fixed ratio to define obstruction . . . the modelresented here is likely to over diagnose obstruction inersons aged 70 years and older.” Unfortunately, neither theitle nor the abstract nor the conclusion includes this caveat.
uch of the striking increase in the rates of “airway ob-truction” at age 60 years (16.0% compared with 7.2% atges 50-59 years) and at age 70 years (28.0%) can bettributed to this alone.
This is well shown by 2 studies in normal nonsmokersomparing an even more stringent fixed cutoff value forEV1/FVC (0.65 for ages � 60 years) with a 95% confi-ence lower limit: Prevalences of obstruction were 22%ersus 7% for male subjects aged 55 years or more in
regon,2 12.2% versus 1.5% for male subjects of all ages in002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
ichigan, and 2.7% versus 0.3% for female subjects of allges in Michigan.3
Other well-known demographic variables associatedith a lower FEV1/FVC ratio include greater height andale sex. The use of a statistically defined lower limit of
ormal for both FEV1/FVC and FEV1 would provide moreeaningful rates of obstruction. This is available in the
oftware options of many spirometers. A reanalysis of theseational Health and Nutrition Examination Survey dataould be of great interest.
Albert Miller, MDNew York Medical College
ValhallaPulmonary Medicine
Caritas Health Care-Mary Immaculate HospitalJamaica, NY
doi:10.1016/j.amjmed.2006.04.026
eferences. Celli BR, Halbert RJ, Nordyke RJ, Schau B. Airway obstruction in
never smokers: results from the Third National Health and NutritionExamination Survey. Am J Med. 2005;118:1364-1372.
. Miller A, Thornton JC, Smith H Jr, Morris JF. Spirometric “abnormal-ity” in a normal male reference population: further analysis of the 1971Oregon Survey. Am J Indust Med. 1980;1:55-68.
. Miller A, Warshaw R, Thornton JC. Prevalence of spirometric abnor-mality in a representative sample of the population of Michigan. Am J
Indust Med. 1991;19:473-485.
L
UH
T
tddppcsa
cncapfoabwotahfrc
smowHtcernTaf
0
The American Journal of Medicine (2006) 120, e21
ETTERS
R
1
sing the Pneumonia Severity Index to Guideospitalization Decisions
o the Editor:
Marrie and Huang1 found that 19% of low-risk pa-ients with community-acquired pneumonia were admittedespite pneumonia severity index (PSI)-based recommen-ations for outpatient treatment and that 19% of these hos-italized low-risk patients had one or more medical com-lications. The authors and the accompanying editorialoncluded that physician judgment is important in the initialite of treatment decision and that rules better than the PSIre needed to guide this decision.1,2
We believe the study findings and conclusions deserveomment. First, the PSI was explicitly recommended to aid,ot replace, provider judgment in the site of treatment de-ision, because no singular decision aid could possiblyccount for all prognostic factors or predict all possibleatient complications.3,4 Second, the complication ratesound by Marrie et al1 might have been inflated by inclusionf low-risk patients with arterial oxygen desaturation. Anlternative strategy we recommended and used in our PSI-ased guideline implementation trial is that this criterionarrants hospitalization regardless of PSI risk class becausef its prognostic implications.3-5 Third, the PSI is not in-ended for use in certain groups of patients who are virtuallylways candidates for inpatient care (eg, prisoners andomeless persons), and hospitalization is not recommendedor low-risk patients with explicit medical and psychosocialeasons that should override outpatient treatment5-8; theseaveats were not followed by Marrie et al.1
Marrie and Huang1 and Siegel2 assert that hospitalizationhould be based on a broader set of medical outcomes thanortality alone. Unfortunately, consensus surrounding seri-
us complications that warrant hospitalization for patientsith pneumonia do not exist. Furthermore, Marrie anduang fail to distinguish complications that warrant hospi-
alization (eg, mechanical ventilation) from an iatrogenicomplication or less serious complications (eg, medicalrrors, rash, or insertion of a urinary catheter). Finally, 3andomized trials have demonstrated the safety and effective-ess of the PSI for guiding hospital admission decisions.5,9,10
hus, the PSI, when implemented as recommended, is a safend effective tool to aid decisions on the initial site of treatment
or patients with community-acquired pneumonia.002-9343/$ -see front matter © 2006 Elsevier Inc. All rights reserved.
Donald M. Yealy, MDThomas E. Auble, PhD
Department of Emergency MedicineUniversity of Pittsburgh
Michael J. Fine, MD, MScVA Center for Health Equity Research and Promotion
VA Pittsburgh Healthcare SystemPittsburgh, Pa
Division of General Internal MedicineDepartment of MedicineUniversity of Pittsburgh
Pittsburgh, Pa
doi:10.1016/j.amjmed.2006.02.017
eferences1. Marrie TJ, Huang JQ. Low-risk patients admitted with community-
acquired pneumonia. Am J Med. 2005;118:1357-1363.2. Siegel RE. Clinical opinion prevails over the pneumonia severity
index. Am J Med. 2005;118:1312-1313.3. Fine MJ, Auble TE, Yealy DM, et al. A prediction rule to identify
low-risk patients with community-acquired pneumonia. New EnglJ Med. 1997;336:243-250.
4. Yealy DM, Auble TE, Stone RA, et al. The emergency departmentcommunity-acquired pneumonia trial: methodology of a quality im-provement intervention. Ann Emerg Med. 2004;43:770-782.
5. Yealy DM, Auble TE, Stone RA, et al. Effectiveness of guidelineimplementation to improve the quality of care for pneumonia: resultsof the Emergency Department Community Acquired Pneumonia(EDCAP) trial. Ann Intern Med. 2005;143:881-894.
6. Mandell LA, Marrie TJ, Grossman RF, Chow AW, Hyland RH.Canadian guidelines for the initial management of community-acquired pneumonia: an evidence-based update by the CanadianInfectious Diseases Society and the Canadian Thoracic Society. TheCanadian Community-Acquired Pneumonia Working Group. ClinInfect Dis. 2000;31:383-421.
7. Mandell LA, Bartlett JG, Dowell SF, File TM Jr, Musher DM,Whitney C. Update of practice guidelines for the management ofcommunity-acquired pneumonia in immunocompetent adults. Clin In-fect Dis. 2003;37:1405-1433.
8. Bartlett JG, Dowell SF, Mandell LA, File TM Jr, Musher DM, FineMJ. Practice guidelines for the management of community-acquiredpneumonia in adults. Infectious Diseases Society of America. ClinInfect Dis. 2000;31:347-382.
9. Marrie TJ, Lau CY, Wheeler SL, Wong CJ, Vandervoort MK, FeaganBG. A controlled trial of a critical pathway for treatment of commu-nity-acquired pneumonia. CAPITAL Study Investigators. Community-Acquired Pneumonia Intervention Trial Assessing Levofloxacin.JAMA. 2000;283:749-755.
0. Carratala J, Fernandez-Sabe N, Ortega L, et al. Outpatient care com-pared with hospitalization for community-acquired pneumonia: a ran-
domized trial in low-risk patients. Ann Intern Med. 2005;142:165-172.
L
PB
Tstnntasocivs
shci4alCo
0
The American Journal of Medicine (2007) 120, e23
ETTER
moitug
R1
2
3
4
neumonia Severity Index (PSI) Lacking inreadth of Applicability
he purpose of a community-acquired pneumonia (CAP)everity of illness scoring system is to assist clinicians withhe hospital admission decision. Patient safety, conve-ience, and cost are the overriding concerns. The Pneumo-ia Severity Index (PSI) is incorporated into many CAPreatment guidelines, but it has not attained wide clinicalpplication. Although the PSI might be helpful in sometudies, as pointed out by Yearly et al, it is perceived to bef limited practical use, particularly in those in whom so-ioeconomic factors may affect outcome but possibly notmmediate mortality. For patients with hypoxia, unstableital signs, shock, or severe medical comorbid illness, nocoring system is needed; inpatient care is required.
It is those in whom underlying psychiatric illness, sub-tance abuse, homelessness, or age-related fragility dictateospital stay that the PSI fails to incorporate risks foromplications. In 2 studies in which clinicians were specif-cally instructed to use the PSI guideline, between 31 to3% of patients in the low-risk categories I-III were stilldmitted.1-3 The present study by Marrie et al confirms thisack of utility, particularly in an urban inner-city setting.4
ost savings are more likely to come from short stay, brief
bservation, and early discharge until pneumonia assess-002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.
ent scoring systems factor in risks for morbidity and poorutcome in addition to death. A pneumonia severity ofllness scoring system will not be useful if it is not used inhe field, and it will not be widely accepted into practicentil one is derived which is more comprehensive in itsuidance for clinicians.
Robert E. Siegel, MDDepartment of Pulmonary and Critical Care Medicine
James J. Peters Veterans Affairs Medical CenterDivision of Pulmonary Medicine, Critical Care and Sleep
Mt. Sinai School of MedicineBronx, NY
doi:10.1016/j.amjmed.2007.01.031
eferences. Niederman MS. What is the prognosis for using the pneumonia severity
index to make site-of-care decisions in community-acquired pneumo-nia? Chest. 2003;124:2051-2053.
. Atlas SJ, Benzer TI, Borowsky LH, et al. Safely increasing the propor-tion of patients with community-acquired pneumonia treated as outpa-tients: an interventional trial. Arch Intern Med. 1998;158:1350-1356.
. Marrie, TJ, Lau CY, Wheeler SL, et al. A controlled trial of a criticalpathway for treatment of community-acquired pneumonia. CAPITALstudy investigators. Community-acquired pneumonia intervention trialassessing levofloxacin. JAMA. 2000;283:749-55.
. Marrie TH, Huang JQ. Low-risk patients admitted with community-
acquired pneumonia. Am J Med. 2005;118:1357-1363.
LHEa
Adgpacfcitvl
scbaaTfrpdicchpc
Mi
APMPerspectives
0d
APM PerspectivesThe Association of Professors of Medicine (APM) is the national organization of departments of internalmedicine at the US medical schools and numerous affiliated teaching hospitals as represented by chairs andappointed leaders. As the official sponsor of The American Journal of Medicine, the association invitesauthors to publish commentaries on issues concerning academic internal medicine.
For the latest information about departments of internal medicine, please visit APM’s website atwww.im.org/APM.
ate to the Feast: Primary Care and USealth Policy
ugene C. Rich, MD,a Anna Maio, MDb
Department of Medicine and bDivision of General Internal Medicine, Creighton University School of Medicine, Omaha, Neb.
EAbttsccu
stacvatifcchm
otwvtr“tt
s health care spending in the United States increases,iscontent among primary care physicians continues torow, and student interest in primary care continues tolummet.1-3 Policymakers struggle with public concernsbout managed care and insurance gaps; physician con-erns about malpractice rates and decreases in Medicareee schedules; and medical school and teaching hospitaloncerns about new technology costs and Graduate Med-cal Education (GME) reimbursements.4-7 The juxtaposi-ion of increasing health care expenditures with highlyisible demands for extra resources has afforded relativelyittle attention to primary care.
Perhaps the difficult position of US primary care phy-icians stems from neither the irrelevance of the primaryare role nor inadequacies of primary care professionals,ut from structural barriers in financing. In other industri-lized nations, primary care physicians remain prominentnd achieve better outcomes with fewer resources.8-12
his commentary explores the evolution of policies af-ecting physician, medical school, and teaching hospitalevenues, as well as potential barriers to sustaining therimary care role in the United States. For purposes of thisiscussion, primary care adheres to the Institute of Med-cine (IOM) definition as a “provision of integrated, ac-essible health care services by clinicians who are ac-ountable for addressing a large majority of personalealth care needs, developing a sustained partnership withatients, and practicing in the context of family andommunity.”13
Requests for reprints should be addressed to Eugene C. Rich,D, Department of Medicine, Creighton University School of Med-
cine, 601 North 30th Street, Suite 5850, Omaha, NE 68131.
sE-mail address: [email protected].002-9343/$ -see front matter © 2007 The Association of Professors of Meoi:10.1016/j.amjmed.2007.03.005
VOLUTION OF PHYSICIAN FEE SCHEDULEScombination of private and public policy decisions,
eginning with the rise of private health insurance inhe 1930s, have influenced the fee-for-service sys-em.14-17 Hospital administrators, general practice phy-icians, and patients began preferring that insuranceover expensive procedures and hospital stays, and notover “little ticket” expenses, such as office-based eval-ation and management services.14-16
With rapid advances in technology, medical andurgical subspecialties manage increasing numbers ofechnically complex and invasive diagnostic and ther-peutic procedures.16,18 The development of these spe-ialized modalities provides important health care ad-ances throughout the industrialized world. However,bility to perform more procedures from improvedechniques and technology does not include a reductionn fees.19 These trends have the effect of creating dif-erent classes of medical practitioners—those who spe-ialize in well-reimbursed technical services (histori-ally provided in hospitals) and those who specialize inistorically under-reimbursed evaluation and manage-ent services.In the 1960s, surgical specialists earned an average
f 40% more than general practitioners and 30% morehan internal medicine physicians.20 When Medicareas created in 1966, the US government adopted pri-ate insurance payment policies guided by billing prac-ices that prevailed in the medical community, thuseinforcing historical imbalances.8,16,17 By extendingthird party” coverage to a large and growing popula-ion of older Americans, Medicare offered new oppor-unities for delivering procedural services. Demand for
urgical and medical subspecialists grew rapidly, anddicine. All rights reserved.
itwipgaidsilm
awasspbpdibiotrteTnrmcpit
TIAbpcbtamihpstc
tRsms
ip
csonsgflawd
F
554 The American Journal of Medicine, Vol 120, No 6, June 2007
ncomes for these specialists grew faster than infla-ion.21 By 1989, a general surgeons’ average incomeas 2.5 times greater and a surgical subspecialists’
ncome three to five times greater than that of a familyhysician.22 Although many specialties still requiredrueling years in residencynd a career of long hoursn the hospital, modest ad-itional training yieldsubstantial advantages inncomes and “controllableifestyle” relative to pri-ary care.23,24
The resource-based rel-tive value scale (RBRVS)as introduced to moder-
te some Medicare feechedule problems, but re-istance by many specialisthysician organizationslunted the impact of pro-osed changes.25,26 Theelayed and greatly alteredmplementation of resource-ased practice expense pol-cies is an important example.26,27 Although RBRVSffered modest improvements in payment for evalua-ion and management services, annual adjustments inecent years have not kept pace with rising administra-ive expenses.28 Many medical specialties did experi-nce an appreciable decrease in Medicare payments.his, when combined with managed care discountsegotiated in the 1990s, contributed to a noticeableeduction in specialty physician income.29 However,ost specialty physicians still earn more than primary
are physicians, and recent analyses suggest training inrimary care does not provide an adequate “return onnvestment” in comparison with advanced subspecialtyraining.30-32
HE EVOLUTION OF PRIVATE HEALTHNSURANCEn alternative to fee-for-service primary care practiceegan quietly developing in the 1930s.16 Prepaid groupractice—multi-specialty practices financed throughapitation payments on behalf of health plan mem-ers—were developed into a primary care infrastruc-ure that was integrated, accessible, and account-ble.8,15,33 By 1990, the spread of these healthaintenance organizations (HMOs) was seen as an
mportant solution to 25 years of startling growth in USealth care costs.28,33 Corporate executives and healthlan CEOs touted the advantages of primary care overpecialty care. Capitated and integrated delivery sys-ems were seen as the new paradigm where primary
PERSPECTIVES VIEW
● The history of physand private healthUnited States hasprimary care.
● Several practice, instural barriers exist toof primary care medi
● The future of primarpends on administraing the distributioexpenditures.
are professionals would manage populations of pa- r
ients and control access to specialized services.34,35
ecruiting and organizing primary care networks tran-iently increased income for primary care physicians inany metropolitan areas, whereas specialists’ incomes
tagnated or decreased.29 Student interest in primarycare greatly increased (Figure) andprimary care residency programsexpanded.3,36,37
Unfortunately, successful prac-tice organizations could not growquickly enough. Primary care prac-tices developed in the fee-for-ser-vice model found that they couldnot rapidly develop the required in-formation systems and administra-tive infrastructure to achieve carecoordination. Specialist physicians,with declining incomes, becamehostile to primary care’s efforts toprovide coordinated and compre-hensive care. Without an opportu-nity to realize the benefits of a ro-bust primary care infrastructure,patients became concerned aboutpotential loss of access to special-
zed services and public support for HMOs droppedrecipitously.4,38
By 2003, physician payment by capitation had de-reased substantially.39-41 Many integrated deliveryystems reorganized, and hospitals divested themselvesf primary care practices.39 Prepaid enrollments stag-ated, and HMOs decreased while discounted fee-for-ervice, open access forms of managed care plansrew.41,42 Thus, the financial rationale for an insurance-unded, physician-based primary care function becameess clear. Rather, health plans identified care coordinations one of their value-added services and began to contractith specialized vendors and national organizations toeliver both disease and case management.40,43
TS
fee schedulesrance in theively affected
onal, and cul-er the success
medicine de-olicies affect-
health care
igure Percent medical graduates choosing generalist ca-
POIN
icianinsu
negat
titutihind
cine.
y caretive pn of
eers (FM, GIM, GPEDs).
PPWdasmre“otPmpottnpiiwa
ccsqubaBctctteb
cocdaihfiovcotdp
oim“cavibpdc
taUmtcpsra
IPTitsstothmttta
isfpctiprtfs
555Rich and Maio Primary Care and US Health Policy
RACTICE BARRIERS TO DELIVERINGRIMARY CAREhile differences in fee schedules created substantial
ifferences in earning potential among US physicians,dditional problems presented barriers to building andustaining practices meeting the IOM definition of pri-ary care.13 Fee-for-service payments could arguably
eward continuity because the provider receives a feeach time the patient makes a visit, but the financing ofaccessibility” has proven more difficult. With highverhead expenses and limited fees, primary care prac-ices have to be careful regarding unused capacity.44
hysicians earn more income with scheduled appoint-ents rather than open slots for urgent access. In the
ast, physicians may have “fit in” extra patients. Familybligations and the cost of employee overtime renderhis a difficult option for modern primary care prac-ices. Coordination of care outside of office visits hasot been paid for at all, because Medicare and privateayers do not reimburse for telephone calls or otherntegrative services.45 Even comprehensiveness of cares discouraged under current fee-for-service incentives,ith insufficient payment to justify time for risk factor
ssessment and preventive care advice.46-48
With the shift from indemnity insurance to managedare, administrative burdens for all physicians in-reased substantially.49 Federal regulations regardingelf-referral for radiology and laboratory services, re-uirements for laboratory certification, new billing doc-mentation requirements, and Health Insurance Porta-ility and Accountability Act compliance imposed newdministrative burdens on many physician offices.50-54
ecause primary care practices had relatively higherosts initially, the regulations had a greater impact onheir infrastructure.30,44 In addition, because most spe-ialized physicians receive their greatest revenuehrough large fee-for-service payments, they can affordo invest much more effort in successfully managingach payment while still realizing a lower cost forilling as a percent of total revenue.31
Nonetheless, promising new techniques for primaryare practice have developed, including “direct access”ffice schedules, group visits, primary care teams, andhronic illness care, but the methods to finance andisseminate these innovations to patients, providers,nd settings remain to be resolved.55-62 Electronic med-cal records and other office informatics innovationsold promise to enhance quality greatly, but the currentnancial environment may limit use in primary careffices.63 There is tremendous potential for growth inarious forms of “asynchronous” physician-patientommunication in primary care, including improvedffice responsiveness to patient telephone calls, elec-ronic outreach, e-mail, patient education, electronicata sharing, and possibly image transmissions between
atients and physicians.64 cManaged care organizations increasingly use theirwn information systems to identify health risks andmplement care coordination and disease manage-ent services.43 Recent trends in plan design offer
consumer-directed” features, including providerhoice, physician profiles, shared decision-making,nd higher out-of-pocket costs.4 The distinct role andalue of primary care is not well articulated in thesensurance products. With primary care under-fundedy private insurance, the “concierge” or “boutique”ractice has emerged where patients pay physiciansirectly to provide personalized medical and care-oordination services.65,66
It is hardly surprising, therefore, that student in-erest in primary care careers has decreased oncegain (Figure), and the physician workforce in thenited States has become unusually weighted towardedical, surgical, and hospital-based subspecial-
ies.3,8,24,67 These specialists manage greater finan-ial resources, influencing not only physician com-ensation but also facilities and staff. Thus, in manyettings, subspecialists have greater access to theesources needed to provide accessible, coordinated,nd patient-oriented care.68,69
NSTITUTIONAL BARRIERS TO SUSTAININGRIMARY CAREeaching hospitals also have played an important role
n developing and sustaining primary care practice inhe United States. Hospitals have strong incentives toupport specialized practice because surgical and otherpecialized procedural services have long providedheir main sources of profit.16,70 With the introductionf the Medicare Prospective Payment System (PPS) inhe 1980s, these incentives became more intense andospitals developed robust infrastructure to support andarket specialized services.13,71,72 Although there was
ransient development of primary care networks duringhe managed care scare of the mid-1990s, many hospi-als now focus on providing specialized services andre investing in specialized “product lines.”41,73,74
Medical schools and teaching hospitals also havenfluenced primary care through emphasis on medicaltudent education.13,24 These institutions have benefitedrom substantial societal investment in specializedractice and hospital-based services. Demand for spe-ialized resident positions and an economic imperativeo leverage federal funding for residents to enhancenvestments in specialized programs resulted in dispro-ortionate growth in specialized training programs untilecent changes in Medicare GME payments.75 Thus, byhe early 1990s, prominent clinical programs, clinicalaculty, and residency positions in many medicalchools and teaching hospitals were related to subspe-
ialty practice.8,13,24
ttrirfeyssmmiss
CPMestehhonircaesp
snhhrtstddncia
EFAl
shmwivic
ctbTmmoiftMtpa
fdNSgutamfcnlvI“Mapncnpvdmpnntm
556 The American Journal of Medicine, Vol 120, No 6, June 2007
Even if financing clinical programs at these institu-ions did not drive them toward a specialized infrastruc-ure, the highly specialized orientation of biomedicalesearch would have presented a challenge for cultivat-ng the primary care perspective. Since the Flexnereport, medical schools and teaching hospitals haveocused on clinical and biomedical scholarship.70 Fed-ral funding for health care research has, for manyears, been substantially directed toward biomedicalcience. Funding opportunities for generalist-orientedcholarship, such as medical education, clinical epide-iology, or health care delivery research, have re-ained modest. Therefore, medical schools and teach-
ng hospitals necessarily invest substantial resources inpecialized biomedical science research rather than re-earch typically conducted by generalists.
ULTURAL BARRIERS TO PRIMARY CARERACTICEuch has been written about “American Culture” in the
volution of the US health care system.76,39 Punditsuggest that Americans emphasize individual, ratherhan community rights and responsibilities.39 This pref-rence has been blamed for the decline of the publicealth infrastructure, the marginalization of safety netealth care provider systems, the tolerance of high ratesf uninsured, the acceptance of for-profit entrepre-eurism in delivery of health services, the emphasis onllness treatment rather than disease prevention, andeliance on “choice” as a proxy for “quality.” Ameri-ans appear fascinated by technical, scientific solutionsnd prefer quick action and straightforward solutions asvidenced by the prominence of scientific advances andpecialized medicine so visible in the news media andopular television shows.16,39
Although the many exciting advances in biomedicalcience do not highlight primary care, they certainlyeed not detract from it. Germany and Switzerland alsoave achieved major advances in biotechnology andave invested in substantial high technology medicalesources.12 Nonetheless, primary care physicians inhese countries are more prevalent and better compen-ated when compared with specialized physicians.12 Inhe United States, however, the pervasive incentivesisadvantaging primary care have led to an unusualegree of prominence of specialist physicians in bothumbers and affluence relative to other industrializedountries. This status likely affects the relative visibil-ty and attractiveness of primary care in the communitynd media.
MERGING TRENDS RELEVANT TO THEUTURE OF PRIMARY CAREs previously discussed, the fee-for-service payments
ong prevalent in the United States do not provide p
upport for key primary care functions such as compre-ensiveness, coordination, or accountability. Further-ore, most physicians do not have the financial where-ithal to develop the information systems and
nterdisciplinary teams required for sophisticated inter-entions to manage chronic illness.77 There are a grow-ng number of problems caused by the resulting lack ofare coordination.78
Recent studies suggest rising health care costs may beomplicated by the decline in primary care infrastruc-ure.11,79-81 The United States faces rapidly growing num-ers of older individuals with multiple chronic illnesses.82
he benefits of increased access to specialized physiciansay be subverted by failures in care coordination amongultiple independent specialist offices. Indeed, a survey
f consumer experiences with patient safety and qualitynformation recently found that two thirds of respondentselt “coordination among the different health professionalshat they see is a problem.”83 Similarly, in a survey of
edicare beneficiaries, investigators found a decline inhe continuity and integration of care by primary carehysicians, as well as in the quality of primary care inter-ctions with patients.84
Securing substantial enhancements to traditionalee-for-service payments for primary care may proveifficult at a time of record health care expenditures.onetheless, the Centers for Medicare & Medicaidervices (CMS) recently published the “Medicare Pro-ram Five-Year Review of Work Relative Value Unitsnder the Physician Fee Schedule,” proposing substan-ive changes to several outpatient evaluation and man-gement codes and offering meaningful relief to pri-ary care physicians.85 However, simply enhancing
ee-for-service payments for traditional face-to-face en-ounters will likely not be sufficient to establish theeeded primary care infrastructure. The American Col-ege of Physicians recently issued a report on the “ad-anced medical home,” and the Society of Generalnternal Medicine has extended this work with its reportRedesigning the Practice Model for General Internaledicine.”76,86 Both reports outline in greater detail the
dministrative and financial rationale for fundamentalayment reforms to support comprehensive, coordi-ated primary medical care in the United States. Poli-ymakers and employers are undertaking a re-exami-ation of traditional fee-for-service and consideringroviding support for inter-visit communication, non-isit-related management, information systems, chronicisease management programs, and quality improve-ent initiatives.61,87,88 Business leaders have initiated
rograms such as Bridges to Excellence to provideon-fee-for-service payments for improved chronic ill-ess care, and CMS is introducing incentive programso support chronic illness care and quality improve-ent.81,87,89 The challenge will be to sustain and ex-
and such reforms in the face of near-term resource
cc
CTSticnpbtcrpfhnb
R
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
557Rich and Maio Primary Care and US Health Policy
onstraints to realize long-term improvements in effi-iency and effectiveness.
ONCLUSIONhe problems confronting primary care in the Unitedtates are longstanding, complex, and not amenable
o easy solution, either by policymakers or academ-cs. Thoughtful, articulate, and evidence-based advo-acy will be needed to address them. The share of theation’s wealth devoted to health care may be ap-ropriate, but it has been distributed inappropriatelyy past and current administrative decisions. It willake courageous leadership to rectify this. Primaryare can be saved and expanded by redistributingeimbursement away from technical specialties to theroviders at the front lines of continuing health careor the nation’s population. In doing so, the USealth care system will mirror other industrializedations’ health care priorities, as well as provideetter outcomes and greater efficiencies.
eferences1. Smith C, Cowan C, Heffler S, Catlin A. National health spending
in 2004: recent slowdown led by prescription drug spending.Health Aff (Millwood). 2006;25:186-196.
2. Landon BE, Aseltine R, Shaul JA, et al. Evolving dissatisfactionamong primary care physicians. Am J Manag Care. 2002;8:890-901.
3. Division of Medical Education. Graduation Questionnaire. AllSchools Reports. Washington, DC: Association of AmericanMedical Colleges; 2006.
4. Rich EC, Oasan A, Maio A. Whatever happened to managedcare? Am J Med. 2003;114:426-430.
5. Studdert DM, Mello MM, Brennan TA. Medical malpractice.N Engl J Med. 2004;350:283-292.
6. MGA Government Affairs Department. 2006 final Medicare rulereduces physician payment by 4.4 percent. MGMA Connex.2006;6:16-18.
7. Committee on the Roles of Academic Health Centers in the 21stCentury. Academic Health Centers: Leading Change in the 21stcentury. Washington, DC: Institute of Medicine; 2004.
8. Starfield B. Primary Care: Concept, Evaluation, and Policy.New York, NY: Oxford University Press; 1992.
9. Macinko J, Starfield B, Shi L. The contribution of primary caresystems to health outcomes within Organization for EconomicCooperation and Development (OECD) countries, 1970-1998.Health Serv Res. 2003;38:831-865.
0. Starfield B, Shi L. Policy relevant determinants of health: aninternational perspective. Health Policy. 2002;60:201-218.
1. Starfield B, Shi L, Grover A, Macinko J. The effects of specialistsupply on populations’ health: Assessing the evidence. HealthAff (Millwood). 2005;(SupplWebExclusives):W5,97-107.
2. Organisation for Economic Co-operation and Development. Healthat a Glance—OECD Indicators 2005. Paris, France: Organisationfor Economic Co-operation and Development; 2005.
3. Committee on the Future of Primary Care, Donaldson MS. Pri-mary Care: America’s Health in a New Era. Washington, DC:Institute of Medicine; 1996.
4. American Society of Internal Medicine. Reimbursement for phy-sician’s cognitive and procedural services: a white paper. Inter-
nist. 1981;22(Suppl):S1-S4.5. Fein R. Medical Care, Medical Costs. The Search for a HealthInsurance Policy. Cambridge, Ma: Harvard University Press;1986.
6. Starr P. The Social Transformation of American Medicine. NewYork, NY: Basic Books; 1982.
7. Delbanco TL, Meyers KC, Segal EA. Paying the physician’s fee:Blue Shield and the reasonable charge. N Engl J Med. 1979;301:1314-1320.
8. Pauly MV. Paying Physicians: Options for Controlling Cost,Volume, and Intensity of Services. Ann Arbor, MI: Health Ad-ministration Press; 1992.
9. Penchansky R, Rosenthal G. Productivity, price, and incomebehavior in the physicians’ services market. Med Care. 1965;3:240-244.
0. Center for Health Services Research and Development. Profile ofMedical Practice. Chicago, IL: American Medical Association;1978.
1. Physician Payment Review Commission. Annual Report to Con-gress. Washington, DC: Physician Payment Review Commis-sion; 1989.
2. Center for Health Policy Research. Socioeconomic Characteris-tics of Medical Practice. Chicago, IL: American Medical Asso-ciation; 1989.
3. Schwartz MD, Linzer M, Babbott D, et al. Medical studentinterest in internal medicine. Initial report of the Society ofGeneral Internal Medicine Interest Group Survey on FactorsInfluencing Career Choice in Internal Medicine. Ann Intern Med.1991;114:6-15.
4. Council on Graduate Medical Education (Third Report). Improv-ing Access to Care through Physician Workforce Reform: Di-rections for the 21st Century. Washington, DC: Council onGraduate Medical Education; 1992.
5. Physician Payment Review Commission. Annual Report to Con-gress. Washington, DC: Physician Payment Review Commis-sion; 1996.
6. Hsiao WC, Dunn DL, Verrilli DK. Assessing the implemen-tation of physician-payment reform. N Engl J Med. 1993;328:928-933.
7. Proudfoot ML. A critique of the practice-expense values ofthe resource-based relative value scale. J Fam Pract. 1993;37:57-67.
8. Medical Services Committee. Revitalizing Internal Medicine:Recommendations for Resolving Payment and Practice Hassles.Philadelphia, PA: American College of Physicians; 2003.
9. Reed MC, Ginsburg PB. Behind the times: physician income, 1995-99. Data Bull (Cent Stud Health Syst Change). 2003;(24):1-2.
0. Center for Health Policy Research. Socioeconomic Characteris-tics of Medical Practice. Chicago, IL: American Medical Asso-ciation; 1993.
1. Medical Group Management Association. MGMA PhysicianCompensation and Production Survey. Englewood, CO: Wiley &Sons; 2003.
2. Weeks WB, Wallace AE. Long-term financial implications ofspecialty training for physicians. Am J Med. 2002;113:393-399.
3. Halvorson GC. Strong Medicine. New York, NY: RandomHouse; 1993.
4. Eisenberg JM. The internist as gatekeeper. Preparing the generalinternist for a new role. Ann Intern Med. 1985;102:537-543.
5. Wartman SA, Wilson M, Kahn N. The generalist health care work-force: Issues and goals. J Gen Intern Med. 1994;9(4 Suppl 1):S7-S13.
6. Kahn NB, Schmittling GT, Garner JG, Graham R. Entry of USmedical school graduates into family practice residencies: 1996-1997 and 3-year summary. Fam Med. 1997;29:544-552.
7. Kahn NB, Garner JG, Schmittling GT, et al. Results of the 1997National Resident Matching Program: family practice. Fam Med.
1997;29:553-558.
3
3
4
4
4
4
44
4
4
4
45
5
5
5
5
5
5
5
5
5
6
6
6
6
6
6
6
6
6
6
7
7
7
7
7
7
7
7
7
7
8
8
8
8
8
558 The American Journal of Medicine, Vol 120, No 6, June 2007
8. Kilborn P. Reality of the H.M.O. system doesn’t live up to thedream. The New York Times. Page A1, October 5, 1998.
9. Morrison JI. Health Care in the New Millennium: Vision, Values,and Leadership. San Francisco, CA: Jossey-Bass Publishers;2000.
0. Coddington DC. Beyond Managed Care: How Consumers andTechnology are Changing the Future of Health Care. San Fran-cisco, CA: Jossey-Bass Publishers; 2000.
1. Kaiser Family Foundation and the Health Research and Educa-tional Trust. Employer Health Benefits 2003 Annual Survey.Washington, DC: Kaiser Family Foundation and the Health Re-search and Educational Trust; 2003.
2. Strunk BC, Reschovsky JD. Kinder and Gentler: Physiciansand Managed Care, 1997-2001 (Tracking Report No. 5).Washington, DC: The Center of Studying Health SystemChange; 2002.
3. Bodenheimer T. Disease management—promises and pitfalls.N Engl J Med. 1999;340:1202-1205.
4. Sox HC. Saving office practice. Ann Intern Med. 2003;139:227-228.5. Pham HH, Devers KJ, May JH, Berenson R. Financial pressures
spur physician entrepreneurialism. Health Aff (Millwood). 2004;23:70-81.
6. Rich EC, Burke W, Heaton CJ, et al. Reconsidering the familyhistory in primary care. J Gen Intern Med. 2004;19:273-280.
7. Yarnall KS, Pollak KI, Ostbye T. Primary care: is there enoughtime for prevention? Am J Public Health. 2003;93:635-641.
8. Ostbye T, Yarnall KS, Krause KM, et al. Is there time formanagement of patients with chronic diseases in primary care?Ann Fam Med. 2005;3:209-214.
9. Meharg JG. The hassle factor. Ann Intern Med. 1992;117:797.0. Moskowitz H, Sunshine J, Grossman D, et al. The effect of imaging
guidelines on the number and quality of outpatient radiographicexaminations. AJR Am J Roentgenol. 2000;175:9-15.
1. Johnson BA, Niederman GA, Bowman LE, McCullough AC.New Stark regulations: key issues for health care decision-mak-ers. Med Group Manage J. 1998;45:10,12-15,50.
2. Born PH, Thran SL. The influence of CLIA ‘88 on physicianoffice laboratories. J Fam Pract. 1998;46:319-327.
3. Kikano GE, Goodwin MA, Stange KC. Evaluation and manage-ment services. A comparison of medical record documentationwith actual billing in community family practice. Arch Fam Med.2000;9:68-71.
4. Darves B. From minor annoyances to treatment delays, physi-cians feeling fallout of HIPAA privacy law. ACP Observer.September 2003.
5. Murray M, Berwick DM. Advanced access: reducing waiting anddelays in primary care. JAMA. 2003;289:1035-1040.
6. Wellington M. Stanford Health Partners: rationale and earlyexperiences in establishing physician group visits and chronicdisease self-management workshops. J Ambul Care Manage.2001;24:10-16.
7. Safran DG. Defining the future of primary care: what can welearn from patients? Ann Intern Med. 2003;138:248-255.
8. Rothman AA, Wagner EH. Chronic illness management: what isthe role of primary care? Ann Intern Med. 2003;138:256-261.
9. Committee on Quality of Health Care in America. Crossing theQuality Chasm: A New Health System for the 21st Century.Washington, DC: Institute of Medicine; 2001.
0. Casalino L, Gillies RR, Shortell SM, et al. External incentives,information technology, and organized processes to improvehealth care quality for patients with chronic diseases. JAMA.2003;289:434-441.
1. Rosenthal MB, Fernandopulle R, Song HR, Landon B. Payingfor quality: providers’ incentives for quality improvement.Health Aff (Millwood). 2004;23:127-141.
2. Larson EB, Fihn SD, Kirk LM, et al. The future of general internal
medicine. Report and recommendations from the Society of GeneralInternal Medicine (SGIM) Task Force on the Domain of GeneralInternal Medicine. J Gen Intern Med. 2004;19:69-77.
3. Bodenheimer T, Grumbach K. Electronic technology: a spark torevitalize primary care? JAMA. 2003;290:259-264.
4. Epstein RM. Virtual physicians, health systems, and the healingrelationship. J Gen Intern Med. 2003;18:404-406.
5. Sandy LG, Schroeder SA. Primary care in a new era: disillusionand dissolution? Ann Intern Med. 2003;138:262-267.
6. Government Accountability Office. Physician Services: Con-cierge Care Characteristics and Considerations for Medicare.Washington, DC: Government Accountability Office; 2005.
7. Schroeder SA. The troubled profession: is medicine‘s glass halffull or half empty? Ann Intern Med. 1992;116:583-592.
8. McAlister FA, Lawson FM, Teo KK, Armstrong PW. A system-atic review of randomized trials of disease management pro-grams in heart failure. Am J Med. 2001;110:378-384.
9. Smetana GW, Landon, BE, Bindman AB, et al. A comparison ofoutcomes resulting from generalist vs specialist care for singlediscrete medical conditions: a systematic review and method-ologic critique. Arch Intern Med. 2007;167:10-20.
0. Ludmerer KM. Time to Heal: American Medical Education fromthe Turn of the Century to the Era of Managed Care. Oxford,NY: Oxford University Press; 1999.
1. Long MJ, Chesney JD, Fleming ST. Profitable and unprofitableDRGs: the implications for access. Health Serv Manage Res.1993;6:61-69.
2. Munoz E, Mulloy K, Goldstein J, et al. Physicians‘ patient loadper DRG, the consumption of hospital resources, and the incen-tives of the DRG prospective payment system. Acad Med. 1990;65:533-538.
3. Rodriguez JL, Jacobs DM, Zera RT, et al. Academic practicegroups: strategy for survival. Surgery. 2000;128:505-512.
4. Berenson RA, Bodenheimer T, Pham HH. Specialty-servicelines: salvos in the new medical arms race. Health Aff (Mill-wood). 2006;25:w337-w343.
5. Eisenberg JM. Local logic and national nonsense: financing ofgraduate medical education encourages expansion of residencies.J Gen Intern Med. 1993;8):639-640.
6. Babbott SF, Bigby JA, Day SC, et al. Redesigning the Practice Modelfor General Internal Medicine: A Proposal for Coordinated Care.Washington, DC: Society for General Internal Medicine; 2006.
7. Wolff JL, Boult C. Moving beyond round pegs and square holes:restructuring Medicare to improve chronic care. Ann Intern Med.2005;143:439-445.
8. Schoen C, Osborn R, Huynh PT, et al. Taking the pulse of healthcare systems: experiences of patients with health problems in sixcountries. Health Aff (Millwood). 2005;(SuppWebExclusives):W5-509-525.
9. Goodman DC, Stukel TA, Chang CH, Wennberg JE. End-of-lifecare at academic medical centers: implications for future work-force requirements. Health Aff (Millwood). 2006;25:521-531.
0. Fisher ES, Wennberg DE, Stukel TA, et al. The implications ofregional variations in Medicare spending. Part 2: health outcomesand satisfaction with care. Ann Intern Med. 2003;138:288-298.
1. Starfield B, Shi L, Macinko J. Contribution of primary care tohealth systems and health. Milbank Q. 2005;83:457-502.
2. Anderson GF. Medicare and chronic conditions. N Engl J Med.2005;353:305-309.
3. The Kaiser Family Foundation-Agency for Healthcare Research andQuality-Harvard School of Public Health. National Survey on Consum-ers’ Experiences with Patient Safety and Quality Information. Wash-ington, DC: The Kaiser Family Foundation-Agency for HealthcareResearch and Quality-Harvard School of Public Health, 2004.
4. Montgomery JE, Irish JT, Wilson IB, et al. Primary care expe-riences of Medicare beneficiaries, 1998 to 2000. J Gen Intern
Med. 2004;19:991-998.
8
8
88
8
559Rich and Maio Primary Care and US Health Policy
5. Medicare Program: Five-Year Review of Work Relative Value Unitsunder the Physician Fee Schedule and Proposed Changes to the PracticeExpense Methodology. Fed Reg. 2006;71:69623-70251.
6. American College of Physicians. The Impending Collapse ofPrimary Care Medicine and Its Implications for the State of theNation’s Health Care. Philadelphia, PA: American College of
Physicians; 2006.7. McClellan MB. Medicare Physician Payments. 2005:1.8. Millenson ML. Demanding Medical Excellence: Doctors and
Accountability in the Information Age. Chicago, IL: Universityof Chicago Press; 1997.
9. Bridges to Excellence. Bridges to Excellence Overview. Avail-able at: http://www.bridgestoexcellence.org/about_us/home.htm.
Accessed May 26, 2006.
MH
AND
OspkoGM
SttpGwyrcoiesrm
oltbiptmpwaa
DMT
0d
The American Journal of Medicine (2007) 120, 560
EDICAL HUMANITIES PERSPECTIVES
elle Mathiasen, Cand. Mag, PhD, Section Editorncient Japanese Medicine in The Tale of Genjiorman A. Desbiens, MD
e of M
en
sFtwM
H
h
r
Mctt
itatittpefl
pthTgawu
R1
2
epartment of Internal Medicine, University of Tennessee Colleg
ne of the major advantages of longer life is the opportunity totudy the great books that, given the pressures of medicalractice, one has heretofore wanted the time, inclination, ornowledge to read. The recommendations of experts can leadne to hidden gems. Professor Harold Bloom’s recent work,enius: A Mosaic of One Hundred Exemplary Creativeinds1 introduced me to an ancient Japanese masterpiece.The Tale of Genji2 was written by Lady Murasaki
hikibu, a denizen of the Japanese emperor’s court duringhe Heian Dynasty (784–897). Centered on the life andimes of Genji, the son of an emperor and his concubine, itortrays the lives and conflicts of its central characters,enji and his love, Murasaki. When she dies at a young age,e feel his anguish. (“The days are numbered for him whoet must mourn. And are they numbered the tears that yetemain?”; p. 732.) Given the theme of life reversal as aause of illness, we are not surprised to read about Genji’swn death in the next chapter. The novel explores thenterplay of love, honor, duty, desire, and sorrow experi-nced by the Japanese royalty, and how they interact overeveral lifetimes. These motifs play out among the strictituals that govern Shintoism, Buddhism, court procedure,usic, and the rhythms of nature.Pertinent to those of us involved in medicine, a leitmotif
f Genji is the evanescence and capriciousness of humanife. Throughout the novel, characters fall ill. The illnesseshemselves lack sufficient detail to allow modern diagnosis,ut the precipitants are often depression, sometimes follow-ng dishonorable actions or life reversals. Holy men andriests, not physicians, are called to perform religious ritualso help the sick and dying. The maladies are attributedostly to malign spirits or moral lapses in the patients’
revious lives. At the end of the novel, an exorcised evilraith confesses: “. . . I wandered here and wandered there,
nd found a house full of beautiful girls. One of them died,nd this one wanted to die too. She said so, every day and
Requests for reprints should be addressed to Norman A. Desbiens, MD,epartment of Internal Medicine, University of Tennessee College ofedicine–Chattanooga Unit, 975 East Third Street, Box 94, Chattanooga,
N 37403.
E-mail address: [email protected]002-9343/$ -see front matter © 2007 Elsevier Inc. All rights reserved.oi:10.1016/j.amjmed.2006.08.024
edicine–Chattanooga Unit, Chattanooga.
very night. I saw my chance and took hold of her one darkight when she was alone” (p. 1050).
A second theme is early death. The novel reflects the sen-ibility of its time—that death was imminent and frequent.orty years is considered to be an advanced age; characters at
his point often begin considering whether to renounce theorld and end their days as monks or nuns. Referring tourasaki, the author tells us that,
er deepest wish, of which she sometimes spoke, had long been to give
erself over entirely to prayers and meditations. But even now Genji
efused to hear of it. Yet he had for some time had similar wishes (p. 712).
ost physicians will find that this acceptance of mortalityontrasts with our contemporary “rage against the dying ofhe light”—perhaps a necessary posture in a time whenherapy was ineffectual.
Similar to practitioners of evidence-based medicine whonterpret patients’ problems in light of the medical literature,he nobles in Genji explain their lives by referencing thebundantly collected poetry of their era. Life events remindhem of relevant verses as their refined sensibilities seek toncorporate well known waka poems. As Lady Murasakiells us, “Old poems, they could see, had much to say abouthe unchanging human heart” (p. 822). As formalized ashysicians’ prescriptions, their poetic messages were alsombellished with various scripts, perfumes, papers, ink, andowers that conveyed additional meaning.
For its insight into ancient Japanese medical concepts, es-ecially the conceptualization of disease causality and attitudeso death and dying, and for its demonstration of universaluman concerns, The Tale of Genji remains a literary classic.his thousand-year-old novel is a wonderful illustration of howreat art can transcend time and the accidentals of life to reachcore of truth: the importance of how to live a dignified andorthy life in the setting of conflicting passions and amid thencertainty and impermanence of life.
eferences. Bloom H. Genius: A Mosaic of One Hundred Exemplary Creative
Minds. New York: Warner Books, Inc.; 2002.. Shikibu M. The Tale of Genji. Seidensticker EG, translator. New York:
Alfred A. Knopf; 1976.