jama issue august 22 (2012)

THIS WEEK IN JAMAAUGUST 22/29, 2012

Medical News& PerspectivesAn upturn in the number of overdosesinvolving methadone has spurredhealth officials to call for more judi-cious prescribing of the drug.SEE PAGE 749

ViewpointsEnding the tobacco epidemicSEE PAGE 767

Preventing patient harms: a systemsapproachSEE PAGE 769

Transforming care and improving out-comes after acute MISEE PAGE 771

Statistical analysis plans in observa-tional researchSEE PAGE 773

A Piece of My Mind“When parents of younger childrendevelop cancer, inevitable questionsfocus on the children. What should wetell them?” From “Poolside.”SEE PAGE 775

Author in the RoomTeleconferenceJoin Philip Greenland, MD, Wednes-day, September 19, from 2 to 3 PMeastern time to discuss novel risk mark-ers to improve assessing patients atintermediate risk of cardiovascular dis-ease. To register, go to http://www.ihi.org/AuthorintheRoom.

Editor’s Audio SummaryDr Bauchner summarizes and com-ments on this week’s issue.

www.jama.com

JAMA Patient PageFor your patients: Information aboutectopic pregnancy.SEE PAGE 829

Biolimus-Eluting Stents and Cardiovascular EventsThe safety and efficacy of drug-eluting stents remain controversial in patients undergo-ing primary percutaneous coronary intervention for ST-segment elevation myocardialinfarction (STEMI). In a multicenter trial, Raber and colleagues randomly assigned 1161patients with STEMI to receive either a stent that elutes biolimus from a biodegradablepolymer or a bare-metal stent. The authors report that use of the biolimus-eluting stentresulted in a lower rate of major adverse cardiac events (a composite of cardiac death,target vessel–related reinfarction, and ischemia-driven target-lesion revascularization) at1 year. In an editorial, Cassese and Kastrati discuss the safety and efficacy of new-generation drug-eluting stents in patients with STEMI.SEE PAGE 777 AND EDITORIAL ON PAGE 814

Cardiovascular Risk AssessmentTwo articles in this issue address whether the addition of novel cardiovascular risk mark-ers to the Framingham Risk Score (FRS) improves cardiovascular risk assessment, particu-larly in patients at intermediate risk. In an analysis of data from 1330 adult participantsin the Multi-Ethnic Study of Atherosclerosis, who were asymptomatic at baseline andwere determined to be at intermediate risk of a cardiovascular event by the FRS, Yeboahand colleagues found that coronary artery calcium score, ankle-brachial index, high-sensitivity C-reactive protein level, and family history were independent predictors ofincident coronary heart disease or cardiovascular disease and that the coronary arterycalcium score provided superior discrimination and risk reclassification compared withthe other examined markers. In a meta-analysis of data from 45 828 participants in 14population-based cohort studies who had baseline assessments of common carotidintima-media thickness (CIMT), Den Ruijter and colleagues found that the addition ofcommon CIMT measurements to the FRS was associated with a small, but likely clinicallyunimportant, improvement in 10-year risk prediction of first-time myocardial infarctionor stroke. In an editorial, Gaziano and Wilson discuss implications of these findings forpatient care.SEE PAGES 788 AND 796, EDITORIAL ON PAGE 816, AND VIDEO INTERVIEW AT

www.jama.com

CLINICIAN’S CORNERLipid-Modifying Therapies and Risk of PancreatitisLipid-modifying therapies—both statins and fibrates—have been associated with a pos-sible increased risk of pancreatitis. To further explore this relationship, Preiss and col-leagues analyzed pooled data from 16 placebo- and standard care–controlled trials ofstatin therapy (113 800 participants), 5 dose-comparison statin trials (39 614 partici-pants), and 7 randomized trials of fibrate therapy (40 162 participants). The authorsreport that use of statin therapy was associated with a lower risk of pancreatitis inpatients with normal or mildly elevated triglyceride levels. Fibrate therapy was not asso-ciated with pancreatitis risk.

SEE PAGE 804

JAMA Clinical ChallengeA 21-year-old man presents withan enlarging nodule on his chest,upper abdominal pain, diarrhea,and recent weight loss. Physicalexamination reveals supraclavicularand cervical lymphadenopathy,and endoscopy reveals gastrointesti-nal ulcers. What would you do next?

SEE PAGE 812

©2012 American Medical Association. All rights reserved. JAMA, August 22/29, 2012—Vol 308, No. 8 739

Downloaded From: http://jama.jamanetwork.com/ by a World Health Organization User on 09/02/2012

VIEWPOINT

ONLINE FIRST

Ending the Tobacco EpidemicHoward K. Koh, MD, MPHKathleen G. Sebelius, MPA

FOR TOO LONG, THE GOAL OF A SOCIETY FREE FROM

tobacco-related disease and disability has been elu-sive. While the prevalence of adult smoking has de-clined from 43% (1964) to about 19% (2010),1 too

many in public health have taken future progress for grantedand turned their attention elsewhere. Heightened, not dimin-ished, attention to the leading preventable cause of death inthe United States is needed. Tobacco dependence still causesmore than 440 000 deaths annually.2 Furthermore, the markedslowing of declines in adult smoking prevalence over the pastdecade creates a renewed sense of urgency. It is time to reaf-firm the commitment to ending the tobacco epidemic.

It is clear what works. Decades of research have docu-mented the effectiveness of a broad strategy involving co-ordinated, multifaceted interventions. Specifically, mass me-dia campaigns, higher prices, smoke-free policies, andincreased access to cessation treatment can reduce tobaccouse individually and collectively.

A reinvigoration of national efforts has occurred. First,President Obama signed 4 new laws—the Children’s HealthInsurance Program Reauthorization Act (2009), the Fam-ily Smoking Prevention and Tobacco Control Act (2009),the Prevent All Cigarette Trafficking Act (2010), and thePatient Protection and Affordable Care Act (2010). Collec-tively, these laws grant federal agencies more authority andfunding to regulate tobacco products, decrease youth ac-cess to tobacco products, and increase access to tobacco de-pendence treatments.

In 2010, the US Department of Health and Human Ser-vices (HHS) unveiled its first-ever national strategic plan fortobacco control, Ending the Tobacco Epidemic: A Tobacco Con-trol Strategic Action Plan.2 The Strategic Plan, featuring 4 pil-lars that guide 21 action steps, has already led to an un-precedented array of actions across HHS and the federalgovernment (FIGURE).

Pillar 1: Lead by ExampleCMS Tobacco Cessation Coverage. In August 2010, theCenters for Medicare & Medicaid Services (CMS) ex-panded coverage of tobacco cessation counseling to ap-proximately 5 million Medicare tobacco users, not just those

with tobacco-related diseases. In June 2011, Medicaid is-sued new guidelines that mandate full cessation coveragefor pregnant women (effective October 2010) and 50% re-imbursement to states providing telephone quitline sup-port to beneficiaries who smoke.

Expanded Cessation Options for Federal Employees anda Tobacco-Free HHS Campus. In January 2011, the Officeof Personnel Management (OPM) ensured that all federalemployees (as well as retirees and dependents covered byan OPM health insurance plan) could access robust evidence-based cessation treatments.3 This then allowed HHS, in July2011, to expand its existing smoke-free policy to prohibitany tobacco product use at any of its facilities (includingbuildings, outdoor spaces, and government vehicles) whileproviding appropriate cessation support to workers.

Pillar 2: Improve the Public’s HealthFood and Drug Administration Regulations to ReduceYouth Access. Because an estimated 4000 youths try ciga-rettes daily, with 1000 becoming lifetime smokers,5 the USFood and Drug Administration (FDA) has pursued strate-gies to protect children. In particular, FDA has awarded $33million in contracts to 37 states (plus the District of Co-lumbia) to enforce tobacco marketing, sale, and distribu-tion laws and regulations at retail locations, with attentionto age and ID verification. As of June 2012, FDA has con-ducted more than 70 000 retail inspections and issued nearly3000 warning letters to retailers, the majority involving vio-lations relating to tobacco sales to minors. Other relevantactions have included a ban of sponsorship of athletic andentertainment events using tobacco product brand namesand logos and a ban on certain candy, fruit, or other fla-vored cigarettes.

Investments in State and Local Tobacco Control Initia-tives. In 2009, HHS directed $200 million from AmericanRecovery and Reinvestment Act funds to launch Commu-nities Putting Prevention to Work, supporting 22 cities andcounties to implement evidence-based strategies to reducetobacco use. In 2011, the Centers for Disease Control andPrevention (CDC) awarded more than $100 million to states,communities, and tribes for tobacco control and other well-

Author Affiliations: Assistant Secretary for Health (Dr Koh) and Secretary of Healthand Human Services (Ms Sebelius), Washington, DC.Corresponding Author: Howard K. Koh, MD, MPH, 200 Independence Ave, Suite716-G, Washington, DC 20201 ([email protected]).



ness initiatives through community transformation grants,funded through the Prevention and Public Health Fund es-tablished by the Affordable Care Act.

Pillar 3: Engage the PublicThe CDC National Media Campaign. In March 2012, theCDC launched a landmark national advertising tobacco con-trol campaign, “Tips from Former Smokers.” The cam-paign depicted the harsh realities of tobacco-related illnessand was disseminated through multimedia channels includ-ing television, radio, newspapers, billboards, and social me-dia. The campaign has generated approximately 200 000 ad-ditional calls to the 1-800-QUIT-NOW tobacco cessationquitline and an additional 417 000 unique visitors to thesmokefree.gov website.

New Warnings to Convey the Health Harms of To-bacco. In 2009, 4 new health warnings were required forsmokeless tobacco products. In June 2011, FDA promul-gated a historic final rule requiring cigarette packaging andadvertisements nationwide to feature graphic images of thenegative health consequences of smoking along with newwarning statements (designed to cover 50% of the front andback of cigarette packages and 20% of each cigarette adver-tisement). The government is currently appealing a courtruling that has halted implementation of the cigarette healthwarnings.

Pillar 4: Advance KnowledgeUS Surgeon General Reports. Since 2010, Surgeon Gen-eral Benjamin has released 2 reports: How Tobacco SmokeCauses Disease: The Biology and Behavioral Basis for Smoking-Attributable Disease4 and Preventing Tobacco Use Among Youthand Young Adults.5 They update dimensions of the epi-demic, especially its current effects on the next generation.

FDA and NIH Cohort Study on Tobacco Use. In Octo-ber 2011, the FDA and the National Institutes of Health (NIH)

launched the largest-ever national cohort study of more than55 000 tobacco users (and those at risk) to monitor and as-sess the health effects and consequences of usage. The study,which evaluates susceptibility to usage and effects of regu-latory changes on initiation, cessation, risk perceptions, andother attitudes, will provide an evidence base for new FDAtobacco product regulations.

In short, under the Obama Administration, a series of re-juvenated tobacco control initiatives has accelerated na-tional action. But more must be done. Approximately 45 mil-lion US residents currently smoke, and many more use otherforms of tobacco. Moreover, society’s most vulnerable mem-bers (eg, the poor, those with coexisting mental health orsubstance use histories) smoke at the highest rates, bear-ing a disproportionately high burden of illness and death.The United States can end the tobacco epidemic by meet-ing these challenges and reaffirming the commitment to ahealthier, tobacco-free future.

Published Online: August 15, 2012. doi:10.1001/jama.2012.9741Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest and none were re-ported.

REFERENCES

1. Centers for Disease Control and Prevention (CDC). Vital signs: current ciga-rette smoking among adults aged �18 years—United States, 2005-2010. MMWRMorb Mortal Wkly Rep. 2011;60(35):1207-1212.2. Ending the Tobacco Epidemic: A Tobacco Control Strategic Action Plan for theU.S. Department of Health and Human Services. http://www.hhs.gov/ash/initiatives/tobacco/tobaccostrategicplan2010.pdf. 2010. Accessed 3 May, 2012.3. Fiore MC, Jaen CR, Baker TB, et al. Treating Tobacco Use and Dependence:2008 Update. Rockville, MD: US Dept of Health and Human Services; 2008.4. How Tobacco Smoke Causes Disease: The Biology and Behavioral Basis forSmoking-Attributable Disease: A Report of the Surgeon General. Atlanta, GA: Centersfor Disease Control and Prevention, National Center for Chronic Disease Preven-tion and Health Promotion, Office on Smoking and Health; 2010.5. Preventing Tobacco Use Among Youth and Young Adults: A Report of theSurgeon General. Atlanta, GA: Centers for Disease Control and Prevention, Na-tional Center for Chronic Disease Prevention and Health Promotion, Office on Smok-ing and Health; 2012.

Figure. Federal Tobacco Prevention and Cessation Initiatives: 2009 to Present

February 2009American Recovery and Reinvestment Act: Communities Putting Prevention to Work grants

February 2009Children’s Health Insurance Program Reauthorization Act (increases the federal cigarette excise tax rate)

March 2010Patient Protection and Affordable Care Act includes tobacco prevention and cessation emphasis

May 2009NIH expands innovative telephone, mobile, and social media initiatives to make quitting easier (eg, smokefree.gov)

June 2009Family Smoking Prevention and Tobacco Control Act (authorizes FDA regulation of tobacco)

August 2010Medicare cessation counseling benefit expanded

December 2010Surgeon General’s Report: How Tobacco Smoke Causes Disease

Jan2009

Jan2010

Jan2011

Jan2012

October 2011FDA and NIH cohort study on tobacco use launched

September 2011CDC awards more than $100 million in Community Transformation Grants, funded by the Affordable Care Act of 2010

March 2012Surgeon General’s Report: Preventing Tobacco Use Among Youth and Young Adults

August 2012CDC awards $21 million to support state telephone quitlines

March 2010Prevent All Cigarette Trafficking Act

October 2010Tobacco cessation coverage for pregnant women by Medicaid

July 2011HHS establishes a tobacco free–campus policy for all its facilities

January 2011Comprehensive cessation coverage for federal employees, their dependents, and retirees mandated by OPM

March 2012CDC national media campaign Tips from Former Smokers launched

VIEWPOINT

768 JAMA, August 22/29, 2012—Vol 308, No. 8 ©2012 American Medical Association. All rights reserved.


VIEWPOINT

Preventing Patient Harms Through Systemsof CarePeter J. Pronovost, MD, PhDGeorge W. Bo-Linn, MD, MHA

THE CENTERS FOR MEDICARE & MEDICAID SERVICES

(CMS) recently launched Partnership for Patients,an ambitious national effort designed to substan-tially reduce 9 types of preventable harm and hos-

pital readmissions.1 These harms include adverse drug events,catheter-associated urinary tract infections, central line–associated bloodstream infections, fall injuries, pressure ul-cers, surgical site infections, venous thromboembolisms, ven-tilator-associated pneumonia, and obstetrical adverse events.Thousands of hospitals have agreed to participate and choseto focus on several harms because it was beyond their ca-pacity to simultaneously address all 9 types of harm.

The dilemma is that most patients are at risk of most ofthe 9 harms and other harms, including loss of dignity anda sense of respect for their values. Yet patients can expectphysicians to focus harm-reduction efforts on just a few ofthese harms. Health care is addressing these harms as if eachtype occurs in isolation. This reaction has occurred be-cause it is too burdensome to attempt to reduce multipleharms at the same time. The “siloing” of preventing patientharm is inefficient. Health care needs a different approachto reducing patient harm.

Harms tend to cluster within patients, making these harmsinterdependent rather than independent. Inpatients oftenexperience multiple harms,2 and patients in the intensivecare unit (ICU) experience multiple complications.3 Thus,the solutions for harm reduction must be integrated and in-terdependent. For instance, patients receiving mechanicalventilation in the ICU are at risk of 8 of the 9 harms on theCMS list, such as central line–associated bloodstream in-fections, ventilator-associated pneumonia, and venous throm-boembolism.1 Mechanically ventilated patients are also sus-ceptible to harms not included on the CMS list, such asdelirium, diagnostic errors, and air embolism. A patient whoexperiences one type of harm is at increased risk of expe-riencing other harms because that patient is further dis-abled and exposed to indwelling catheters and other de-vices and spends more time in the ICU and the hospital.2

To mitigate each type of harm, patients require a check-list of interventions, with some checklists for some pa-tients used multiple times daily. For example, patients must

receive 5 separate interventions multiple times a day to re-duce their risk of developing ventilator-associated pneu-monia, and they should receive subcutaneous heparin mul-tiple times a day and continuously wear sequentialcompression devices (SCDs) to reduce their risk of devel-oping deep venous thrombosis. Each type of harm requiresits own list of interventions. To implement every interven-tion multiple times a day to keep a mechanically ventilatedpatient safe from all harms, the patient must reliably re-ceive scores of interventions. The actual number of inter-ventions, however, depends on the frequency of patient moni-toring for continuous interventions (eg, SCDs) and thethoroughness of checklists (eg, oral care can be 1 or mul-tiple items). The complexity makes achieving these goalschallenging, if not impossible.

Technology has not provided an effective solution. In manyinstances, clinicians rely on memory and checklists to com-plete interventions, but no list has all the required inter-ventions. Moreover, clinician competency in performing in-terventions is not tested, and ambiguity increases variation.4

There is no continuously refreshed visual display of com-pleted, duplicated, or missed interventions and no auto-mated prompts to perform interventions or e-uploading ofcompleted tasks into the electronic medical record (EMR).From a patient’s perspective, success does not equate to avoid-ing 2 CMS-listed harms but experiencing all other harms.

Clinicians can reduce patient harm in the same way thatengineers manage dynamic complexity: as a systems prob-lem that requires an interdisciplinary systems solution. Aholistic view is necessary when analyzing patient require-ments. To protect a mechanically ventilated patient, for in-stance, requires compiling a list of that patient’s risks of harmand the evidence-based practices to reduce those harms.Those practices should be implemented efficiently by or-ganizing them as a system of care, automating as many prac-tices as possible, showing when practices are needed or com-pleted, ensuring integration across devices and with the EMR,

Author Affiliations: Armstrong Institute for Patient Safety and Quality, Johns Hop-kins Medicine, Departments of Anesthesiology and Critical Care Medicine and Sur-gery, Johns Hopkins University School of Medicine, Department of Health Policyand Management, Johns Hopkins Bloomberg School of Public Health, Johns Hop-kins University School of Nursing, and Johns Hopkins Carey Business School, Bal-timore, Maryland (Dr Pronovost); and Patient Care Program, Gordon and BettyMoore Foundation, Palo Alto, California (Dr Bo-Linn).Corresponding Author: Peter J. Pronovost, MD, PhD, Department of Anesthesi-ology and Critical Care Medicine, Johns Hopkins Medical Institutions, 1909 ThamesSt, Second Floor, Baltimore, MD 21231 ([email protected]).



and continuously learning how to improve performance. Forexample, patients receiving narcotic infusions may experi-ence respiratory depression and possibly respiratory ar-rest. However, most infusion pumps do not interface withdevices that monitor a patient’s respiratory rate. A systems-based solution would integrate the devices: the pump wouldstop infusing the narcotic if the patient’s respiratory rate de-creased below a specific threshold and the pump would beprogrammed to send a message to the clinician.

Another example is one therapy known to reduce mor-tality among patients with acute lung injury: lung-protective ventilation. Less than half of patients receive thislifesaving therapy.5 If the ventilator were programmablethrough the EMR, the system could automatically recog-nize patients with acute lung injury and initiate lung pro-tective ventilation. Another system-based integration couldautomatically gather the multiple data elements needed todiagnose sepsis, decreasing the usual delay in diagnosis andtreatment. Such a systems solution could address alarm fa-tigue and decrease reliance on clinician vigilance. The sys-tems approach would also include a machine-human inter-face, requiring data transparency, actionable feedback,teamwork, and continuous learning.

The Institute of Medicine and the National Academy ofEngineering have highlighted the benefits of using engi-neering principles to deliver reliable and affordable high-quality health care.6 Examples of successful collaborationbetween engineering and medicine are rare, and none havebeen replicated or comprehensive or have incorporated thepatient perspective. As a result, patients continue to expe-rience preventable harms that cost billions of dollars in avoid-able complications and needless care.7

Academic health systems, partnering with industry andother disciplines such as engineering, physics, and business,could develop and implement such a systems approach bycollaborating, for example, to design a systems model to pro-tect ICU patients from various types of preventable harm thatcould occur simultaneously. Such a model could reliably de-liver the best practices to reduce harm, which also could po-tentially reduce costs and enhance productivity.

This systems approach to reduce preventable harm couldbe applied in primary care settings, where preventable harmalso occurs. For example, care of patients with complex medi-cal problems is managed by many physicians who recom-mend multiple therapies, often with limited communica-tion among physicians, limited understanding of patients’ability to adhere to recommendations, and limited feed-back to physicians about adherence or performance.

Developing and implementing a systems approach to man-aging patient care will be difficult to achieve. Doing so re-

quires an engineering mind-set that most clinicians are notused to, a melding of different professional cultures, lan-guages, and styles. It also requires clearly defined goals toreduce preventable harm and costs and to meet those goals,just as an engineer would manage a project. A pilot test ofthis systems approach for mechanically ventilated patientsshowed that it resonated immediately with patients, clini-cians, engineers, social scientists, and health services re-searchers. This approach will not replace the art of medi-cine or the science of experience, intuition, collegialrelationships, and communication. The art of medical prac-tice will always be vitally important, but whenever pos-sible, science should guide decisions and actions. No doubtthere will be gaps in the evidence when developing sys-tems models, and when there are, clinicians can pool theirexperience and wisdom of practice to decide what to do.

It is time for the science of health care delivery to matureand embrace systems engineering. It is time for health careto embrace the compelling goal of reducing preventable pa-tient harm. By systematically addressing all the known harmspatients may experience, clinicians may realize this goal and,by doing so, improve health care and reduce costs of care,ultimately improving value for the nation and its citizens.Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Pronovost reportsreceiving grant or contract support from the Agency for Healthcare Research andQuality, the National Institutes of Health, RAND, and the Commonwealth Fundfor research related to measuring and improving patient safety; honoraria fromvarious hospitals and health care systems and the Leigh Bureau to speak on qual-ity and safety; consultancy with the Association for Professionals in Infection Con-trol and Epidemiology Inc; and book royalties for authoring Safe Patients, SmartHospitals: How One Doctor’s Checklist Can Help Us Change Health Care Fromthe Inside Out. Dr Bo-Linn reports no conflicts of interest.Additional Contributions: We thank Christine G. Holzmueller, BLA, Armstrong In-stitute of Patient Safety and Quality, Johns Hopkins Medicine, and Departmentof Anesthesiology and Critical Care Medicine, Johns Hopkins University School ofMedicine, for her thoughtful review and assistance in editing the manuscript. Nocompensation was received.

REFERENCES

1. US Department of Health and Human Services. Partnership for patients:learn about our areas of focus. http://www.healthcare.gov/compare/partnership-for-patients/safety/index.html. Accessed June 22, 2012.2. Classen DC, Resar R, Griffin F, et al. “Global trigger tool” shows that adverseevents in hospitals may be 10 times greater than previously measured. Health Aff(Millwood). 2011;30(4):581-589.3. Pronovost PJ, Dang D, Dorman T, et al. Intensive care unit nurse staffing andthe risk for complications after abdominal aortic surgery. Eff Clin Pract. 2001;4(5):199-206.4. Gurses AP, Seidl KL, Vaidya V, et al. Systems ambiguity and guideline compli-ance: a qualitative study of how intensive care units follow evidence-based guide-lines to reduce healthcare-associated infections. Qual Saf Health Care. 2008;17(5):351-359.5. Pronovost PJ, Murphy DJ, Needham DM. The science of translating researchinto practice in intensive care. Am J Respir Crit Care Med. 2010;182(12):1463-1464.6. Grossmann C, Goolsby WA, Olsen L, et al. Engineering a Learning HealthcareSystem: A Look at the Future. Washington, DC: National Academies Press; 2011.7. Emanuel EJ. Where are the health care cost savings? JAMA. 2012;307(1):39-40.

VIEWPOINT



VIEWPOINT

Transforming Quality of Care and ImprovingOutcomes After Acute MILessons From the National Registry of Myocardial InfarctionWilliam J. French, MDVanessa S. Reddy, MSHal V. Barron, MD

THE NATIONAL REGISTRY OF MYOCARDIAL INFARCTION

(NRMI) was the leading cardiovascular disease (CVD)patient registry from 1990 to 2006 in both size andduration. Notably, the various analyses and publica-

tions that resulted from the NRMI had direct and substantialeffects in improving quality care for acute myocardial infarc-tion (AMI) treatment in the United States.1

The NRMI has provided important lessons regarding effec-tiveobservational studydesign, implementationandanalysis,and the use and communication of findings pertinent to thebroader clinical and research community. At the intersectionbetween comparative effectiveness research and clinical qual-ity improvement, the NRMI serves as a template for the devel-opment of future registries and large observational studies.

Although CVD continues to be the leading cause of deathin the United States, major advances in AMI management, suchas the introduction of thrombolytic therapy (1980s) and ad-vanced invasive primary angioplasty techniques (1990s), havemitigated the disease burden over recent decades. These thera-peutic milestones coincided with important changes in the UShealth care system, with demands for higher-quality and moreuniform care across hospitals. However, clinical observa-tional tools and data infrastructure necessary for quality mea-surement, such as databases and registries, were lacking at lo-cal and national levels. Development of the NRMI was a steptoward clinical measurement required to quantify this signifi-cant unmet medical need and evaluate effectiveness of exist-ing and emerging AMI therapies.

History of NRMI: Evolving Research ObjectivesThe NRMI, initiated by Genentech Inc in 1990, began as apostmarketingcommitmenttomonitorthesafetyandeffective-nessof recombinant tissueplasminogenactivator.Anindepen-dent, external scientific advisory board, composed of practic-ing emergency physicians and cardiologists, was established.Theseadvisorswerecritical ingeneratingclinicallyrelevantandcontemporaryresearchquestionsapplicabletoparticipatinghos-pitals and the medical community and in providing strategicdirection over the NRMI’s 17-year duration (NRMI 1-5). Alsofundamental to the NRMI model was the participation of hos-pitalsitesandstaff indatacollectionandprovisionofnovelquar-terlyperformancereports thathelpedmonitorandbenchmarkagainstguidelinerecommendationsandbestpractices.Thisdata

collectionmodelwasrevolutionaryinscope,anditsgrowthwasexponential, enrolling more than 2.5 million patients hospi-talized for AMI from more than 1700 sites, representing morethan 25% of US hospitals providing AMI care.

Ultimately evolving into a national observational tool, theNRMI helped transform the quality of AMI care through mea-surement and analysis of treatment patterns and patient out-comes. A major observation from the NRMI was the relation-ship between the time to reperfusion strategy and subsequentadverse in-hospital outcomes, which was previously uncleardue toconflicting studyresults (TABLE).2 Tracked throughqual-ity improvement initiatives using NRMI reports, door-to-needle (DTN) and door-to-balloon (DTB) times decreasedsignificantly over time and were associated with improved in-hospital mortality.3 By study end, median DTN and DTB timesfell below guideline recommendations (Table).3

The NRMI findings have been communicated throughoutthe clinical community by regional presentations and by morethan 130 published articles and 170 abstracts. Widespread andaccelerated dissemination of results and best practices helpedphysicians and hospitals achieve remarkable improvementsin AMI treatment delivery and outcomes. The AMI treat-ment knowledge base has also expanded through the mul-tiple iterations of NRMI findings incorporated within mul-tiple iterations of AMI clinical practice guidelines, reflectingthe effect of the NRMI on clinical practice.4

Furthermore, the influence of the NRMI is also evidencedbyits integrationasahospitalquality improvementandreport-ing tool, and the enduring legacy with its adaption into theACTION-GWTG registry. Of note, the approved use of NRMIdata for the JointCommissionAMImeasure reporting5 furtherembeddedthestudywithinparticipatingsitesandencouragedadditional hospital involvement in quality reporting.

TheNRMIunderscoredtheimportanceofamultistakeholderresearchpartnershipamongpracticingclinicians,hospitalstaff,andthe industrysponsor.Thiscollaborationcontributedtothesuccessandlongevityof thestudy, fulfillingdataneedsforregu-latory requirements while integrating clinical considerations.TheNRMIexemplifies theutilityofobservational researchandits potential as a tool to monitor drug effectiveness, while col-lecting data necessary to evaluate and enact improvements inthe health care system through a research partnership.

Author Affiliations: Department of Medicine, David Geffen School of Medicineat University of California, Los Angeles (UCLA), Harbor-UCLA Medical Center,Torrance (Dr French); and Genentech Inc, South San Francisco, California (Ms Reddyand Dr Barron).Corresponding Author: William J. French, MD, Department of Medicine, DavidGeffen School of Medicine at UCLA, Harbor-UCLA Medical Center, Torrance, CA90509 ([email protected]).



Looking Forward: Building on the NRMI ModelThe substantial effect of the NRMI analyses demonstrates howobservational studies have become a critical tool for answer-ing research questions often not feasible or ethical to ad-dress using randomized controlled trials (RCTs). As illus-trated by the NRMI, observational research should be viewedas complementary to other study designs, with acknowledg-ment of its limitations, such as confounding and bias. Fur-thermore, patient registries may address certain research ques-tions better than RCTs, such as the natural history of disease,changing practice patterns, treatment effectiveness in di-verse populations, regional variations, and medical resourceutilization while enabling the study of larger numbers of pa-tients within the usual care setting.6 RCTs, by their nature,may not have been feasible for evaluating many of the NRMIresearch questions and would not have similarly catapultedimprovements in patient care and outcomes through the mea-surement and analysis of near real-time data. Moreover, find-ings from registries may contribute to the success of RCTsby generating hypotheses for future testing and providing eventrate estimates to improve power calculations.

As health care practice and treatment options evolve, pa-tient registries remain an important tool for evaluating theinterface among patients, physicians, and sites of care. TheNRMI was a remarkable achievement and unexpectedly suc-

cessful in transforming CVD care, in part because it set ahigh standard for observational research and met an emerg-ing need in the US health care system. The legacy of the NRMIis a model to build upon.

Conflict of Interest Disclosures: All authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Ms Reddy and Dr Bar-ron are employees and own stock in Hoffmann-La Roche.Additional Contributions: We thank Susan Begelman, MD, Patricia Reilly, RN, MSN,Charles Barr, MD, MPH, and Eshita Patel, MBA, MS, for their assistance (withoutcompensation) and Cindy Macpherson, MD, MediTech Media, for editorial sup-port, as part of her regular duties.

REFERENCES

1. Peterson ED, Shah BR, Parsons L, et al. Trends in quality of care for patientswith acute myocardial infarction in the National Registry of Myocardial Infarctionfrom 1990 to 2006. Am Heart J. 2008;156(6):1045-1055.2. McNamara RL, Herrin J, Bradley EH, et al; NRMI Investigators. Hospital im-provement in time to reperfusion in patients with acute myocardial infarction, 1999to 2002. J Am Coll Cardiol. 2006;47(1):45-51.3. Gibson CM, Pride YB, Frederick PD, et al. Trends in reperfusion strategies, door-to-needle and door-to-balloon times, and in-hospital mortality among patients withST-segment elevation myocardial infarction enrolled in the National Registry ofMyocardial Infarction from 1990 to 2006. Am Heart J. 2008;156(6):1035-1044.4. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the man-agement of patients with ST-elevation myocardial infarction—executive summary.Circulation. 2004;110(5):588-636.5. French WJ. Trends in acute myocardial infarction management. Am J Cardiol.2000;85(5A):5B-9B.6. Canto JG, Kiefe CI, Williams OD, Barron HV, Rogers WJ. Comparison of out-comes research with clinical trials using preexisting data. Am J Cardiol. 1999;84(8):923-927, A6.

Table. Summary of Key National Registry of Myocardial Infarction Findings

Research Question Results Effect Guideline Use Source

Reperfusion times andoutcomes, year:median, min

DTN (thrombolytics) 1990: 592006: 29

16.3% Relative in-hospitalmortality improvement Time to reperfusion

standards createdGibson CM, et al. Am Heart J.

2008;156(6):1035-1044DTB (primary PCI) 1990: 1112006: 79

7.5% Relative in-hospitalmortality improvement

Procedural volume andclinical outcomes

28% Lower in-hospital mortalityin high-volume angioplastyhospitals

Focused angioplastyprocedures in largervolume hospitals

Minimum proceduralvolumerecommendationsestablished

Canto JG, et al. N Engl J Med.2000;342(21):1573-1580

Characteristics of patientseligible for reperfusiontherapy but untreated

24% received no therapy;Independent predictors:LBBB, no angina, �75 y,womena

More rigorous assessment inundertreated patients

Reflected in clinicalpractice guidelines

Barron HV, et al. Circulation.1998;97(12):1150-1156

Minority population treatment,median time, min

DTN time Black: 41Hispanic: 36White: 34 Initiated a call to action to

improve treatment andoutcomes

Reflected in clinicalpractice guidelines

Bradley EH, et al. JAMA.2004;292(13):1563-1572DTB time Black: 122

Hispanic: 115White: 103

Introduction of Gp IIb-IIIainhibitors

Only 25% of patients treated;Use within 24 h loweredmortality by 12%

Increased use of Gp IIb-IIIainhibitors to improveclinical outcomes

Data used to supportupdates to clinicalpractice guidelines

Peterson ED, et al. J Am CollCardiol. 2003;42:45-53

AMI performance and qualitymeasurement

NRMI data quality comparablewith government-sponsoreddatabase (CCP); usingNRMI variations in treatmentand patient outcomesmeasured and reported

Approval of NRMI data for theJoint Commission AMImeasure reporting

Encouraged use ofregistries forcontinuous qualitymeasurement

Every NR, et al. J Am CollCardiol.1999;33(7):1886-1894

Abbreviations: AMI, acute myocardial infarction; CCP, Cooperative Cardiovascular Project; DTB, door to balloon; DTN, door to needle; Gp IIb-IIIa, glycoprotein IIb/IIIa; NRMI, NationalRegistry of Myocardial Infarction; PCI, percutaneous coronary intervention.

aOdds ratio: 0.22 for left bundle-branch block (LBBB); 0.22 for no angina, 0.40 for women, and 0.88 for older than 75 years.

VIEWPOINT



VIEWPOINT

The Value of Statistical Analysis Plansin Observational ResearchDefining High-Quality Research From the StartLaine Thomas, PhDEric D. Peterson, MD, MPH

THE INCREASING AVAILABILITY OF ELECTRONIC HEALTH

data combined with federal investment has stimu-lated an expansion in observational clinical re-search.1 Observational studies can complement clini-

cal trials and provide important information aboutcomparative safety and effectiveness in populations not wellstudied in clinical trials. However, there are numerous ex-amples in which the findings from observational studies havefailed to be replicated.2 These failures may be due to sev-eral factors, including the exploratory nature of observa-tional questions, failure to fully account for treatment se-lection bias, known publication biases, and the tendency topursue post hoc hypotheses. This later problem, termed datadredging, is facilitated by the lack of fidelity to a prespeci-fied hypothesis and inadequate reporting of the actual ana-lytic process.

In contrast to observational research, clinical trials ordi-narily operate under strict standards at every step of studyplanning and data analysis. A detailed protocol, includingthe definition of end points, hypotheses, and all analyticalprocedures, is submitted to the US Food and Drug Admin-istration and registered in various data repositories, such asclinicaltrials.gov, prior to enrollment of patients. Trial reg-istration helps ensure that both positive and negative find-ings are publicly known. Prespecification of trial protocolscreates an incentive to understand the biological functionof the intervention, carefully define the population of in-terest, target the most appropriate end points, and achievecertainty about the statistical approach. Prespecification ofhypotheses and minimal testing means that standard er-rors and P values are accurate measures of uncertainty andstatistical evidence is rigorous. Trial protocols can also bereferred to and reviewed to understand the questions, endpoints, and subgroup analyses that were defined ahead oftime and those that were post hoc and in need of replica-tion for validation.

A natural question arises as to whether elements of thisrigorous process should be applied to observational re-search. While select observational studies are already reg-istered in clinicaltrials.gov,3 some argue that observationalresearch is, by its nature, exploratory and requires substan-tial flexibility to investigate novel findings and unexpectedsignals in the data.4,5 Yet interpretation of statistical evi-dence (P values and confidence intervals) can be made po-tentially meaningless when multiple hypotheses are gener-ated by exploring the available data. Hence, a balance mustbe achieved to promote some flexibility but also encouragea rigorous, efficient analytical process that minimizes un-necessary data dredging.

Aside from considering the advantages of preregistra-tion, substantial progress has been made to define stan-dards for reporting observational research.6,7 The Strength-ening the Reporting of Observational Studies in Epidemiology(STROBE) recommendations provide a checklist of itemsto include in reporting of cohort, case-control, and cross-sectional studies.6 Good Research for Comparative Effec-tiveness (GRACE) principles similarly reflect a consensuson good practice for design and evaluation in comparativeeffectiveness research.7 Despite these standards for high-quality observational research and reporting, such guide-lines are not consistently adopted, in part because of theircomplexity and the difficulty of including all componentsin published articles.

The concepts for improving observational research canbe operationalized via use of a formal, prospectively de-fined statistical analysis plan (SAP). The SAP should in-clude enough detail that another statistician familiar withthe data set (or their own independent data) could repli-cate the analysis. This implies that the SAP should delin-eate populations (exclusion criteria); end points; descrip-tive objectives; testable hypotheses; modifications orderivations of standard variables; statistical methods, in-cluding handling of missing data, correlated data, bias, andconfounding; subgroups; interactions; and sensitivity analy-

See also p 771.

Author Affiliations: Department of Biostatistics and Bioinformatics, Duke Uni-versity School of Medicine (Dr Thomas), and Duke Clinical Research Institute(Dr Peterson), Durham, North Carolina. Dr Peterson is also Contributing Editor,JAMA.Corresponding Author: Eric D. Peterson, MD, MPH, Duke Clinical Research In-stitute, 2400 Pratt St, Durham, NC 27705 ([email protected]).



sis. While the SAP should be finalized prior to data analy-sis, authors may make changes to the analytic plan in re-sponse to subsequent findings. These changes to methods,hypotheses, or both should be noted in the SAP to capturewhen and why components were modified during analysis.Once created, the SAP should be cited in the methods, sub-mitted along with the manuscript for review, and poten-tially made available as an online appendix to a publishedarticle.

The benefits of this process are numerous. First, it pro-motes good planning rather than haphazard data analysisand communicates this distinction to reviewers and read-ers. Second, it optimizes statistical resources to focus thebest methods on good questions, those with the potentialfor important findings, either negative or positive. When keyhypothesis are defined at the outset, they can be carefullyaddressed. Third, it facilitates transparency. The existingSTROBE recommendations for reporting are quite compre-hensive but almost impossible to address in the limited spacethat is afforded to a published methods section, particu-larly if the statistical methods are to be replicable. The sub-mission of an SAP provides reviewers with a complete de-scription of what was done, not limited by space. Fourth,this approach may increase efficiency by avoiding distract-ing messages, maintaining focus on the a priori hypotheseswith room for post hoc and sensitivity analysis to be re-ported.

This suggested process for the SAP in observationalresearch is feasible for real-world adoption. For instance,the Duke Clinical Research Institute, an analytical centerfor the American College of Cardiology National CardiacData Registries, has implemented such an SAP processwithin their ACTION and Cath/PCI registries. A formalproposal is submitted by clinical researchers, includingdetailed background and hypotheses, and prioritized by apublications committee. From the proposal, a primary stat-istician works with the clinical researcher to develop theSAP, translating the clinical questions into descriptiveobjectives and testable hypotheses. Statistical methods areproposed to address each major objective, including thedetails mentioned above, table shells for intended output,and the corresponding potential conclusions or takeawaymessage. A senior statistician and coauthors review theSAP, and it is revised iteratively until all parties supportthe aims and approach. The review of table shells andcross-checking of potential conclusions with technicallystated hypotheses help avoid misunderstandings betweenthe clinician and statistician. These details allow theauthors to visualize the project and anticipate issues.While developing the SAP, the statistician may investigate

the data availability, extent of missingness, colinearity, andother issues. However, the analysis begins once the SAP isfinalized and a single report containing all information isprovided to the primary author. Some revisions are nearlyalways necessary; unforeseen issues with the data mayindicate alternative statistical methods or unexpectedresults may require new analysis. Both the SAP and reportare revised to reflect changes.

The process of writing and submitting an SAP capturesmany of the attributes of clinical trials, without excessiverigidity that would inhibit exploratory research. It requiresextra work on the front end but greater efficiency and clar-ity in producing and reporting results. Current practicemay be augmented by making the SAP publicly available asonline ancillary material.8 This would allow readers to con-firm and, if desired, replicate the methods used in thestudy. Some authors have expressed concerns that readersof observational research may become too rigid and dismissan important finding just because it was not prespecified.5

However, this can be mitigated if authors make a strongbiological case to support post hoc findings and readersmay, appropriately, require more confirmatory evidence.The gains in public and academic trust associated withtransparency outweigh this concern.9 Thus, investigatorsconducting observational research should develop and useprespecified SAPs and should submit these to journals,along with their manuscripts, for review and ultimateonline publication.

Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest and none were re-ported.

REFERENCES

1. Patient-Centered Outcomes Research Institute. http://www.pcori.org/what-we-do/. Accessed July 27, 2012.2. Lauer MS. Will academia embrace comparative effectiveness research? AcadMed. 2011;86(6):671-673.3. Williams RJ, Tse T, Harlan WR, Zarin DA. Registration of observational studies:is it time? CMAJ. 2010;182(15):1638-1642.4. Vandenbroucke JP. Observational research, randomised trials, and 2 views ofmedical science. PLoS Med. 2008;5(3):e67.5. Pearce N. Registration of protocols for observational research is unnecessaryand would do more harm than good. Occup Environ Med. 2011;68(2):86-88.6. von Elm E, Altman DG, Egger M, Pocock SJ, Gøtzsche PC, Vandenbroucke JP;STROBE Initiative. Strengthening the Reporting of Observational Studies in Epi-demiology (STROBE) statement: guidelines for reporting observational studies. BMJ.2007;335(7624):806-808.7. Dreyer NA, Schneeweiss S, McNeil BJ, et al; GRACE Initiative. GRACE prin-ciples: recognizing high-quality observational studies of comparative effectiveness.Am J Manag Care. 2010;16(6):467-471.8. Sox HC, Helfand M, Grimshaw J, et al; PLoS Medicine Editors. Comparativeeffectiveness research: challenges for medical journals. PLoS Med. 2010;7(4):e1000269.9. Ioannidis JPA. The importance of potential studies that have not existed andregistration of observational data sets. JAMA. 2012;308(6):575-576.

VIEWPOINT



A PIECE OF MY MIND

Poolside

FOR MANY YEARS I’VE WORKED WITH CANCER PATIENTS

and their families. When parents of younger childrendevelop cancer, inevitable questions focus on the chil-

dren. What should we tell them? How can we help themeven as we don’t know how to help ourselves?

We all have algorithms in our heads, abbreviated formu-las for dealing with common situations. Be transparent, I’vealways said. Be developmentally appropriate. Children aresponges, they’ll pick up on what’s happening. Better theyunderstand that a parent has an illness. Make things pre-dictable for them. Given them attention, structure, love.

But I’d never stopped to think about how these sugges-tions really play out in a family. How a diagnosis really im-pacts children or, for that matter, a spouse. That is, untilmy wife was diagnosed with an aggressive breast cancer. Itall came home to me on an ordinary day. We were actuallydoing quite well, I thought. I’d shared the diagnosis withour 8-year-old, but not yet broached the topic in detail withthe 5-year-old. On this particular day, a few months intoher treatment, Terry was recovering from chemotherapy, andto give her some space, I took our two daughters to the pool.

I parked the minivan, and the girls and I made the longwalk around the fence line, each of us carrying towels. Alex,8, and Abby, 5, had backpacks containing their clothing tochange into after we swam. From the sidewalk we could seeinto the swimming area, where a number of families had al-ready staked out chairs and children’s shouts emanated fromthe kid’s area beneath the flapping tarp. The girls walked alittle faster; Abby ran a hand along the fence, her towel drag-ging in the dust.

“Abby, pick up your towel,” I told her.We entered the local community center, I swiped my mem-

bership card, and Alex led Abby into the girls’ room whileI walked through the boys’. I stood outside waiting for themto emerge. It’s always a little awkward, this part—becausethe girls’ room is large and has a separate shower and bath-room and locker room, not to mention a hot tub. My girlshave occasionally wandered into the wrong section and got-ten lost. Alex panicked once, but lately she’s rememberedhow to get out.

I was standing, waiting for them, when I saw movementover my shoulder. Looking back, I could see two motherssitting on the lawn chairs looking at me, their faces dartedaway quickly, just as Alex pushed open the door and Abbyfollowed her out.

“I told you to pick up your towel!” Alex implores.“I don’t have to listen to you. Leave me alone!” her little

sister snaps back.

“Girls!” I say, louder than intended.“She’s bossy!” Abby proclaims.“She’s dragging her towel through the pee water. It’s gross.”“You’re gross,” Abby yells. And she lowers her forehead

and growls. A real growl. Like a rabid raccoon. The womenare looking at us. I want us to look normal, above the fray,totally fine. But I know right now I am ABCD for them—ABreast Cancer Dad, in totally over my head and unable tosoothe and control my own children. I know I shouldn’t carewhat they think, I know this is normal, I know we are do-ing as well as we can, right? But right now, Abby is pushingher sister.

“Abby!” Alex is saying. She towers over her little sisterbut is, by temperament, a pacifist. Or so I think, becausenow she grimaces and throws her hands forward, catchingher little sister’s shoulders, toppling her backwards onto theconcrete. I know that Abby has been provoking her big sis-ter lately, but the only thing the women see is the little girlgetting knocked over by the big girl, and the father stand-ing there motionless.

The mothers stand, hesitating before breaking into an-other family’s tensions, but then one of them says some-thing to the other and they rush to us as if they’ve seen enoughof an adolescent driver. That’s me, the adolescent driver, Irealize, unable to control the massive vehicle that is my fam-ily. Abby is bawling, there are tears in Alex’s eyes. Right nowmy wife is either asleep or throwing up, alone, two milesfrom here and all I want is to be there and for all of this togo away.

They pick up Abby, whose back is scraped; the womenfawn over her, Abby’s big blue eyes are filled with tears. Alexquietly walks around them and sits on one of the loungechairs, still clutching her towel, and stares vacantly out intothe pool. One of the women gives Alex a dirty look and, Ithink, contemplates putting her in time out, but I sit downnext to Alex like a guard dog.

“You can go in,” I say to her, gesturing to the pool. Thissurprises the women fawning over Abby because they ex-pect me to chastise her, I suppose. Alex looks at me, blink-ing into the sun, but doesn’t speak. She clutches her towelto her belly and looks out over the pool, rocking a little.

It’s a look I’ve seen before in veterans, the hollowed-outstare. Last night she found Terry in our bathroom on herknees in the act of throwing up. It was one of the first timesshe’d seen the evidence of the illness. Unlike her little sis-ter, who can sense that things aren’t normal, but probably

A Piece of My Mind Section Editor: Roxanne K. Young, Associate Senior Editor.



can’t understand the details, Alex is observant and tunedto the unseen emotional frequencies in the house. Beforethis I could tell that she was upset, but largely based on ourchaotic adjustments to the new schedule, and perhaps, myfear. But this was the first time she could see it. Her mom,kneeling in sweatpants and a baggy T-shirt, on the floor,her hair stringy and bunched. Small globes of sweat on herneck.

I heard Terry spit and then rouse a passable, “Hi, honey.What do you need?”

“Are you okay?” Alex asked.“Oh, sure. You hungry?” Terry asked, as if she was stand-

ing in the kitchen instead of on her knees.

I arrived in the bathroom in time to say, “Hey, you wantsome mac and cheese?” But Alex just shook her head. Ab-solutely not.

And now she stares at the other children diving and chas-ing, giggling and floating across the pool, and her eyes areindifferent and dead.

“You can go in,” I say to her again. She doesn’t respond,her eyes locked on the silver reflections in the rippling water.

Daniel Shapiro, PhD

Author Affiliation: Department of Humanities, Penn State College of Medicine,Hershey, Pennsylvania ([email protected]).Conflict of Interest Disclosures: The author has completed and submitted the ICMJEForm for Disclosure of Potential Conflicts of Interest and none were reported.

The roots of reason are embedded in feelings—feelings that have formed and accumulated and de-veloped over a lifetime of personality-shaping. Thesefeelings are not a source of weakness but a resourceof strength. They are not there for occasional usingbut are inescapable. To know what we think, we mustknow how we feel. It is feeling that shapes belief andforms opinion. It is feeling that directs the strategy ofargument. It is our feelings, then, with which we mustcome to honorable terms.

—James E. Miller Jr (1920-2010)

A PIECE OF MY MIND



ORIGINAL CONTRIBUTION

Scan for AuthorVideo InterviewEffect of Biolimus-Eluting Stents With

Biodegradable Polymer vs Bare-Metal Stentson Cardiovascular Events Among PatientsWith Acute Myocardial InfarctionThe COMFORTABLE AMI Randomized TrialLorenz Raber, MDHenning Kelbæk, MDMiodrag Ostoijc, MDAndreas Baumbach, MDDik Heg, PhDDavid Tuller, MDClemens von Birgelen, MD, PhDMarco Roffi, MDAris Moschovitis, MDAhmed A. Khattab, MDPeter Wenaweser, MDRobert Bonvini, MDGiovanni Pedrazzini, MDRan Kornowski, MDKlaus Weber, MDSven Trelle, MDThomas F. Luscher, MDMasanori Taniwaki, MDChristian M. Matter, MDBernhard Meier, MDPeter Juni, MDStephan Windecker, MDfor the COMFORTABLE AMI TrialInvestigators

PRIMARY PERCUTANEOUS CORO-nary intervention (PCI) is thereperfusion therapy of choiceamong patients with ST-

segment elevation myocardial infarc-tion (STEMI) owing to a lower risk of re-

infarction and improved survivalcompared with fibrinolysis.1,2 Bare-metal stents minimize the risk of in-farct vessel reocclusion and reinfarc-tion compared with balloon angioplasty,but are associated with restenosis due toneointimal hyperplasia.3,4 Early genera-tion drug-eluting stents releasing siroli-mus or paclitaxel from durable poly-

mers reduce the need for repeatrevascularization compared with bare-metal stents.5-7 However, vessel healingis delayed with evidence of chronic in-

For editorial comment see p 814.

Author Video Interview available atwww.jama.com.

Author Affiliations are listed at the end of this ar-ticle.Corresponding Author: Stephan Windecker, MD, De-partment of Cardiology, Bern University Hospital, 3010Bern, Switzerland ([email protected]).

Context The efficacy and safety of drug-eluting stents compared with bare-metalstents remains controversial in patients with ST-segment elevation myocardial infarc-tion (STEMI) undergoing primary percutaneous coronary intervention (PCI).

Objective To compare stents eluting biolimus from a biodegradable polymer withbare-metal stents in primary PCI.

Design, Setting, and Patients A prospective, randomized, single-blinded, con-trolled trial of 1161 patients presenting with STEMI at 11 sites in Europe and Israelbetween September 19, 2009, and January 25, 2011. Clinical follow-up was per-formed at 1 and 12 months.

Intervention Patients were randomized 1:1 to receive the biolimus-eluting stent(n=575) or the bare-metal stent (n=582).

Main Outcome Measures Primary end point was the rate of major adverse car-diac events, a composite of cardiac death, target vessel–related reinfarction, and ischemia-driven target-lesion revascularization at 1 year.

Results Major adverse cardiac events at 1 year occurred in 24 patients (4.3%) re-ceiving biolimus-eluting stents with biodegradable polymer and 49 patients (8.7%)receiving bare-metal stents (hazard ratio [HR], 0.49; 95% CI, 0.30-0.80; P=.004). Thedifference was driven by a lower risk of target vessel–related reinfarction (3 [0.5%] vs15 [2.7%]; HR, 0.20; 95% CI, 0.06-0.69; P=.01) and ischemia-driven target-lesionrevascularization (9 [1.6%] vs 32 [5.7%]; HR, 0.28; 95% CI, 0.13-0.59; P� .001) inpatients receiving biolimus-eluting stents compared with those receiving bare-metalstents. Rates of cardiac death were not significantly different (16 [2.9%] vs 20 [3.5%],P=.53). Definite stent thrombosis occurred in 5 patients (0.9%) treated with biolimus-eluting stents and 12 patients (2.1%; HR, 0.42; 95% CI, 0.15-1.19; P=.10) treatedwith bare-metal stents.

Conclusion Compared with a bare-metal stent, the use of biolimus-eluting stentswith a biodegradable polymer resulted in a lower rate of the composite of major ad-verse cardiac events at 1 year among patients with STEMI undergoing primary PCI.

Trial Registration clinicaltrials.gov Identifier: NCT00962416JAMA. 2012;308(8):777-787 www.jama.com



flammation related at least in part to thepersistence of durable polymer compo-nents in patients with acute STEMI.8

In addition, acute myocardial infarc-tion (MI) is a predictor of thromboticstent complications occurring late afterdrug-eluting stent implantation, particu-larly in the presence of a high thrombusburden, raising concerns regarding thebalance of risks and benefits of these de-vices in this clinical setting.9,10 Twometa-analyses in patients with acute MIconfirmed a lower risk of repeat revas-cularization with early generation drug-eluting stents compared with bare-metal stents, however, at the expense ofa 2-fold increased risk of very late stentthrombosis.11,12

Newer-generation drug-eluting stentswith biodegradable polymers providecontrolled drug release with subse-quent degradation of the polymer ren-dering the stent surface more closely toa bare-metal stent after the period of bio-degradation. The unrestricted use ofstents eluting biolimus, an equipotent si-rolimus analogue, from biodegradablepolylactic acid was noninferior and po-tentially better than sirolimus-elutingstents in terms of major adverse clinicalevents in a large clinical trial with fol-low-up of 4 years, with a substantially re-duced risk (80%) of stent thrombosis oc-curring beyond 1 year.13,14 A stratifiedanalysis suggested a particularly pro-nounced benefit among patients withacute STEMI. We therefore performed amulticenter, randomized trial to com-pare the efficacy and safety of stents elut-ing biolimus from a biodegradable poly-mer with bare-metal stents of otherwiseidentical design.

METHODSStudy Design

The COMFORTABLE AMI trial (Com-parison of Biolimus Eluted From anErodible Stent Coating With Bare MetalStents in Acute ST-Elevation Myocar-dial Infarction) was a multicenter, ran-domized, assessor-blinded, superior-ity trial in patients presenting withSTEMI undergoing primary PCI. Thefull study design was reported else-where.15 The study complied with the

Declaration of Helsinki and was ap-proved by all institutional ethics com-mittees. All patients provided writteninformed consent.

Patient Population

Patients aged 18 years or older withsymptom onset within 24 hours and ST-segment elevation of at least 1 mm in2 or more contiguous leads, true pos-terior MI, or new left bundle branchblock were eligible for randomizationin the presence of at least 1 culprit le-sion within the infarct vessel. There wasno limit regarding the number of treatedlesions, vessels, or complexity. Exclu-sion criteria were presence of mechani-cal complications of acute MI, knownallergy to any study medication, use ofvitamin K antagonists, planned sur-gery unless dual antiplatelet therapycould be maintained throughout theperisurgical period, history of bleed-ing diathesis or known coagulopathy,pregnancy, participation in another trialbefore reaching the primary end point,inability to provide informed consent,and noncardiac comorbid conditionswith life expectancy of less than 1 year.Patients were recruited between Sep-tember 19, 2009, and January 25, 2011,in 11 centers throughout Europe andIsrael (Denmark [n=1], Israel [n=1],the Netherlands [n=1], Serbia [n=1],Switzerland [n=6], and the UnitedKingdom [n=1]).

Procedures

Randomization was performed via aweb-based system after diagnostic an-giography. The allocation sequence wascomputer generated in randomly vary-ing blocks of 2, 4, and 6, and stratifiedby center. Patients were randomly al-located on a 1:1 basis to treatment withstents eluting biolimus from a biode-gradable polylactic acid polymer (Bio-Matrix, Biosensors Europe SA) or bare-metal stents of otherwise identicaldesign (Gazelle, Biosensors Europe SA).

Both stent types were available in di-ameters of 2.25, 2.50, 2.75, 3.00, 3.50,and 4.00 mm and in lengths of 8, 11,18, 24, and 28 mm. Before stent im-plantation, thrombus aspiration was

recommended in all patients when-ever aspiration was deemed techni-cally feasible. Predilation of the cul-prit lesion was left to the discretion ofthe operator. Complete revasculariza-tion of all lesions within the infarct ves-sel had to be performed with the ran-domly allocated study stent. Nonculpritvessels were treated by default with bio-limus-eluting stents at baseline and dur-ing follow-up and did not contribute tothe primary end point. Staged proce-dures for the treatment of nonculpritvessels were permitted within 3 monthswith the uniform use of biolimus-eluting stent in all lesions.

Acetylsalicylic acid (�250 mg) wasadministered before the procedure. Incenters where prasugrel was available,an initial dose of 60 mg (including pa-tients preloaded with clopidogrel) wasadministered followed up with a dailydose of 10 mg. If prasugrel was notavailable or contraindicated, clopido-grel was administered at a loading doseof 600 mg, followed up with a dose of75 mg twice daily for 7 days, followedup with a maintenance dose of 75 mgonce daily. Dual antiplatelet therapywas prescribed for the duration of atleast 1 year in all patients. Unfraction-ated heparin was routinely adminis-tered with a minimal dose of 5000 IEor a dose of 70 to 100 IU/kg to main-tain an activated clotting time of 250seconds. Bivalirudin was adminis-tered at a dose of 0.75 mg/kg intrave-nously followed up with an infusion of1.75 mg/kg per hour during the dura-tion of the procedure. The use of gly-coprotein IIb/IIIa inhibitors was left tothe discretion of the operator.

We assessed creatinine kinase, cre-atinin kinase-MB, and troponin at ad-mission and thereafter every 8 hoursuntil the peak creatinine kinase hadbeen reached. A 12-lead electrocardio-gram was performed before the proce-dure, within 24 hours after the proce-dure, before discharge, and in case ofrecurrent signs of ischemia.

Data Management

Independent study monitors verifiedsource data according to a prespeci-

BIOLIMUS-ELUTING VS BARE-METAL STENTS AND CARDIOVASCULAR EVENTS



fied monitoring plan.15 Data were storedin a central database (Cardiobase, Clini-cal Trials Unit, and Department of Car-diology, Bern University Hospital, Swit-zerland, and 2mT, Ulm, Germany).Follow-up appointments were sched-uled at 30 days and 1 year, and pa-tients were questioned about the oc-currence of angina, any adverse events,hospitalization, and cardiovascularmedication intake. All serious adverseevents were submitted to Clinical TrialsUnit, University of Bern, Bern, Swit-zerland, in a blinded fashion. Any death,reinfarction, revascularization, stentthrombosis, cerebrovascular accident,and bleeding event was indepen-dently adjudicated by a blinded clini-cal event committee. An independentdata and safety monitoring boardblinded to treatment groups periodi-cally reviewed all event information and

compared safety outcomes betweengroups.

Angiograms were centrally assessedby the core laboratory at Bern Univer-sity Hospital by blinded personnel. Thebare-metal stent and biodegradablepolymer-based drug-eluting stent wereindistinguishable on angiography. Thecore laboratory assessment encom-passed quantitative coronary angiog-raphy using dedicated software (XAan-gio XA 7.1, Medis) and the assessmentof the SYNTAX MI score,16 and throm-bolysis in MI (TIMI) flow grade.

Study End Points

The prespecified primary end point wasthe device-oriented composite of car-diac death, target vessel–related rein-farction, and ischemia-driven target-lesion revascularization at 1 year.Secondary end points included the pa-

tient-oriented composite of death, anyreinfarction, and any revasculariza-tion, as well as target vessel–related re-infarction and any revascularization(percutaneous and surgical proce-dures), cardiac death, all-cause mor-tality, Q-wave and non–Q-wave rein-farction, stroke, stent thrombosisaccording to the definitions of the Aca-demic Research Consortium, and de-vice and lesion success. Detailed defi-nitions of all primary and secondary endpoints were reported elsewhere.15

Target-lesion revascularization wasdefined as a repeated revasculariza-tion due to a stenosis within the stentor within the 5-mm borders proximalor distal to the stent. Target-vessel re-vascularization was defined as any re-peat PCI or surgical bypass of any seg-ment within the entire major coronaryvessel proximal and distal to a target le-

Figure 1. Trial Profile and Flow of Patients

575 Included in primary analysis15 Censored at time point of refusal or

lost to follow-up

582 Included in primary analysis16 Censored at time point of refusal or

lost to follow-up

2575 Patients with STEMI were treated at 11centers during study inclusion period

575 Randomized to receive biolimus-eluting stent (629 lesions)570 Received ≥1 biolimus-eluting stent (624 lesions)

3 Did not receive biolimus-eluting stent (3 lesions)

2 Had PCI but did not receive a stent (2 lesions)

1 Received study bare-metal stent only (1 lesion)2 Received nonstudy bare-metal stent only (2 lesions)

568 Received biolimus-eluting stent only (621 lesions)1 Received biolimus-eluting stent (1 lesion) and

nonstudy drug-eluting stent (1 lesion)1 Received biolimus-eluting stent and nonstudy

bare-metal stent in same lesion (1 lesion)

578 Randomized to receive biolimus-eluting stent3 Provided preliminary consent but refused definite

consent after randomization

583 Randomized to receive bare-metal stent1 Provided preliminary consent but refused definite

consent after randomization

582 Randomized to receive bare-metal stent (648 lesions)572 Received ≥1 bare-metal stent (638 lesions)

6 Did not receive bare-metal stent (6 lesions)

3 Had PCI but did not receive a stent (3 lesions)1 Had no PCI (1 lesion assumed) b

2 Received study biolimus-eluting stent only (2 lesions)2 Received nonstudy drug-eluting stent only (2 lesions)2 Received nonstudy bare-metal stent only (2 lesions)

563 Received bare-metal stent only (621 lesions)9 Received bare-metal stent and biolimus-eluting

stent (17 lesions)

11 Lost to follow-up before 1 year4 Refused follow-up

542 Alive18 Died

560 Follow-up for primary clinical end pointavailable up to 1 year

11 Lost to follow-up before 1 year5 Refused follow-up

543 Alive23 Died

566 Follow-up for primary clinical end pointavailable up to 1 year

1414 Excluded a

1161 Randomized

STEMI indicates ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention.aA total of 2575 patients underwent primary PCI for treatment of STEMI at 11 international sites during the inclusion period. No reliable data for patients assessed foreligibility are available.bOne patient treated with bare-metal stent did not undergo PCI and was assumed to have 1 lesion.




sion, including upstream and down-stream branches and the target lesionitself. A revascularization was consid-ered ischemia-driven if the diameter ste-

nosis of the treated lesion was at least50% on the basis of quantitative coro-nary angiography in the presence of is-chemic signs or symptoms, or if there

was a diameter stenosis of at least 70%irrespective of the presence of ische-mic signs or symptoms.

Statistical Analysis

Our trial was powered for superiority onthe primary clinical end point at 1 year.On the basis of the HORIZON-AMI6 andLEADERS trials,13 we assumed an inci-dence of major adverse cardiac events of14% within 1 year in the bare-metal stentgroup and a 40% relative risk reductionassociated with the biolimus-elutingstent, corresponding to a rate of majoradverse cardiac events of 8.4%. Enroll-ment of 1064 patients would thereforeprovide 80% power to detect a relativerisk of 0.60 with 2-sided �=.05.

All analyses were performed accord-ing to the intention-to-treat principle,with inclusion of all randomized pa-tients in the analysis according to thegroup they were originally allocated.Cox proportional hazards regressionmodels were used to compare clinicaloutcomes between the groups, with pa-tients censored at the time of their lastknown contact. Correspondingly, weconstructed time-to-event curves usingKaplan-Meier estimates and the pro-portional hazards assumption wastested and met in each case.

Categorical variables are reported asnumbers and percentages, and groupsare compared using �2 or Fisher exacttests (low counts). Continuous vari-ables are reported as means ±standarddeviations, and groups are comparedusing unpaired t tests. Times are re-ported as medians (interquartile ranges[IQRs]), and groups are comparedusing Wilcoxon Mann-Whitney ranksum test.

We prespecified stratified analyses ofthe primary end point at 1 year accord-ing to age, sex, diabetes, renal failure, le-sion length, and vessel size. In addition,we performed post hoc analyses strati-fied according to left ventricular ejec-tion fraction, left anterior descending ar-tery lesion localization, preproceduralTIMI flow, time from pain onset to bal-loon time, thrombus aspiration, and mul-tivessel treatment. All stratified analy-ses were accompanied by tests for

Table 1. Baseline Clinical Characteristics of Patients Receiving Either Biolimus-Eluting Stentsor Bare-Metal Stentsa

CharacteristicsBiolimus-Eluting Stents

(n = 575)Bare-Metal Stents

(n = 582)

PatientsAge, mean (SD), y 60.7 (11.6) 60.4 (11.9)

Male sex 463 (80.5) 455 (78.2)

BMI, mean (SD) 27.3 (4.5) 27.2 (4.0)

Cardiovascular risk factorsDiabetes 84 (14.6) 90 (15.5)

Hypertension 279 (48.5) 265 (45.5)

Hyperlipidemia 324 (56.6) 328 (56.7)

Current smoker 272 (47.9) 301 (52.3)

Family history of CAD 193 (34.3) 179 (31.3)

Renal failure 79 (14.1) 86 (15.1)

Previous MI 31 (5.4) 32 (5.5)

Previous PCI 19 (3.3) 27 (4.6)

Previous CABG surgery 10 (1.7) 4 (0.7)

Laboratory findingsAnemia 87 (15.6) 79 (13.9)

Thrombocytopenia 22 (4.0) 30 (5.3)

Clinical presentationTime to balloon inflation, median (IQR), min

From symptom onset 232 (164-380) 236 (163-400)

0-6 421 (73.2) 421 (72.6)

6-12 109 (19.0) 100 (17.2)

12-24 45 (7.8) 59 (10.2)

From hospital admission 44 (32-70) 44 (32-74)

Killip class II, III, or IV, No./total No. (%) 40 (7.0) 37 (6.4)

Resuscitation before hospital arrival 16 (2.8) 9 (1.5)

LVEF, mean (SD), %b 49 (11) 50 (10)

Electrocardiographic localization of MIAnterior 211 (36.9) 213 (36.9)

Lateral 18 (3.1) 13 (2.2)

Inferior 236 (41.3) 249 (43.1)

Posterior 29 (5.1) 23 (4.0)

Inferior and posterior 78 (13.6) 80 (13.8)

Right ventricular MI 52 (9.1) 55 (9.5)

Lesion complexityc

Bifurcation lesion 52 (9.0) 49 (8.4)

Small vessel 74 (12.9) 79 (13.7)

Long lesion 204 (35.7) 183 (31.7)

SYNTAX MI score, mean (SD)d 15.1 (8.2) 14.8 (8.1)Abbreviations: ACE, angiotensin-converting enzyme; BMI, body mass index, calculated as weight in kilograms divided

by height in meters squared; CABG, coronary artery bypass graft; CAD, coronary artery disease; IQR, interquartilerange; LVEF, left ventricular ejection fraction; MI, myocardial infarction; PCI, percutaneous coronary intervention.

aData are expressed as No. (%) unless otherwise specified. Anemia was defined as a hemoglobin concentration ofless than 120 g/L for women and less than 130 g/L for men, according to the definition of the World Health Orga-nization. Thrombocytopenia was defined as less than 150 000 platelets/µL. Renal failure was defined as glomerularfiltration rate of less than 60 mL/min.

bLeft ventricular function as assessed by angiography or by echocardiography, in case angiography was not available.cSmall vessel indicates reference vessel diameter of 2.5 mm or less and long lesion is a lesion length of 20 mm or

more.dSYNTAX score was assessed before recanalization according to Magro et al.16 The SYNTAX score provides a nu-

merical score summarizing a comprehensive angiographic assessment of the entire coronary artery tree, with higherscores indicating increasing complexity of CAD. In the absence of significant CAD (absence of a stenosis �50% ina vessel with a reference diameter �1.5 mm), the score amounts to 0. The score increases with more complex CAD.The highest score as reported in the SYNTAX trial amounted to 84.17




interaction between stent type and sub-group. There were positive treatment �patient characteristics interactions for ageand sex. In view of a correlation be-tween age and sex (P� .001), we ex-plored treatment � age interactions sepa-rately in men and women. All analyseswere performed with STATA version12.1 (StataCorp) and 2-sided P�.05 wasconsidered statistically significant.

RESULTSA total of 1161 patients were ran-domly assigned to receive biolimus-eluting stents with biodegradable poly-mer (578 patients) or bare-metal stents(583 patients). Three patients allo-cated to the biolimus-eluting stent and1 patient allocated to the bare-metalstent did not confirm their initial writ-ten consent and had to be excluded, re-sulting in 575 patients with 629 infarct-vessel lesions randomly assigned tobiolimus-eluting stents and 582 pa-tients with 648 infarct-vessel lesionsrandomly assigned to bare-metal stentsfor final analyses (FIGURE 1). A total of31 patients refused or were lost to fol-low-up at a median of 31 days in thebiolimus-eluting stent group and 32days in the bare-metal stent group.

Baseline medications and clinical, an-giographic, and procedural character-istics were similar in both groups(TABLE 1, TABLE 2, and TABLE 3). Themean (SD) age of patients was 60.6(11.8) years and 79% were men. Themedian (IQR) time from symptomonset to balloon inflation was 234 (164-386) minutes and from hospital admis-sion to balloon inflation was 44 (32-72) minutes. Thrombus aspiration wasperformed in 62% of patients and 47%received a glycoprotein IIb/IIIa antago-nist during the procedure. No differ-ences were observed in lesion complex-ity between both groups including theSYNTAX MI score (mean, 15; SD, 8).At discharge, 43% of patients receivedprasugrel and 57% of patients re-ceived clopidogrel. The use of dual an-tiplatelet therapy was high and bal-anced in both treatment groupsthroughout the entire follow-up pe-riod up to 1 year (Table 2).

Clinical outcomes during follow-upare shown in TABLE 4. At 1 year, the pri-mary end point of major adverse car-diac events (cardiac death, target vessel–related reinfarction, and ischemia-driven target-lesion revascularization)occurred in 4.3% of patients receivingbiolimus-eluting stents and 8.7% of pa-tients receiving bare-metal stents (haz-ard ratio [HR], 0.49; 95% CI, 0.30-0.80; P=.004) (FIGURE 2A). For cardiacdeath alone, the percentages weresmaller (2.9% of patients received bio-limus-eluting stents and 3.5% of pa-tients received bare-metal stents; HR,0.81; 95% CI, 0.42-1.56; P = .53)

(Figure 2B). The treatment effect in fa-vor of patients receiving biolimus-eluting stents was attributable to botha lower risk of target vessel–related re-infarction (0.5% vs 2.7%; HR, 0.20; 95%CI, 0.06-0.69; P=.01) (Figure 2C) andischemia-driven target-lesion revascu-larization (1.6% vs 5.7%; HR, 0.28; 95%CI, 0.13-0.59; P� .001) (Figure 2D).Differences between stent types with re-spect to the primary outcome emergedearly and continued throughout thestudy period.

Among patients treated with bioli-mus-eluting stents, 3 target vessel–related reinfarctions resulted from defi-

Table 2. Medication Use Among Patients Receiving Either Biolimus-Eluting Stents orBare-Metal Stentsa

MedicationsBiolimus-Eluting Stents


(n = 582)

During primary PCIUnfractionated heparin 510 (88.7) 523 (89.9)

Bivalirudin 74 (12.9) 67 (11.5)

Low molecular weight heparin 19 (3.3) 19 (3.3)

Glycoprotein IIb/IIIa antagonists 276 (48.0) 266 (45.7)

Loading dose of clopidogrel and prasugrelNone 2 (0.3) 6 (1.0)

Clopidogrel only (600 mg) 320 (55.8) 330 (57.0)

Prasugrel only (60 mg) 104 (18.2) 110 (19.0)

Both 126 (22.0) 128 (22.1)

At dischargeAcetylsalicylic acid 568 (99.8) 576 (99.7)

Clopidogrel 323 (56.8) 327 (56.6)

Prasugrel 245 (43.1) 248 (42.9)

Any dual antiplatelet therapy 567 (99.6) 574 (99.3)

�-Blockers 491 (86.3) 476 (82.4)

ACE inhibitors or receptor blockers 410 (72.1) 411 (71.1)

Statins 559 (98.2) 571 (98.8)

At 30 dAcetylsalicylic acid 555 (99.3) 565 (99.1)

Clopidogrel 323 (57.6) 322 (56.5)

Prasugrel 241 (43.0) 245 (43.0)


�-Blockers 487 (87.0) 482 (84.6)


Statins 534 (95.5) 550 (96.5)

At 1 yAcetylsalicylic acid 529 (97.6) 524 (96.3)

Clopidogrel 284 (52.4) 266 (48.9)

Prasugrel 214 (39.6) 222 (40.8)


�-Blockers 435 (80.4) 428 (78.8)


Statins 496 (91.7) 506 (93.2)Abbreviations: ACE, angiotensin-converting enzyme; PCI, percutaneous coronary intervention.aData are expressed as No. (%). Two-sided P values were calculated using �2 or Fisher exact texts (all P� .07).




nite stent thrombosis in 2 patients andrestenosis in 1 patient. Among pa-tients treated with bare-metal stents, 15target vessel–related reinfarctions re-sulted from definite stent thrombosisin 10 patients, restenosis in 4 patients,and spontaneous MI in 1 patient. Therisk of target vessel–related reinfarc-tion associated with stent thrombosisor restenosis was lower among pa-tients treated with biolimus-elutingstents vs bare-metal stents (HR, 0.22;95% CI, 0.06-0.75; P=.02).

The findings for the primary endpoint were consistent across stratifiedanalyses for diabetes, renal failure, leftventricular ejection fraction, left ante-rior descending artery, thrombus aspi-ration, time from pain onset to bal-loon inflation, multivessel treatment,small vessel disease, and lesion length(FIGURE 3). A significant interactionwith stent type was observed for age andsex. Men were on average 6.5 yearsyounger than women. In exploratoryanalyses, we found HRs below the pointestimate of the primary end point in theoverall cohort (HR=0.49) in womenyounger than 65 years (HR, 0.40; 95%CI, 0.80-1.95), in men younger than 65years (HR, 0.25; 95% CI, 0.10-0.61),and in men 65 years or older (HR, 0.43;95% CI, 0.16-1.11), but not in women65 years or older (HR, 1.89; 95% CI,0.65-5.54).

At 1 year, rates of definite stent throm-bosis amounted to 0.9% among pa-tients receiving biolimus-eluting stentsand 2.1% among patients receiving bare-metal stents (HR, 0.42; 95% CI, 0.15-1.19; P=.10). Five patients treated withbiolimus-eluting stents experienced defi-nite stent thrombosis while receivingdual antiplatelet therapy, whereas 12 pa-tients treated with bare-metal stents ex-perienced definite stent thrombosis with11 patients receiving dual antiplatelettherapy and 1 patient not taking acetyl-salicylic acid and clopidogrel. We ob-served no differences in all-cause and car-diac mortality between the groups at 1year. In addition to the device-orientedprimary outcome measure, we re-corded a lower risk of the comprehen-sive patient-oriented composite of death,

Table 3. Angiographic and Procedural Characteristics of Patients Receiving EitherBiolimus-Eluting Stents or Bare-Metal Stentsa

CharacteristicsBiolimus-Eluting Stents


(n = 648)P

ValueLesions

Treated in infarct vessel, No. 629 648b

Treated per patient, mean (SD) 1.1 (0.3) 1.1 (0.4) .61�2 treated in infarct vessel 49 (8.5) 59 (10.1) .34

Infarct vessel localizationLeft main coronary artery 2 (0.3) 1 (0.2)Left anterior descending artery 226 (39.3) 230 (39.6)Left circumflex artery 82 (14.3) 90 (15.5) .74Right coronary artery 264 (45.9) 259 (44.6)Saphenous vein graft 1 (0.2) 1 (0.2)

Baseline TIMI flow0 or 1 437 (69.6) 423 (65.6)2 81 (12.9) 95 (14.7) .313 110 (17.5) 127 (19.7)

Baseline quantitative coronary angiographic dataReference vessel diameter, mm

Mean (SD) 3.04 (0.47) 3.01 (0.46) .17Median (IQR) 3.02 (2.72-3.33) 2.97 (2.67-3.29) .17

Minimum lumen diameter, mmMean (SD) 0.42 (0.59) 0.44 (0.57) .47Median (IQR) 0 (0-0.80) 0 (0-0.88) .41

Diameter stenosis, %Mean (SD) 86.44 (18.37) 85.24 (18.37) .25Median (IQR) 100.0 (72.6-100.0) 100.0 (70.4-100.0) .21

Lesion length, mmMean (SD) 18.19 (9.73) 17.77 (9.57) .44Median (IQR) 15.46 (11.79-21.97) 15.63 (11.52-21.30) .46

Primary PCI procedureNo. of stents per lesion, mean (SD) 1.32 (0.61) 1.26 (0.60) .16

Type of stentc

Study biolimus-eluting 623 (99.4) 12 (1.9) �.001Other drug-eluting 0 2 (0.3) .50Study bare-metal 1 (0.2) 633 (98.3) �.001Other bare-metal 4 (0.6) 2 (0.3) .44No stent implanted 2 (0.3) 4 (0.6) .69

Stent length per lesion, mean (SD), mm 25.2 (12.7) 24.1 (12.3) .10Stent diameter per lesion, mean (SD), mm 3.2 (0.4) 3.2 (1.1) .42Direct stenting 236 (37.6) 240 (37.3) .89Maximal balloon pressure, mean (SD), atm 15.2 (3.5) 15.1 (3.4) .50Overlapping stents 148 (23.6) 120 (18.7) .03Thrombus aspiration 350 (60.9) 374 (64.4) .22Intraaortic balloon counterpulsation 14 (2.4) 15 (2.6) .88Intravenous vasopressors 18 (3.1) 19 (3.3) .90TIMI flow

0 or 1 3 (0.5) 3 (0.5)2 25 (4.0) 32 (5.0) .703 601 (95.5) 611 (94.6)

Postprocedural QCA, mean (SD)Reference vessel diameter, mm 3.06 (0.50) 3.02 (0.48) .07Minimum lumen diameter, mm 2.59 (0.50) 2.56 (0.48) .19Diameter stenosis, % 15.53 (8.35) 15.22 (7.81) .49

Abbreviations: IQR, interquartile range; PCI, percutaneous coronary intervention; QCA, quantitative coronary angiography;TIMI, thrombolysis in myocardial infarction.

aData are expressed as No. (%) unless otherwise specified. Two-sided P values were calculated using a �2 test for cat-egorical variables and using t test for continuous variables.

b Includes 1 patient randomly allocated to bare-metal stents who had no PCI and where 1 lesion was assumed; 2 patientsrandomly allocated to biolimus-eluting stents and 1 patient randomly allocated to bare-metal stents did not have QCAdue to missing angiography cardiovascular disease.

c In 1 patient who was randomly allocated to the biolimus-eluting stent group, a biolimus-eluting stent and a bare-metalstent were implanted within the same lesion; and in 5 patients who were randomly allocated to the bare-metal stentgroup, a biolimus-eluting stent and a bare-metal stent were implanted within the same lesion.




any reinfarction, and any revasculariza-tion in favor of biolimus-eluting stents(8.4% vs 12.2%; HR, 0.68; 95% CI, 0.47-0.98; P=.04).

Staged procedures were performedafter a median duration of 12.0 days(IQR, 4.0-41.5 days) among patientstreated with biolimus-eluting stents andafter a median duration of 6 days (IQR,3-33 days; P=.25). A total of 1 ischemia-driven target-lesion revascularizationwas associated with a staged proce-dure in the biolimus-eluting stent groupcompared with 2 ischemia-driven tar-get-lesion revascularization events inthe bare-metal stent group.

COMMENTIn this randomized, multicenter, asses-sor-blinded trial in patients with STEMI,compared with the use of bare-metalstents, the use of biolimus-eluting stentswith a biodegradable polymer was as-sociated with a significant 4.4% abso-lute reduction and 51% relative reduc-tion in the risk of major adverse cardiacevents at 1 year, which prevents 42events per 1000 patients treated withbiolimus-eluting stents compared withbare-metal stents at 1 year. Findingswere also robust for the more compre-hensive patient-oriented composite ofany death, reinfarction, or revascular-ization. Accordingly, our results sug-gest better clinical outcomes in termsof major adverse cardiac events of astent releasing biolimus from a biode-gradable polymer compared with abare-metal stent for the treatment of pa-tients with STEMI.

In the single largest trial enrolling pa-tients with STEMI,6 paclitaxel-elutingstents resulted in a 41% lower risk oftarget-lesion revascularization com-pared with bare-metal stents. In ourtrial, biolimus-eluting stents were as-sociated with a 4.9% absolute reduc-tion and a 72% relative reduction in therisk of ischemia-driven target-lesion re-vascularization compared with bare-metal stents. This risk reduction is no-table as repeat revascularizations weredue to recurrent ischemia in the ab-sence of protocol-mandated angio-graphic follow-up before assessment of

the primary end point at 1 year. Ratesof revascularization with the biolimus-eluting stent in our study were lowerthan in the LEADERS trial,13 which en-rolled patients with a broad spectrumof indications and lesions. Explana-tions for this finding include the largerreference vessel diameter in vesselscausing acute MI (3.0 vs 2.6 mm) andthe lack of routine angiographic follow-up, as well as less ischemia in vesselssubtending previously infarcted myo-cardium.13

Despite a similar risk profile andmortality in our study compared withpatients with STEMI enrolled into pre-vious large-scale randomized trials,5,6 weobserved a lower absolute rate of re-peat revascularization in both treat-ment groups. This observation is con-sistent with a recent trial comparingnewer-generation everolimus-elutingstents with bare-metal stents among pa-tients with STEMI18 and is potentially

related to improved lesion prepara-tion due to thrombus aspiration19 andmore potent antithrombotic medica-tions, such as prasugrel,20 reducing therisk of stent thrombosis–related revas-cularization. Notwithstanding, the ab-solute risk reduction of 4.1% in our trialmeans that 24 patients need to betreated with biolimus-eluting stents toprevent 1 major adverse cardiac event.

Differences in favor of biolimus-eluting stents over bare-metal stents inour study with respect to the primaryend point were not limited to efficacybut also driven by an 80% lower riskof target vessel–related reinfarction.This difference in safety has not beenobserved in previous randomized trialscomparing drug-eluting and bare-metal stents among patients withSTEMI,5,6,21-26 but is consistent with thefindings of a recent meta-analysis12 re-porting a lower risk of reinfarction dur-ing the first year with a number needed

Table 4. Clinical Outcomes at 1 Year of Patients Receiving Either Biolimus-Eluting Stents orBare-Metal Stentsa

Clinical Outcomes

No. (%) of Patients

Hazard Ratio(95% CI)

PValue

Biolimus-ElutingStents

(n = 575)

Bare-MetalStents

(n = 582)

Death 18 (3.2) 23 (4.1) 0.79 (0.43-1.47) .46

Cardiac death 16 (2.9) 20 (3.5) 0.81 (0.42-1.56) .53

Reinfarction 11 (2.0) 21 (3.7) 0.53 (0.25-1.09) .08

Q-wave 2 (0.4) 7 (1.2) 0.29 (0.06-1.39) .12

Non–Q-wave 9 (1.6) 14 (2.5) 0.65 (0.28-1.50) .31

Target vessel–related reinfarction 3 (0.5) 15 (2.7) 0.20 (0.06-0.69) .01

Q-wave 1 (0.2) 7 (1.2) 0.14 (0.02-1.17) .07

Non–Q-wave 2 (0.4) 8 (1.4) 0.25 (0.05-1.19) .08

Cardiac death or target vessel–relatedreinfarction

19 (3.4) 32 (5.7) 0.60 (0.34-1.05) .07

Any target-lesion revascularization 9 (1.6) 34 (6.0) 0.26 (0.13-0.55) �.001

Ischemia-driven target-lesionrevascularization

9 (1.6) 32 (5.7) 0.28 (0.13-0.59) �.001

Any target-vessel revascularization 11 (2.0) 37 (6.5) 0.30 (0.15-0.58) �.001

Ischemia-driven target-vesselrevascularization

11 (2.0) 35 (6.2) 0.31 (0.16-0.62) �.001

Major adverse cardiac eventsb 24 (4.3) 49 (8.7) 0.49 (0.30-0.80) .004

Death, any reinfarction, anyrevascularization

47 (8.4) 69 (12.2) 0.68 (0.47-0.98) .04

Stroke 6 (1.1) 4 (0.7) 1.52 (0.43-5.40) .51

Stent thrombosisDefinite 5 (0.9) 12 (2.1) 0.42 (0.15-1.19) .10

Definite or probable 14 (2.5) 21 (3.7) 0.67 (0.34-1.32) .25aHazard ratios are derived from Cox proportional hazard regression models. P values are 2-sided from superiority testing

with a �2 test.bComposite of cardiac death, target vessel–related reinfarction, and ischemia-driven target-lesion revascularization.




to treat of 79 compared with 45 in ourstudy.

In exploring the mechanism for thelower risk of target vessel–related rein-farction, we observed that the device-related adverse events definite stentthrombosis or target-lesion revascular-ization due to restenosis were respon-sible for target vessel–related reinfarc-tion in 17 of 18 cases and were lesscommon among patients receiving bio-limus-eluting stents than patients re-ceiving bare-metal stents (3 vs 14, re-spectively; P=.01). In contrast with mostprevious trials among patients withSTEMI, dual antiplatelet therapy was bal-anced between patients receiving bioli-mus-eluting stents and those receiving

bare-metal stents throughout the en-tire study period rendering differencesin antiplatelet therapy unlikely as an ex-planation for the differential in targetvessel–related reinfarction.

We found positive interactions be-tween stent type and age and sex. Be-cause age and sex were correlated, wefurther explored these interactions andfound estimated HRs below the pointestimate of the primary end point forthe overall cohort (HR = 0.49) inyounger women and men irrespectiveof age, but not in women aged 65 yearsor older. It is unclear whether our re-sults reflect a lack of benefit in womenor in those aged 65 years or older, orin the subgroup of elderly women only.

We are unaware of biological mecha-nisms that might explain interactionswith age or sex, and in view of the lackof mechanisms and the large numberof stratified analyzes, chance should alsobe considered as an explanation of ourfindings.

A numerically lower rate of definitestent thrombosis was observed (0.9%vs 2.1%, P=.10) with the use of bioli-mus-eluting stents vs bare-metal stentsat 1 year, with most events occurringduring the peri-interventional period.Although this finding has to be inter-preted cautiously, a similar statisti-cally nonsignificant reduction at up to1 year among patients with STEMI hasbeen observed in a recent meta-

Figure 2. Kaplan-Meier Curves for Major Adverse Cardiac Events, Cardiac Death, Target Vessel–Related Reinfarction, and Ischemia-DrivenTarget-Lesion Revascularization Among Patients Randomized to Receive Either the Biolimus-Eluting Stent or the Bare-Metal Stent

Major Adverse Cardiac EventsA

10

9

8

4

5

7

6

3

2

1

0

No. at riskBare-metal stentBiolimus-eluting stent

Bare-metal stent

HR, 0.49; 95% CI, 0.30-0.80; P = .004

Biolimus-eluting stent

0

575582

30

555562

60

543546

90

543542

120

541539

150

540535

180

540531

210

539526

240

537525

270

535522

300

534519

330

533517

365

514530

Days Since Index Procedure

Per

cent

age

Target Vessel–Related ReinfarctionC

10

9

8

4

5

7

6

3

2

1

0


Bare-metal stent

HR, 0.20; 95% CI, 0.06-0.69; P = .01


0

575582

30

558564

60

547548

90

547546

120

545545

150

544542

180

544539

210

543538

240

542538

270

540535

300

539534

330

538532

365

530535


Per

cent

age

Cardiac DeathB

10

9

8

4

5

7

6

3

2

1

0


Bare-metal stent

HR, 0.81; 95% CI, 0.42-1.56; P = .53


0

575582

30

560573

60

549558

90

549557

120

547556

150

546553

180

546551

210

545549

240

544549

270

543546

300

542545

330

541543

365

542538


Per

cent

age

Target-Lesion RevascularizationD

10

9

8

4

5

7

6

3

2

1

0


Bare-metal stent

HR, 0.28; 95% CI, 0.13-0.59; P <.001


0

575582

30

555563

60

543547

90

543543

120

541540

150

540536

180

540532

210

539527

240

537526

270

535523

300

534520

330

533518

365

515530


Per

cent

age

Major adverse cardiac events included a composite of cardiac death, target vessel–related reinfarction, and ischemia-driven target-lesion revascularization. P values are2-sided from Cox proportional hazards regression models �2 test. HR indicates hazard ratio.




analysis12 comparing early generationdrug-eluting stents with bare-metalstents. Moreover, a significant reduc-tion in the risk of stent thrombosis hasbeen reported in the EXAMINATIONtrial18 comparing newer-generationeverolimus-eluting stents with bare-metal stents (0.5% vs 1.9%, P=.01). Ex-perimental data indicate lower throm-bogenicity of drug-eluting stentscompared with bare-metal stents sug-gesting a possible thromboresistant ef-fect of polymer coatings during the im-

mediate peri-interventional period.27

The latter may be particularly impor-tant among patients with STEMI whocarry a higher baseline risk of stentthrombosis due to a large thrombusburden9 and increased platelet activa-tion.28

In addition, biolimus is the limusanalogue with the highest lipophilic-ity used for drug elution on currentlyavailable stent platforms.29 Among pa-tients with STEMI, the acute coronarylesions predominantly consist of lipid-

rich, ruptured plaques with large ne-crotic cores.30 Theoretically, the in-creased lipophilicity of the drugbiolimus may provide a more rapid andhomogeneous drug distribution, po-tentially leading to a more potent anti-inflammatory and antithrombotic lo-cal effect. However, this hypothesisrequires validation in dedicated stud-ies assessing the properties of variousdrugs used for elution on drug-eluting stents in the presence of lipid-rich plaques.

Figure 3. Subgroup Analyses of the 1-Year Rates of Major Adverse Cardiac Events Among Patients Randomized to Receive Either theBiolimus-Eluting Stent or the Bare-Metal Stent

FavorsBiolimus-Eluting

Stents

FavorsBare-MetalStents

0.05 41.00.1

Hazard Ratio (95% CI)

Major Adverse CardiacEvents, No.


Bare-MetalStents

Major Adverse CardiacEvents, Total No.


Bare-MetalStents P Value

Hazard Ratio(95% CI)

P forInteraction

Age, y8 30 370 381<65 .0010.27 (0.12-0.59)

16 19 205 201≥65 .580.83 (0.43-1.61).03

BMI18 30 447 441<30 .070.58 (0.33-1.05)5 17 122 136≥30 .030.32 (0.12-0.87)

.31

LVEF, %17 22 248 215<50 .220.67 (0.36-1.26)6 21 290 322≥50 .010.31 (0.13-0.77)

.17

Reference vessel diameter, mm5 8 74 79≤2.50 .500.68 (0.22-2.07)

19 41 498 501>2.50 .0050.46 (0.27-0.79).53

Lesion length, mm10 19 204 184≥20 .040.46 (0.21-0.98)14 30 368 396<20 .030.50 (0.27-0.95)

.85

Sex12 37 463 455Male <.0010.31 (0.16-0.60)12 12 112 127Female .741.14 (0.51-2.55)

.01

Diabetes8 9 84 90Yes .920.95 (0.37-2.47)

16 40 491 492No .0020.39 (0.22-0.70).13

Infarct lesion in left anterior descending artery12 24 238 239Yes .0480.50 (0.25-0.99)12 25 337 342No .040.48 (0.24-0.96)

.95

Thrombus aspiration13 34 350 374Yes .0050.40 (0.21-0.76)11 15 225 207No .310.67 (0.31-1.46)

.32

Multivessel treatment2 5 32 35Yes .280.41 (0.08-2.10)

22 44 543 546No .0070.50 (0.30-0.83).83

Preprocedural TIMI flow20 33 430 4070-1 .0470.57 (0.33-0.99)4 16 144 1722-3 .030.29 (0.10-0.86)

.28

Pain onset to balloon, h21 41 530 5210-12 .0090.50 (0.29-0.84)3 8 45 5912-24 .310.50 (0.13-1.89)

.99

Renal failure2 7 79 86Yes .140.30 (0.06-1.46)

21 42 483 483No .0080.49 (0.29-0.83).56

BMI indicates body mass index (calculated as weight in kilograms divided by height in meters squared); LVEF, left ventricular ejection fraction; TIMI, thrombolysis inmyocardial infarction. Two patients randomized to receive the biolimus-eluting stent and 1 patient randomized to receive the bare-metal stent could not be includedin the stratified analyses for lesion characteristics (reference vessel diameter and lesion length) due to missing angiography.




Our results have to be interpreted inview of the following limitations. First,the trial indicated superiority on the pri-mary composite outcome but was notpowered to address individual compo-nents of efficacy or safety. Moreover,observed event rates were lower thananticipated. In view of the size of theobserved treatment effect and results ofprevious trials, we consider it unlikelythat estimates of efficacy would sub-stantially differ in a larger patient co-hort. The inclusion of safety out-comes in the primary compositeoutcome is meaningful as cardiac deathor target vessel–related reinfarction maybe device related. Event rates of car-diac death or target vessel–related re-infarction were of similar magnitude asischemia-driven target-lesion revascu-larization in our trial providing a simi-lar weight of efficacy and safety param-eters within the composite end point.

Second, the biolimus-eluting stentused in our study is currently not ap-proved by the US Federal Drug Admin-istration and not considered as stan-dard of care in the United States. Thebiolimus-eluting stent has been shownto be noninferior compared with the si-rolimus-eluting CYPHER stent in a ran-domized trial of 1707 all-comer pa-tients for the composite clinical endpoint of major adverse events at 9months and 4 years.13,14 On the basis ofthese data, the biolimus-eluting stentis recommended as one of a few drug-eluting stents for clinical use in the Eu-ropean guidelines on myocardial re-vascularization.31 It remains to bedetermined how this stent platform per-forms compared with newer-genera-tion durable polymer-based drug-eluting stents. Similarly, the optimalduration of dual antiplatelet therapy af-ter implantation of biolimus-elutingstents with a biodegradable polymer hasnot been established.

Third, although our trial had very fewexclusion criteria, the results apply onlyto patients with characteristics similarto those enrolled. Patients who were un-able to provide written informed con-sent before the procedure had to be ex-cluded from participation in this trial

introducing an element of selection bias.Because no reliable data for reasonsleading to patient exclusion were col-lected, we cannot determine the pro-portion of patients excluded due to poorclinical condition and those refusingparticipation in the trial.

Fourth, the P2Y12 inhibitor prasu-grel was administered instead of clopi-dogrel in 40% of patients and may havecontributed to the low overall eventrates in our study. Although the use ofprasugrel was higher than in previoustrials comparing drug-eluting stentswith bare-metal stents among patientswith STEMI, it conforms to the recom-mendations of the American College ofCardiology/American Heart Associa-tion guidelines for the management ofSTEMI and reflects contemporary prac-tice.

Fifth, our study does not address lateevents beyond 1 year. However, in aprevious study,14 biolimus-elutingstents were shown to reduce the risk ofstent thrombosis beyond 1 year by 80%compared with early generation siro-limus-eluting stents providing sup-port for the improved long-term bio-compatibility of drug-eluting stents withbiodegradable polymer coatings.

In conclusion, compared with a bare-metal stent, the use of a biolimus-eluting stent with a biodegradable poly-mer resulted in a lower rate of thecomposite of major adverse cardiacevents at 1 year among patients withSTEMI undergoing primary PCI.

Author Affiliations: Department of Cardiology, BernUniversity Hospital, Bern, Switzerland (Drs Raber,Moschovitis, Khattab, Wenaweser, Taniwaki, Meier,and Windecker); Cardiac Catheterization Laboratory,Rigshospitalet, Copenhagen, Denmark (Dr Kelbæk);Department of Cardiology, Clinical Center of Serbia,Belgrade, Serbia (Dr Ostoijc); Bristol Heart Institute,Bristol, England (Dr Baumbach); Institute of Socialand Preventive Medicine (Drs Heg and Juni) andClinical Trials Unit, Department of Clinical Research(Drs Trelle, Juni, and Windecker), University of Bern,Bern, Switzerland; Cardiology Department, Triem-lispital, Zurich, Switzerland (Dr Tuller); Thoraxcen-trum Twente, Twente University, Enschede, theNetherlands (Dr von Birgelen); Division of Cardiol-ogy, University Hospital, Geneva, Switzerland (DrsRoffi and Bonvini); Cardiocentro, Lugano, Switzer-land (Dr Pedrazzini); Rabin Medical Center, PetachTikva, and Tel Aviv University, Tel Aviv, Israel (DrKornowski); Herzzentrum Bodensee, Kreuzlingen,Switzerland (Dr Weber); and Cardiology Depart-ment, University Hospital Zurich, Zurich, Switzerland(Drs Luscher and Matter).

Author Contributions: Drs Raber and Windecker hadfull access to all of the data in the study and take re-sponsibility for the integrity of the data and the ac-curacy of the data analysis.Study concept and design: Raber, Kelbæk, Trelle,Luscher, Meier, Juni, Windecker.Acquisition of data: Raber, Kelbæk, Ostoijc, Baumbach,Tuller, von Birgelen, Roffi, Moschovitis, Khattab,Wenaweser, Bonvini, Pedrazzini, Kornowski, Weber,Luscher, Taniwaki, Meier, Windecker.Analysis and interpretation of data: Raber, Kelbæk,Ostoijc, Baumbach, Heg, Tuller, von Birgelen, Roffi,Moschovitis, Khattab, Wenaweser, Bonvini, Pedrazzini,Kornowski, Weber, Trelle, Luscher, Taniwaki, Matter,Meier, Juni, Windecker.Drafting of the manuscript: Raber, Windecker.Critical revision of the manuscript for important intel-lectual content: Raber, Kelbæk, Ostoijc, Baumbach, Heg,Tuller, von Birgelen, Roffi, Moschovitis, Khattab,Wenaweser, Bonvini, Pedrazzini, Kornowski, Weber,Trelle, Luscher, Taniwaki, Matter, Meier, Juni, Windecker.Statistical analysis: Heg, Juni, Windecker.Obtained funding: Ostoijc, Matter, Juni, Windecker.Administrative, technical, or material support: Raber,Kelbæk, Ostoijc, Baumbach, Heg, Tuller, von Birgelen,Roffi, Moschovitis, Khattab, Wenaweser, Bonvini,Pedrazzini, Kornowski, Weber, Trelle, Luscher,Taniwaki, Matter, Meier, Juni, Windecker.Study supervision: Raber, Juni, Windecker.Conflict of Interest Disclosures: All authors have com-pleted and submitted the ICMJE Form for Disclosureof Potential Conflicts of Interest. Clinical Trials Unit(CTU Bern), which is part of the University of Bern,Bern, Switzerland, has a staff policy of not acceptinghonoraria or consultancy fees. However, CTU Bern isinvolved in design, conduct, or analysis of clinical stud-ies funded by Abbott Vascular, Ablynx, Amgen,AstraZeneca, Biosensors, Biotronic, Boehringer Ingel-heim, Eisai, Ei Lilly, Exelixis, Geron, Gilead Sciences,Nestle, Novo nordisc, Padma, Roche, Schering-Plough, St. Jude Medical, and Swiss Cardio Technolo-gies. Dr Baumbach reported being on advisory boardsand receiving consultancy fees from Boston Scien-tific, Medicines Company, and Abbott Vascular; andreceiving payment for lectures from Medicines Com-pany and Japan Stent Inc. Dr Tuller reported receiv-ing travel expenses from Biotronik, Biosensors, Terumo,and Medtronic. Dr von Birgelen reported board mem-berships and receiving lecture fees from Abbott Vas-cular, Medtronic, and Boston Scientific; receiving con-sultancy fees from Medtronic; unpaid consultanciesfrom Abbott Vascular, Boston Scientific, Biosensors,Biotronik, and Cordis; receiving grants from AbbottVascular, Boston Scientific, Biosensors, Biotronik, Cordis,Medtronic, and St Jude Medical; payment for lec-tures from Abbott Vascular, Boston Scientific,Medtronic, and MSD; and receiving payment for de-velopment of educational presentations from Cordis. DrRoffi reported receiving grants from Boston Scientific,Abbott Vascular, Medtronic, and Biopsensor; and pay-ment for lectures from Lilly-Daiichy Sankyo. Dr Khat-tab reported receiving payment for lectures from Bio-sensors. Dr Wenaweser reported receiving consultancyfees from NVT, grants from Medtronic, and paymentfor lectures from Medtronic, Edwards Lifesciences, andCordis, and payment for development of educationalpresentations from Medtronic. Dr Luscher reported re-ceiving research grants to the institution from Abbott,Biosensors, Biotronik, Boston Scientific, and Medtronic,and consultant payments from AstraZeneca, Boeh-ringer Ingelheim, Bayer, Merck, and Pfizer. Dr Matterreported receiving grants from MSD, Eli Lilly,AstraZeneca, and Bayer; expert testimony from MSD;payment for lectures from MSD, AstraZeneca, andRoche; and having patents from Mabimmune, CH.Dr Meier reported receiving research contracts to theinstitution from Abbott, Boston Scientific, Biosen-sors, and Cordis. Dr Juni is an unpaid steering com-




mittee or statistical executive committee member oftrials funded by Abbott Vascular, Biosensors,Medtronic, and St. Jude Medical. Dr Windecker re-ported receiving research contracts to the institutionfrom Abbott, Boston Scientific, Biosensors, Biotronik,Cordis, Medtronic, and St. Jude Medical. All other au-thors reported no conflicts of interest.Funding/Support: The COMFORTABLE AMI trial wasinvestigator-initiated, managed by the Clinical TrialsUnit, University of Bern, Bern, Switzerland, and sup-ported by the Swiss National Science Foundation (grant33CM30-124112), and an unrestricted research grantfrom Biosensors Europe SA, Morges, Switzerland (DrsJuni and Windecker). Dr Raber is the recipient of a re-search fellowship (SPUM) funded by the Swiss Na-tional Science Foundation.Role of the Sponsor: The sponsors had no role in thedesign and conduct of the study; in the collection,analysis, and interpretation of the data; or in the prepa-ration, review, or approval of the manuscript.Statistical Analysis: The primary statistical analysis wasperformed by statisticians Dik Heg, PhD, and Peter Juni,MD, of the Clinical Trials Unit, Bern University Hos-pital, Bern, Switzerland.The COMFORTABLE AMI Trial Investigators: Steer-ing Committee: Lorenz Raber, Peter Juni, Stephan Win-decker; Data and Safety Monitoring Board: Jona-than Sterne, Bristol, United Kingdom (chair), MichelBertrand, Lille, France; Philip Urban, Geneva, Swit-zerland; Clinical Events Committee: Pascal Vranckx,Hasselt, Belgium (chair), Gerrit Hellige, Aarau, Swit-zerland, Igal Moarof, Aarau, Switzerland; Data Man-agement and Monitoring: Clinical Trials Unit, Univer-sity of Bern, Switzerland-Brigitte Wanner, Anna Plym,Lucia Kacina, Sandro Baumgartner, Timon Sporri; Elec-tronic Data Capture: 2mT GmbH, Ulm, Germany–Jurgen Nagler-Ihlein, Thorsten Ihlmann; Core Angio-graphic Laboratory: Bern University Hospital, Bern,Switzerland.Study Sites, Collaborators: Catheterization Labora-tory, Rigshospitalet, Copenhagen, Denmark: ThomasEngstrøm, Lene Holmvang, Erik Jørgensen, Maria D.Radu, Kari Saunamaki; Department of Cardiology, Clini-cal Center of Serbia, Belgrade, Serbia: Branko Beleslin,Dejan Orlic, Jelena Kostic, Vladan Vukcevic; Bristol HeartInstitute, Bristol, United Kingdom: Sujatha Kesavan, Ju-lian Strange, Tom W. Johnson; Cardiology Depart-ment, Triemlispital, Zurich, Switzerland: Franz Eberli,David Kurz, Rainer Zbinden; Thoraxcentrum Twente,Twente University, Enschede, the Netherlands: K. Gertvan Houwelingen, Martin G. Stoel, J. Hans W. Lou-werenberg; Cardiocentro, Lugano, Switzerland: TizianoMocetti, Maria Grazia Rossi; Division of Cardiology, Uni-versity Hospital, Geneva, Switzerland: David Carballo,Francois Mach, Pierre-Frederic Keller; Cardiology De-partment, University Hospital Zurich, Zurich, Switzer-land: Roberto Corti, Roland Klingenberg, Ulf Land-messer; Rabin Medical Center, Petach Tikva, Israel, andTel Aviv University, Tel Aviv, Israel: Abid Assali, HanaVaknin-Assa; Herzzentrum Bodensee, Kreuzlingen, Swit-zerland: Michael Pieper; Cardialysis B.V., Rotterdam,the Netherlands: Hector M. Garcia Garcia; Institute ofSocial andPreventiveMedicine,UniversityofBern,Bern,Switzerland: Bindu Kalesan.Online-Only Material: The Author Video Interviewis available at http://www.jama.com.

REFERENCES

1. Grines CL, Browne KF, Marco J, et al; The PrimaryAngioplasty in Myocardial Infarction Study Group. Acomparison of immediate angioplasty with thrombo-lytic therapy for acute myocardial infarction. N Engl JMed. 1993;328(10):673-679.2. Zijlstra F, de Boer MJ, Hoorntje JCA, Reiffers S, Reiber

JHC, Suryapranata H. A comparison of immediate coro-nary angioplasty with intravenous streptokinase inacute myocardial infarction. N Engl J Med. 1993;328(10):680-684.3. Grines CL, Cox DA, Stone GW, et al; Stent Pri-mary Angioplasty in Myocardial Infarction Study Group.Coronary angioplasty with or without stent implan-tation for acute myocardial infarction. N Engl J Med.1999;341(26):1949-1956.4. Stone GW, Grines CL, Cox DA, et al; ControlledAbciximab and Device Investigation to Lower Late An-gioplasty Complications (CADILLAC) Investigators.Comparison of angioplasty with stenting, with or with-out abciximab, in acute myocardial infarction. N EnglJ Med. 2002;346(13):957-966.5. Spaulding C, Henry P, Teiger E, et al; TYPHOONInvestigators. Sirolimus-eluting versus uncoated stentsin acute myocardial infarction. N Engl J Med. 2006;355(11):1093-1104.6. Stone GW, Lansky AJ, Pocock SJ, et al; HORIZONS-AMI Trial Investigators. Paclitaxel-eluting stents ver-sus bare-metal stents in acute myocardial infarction.N Engl J Med. 2009;360(19):1946-1959.7. Brar SS, Leon MB, Stone GW, et al. Use of drug-eluting stents in acute myocardial infarction: a sys-tematic review and meta-analysis. J Am Coll Cardiol.2009;53(18):1677-1689.8. Nakazawa G, Finn AV, Joner M, et al. Delayed ar-terial healing and increased late stent thrombosis atculprit sites after drug-eluting stent placement for acutemyocardial infarction patients: an autopsy study.Circulation. 2008;118(11):1138-1145.9. Sianos G, Papafaklis MI, Daemen J, et al. Angio-graphic stent thrombosis after routine use of drug-eluting stents in ST-segment elevation myocardial in-farction: the importance of thrombus burden. J AmColl Cardiol. 2007;50(7):573-583.10. Daemen J, Wenaweser P, Tsuchida K, et al. Earlyand late coronary stent thrombosis of sirolimus-eluting and paclitaxel-eluting stents in routine clini-cal practice: data from a large two-institutional co-hort study. Lancet. 2007;369(9562):667-678.11. De Luca G, Dirksen MT, Spaulding C, et al; Drug-Eluting Stent in Primary Angioplasty (DESERT)Cooperation. Drug-eluting vs bare-metal stents in pri-mary angioplasty: a pooled patient-level meta-analysis of randomized trials. Arch Intern Med. 2012;172(8):611-621.12. Kalesan B, Pilgrim T, Heinimann K, et al. Com-parison of drug-eluting stents with bare metal stentsin patients with ST-segment elevation myocardialinfarction. Eur Heart J. 2012;33(8):977-987.13. Windecker S, Serruys PW, Wandel S, et al. Bio-limus-eluting stent with biodegradable polymer ver-sus sirolimus-eluting stent with durable polymer forcoronary revascularisation (LEADERS): a randomisednon-inferiority trial. Lancet. 2008;372(9644):1163-1173.14. Stefanini GG, Kalesan B, Serruys PW, et al. Long-term clinical outcomes of biodegradable polymer bio-limus-eluting stents versus durable polymer sirolimus-eluting stents in patients with coronary artery disease(LEADERS): 4 year follow-up of a randomised non-inferiority trial. Lancet. 2011;378(9807):1940-1948.15. Raber L, Kelbaek H, Ostoijc M, et al. Compari-son of biolimus eluted from an erodible stent coatingwith bare metal stents in acute ST-elevation myocar-dial infarction (COMFORTABLE AMI trial): rationaleand design. EuroIntervention. 2012;7(12):1435-1443.16. Magro M, Nauta S, Simsek C, et al. Value of theSYNTAX score in patients treated by primary percu-taneous coronary intervention for acute ST-elevationmyocardial infarction: The MI SYNTAXscore study. AmHeart J. 2011;161(4):771-781.

17. Serruys PW, Onuma Y, Garg S, et al. Assess-ment of the SYNTAX score in the Syntax study.EuroIntervention. 2009;5(1):50-56.18. Sabate M. A randomised comparison betweeneverolimus-eluting stents and cobalt-chromium bare-metal stents in ST-elevation myocardial infarction.Presented at European Society of Cardiology Con-gress; August 30, 2011; Paris, France.19. Vlaar PJ, Svilaas T, van der Horst IC, et al. Car-diac death and reinfarction after 1 year in the Throm-bus Aspiration during Percutaneous coronary inter-vention in Acute myocardial infarction Study (TAPAS):a 1-year follow-up study. Lancet. 2008;371(9628):1915-1920.20. Montalescot G, Wiviott SD, Braunwald E, et al;TRITON-TIMI 38 investigators. Prasugrel comparedwith clopidogrel in patients undergoing percutane-ous coronary intervention for ST-elevation myocar-dial infarction (TRITON-TIMI 38): double-blind, ran-domised controlled trial. Lancet. 2009;373(9665):723-731.21. Laarman GJ, Suttorp MJ, Dirksen MT, et al.Paclitaxel-eluting versus uncoated stents in primary per-cutaneous coronary intervention. N Engl J Med. 2006;355(11):1105-1113.22. Valgimigli M, Percoco G, Malagutti P, et al;STRATEGY Investigators. Tirofiban and sirolimus-eluting stent vs abciximab and bare-metal stent foracute myocardial infarction: a randomized trial. JAMA.2005;293(17):2109-2117.23. Valgimigli M, Campo G, Percoco G, et al; Mul-ticentre Evaluation of Single High-Dose Bolus Tirofi-ban vs Abciximab With Sirolimus-Eluting Stent or BareMetal Stent in Acute Myocardial Infarction Study(MULTISTRATEGY) Investigators. Comparison of an-gioplasty with infusion of tirofiban or abciximab andwith implantation of sirolimus-eluting or uncoatedstents for acute myocardia l infarct ion: theMULTISTRATEGY randomized trial. JAMA. 2008;299(15):1788-1799.24. Di Lorenzo E, Sauro R, Varricchio A, et al. Ben-efits of drug-eluting stents as compared to bare metalstent in ST-segment elevation myocardial infarction:four year results of the PaclitAxel or Sirolimus-Elutingstent vs bare metal stent in primary angiOplasty(PASEO) randomized trial. Am Heart J. 2009;158(4):e43-e50.25. Dıaz de la Llera LS, Ballesteros S, Nevado J, et al.Sirolimus-eluting stents compared with standard stentsin the treatment of patients with primary angioplasty.Am Heart J. 2007;154(1):164-, e1-e6.26. van der Hoeven BL, Liem SS, Jukema JW, et al.Sirolimus-eluting stents versus bare-metal stents in pa-tients with ST-segment elevation myocardial infarc-tion: 9-month angiographic and intravascular ultra-sound results and 12-month clinical outcome resultsfrom the MISSION! Intervention Study. J Am CollCardiol. 2008;51(6):618-626.27. Kolandaivelu K, Swaminathan R, Gibson WJ, et al.Stent thrombogenicity early in high-risk interven-tional settings is driven by stent design anddeployment and protected by polymer-drugcoatings. Circulation . 2011;123(13):1400-1409.28. Davì G, Patrono C. Platelet activation andatherothrombosis. N Engl J Med. 2007;357(24):2482-2494.29. Raber L, Windecker S. Current status of drug-eluting stents. Cardiovasc Ther. 2011;29(3):176-189.30. Virmani R, Burke AP, Farb A, Kolodgie FD. Pa-thology of the vulnerable plaque. J Am Coll Cardiol.2006;47(8)(Suppl):C13-C18.31. Wijns W, Kolh P, Danchin N, et al. Guidelines onmyocardial revascularization. Eur Heart J. 2010;31(20):2501-2555.





Comparison of Novel Risk Markersfor Improvement in Cardiovascular RiskAssessment in Intermediate-Risk IndividualsJoseph Yeboah, MD, MSRobyn L. McClelland, PhDTamar S. Polonsky, MD, MSCIGregory L. Burke, MD, MSChristopher T. Sibley, MDDaniel O’Leary, MDJeffery J. Carr, MD, MScDavid C. Goff Jr, MD, PhDPhilip Greenland, MDDavid M. Herrington, MD, MHS

CURRENT TRENDS IN PRIMARY

prevention of cardiovasculardisease (CVD) emphasize theneed to treat individuals based

on their global cardiovascular risk.1 Ac-cordingly, practice guidelines recom-mend approaches to classify individu-als as high, intermediate, or low riskusing the Framingham Risk Score (FRS)or other similar CVD risk predictionmodels.2,3 However, there is increasingrecognition of the imprecision of theseclassifications such that the intermediate-risk group actually represents a compos-ite of higher-risk individuals for whommore aggressive (ie, drug) therapy mightbe indicated.The intermediate-riskgroupalso contains lower-risk individuals inwhom CVD might be managed with life-style measures alone. This recognitionhas motivated researchers to identifymarkers that could offer greater discrimi-nation of higher- and lower-risk pa-tients within the intermediate-risk group.

Risk markers that have shown prom-ise in improving risk discrimination in-clude carotid intima–media thickness(CIMT), coronary artery calcium (CAC)

Author Affiliations: Departments of Internal Medicine/Cardiology (DrsYeboahandHerrington),andRadiology(DrCarr),andDivisionofPublicHealthSciences(DrsBurkeand Carr), Wake Forest University School of Medicine,Winston-Salem, North Carolina; Department of Biosta-tistics,UniversityofWashington,Seattle (DrMcClelland);SectionofCardiology,Departmentof InternalMedicine,University of Chicago, Chicago, Illinois (Dr Polonsky);National Institutes of Health, Bethesda, Maryland (Dr

Sibley); Tufts Medical Center, Brookline, Massachusetts(Dr O’Leary); University of Colorado School of PublicHealth, Aurora (Dr Goff ); and Department of Preven-tiveMedicine,NorthwesternUniversityFeinbergSchoolof Medicine, Chicago, Illinois (Dr Greenland).Corresponding Author: Joseph Yeboah, MD, MS, De-partment of Internal Medicine/Cardiology, Wake For-est Baptist Health, Medical Center Boulevard, Winston-Salem, NC 27157 ([email protected]).

Context Risk markers including coronary artery calcium, carotid intima–media thick-ness, ankle-brachial index, brachial flow–mediated dilation, high-sensitivity C-reactive protein (CRP), and family history of coronary heart disease (CHD) have beenreported to improve on the Framingham Risk Score (FRS) for prediction of CHD, butthere are no direct comparisons of these markers for risk prediction in a single cohort.

Objective We compared improvement in prediction of incident CHD/cardiovascular disease (CVD) of these 6 risk markers within intermediate-risk partici-pants (FRS �5%-�20%) in the Multi-Ethnic Study of Atherosclerosis (MESA).

Design, Setting, and Participants Of 6814 MESA participants from 6 US fieldcenters, 1330 were intermediate risk, without diabetes mellitus, and had complete dataon all 6 markers. Recruitment spanned July 2000 to September 2002, with follow-upthrough May 2011. Probability-weighted Cox proportional hazard models were usedto estimate hazard ratios (HRs). Area under the receiver operator characteristic curve(AUC) and net reclassification improvement were used to compare incremental con-tributions of each marker when added to the FRS, plus race/ethnicity.

Main Outcome Measures Incident CHD defined as myocardial infarction, anginafollowed by revascularization, resuscitated cardiac arrest, or CHD death. Incident CVDadditionally included stroke or CVD death.

Results After 7.6-year median follow-up (IQR, 7.3-7.8), 94 CHD and 123 CVD eventsoccurred. Coronary artery calcium, ankle-brachial index, high-sensitivity CRP, and fam-ily history were independently associated with incident CHD in multivariable analyses(HR, 2.60 [95% CI, 1.94-3.50]; HR, 0.79 [95% CI, 0.66-0.95]; HR, 1.28 [95% CI,1.00-1.64]; and HR, 2.18 [95% CI, 1.38-3.42], respectively). Carotid intima–mediathickness and brachial flow–mediated dilation were not associated with incident CHDin multivariable analyses (HR, 1.17 [95% CI, 0.95-1.45] and HR, 0.95 [95% CI, 0.78-1.14]). Although addition of the markers individually to the FRS plus race/ethnicityimproved AUC, coronary artery calcium afforded the highest increment (0.623 vs 0.784),while brachial flow–mediated dilation had the least (0.623 vs 0.639). For incident CHD,the net reclassification improvement with coronary artery calcium was 0.659, brachialflow–mediated dilation was 0.024, ankle-brachial index was 0.036, carotid intima–media thickness was 0.102, family history was 0.160 and high-sensitivity CRP was 0.079.Similar results were obtained for incident CVD.

Conclusions Coronary artery calcium, ankle-brachial index, high-sensitivity CRP, andfamily history were independent predictors of incident CHD/CVD in intermediate-riskindividuals. Coronary artery calcium provided superior discrimination and risk reclas-sification compared with other risk markers.JAMA. 2012;308(8):788-795 www.jama.com




scores, brachial flow–mediated dilation(FMD),ankle-brachial index(ABI),high-sensitivityC-reactiveprotein(CRP),andfamily history of coronary heart disease(CHD).4-9 A recent American College ofCardiologyFoundation/AmericanHeartAssociation statement on the use ofmarkers to improve cardiovascular riskprediction beyond the FRS gave fam-ily history a class I recommendation;CIMT, CAC, ABI, and high-sensitivityCRPreceivedclass II recommendations,whiletheAmericanCollegeofCardiologyFoundation/AmericanHeartAssociationrecommendedagainst theuseofbrachialFMD(class III).10 However, theserecom-mendations were limited by the relativepaucityofpublisheddataandthefact thatthe published studies of individual riskmarkers were performed in different co-hortswithdifferentcompositeoutcomes,analytic methodologies, and inadequatestatisticalpower todetect improvementsofriskpredictionbeyondcommonlyusedrisk-prediction algorithms. Moreover,therearenocomprehensivehead-to-headcomparisons of these risk markers in asinglepopulationcohortsimilartotheUSpopulation.Determiningtherelative im-provementsinpredictionaffordedbyvari-ousriskmarkers,especiallywhenconsid-ering intermediate-risk individuals asclassifiedbytheFRS,couldhelpdeterminethemostefficient strategy for identifyingselect, intermediate-riskparticipants formore aggressive primary prevention in-terventions, including the use of aspirinand lower targets for drug treatments oflow-densitylipoprotein(LDL)cholesteroland blood pressure.

In this report, we assess the improve-ments in prediction accuracy and reclas-sification to high- and low-risk categoriesusing CIMT, CAC, FMD, ABI, high-sensitivity CRP, and family history ofCHD in asymptomatic adults classifiedas intermediate risk (FRS standards) whoparticipated in the Multi-Ethnic Study ofAtherosclerosis (MESA).

METHODSStudy Populationand Data Collection

The study design for the MESA studyhas been published elsewhere.11 In brief,

MESA is a prospective cohort study toinvestigate the prevalence, correlates,and progression of subclinical CVD inindividuals without known CVD atbaseline.

The full cohort includes 6814women and men aged 45 to 84 yearswithout known CVD, recruited from6 US communities (Baltimore, Mary-land; Chicago, Illinois; ForsythCounty, North Carolina; Los AngelesCounty, California; northern NewYork City, New York; and St Paul,Minnesota). Self-reported race/ethnicity was collected to explore thepossible racial/ethnic differences inthe development and progression ofatherosclerosis. The race/ethnicbreakdown of MESA participants was38% white, 28% black, 22% His-panic, and 12% Chinese adults. Par-ticipants with diabetes were excludedbecause it is considered to be a CHDrisk equivalent. Diabetes was definedas self-reported history of diabetesmellitus, diabetes medication use, ora fasting glucose level of 126 mg/dLor greater (for mmol/L, multiply by0.0555). Demographics, medical his-tory, and anthropometric and labora-tory data for the present study weretaken from the first examination(July 2000 to August 2002). Currentsmoking was defined as havingsmoked a cigarette in the last 30days. Use of antihypertensive andother medications was based onreview of prescribed medication con-tainers. Resting blood pressure wasmeasured 3 times in the seated posi-tion and the average of the secondand third readings was recorded.Hypertension was defined as a sys-tolic blood pressure of at least 140mm Hg, diastolic blood pressure of atleast 90 mm Hg, or use of medicationprescribed for hypertension. Bodymass index was calculated as weightin kilograms divided by height inmeters squared. Total cholesterol andhigh-density lipoprotein (HDL) cho-lesterol were measured from bloodsamples obtained after a 12-hour fast.LDL cholesterol was estimated usingthe Friedewald equation.12 High-

sensitivity CRP was measured usingthe BNII nephelometer (N High Sen-sitivity CRP; Dade Behring Inc) atthe Laboratory for Clinical Biochem-istry Research (University of Ver-mont , Bur l ington) . Ana ly t ica lintra-assay coefficient of variationsranged from 2.3% to 4.4%, and inter-assay coefficient of variation rangedfrom 2.1% to 5.7% with a detectionlevel of 0.18 mg/L. Family history ofCHD was obtained by asking partici-pants whether any member in theirimmediate family (parents, siblings,and children) experienced fatal ornonfatal myocardial infarction. TheMESA study was approved by theinstitutional review boards of eachstudy site and written informed con-sent was obtained from all partici-pants.

Measurement of ABI

Protocol for ankle-brachial index mea-surement in MESA was reported byCriqui et al.9 Briefly, systolic blood pres-sure measurements in the bilateral bra-chial, dorsalis pedis, and posterior tibialarteries were obtained in the supine po-sition using a handheld Doppler in-strument with a 5-mHz probe. To avoidpotential bias from subclavian steno-sis, the higher of the brachial arterypressures was used as the denomina-tor. For each lower extremity, the ABInumerator used was the highest pres-sure (dorsalis pedis or posterior tibial)from that leg. Reproducibility of ABIwas evaluated using measurements of43 participants by 2 technicians. Theinterreader and intrareader correla-tion coefficients were 0.845 and 0.937,with the intrareader and interreader co-efficients of variation being 5.14% and3.27%.

Measurement of the CAC Score

CT scanning and interpretation meth-ods in MESA have were reported byCarr et al.13 Scanning centers assessedCAC by chest computed tomography(CT) with either a cardiac-gated elec-tron-beam CT scanner (Chicago, LosAngeles County, and New York Cityfield centers) or a multidetector CT sys-

CARDIOVASCULAR RISK ASSESSMENT IN INTERMEDIATE-RISK INDIVIDUALS



tem (Baltimore, Forsyth County, andSt Paul field centers). Certified tech-nologists scanned all participants twiceover phantoms of known physical cal-cium concentration. A radiologist or

cardiologist read all CT scans at a cen-tral reading center (Los Angeles Bio-medical Research Institute at Harbor–UCLA, Torrance, California). We usedthe mean Agatston score for the 2 scansin all analyses14 Intraobserver and in-terobserver agreements were excel-lent (�=0.93 and �=0.90).

Measurement of Brachial FMD

Methods for measuring and interpret-ing brachial FMD in MESA were re-ported by Yeboah et al.8 Intrareader re-producibility for baseline diameter,maximum diameter, and percent FMDwas evaluated by comparing an origi-nal and a blinded quality control re-read of ultrasounds from 40 MESA par-ticipants. The intraclass correlationcoefficients were 0.99, 0.99, and 0.93,respectively. Intraparticipant variabil-ity was evaluated by comparing re-sults from repeated examinations of 19participants on 2 days, 1 week apart.The intraclass correlation coefficientsfor baseline diameter, maximum diam-eter, and percent FMD were 0.90, 0.90,and 0.54, respectively. Percent techni-cal error of measurement was 1.39% forbaseline diameter measurement, 1.47%for maximum diameter measurement,and 28.4% for percent FMD measure-ment.

Measurement of CIMT

Methods for measuring and interpret-ing CIMT were reported by Polak et al.15

The mean of maximum intima-mediathickness of the common carotid ar-tery was used. Reproducibility was as-sessed by blinded replicate readings ofCIMT performed by 2 readers. Onereader reread 66 studies for a between-reader correlation coefficient of 0.84(n=66), and a second reader reread 48studies for a correlation coefficient of0.86. The rescan and the reread coef-ficients of variation were 7.07% and3.48%.

Ascertainment of Incident CHDand CVD

Follow-up took place through May2011. CVD events were adjudicated bya MESA study committee that in-

cluded cardiologists, physician epide-miologists, and neurologists. A de-tailed description of the adjudicationprocess was previously reported.8 Forthe purposes of this study, we defineincident CHD as myocardial infarc-tion, CHD death, resuscitated cardiacarrest, or definite or probable angina iffollowed by coronary revasculariza-tion. Incident CVD additionally in-cluded stroke or CVD death as de-fined by the MESA protocol (availableat http://www.mesa-nhlbi.org). ThusCHD is a subset of CVD.


The study population was limited toMESA participants classified as inter-mediate risk (estimated 10-y CHD riskof �5%-�20%) based on the Framing-ham risk equation.3 The intermediate-risk range was chosen to make our re-sults comparable to other studies thathave reported data on intermediate-risk participants using some of the riskmarkers under consideration.4-6 De-scriptive data are presented as mean(SD) for continuous variables or fre-quencies of participants for categori-cal variables. CAC scores were ex-pressed as In (CAC�1). Family historywas entered into models as a categori-cal variable (yes/no), high-sensitivityCRP had a highly skewed distributionand was log transformed; all other vari-ables were expressed as continuousvariables. Weighted analyses were doneto reflect the sampling from the over-a l l MESA cohort . Probabi l i ty -weighted Cox proportional hazardanalysis with robust variance esti-mates was used to assess the associa-tion between each of the markers (CAC,FMD, CIMT, ABI, high-sensitivity CRP,and family history of CHD) and inci-dent CHD or CVD in univariable andmultivariable models while adjustingfor age, sex, race/ethnicity, total andHDL cholesterol, cigarette smoking sta-tus, body mass index, blood pressuremedication use, and hydroxmethyl glu-taryl coenzyme A–reductase inhibitoruse. These confounders were chosenbased on their association with the out-comes of interest (incident CHD/

Table 1. Baseline Characteristics ofParticipants in the Multi-Ethnic Study ofAtherosclerosis at Intermediate FraminghamRisk (N = 1330)a

Variables Mean (SD)

Age, y 63.8 (9.5)

Women, No. (%) 443 (33.3)

Race/ethnicity, No. (%)White 475 (35.7)

Chinese 225 (16.9)

Black 292 (22.0)

Hispanic 338 (25.4)

Body mass indexb 27.9 (4.7)

Cigarette smokingstatus, No. (%)

Never 616 (46.3)

Former 494 (37.1)

Current 220 (16.5)

Cholesterol, mg/dLTotal 196.9 (34.6)

LDL 122.4 (30.3)

HDL 46.5 (11.9)

Triglycerides, mg/dL 140.7 (77.0)

Blood pressure, mm HgSystolic 129.9 (19.8)

Diastolic 74.4 (9.8)

Heart rate, beats/min 61.9 (9.2)

Statin use, No. (%) 187 (14.1)

Blood pressuremedication use, No. (%)

508 (38.2)

Coronary calcium score,median (IQR)c

7.0 (0-111.7)

Brachial flow–mediateddilation, median (IQR), %

3.60 (2.1-5.4)

Carotid intima–mediathickness, median (IQR),mm

0.86 (0.76-0.98)

Ankle-brachial index,median (IQR)

1.14 (1.07-1.20)

High-sensitivity CRP,median (IQR), mg/L

1.62 (0.79-3.68)

Family history ofpremature CHD, No. (%)

567 (42.6)

Framingham Risk Score,median (IQR), %

8.8 (6.5-12.2)

Abbreviations: CHD, coronary heart disease; CRP,C-reactive protein; HDL, high-density lipoprotein; IQR,interquartile range; LDL, low-density lipoprotein.

SI conversion factors: To convert CRP from mg/L to nmol/L,multiply by 9.524; HDL, LDL, and total cholesterol frommg/dL to mmol/L, multiply by 0.0259; triglyceridesfrom mg/dL to mmol/L, multiply by 0.0113.

aData are reported as mean (SD) unless otherwise indi-cated.

bBody mass index is calculated as weight in kilograms di-vided by height in meters squared.

c Indicates Agatston.




CVD) in the current analysis and priorpublished data.

We assessed the improvement in dis-crimination by comparing the area un-der the receiver operator characteristiccurves (AUC) in models with and with-out each novel risk marker, using themethod of DeLong et al.16 The FRS (de-rived using age, sex, total cholesterol,HDL cholesterol, smoking status, sys-tolic blood pressure, and blood pres-sure medication use), a general clinicalpractice tool, plus race/ethnicity, servedas the baseline model. ROC curves weredeveloped using a probabil i ty-weighted Cox model. We assessed theclassification of risk using the net re-classification improvement (NRI):

NRI=[Prob (being correctly reclassifiedto a higher-risk category/event)

−Prob (being incorrectly reclassified toa lower-risk category/event)]

�[Prob (being correctly reclassified toa lower-risk category/nonevent)

−Prob (being incorrectly classified to ahigher-risk category/nonevent)].17

The NRI captures the relative im-provement in classification associatedwith the additional predictive vari-able, while explicitly balancing trade offbetween changes in sensitivity andspecificity. NRI is the sum of 2 percent-ages with different denominators andhence, is reported as a proportion (pos-sible range, −2.0 to 2.0). At the time ofthese analyses, the mean observed fol-low-up in MESA was 7.5 years (maxi-mum follow-up, 9 y). To account forthe fact that actual follow-up was lessthan 10 years, we redefined the risk interms of 7.5-year risk when calculat-ing the NRI, using a logistic regres-sion model with probability weight-ing to reflect the sampling from theoverall cohort. Based on the new model,intermediate 7.5-year risk categories forCHD and CVD were defined as 2.0% to15.4% and 3.4% to 21.1%. With the ad-dition of each novel risk marker to thebase model, participants were consid-ered to be reclassified to high risk if theirestimated risks for CHD and CVD were

greater than 15.4% and 21.1%, and re-classified to low risk if their estimatedrisks were lower than 2.0% and 3.4%for CHD and CVD. As a sensitivityanalysis, we repeated our evaluation ofABI and the imaging markers using theReynolds score (RS, calculated sepa-rately for men and women)6,18 insteadof the FRS to define these risk groups.This score incorporates family historyand log-transformed high-sensitivityCRP in addition to other risk factors.A 2-tailed P value of less than .05 wasconsidered significant. All statisticalanalyses were performed using SAS ver-sion 9.2 (SAS Institute).

RESULTSStudy Cohort

The final study population included1330 participants without diabetes melli-tus, with an FRS of more than 5% to lessthan 20%, and with complete data on all6 of the novel risk markers. The me-dian (IQR) for the FRS of the cohort atbaseline was 8.8% (6.5%-12.2%). Thebaseline characteristics of the studypopulation are shown in TABLE 1.

After a median follow-up of 7.6 years(IQR, 7.3-7.8; maximum follow-up, 9y), 94 participants (7.1%) experi-enced a CHD event and 123 (9.2%) ex-perienced a CVD event. Specific eventswere: 43 with myocardial infarction, 3with resuscitated cardiac arrest, 14 withCHD death, 44 with angina followed by

revascularization, and 31 participantswith stroke.

Association of Risk MarkersWith Incident CHD and CVD

In the univariable probabil i ty-weighted Cox proportional hazardanalyses, each of the novel risk mark-ers was associated with incident CHD;however, after adjusting for confound-ers, the associations with CIMT andFMD were no longer significant(TABLE 2). Among all of the risk mark-ers, CAC had the strongest association.Similarly, for incident CVD in univari-able analyses, each of the novel riskmarkers was associated with events ex-cept high-sensitivity CRP. However, af-ter adjusting for confounders, the asso-ciations between CIMT and FMD wereno longer significant (eTable 1). CACalso had the strongest association in themultivariable models for CVD.

Improvement of Discriminationby Addition of Novel Risk Markersto the FRS

For CHD/CVD events, the addition ofeach of the 6 risk markers to the base-line model improved the AUC. CACshowed the highest increment whileFMD showed the least increment(FIGURE) for incident CHD. CAC alsoshowed the highest increment whilehigh-sensitivity CRP showed the leastincrement for incident CVD (Figure).

Table 2. Association of Several Novel Risk Markers With Incident Coronary Heart Disease(No. of Events = 94)a

Marker

Univariable Multivariableb

HR (95% CI)c P Value HR (95%CI)c P Value

ABI 0.78 (0.66-0.93) .005 0.79 (0.66-0.95) .01

Brachial FMD 0.82 (0.66-1.03) .09 0.93 (0.74-1.16) .52

CACd 2.72 (2.09-3.55) �.001 2.60 (1.94-3.50) �.001

Carotid IMT 1.33 (1.12-1.59) .001 1.17 (0.95-1.45) .13

Family history 2.39 (1.54-3.70) �.001 2.18 (1.38-3.42) .001

High-sensitivity CRPd 1.26 (1.01-1.57) .05 1.28 (1.00-1.64) .05Abbreviations: ABI, ankle-brachial index; CAC, coronary calcium score; CRP, C-reactive protein; FMD, flow-mediated

dilation; IMT, intima-media thickness; HR, hazard ratio.aThere were 43 myocardial infarctions, 3 resuscitated cardiac arrests, 44 cases of angina followed by revascularization,

and 14 deaths due to coronary heart disease that occurred during follow-up. An individual may have experienced morethan 1 event but only the first event was included for the composite number of events used in this analysis.

bMultivariable models were adjusted for age, sex, race/ethnicity, systolic blood pressure, total cholesterol, high-densitylipoprotein cholesterol, smoking status, body mass index (calculated as weight in kilograms divided by height in meterssquared), blood pressure medication use, and statin use.

cFor continuous risk markers, HRs are standardized per unit SD change in the marker.dCAC was expressed as In (CAC�1) and high-sensitivity CRP was log transformed.




Classification of RiskUsing TABLE 3 (FRS�CAC variable) asan example, 51.1% of participants whohad CHD (events) and 54.9% of thosewho did not have CHD (nonevents)during follow-up period were reclassi-fied either to low or high risk by the ad-dition of CAC to the FRS(�race/ethnicity). Applying the NRI formulathat considers both those correctly re-classified as well as those incorrectly re-classified, a net 25.5% of the eventsgroup were reclassified to high risk ap-propriately, while a net 40.4% of thenonevents group were appropriately re-classified into the low-risk group by theaddition of CAC to FRS(�race/ethnicity). The NRI for the addition ofCAC to FRS (�race/ethnicity) is there-fore calculated by adding 0.255 to 0.404(0.255�0.405=0.659).

The addition of CAC to the FRS(�race/ethnicity) resulted in the high-est NRI (0.659) and the greatest abso-lute number of correctly reclassifiedparticipants (n=625) (Table 3), whilethe addition of FMD resulted in the low-est NRI (0.024) and the fewest total

number of correctly reclassified indi-viduals. Carotid IMT, ABI, CRP, andfamily history afforded modest NRIs forCHD events (Table 3). CAC also pro-vided the greatest NRI and total num-ber of correctly reclassified partici-pants for CVD events (Figure andTABLE 4). Among the non-CAC riskmarkers, family history performed thebest for CHD risk reclassification(NRI=0.160), while ABI performed thebest for CVD risk reclassification(NRI=0.068).

The respective AUCs were 0.642 forthe RS alone, 0.766 for RS plus CAC,0.643 for RS plus CIMT, 0.648 for RSplus ABI, and 0.642 for RS plus FMDfor incident CHD in the present co-hort. The NRIs were 0.528 for RS plusCAC, 0.003 for RS plus CIMT, 0.002for RS plus ABI, and 0 for RS plus FMDover RS alone for incident CHD. Simi-larly the AUCs were 0.645 for FRSalone, 0.742 for RS plus CAC, 0.645 forRS plus CIMT, 0.656 for RS plus ABI,and 0.646 fpr RS plus FMD for inci-dent CVD. The NRIs were 0.415 for RSplus CAC, 0 for RS plus CIMT, 0.008

for RS plus ABI, and 0.007 for RS plusFMD over RS alone for incident CVD.

COMMENTThe current study shows that among 6of the most promising novel risk mark-ers, CAC provides the highest improve-ment in discrimination over the FRSand RS in individuals classified as in-termediate risk. The present study pro-vides additional support for the use ofCAC as a tool for refining cardiovas-cular risk prediction in individuals clas-sified as intermediate risk by the FRSor the RS. To our knowledge, this is thefirst study to compare directly the im-provement in risk prediction pro-vided by the novel markers included inthe present study in a multiethnic co-hort with intermediate Framingham orReynolds risk.

Previous studies showed that CIMT,CAC, brachial FMD, ABI, high-sensitivity CRP and family history ofCHD improve the classification of riskover the FRS, but to varying degrees.Direct comparisons between studiesshould be made with caution because

Figure. Receiver Operator Characteristic Curves Showing Area Under the Curve for Incident Coronary Heart Disease and IncidentCardiovascular Disease in Intermediate-Risk MESA Participants

1.0

0.8

0.6

0.4

0.2

0

0 0.2

1.0 0.8 0.6 0.20.4 0.0

0.4 0.80.6 1.01–Specificity

Specificity

Sen

sitiv

ity

1.0

0.8

0.6

0.4

0.2

0

0 0.2

1.0 0.8 0.6 0.20.4 0.0

0.4 0.80.6 1.01–Specificity

Specificity

Sen

sitiv

ity

Incident coronary heart diseaseA Incident cardiovascular diseaseB

FRS plus brachial flow-mediateddilation

Framingham Risk Score (FRS)alone (reference)FRS plus coronary arterycalciumFRS plus carotid intima-mediathickness

FRS plus C-reactive proteinFRS plus family historyFRS plus ankle-brachial index

A, Receiver operator characteristic curves showing area under the curve for FRS alone, 0.623; FRS plus coronary artery calcium, 0.784 (P�.001); FRS plus intima-mediathickness, 0.652 (P=.01); FRS plus flow-mediated dilation, 0.639 (P=.06); FRS plus high-sensitivity C-reactive protein, 0.640 (P=.03); FRS plus family history, 0.675(P=.001); and FRS plus ankle-brachial index, 0.650 (P=.01). B, Receiver operator characteristic curves showing area under the curve for FRS alone, 0.623; FRS pluscoronary artery calcium, 0.784 (P�.001); FRS plus intima-media thickness, 0.652 (P=.01); FRS plus flow-mediated dilation, 0.639 (P=.06); FRS plus high-sensitivityC-reactive protein, 0.640 (P=.03); FRS plus family history, 0.675 (P=.001); and FRS plus ankle-brachial index, 0.650 (P=.01). MESA indicates Multi-Ethnic Study ofAtherosclerosis.




they were conducted in different co-horts, did not have uniform defini-tions of the primary outcome, and hadvarying duration of follow-up. Nambiet al4 showed in the Atherosclerosis Risk

In Communities study that CIMT im-proved the AUC from 0.742 to 0.750,and had a net clinical NRI of 0.167. Po-lonsky et al5 showed in a larger subsetof the MESA study that CAC is an in-

dependent predictor of CHD, im-proved the AUC from 0.76 to 0.81, andhas a net clinical NRI of 0.55 in the in-termediate-risk stratum. Yeboah et al8

showed that brachial FMD is an inde-

Table 3. Net Reclassification Improvement (NRI) for Incident Coronary Heart Disease Events With Addition of Novel Risk Markers to theFramingham Risk Score in Intermediate-Risk MESA Participants (N = 1330)

Variable%

Reclassified

Risk Category, No. of EventsFRS Events (n = 94)

FRS Nonevents (n = 1236)% Net Correct

Reclassification NRILow Intermediate High

FRS plus carotid IMT .102Events 7.4 0 87 7 7.4

Nonevents 5.3 50 1170 16 2.8

FRS plus CAC .659Events 51.1 12 46 36 25.5 .

Nonevents 54.9 589 557 90 40.4

FRS plus brachial FMD .024Events 0.0 0 94 0 0

Nonevents 3.2 35 1196 5 2.4

FRS plus ABI .036Events 4.3 1 90 3 2.1

Nonevents 4.0 34 1186 16 1.5

FRS plus high-sensitivity CRP .079Events 4.3 0 90 4 4.3

Nonevents 5.2 54 1172 10 3.6

FRS plus family history .160Events 8.5 0 86 8 8.5

Nonevents 11.2 116 1097 23 7.5Abbreviations: ABI, ankle-brachial index; CAC, coronary calcium score; CRP, C-reactive protein; FMD, flow-mediated dilation; FRS, Framingham Risk Score; IMT, intima-media

thickness; MESA, Multi-Ethnic Study of Atherosclerosis.

Table 4. Net Reclassification Improvement (NRI) for Incident Cardiovascular Disease Events With Addition of Novel Risk Markers to theFramingham Risk Score in Intermediate-Risk MESA Participants (N = 1330)

Variable%

Reclassified

Risk Category, No. of EventsFRS Events (n = 123)

FRS Nonevents (n = 1207)% Net Correct

Reclassification NRILow Intermediate High

FRS plus carotid IMT .060Events 3.3 0 119 4 3.3

Nonevents 3.8 39 1161 7 2.7

FRS plus CAC .466Events 36.6 16 78 29 10.6

Nonevents 45.7 493 655 59 36.0

FRS plus brachial FMD .023Events 2.4 3 120 0 −2.4

Nonevents 5.6 62 1140 5 4.7

FRS plus ABIEvents 4.1 0 118 5 4.1 .068

Nonevents 4.6 44 1151 12 2.7

FRS plus high-sensitivity CRP .037Events 1.6 0 121 2 1.6

Nonevents 3.2 32 1168 7 2.1

FRS plus family history .040Events 2.4 1 120 2 0.8

Nonevents 4.9 49 1148 10 3.2Abbreviations: ABI, ankle-brachial index; CAC, coronary calcium score; CRP, C-reactive protein; FMD, flow-mediated dilation; FRS, Framingham Risk Score, IMT, intima-media

thickness; MESA, Multi-Ethnic Study of Atherosclerosis.




pendent predictor of incident CVD, andthat it did not improve the AUC of theFRS (0.74) but has a net clinical NRI,defined using the 10% to 20% Framing-ham risk, of 0.28. Fowkes et al19 showedin a meta-analysis that ABI is an inde-pendent predictor of incident CVD, andimproves the AUC from 0.646 to 0.655and reclassification of the risk cat-egory and modification of treatmentrecommendations in approximately19% of men and 36% of women. Wil-son et al20 showed in the FraminghamHeart Study that high-sensitivity CRPis an independent predictor of inci-dent CHD/CVD, it improved AUC from0.795 to 0.865 and to 0.799, and it hadan NRI of 0.118 and 0.056, respec-tively. Sivapalaratnam et al7 showed inthe EPIC-Norfolk study that family his-tory of CHD is an independent predic-tor of incident CHD with a net clinicalNRI of 0.021.

Recently, investigators from the Rot-terdam study also performed a directcomparison of several novel risk mark-ers, and also found that CAC pro-vided the most robust improvement inrisk prediction. Kavousi et al21 com-pared the N-terminal fragment of pro-hormone B-type natriuretic peptide,Von Willebrand factor antigen, fibrino-gen levels, homocysteine levels, uricacid levels, high-sensitivity CRP, leu-kocyte count, chronic kidney disease,CAC, CIMT, peripheral artery dis-ease, and pulse wave velocity to the FRSusing a similar definition of CHD andstatistical approach to the current study.The authors found that CAC providedthe highest increment in AUC and NRIover the FRS. It is noteworthy that theRotterdam investigators found similarresults to the current study, despite dif-ferences in the 2 study populations. Allof the Rotterdam Study participantswere white and approximately 13% ofthe cohort had diabetes mellitus.

Even though our study indicates con-siderable superiority of CAC over sev-eral risk markers for risk prediction ofCHD and CVD, several other factorsshould be considered before makingbroad recommendations about incor-poration of CAC into primary preven-

tion screening strategies. One notableconcern is that measurement of CACexposes individuals to a small but non-trivial amount of ionizing radiation (ap-proximately 0.9-1.1 mSv). Recent ef-forts have been made to standardizeequipment and imaging protocols to re-duce radiation exposure during CACimaging22; however, the extent to whichthese recommendations have beenimplemented in general medical prac-tice is not known. Previous studies sug-gest wide variations in radiation doseduring CAC imaging by region/type ofinstitution/protocol.23 Even with thelowest possible radiation dose, there re-mains uncertainty about the magni-tude of long-term cancer risks.24 A smallrisk associated with the lowest pos-sible radiation dose during CACimaging could translate into a largenumber of avoidable cancers if CACwere to be uniformly applied to the es-timated 23 million people in the UnitedStates currently classified as interme-diate risk by the FRS.25 Similarly, thebenefits and risks associated with in-cidental findings detected during CACimaging remain unclear. These indi-rect costs, in addition to the direct fi-nancial costs of CAC imaging, need tobe weighed against the presumed ben-efits from better discrimination of par-ticipants at high risk for CHD and CVDevents to best determine the role of CACscreening of patients with an interme-diate risk for a CHD/CVD event. Thus,the ultimate decision regarding the op-timum test to order should not be basedsolely on improvement in risk predic-tion afforded by a test but also cost ef-fectiveness, acceptability to patients,and the potential risk and benefits as-sociated with the test.26

The current study has limitations.We limited our analysis to the subsetof MESA participants with completedata on all 6 risk markers, which de-creased our sample size. Nevertheless,there were sufficient numbers of eventsto demonstrate clearly the superiorityof CAC over the other measures inhead-to-head ROC and NRI analyses.Finally, in MESA we did not specifi-cally define family history of CHD as

premature (ie, before the age of 55 formen and 65 for women). This may haveinfluenced the association of family his-tory with CHD and CVD.

CONCLUSIONSCAC, ABI, high-sensitivity CRP, andfamily history were independent pre-dictors of incident CHD/CVD beyondtraditional risk factors, but had vary-ing degrees of improvement in discrimi-nation and classification of risk withinintermediate-risk individuals. CAC hadthe highest improvement in both AUCand NRI when added to the FRS/RS.Additional research is warranted toexplore further both the costs and ben-efits of CAC screening in intermediate-risk individuals.

Author Contributions: Dr Yeboah had full access toall of the data in the study and takes responsibility forthe integrity of the data and the accuracy of the dataanalysis.Study concept and design: Yeboah, Burke, O’Leary,Carr, Herrington.Acquisition of data: Burke, O’Leary, Carr, Greenland.Analysis and interpretation of data: Yeboah,McClelland, Polonsky, Sibley, Carr, Goff, Greenland,Herrington.Drafting of the manuscript: Yeboah.Critical revision of the manuscript for important in-tellectual content: Yeboah, McClelland, Polonsky,Burke, Sibley, O’Leary, Carr, Goff, Greenland,Herrington.Statistical analysis: Yeboah, McClelland, Herrington.Obtained funding: Burke, O’Leary, Carr, Greenland.Administrative, technical, or material support:Polonsky, Burke, O’Leary, Carr.Study supervision: O’Leary, Carr, Greenland,Herrington.Conflict of Interest Disclosures: Dr McClelland re-ports receipt of an institutional grant or pending grantfrom the National Institutes of Health (NIH) for con-tracts that fund the Multi-Ethnic Study of Atheroscle-rosis (MESA) study. Dr Burke reports receipt of an in-stitutional grant from the National Heart, Lung, andBlood Institute (NHLBI). Dr O’Leary reports receipt ofan institutional grant and support (to the institution)for travel to meetings for the study or other purposesfrom NIH. Dr Carr reports receiptof an institutionalgrant from NIH/NHLBI and receipt of consultancy feesfrom Merck & Co. Dr Goff reports receipt of an insti-tutional grant and support (to the institution) for travelto meetings for the study or other purposes fromNHLBI, board membership on the data and safetymonitoring committee for a clinical trial from Takeda,and receipt of consultancy (operations committeemember for a clinical trial) fees from Merck & Co. DrGreenland reports receipt of an institutional grant andsupport (to the institution) for travel to meetings forthe study or other purposes) from NHLBI, receipt ofa grant or pending grant from NIH, and payment forlectures including service on speakers bureaus fromuniveristies and NIH, Dr Herrington reports receipt ofan institutional grant from NIH. Drs Yeboah, Polon-sky, and Sibley report no disclosures. All authors havecompleted and submitted the ICMJE Form for Disclo-sure of Potential Conflicts of Interest.Funding/Support: This research was supported by con-tracts N01-HC-95159. 60, 61, 62, 63, 64, 65, 66, and




N01-HC-95167 and Divers i ty SupplementR01HL098445 (primary investigator, Dr Carr). A fulllist of participating MESA investigators and institu-tions can be found at http://www.mesa-nhlbi.org.Role of the Sponsor: The NHLBI participated in thedesign and conduct of MESA. A member of the NHLBIstaff served as a coauthor and had input into the col-lection, management, analysis, and interpretation ofthe data and in preparation of the manuscript, as didthe other coauthors. Although additional members ofthe NHLBI staff were able to view the manuscript priorto submission, they did not participate in the decisionto submit the manuscript or approve it prior to pub-lication. The National Center for Research Resourceshad no role in the design and conduct of the study,in the collection, analysis, and interpretation of the data,or in the preparation, review, or approval of the manu-script.Online-Only Material: The eTable is available at http://www.jama.com.Additional Contributions: We thank the investiga-tors, the staff, and the participants of MESA for theirvaluable contributions.

REFERENCES

1. Hayward RA, Krumholz HM, Zulman DM, TimbieJW, Vijan S. Optimizing statin treatment for primaryprevention of coronary artery disease. Ann Intern Med.2010;152(2):69-77.2. Collins GS, Altman DG. An independent externalvalidation and evaluation of QRISK cardiovascular riskprediction: a prospective open cohort study. BMJ.2009;339:b2584.3. Expert Panel on Detection, Evaluation, and Treat-ment of High Blood Cholesterol in Adults. ExecutiveSummary of The Third Report of The National Cho-lesterol Education Program (NCEP) Expert Panel onDetection, Evaluation, And Treatment of High BloodCholesterol In Adults (Adult Treatment Panel III).JAMA. 2001;285(19):2486-2497.4. Nambi V, Chambless L, Folsom AR, et al. Carotidintima-media thickness and presence or absence ofplaque improves prediction of coronary heart diseaserisk: the ARIC study. J Am Coll Cardiol. 2010;55:1600-1607.5. Polonsky TS, McClelland RL, Jorgensen NW, et al.Coronary artery calcium score and risk classificationfor coronary heart disease prediction. JAMA. 2010;303(16):1610-1616.6. Ridker PM, Paynter NP, Rifai N, Gaziano JM, CookNR. C-reactive protein and parental history improveglobal cardiovascular risk prediction: the Reynolds Risk

Score for men. Circulation. 2008;118(22):2243-2251, 4p, 2251.7. Sivapalaratnam S, Boekholdt SM, Trip MD, et al.Family history of premature coronary heart disease andrisk prediction in the EPIC-Norfolk prospective popu-lation study. Heart. 2010;96(24):1985-1989.8. Yeboah J, Folsom AR, Burke GL, et al. Predictivevalue of brachial flow-mediated dilation for incidentcardiovascular events in a population-based study: themulti-ethnic study of atherosclerosis. Circulation. 2009;120(6):502-509.9. Criqui MH, McClelland RL, McDermott MM, et al.The ankle-brachial index and incident cardiovascularevents in the MESA (Mult i-Ethnic Study ofAtherosclerosis). J Am Coll Cardiol. 2010;56(18):1506-1512.10. Greenland P, Alpert JS, Beller GA, et al; Ameri-can College of Cardiology Foundation/American HeartAssociation Task Force on Practice Guidelines. 2010ACCF/AHA guideline for assessment of cardiovascu-lar risk in asymptomatic adults: a report of the Ameri-can College of Cardiology Foundation/American HeartAssociation Task Force on Practice Guidelines.Circulation. 2010;122(25):e584-e636.11. Bild DE, Bluemke DA, Burke GL, et al. Multi-ethnic study of atherosclerosis: objectives and design.Am J Epidemiol. 2002;156(9):871-881.12. Friedewald WT, Levy RI, Fredrickson DS. Estima-tion of the concentration of low-density lipoproteincholesterol in plasma, without use of the preparativeultracentrifuge. Clin Chem. 1972;18(6):499-502.13. Carr JJ, Nelson JC, Wong ND, et al. Calcified coro-nary artery plaque measurement with cardiac CT inpopulation-based studies: standardized protocol ofMulti-Ethnic Study of Atherosclerosis (MESA) andCoronary Artery Risk Development in Young Adults(CARDIA) study. Radiology. 2005;234(1):35-43.14. Agatston AS, Janowitz WR, Hildner FJ, ZusmerNR, Viamonte M Jr, Detrano R. Quantification of coro-nary artery calcium using ultrafast computedtomography. J Am Coll Cardiol. 1990;15(4):827-832.15. Polak JF, Pencina MJ, O’Leary DH, D’AgostinoRB. Common carotid artery intima-media thickness pro-gression as a predictor of stroke in multi-ethnic studyof atherosclerosis. Stroke. 2011;42(11):3017-3021.16. Delong ER, Delong DM, Clarke-Pearson DL. Com-paring the area under two or more correlatedreceiver operating characteristic curves. a non-parametric approach. Biometrics. 1988;44:837-845.17. Pencina MJ, D’Agostino RB, Vasan RS. Statisti-cal methods for assessment of added usefulness of new

biomarkers. Clin Chem Lab Med. 2010;48(12):1703-1711.18. Ridker PM, Buring JE, Rifai N, Cook NR. Devel-opment and validation of improved algorithms for theassessment of global cardiovascular risk in women: theReynolds Risk Score. JAMA. 2007;297(6):611-619.19. Fowkes FG, Murray GD, Butcher I, et al; AnkleBrachial Index Collaboration. Ankle brachial index com-bined with Framingham Risk Score to predict cardio-vascular events and mortality: a meta-analysis. JAMA.2008;300(2):197-208.20. Wilson PW, Pencina M, Jacques P, Selhub J,D’Agostino R Sr, O’Donnell CJ. C-reactive protein andreclassification of cardiovascular risk in the Framing-ham Heart Study. Circ Cardiovasc Qual Outcomes.2008;1(2):92-97.21. Kavousi M, Elias-Smale S, Rutten JH, et al. Evalu-ation of newer risk markers for coronary heart dis-ease risk classification: a cohort study. Ann Intern Med.2012;156(6):438-444.22. Voros S, Rivera JJ, Berman DS, et al; Society forAtherosclerosis Imaging and Prevention Tomo-graphic Imaging and Prevention Councils; Society ofCardiovascular Computed Tomography. Guideline forminimizing radiation exposure during acquisition ofcoronary artery calcium scans with the use of multi-detector computed tomography: a report by the So-ciety for Atherosclerosis Imaging and Prevention To-mographic Imaging and Prevention Councils incollaboration with the Society of Cardiovascular Com-puted Tomography. J Cardiovasc Comput Tomogr.2011;5(2):75-83.23. Kim KP, Einstein AJ, Berrington de Gonzalez A.Coronary artery calcification screening: estimated ra-diation dose and cancer risk. Arch Intern Med. 2009;169(13):1188-1194.24. Berrington de Gonzalez A, Mahesh M, Kim KP,et al. Projected cancer risks from computed tomo-graphic scans performed in the United States in 2007.Arch Intern Med. 2009;169(22):2071-2077.25. Ford ES, Giles WH, Mokdad AH. The distribu-tion of 10-Year risk for coronary heart disease amongUS adults: findings from the National Health and Nu-trition Examination Survey III. J Am Coll Cardiol. 2004;43(10):1791-1796.26. Hlatky MA, Greenland P, Arnett DK, et al; Ameri-can Heart Association Expert Panel on Subclinical Ath-erosclerotic Diseases and Emerging Risk Factors andthe Stroke Council. Criteria for evaluation of novelmarkers of cardiovascular risk: a scientific statementfrom the American Heart Association. Circulation.2009;119(17):2408-2416.

Join Dr Greenland, author of this article, on Wednesday, September 19, from 2 to 3 PM

eastern time for “Author in the Room,” an interactive teleconference aimed at closingthe gap between knowledge—what is published in this article—and action—how muchof this knowledge can be put into your actual practice. This teleconference, facilitatedby clinical experts, should help readers answer their questions and consider the impli-cations of the article for their practice.

Author in the Room is brought to you by JAMA and the Institute for Healthcare Improve-ment. To register for Author in the Room, please visit http://www.ihi.org/authorintheroom. You can listen to past conferences or subscribe to the podcast at http://jama.jamanetwork.com/multimedia.aspx#Teleconferences.





Common Carotid Intima-Media ThicknessMeasurements in Cardiovascular Risk PredictionA Meta-analysis

Hester M. Den Ruijter, PhD; Sanne A. E.Peters, MSc; Todd J. Anderson, MD;Annie R. Britton, PhD; Jacqueline M.Dekker, PhD; Marinus J.Eijkemans, PhD; Gunnar Engstrom, MD,PhD; Gregory W. Evans, MA; Jacquelinede Graaf, MD, PhD; Diederick E.Grobbee, MD, PhD; Bo Hedblad, MD,PhD; Albert Hofman, MD, PhD; SuzanneHolewijn, PhD; Ai Ikeda, PhD; MaryamKavousi, MD, MSc; KazuoKitagawa, MD; Akihiko Kitamura, MD,PhD; Hendrik Koffijberg, PhD; Eva M.Lonn, MD; Matthias W. Lorenz, MD;Ellisiv B. Mathiesen, MD; GielNijpels, MD, PhD; Shuhei Okazaki, MD;Daniel H. O’Leary, MD; Joseph F.Polak, MD; Jackie F. Price, MD;Christine Robertson, MBChB;Christopher M. Rembold, MD; MariaRosvall, MD, PhD; Tatjana Rundek, MD,PhD; Jukka T. Salonen, MD, PhD;Matthias Sitzer, MD; Coen D. A.Stehouwer, MD, PhD; Jacqueline C.Witteman, PhD; Karel G. Moons, PhD;Michiel L. Bots, MD, PhD

CARDIOVASCULAR DISEASE IS

among the leading causes ofmorbidity and mortalityworldwide. Preventive treat-

ment of high-risk asymptomatic indi-viduals depends on accurate predic-tion of a person’s risk to develop acardiovascular event. Currently, car-diovascular risk prediction in asymp-

tomatic individuals is based on the levelof cardiovascular risk factors incorpo-rated in scoring equations.1 Severalscores are available, with the Framing-ham Risk Score among the most widelyused.1,2 These risk equations performreasonably well, yet there remains con-siderable overlap in estimated risk be-tween those who are affected by a car-diovascular event and those who are

not.3 Improvement in cardiovascularrisk prediction is needed and may beestablished by including a measure ofpreclinical atherosclerosis in the riskprediction algorithms4 because athero-


Author Affiliations are listed at the end of this article.Corresponding Author: Hester M. Den Ruijter, PhD,Julius Center for Health Sciences and Primary Care,University Medical Center Utrecht, Heidelberglaan 100,3584 CX, Utrecht, the Netherlands ([email protected]).

Context The evidence that measurement of the common carotid intima-media thick-ness (CIMT) improves the risk scores in prediction of the absolute risk of cardiovas-cular events is inconsistent.

Objective To determine whether common CIMT has added value in 10-year riskprediction of first-time myocardial infarctions or strokes, above that of the Framing-ham Risk Score.

Data Sources Relevant studies were identified through literature searches of data-bases (PubMed from 1950 to June 2012 and EMBASE from 1980 to June 2012) andexpert opinion.

Study Selection Studies were included if participants were drawn from the generalpopulation, common CIMT was measured at baseline, and individuals were followedup for first-time myocardial infarction or stroke.

Data Extraction Individual data were combined into 1 data set and an individualparticipant data meta-analysis was performed on individuals without existing cardio-vascular disease.

Results We included 14 population-based cohorts contributing data for 45 828 in-dividuals. During a median follow-up of 11 years, 4007 first-time myocardial infarc-tions or strokes occurred. We first refitted the risk factors of the Framingham Risk Scoreand then extended the model with common CIMT measurements to estimate the ab-solute 10-year risks to develop a first-time myocardial infarction or stroke in both mod-els. The C statistic of both models was similar (0.757; 95% CI, 0.749-0.764; and 0.759;95% CI, 0.752-0.766). The net reclassification improvement with the addition of com-mon CIMT was small (0.8%; 95% CI, 0.1%-1.6%). In those at intermediate risk, thenet reclassification improvement was 3.6% in all individuals (95% CI, 2.7%-4.6%)and no differences between men and women.

Conclusion The addition of common CIMT measurements to the Framingham RiskScore was associated with small improvement in 10-year risk prediction of first-time myo-cardial infarction or stroke, but this improvement is unlikely to be of clinical importance.JAMA. 2012;308(8):796-803 www.jama.com



sclerosis underlies the occurrence ofcardiovascular events, develops over de-cades, and has a prolonged asymptom-atic phase during which it is possibleto modify the course of the disease.5

Measurementofcarotid intima-mediathickness (CIMT) has been proposed tobeaddedtocardiovascular risk factors toimprove individual riskassessment.6,7 Sofar, individual studies reported on theadded value of CIMT measurements incardiovascularriskprediction,buttheevi-dence isnotconsistentacross studies.8-14

Furthermore,guidelinesdifferintheirrec-ommendationsforusingCIMTmeasure-ments in primary prevention and whichpatients to consider, ranging from mea-surement inall individuals15 tomeasure-ment in only those at intermediate risk.4

Therefore,solidandvalidevidenceonthisissue is needed. The USE Intima-MediaThickness (USE-IMT) collaboration is aglobal meta-analysis project using indi-vidualparticipantdata fromprospectivecohort studies to determine the addedvalueof theCIMTtocurrent riskpredic-tion models in asymptomatic individu-als at risk for cardiovascular disease.

METHODSThe USE-IMT project is an ongoingmeta-analysis of individual partici-pant data. Eligible cohorts are identi-fied through literature searches of da-tabases and through expert suggestion(the current analysis used PubMed from1950 to June 2012 and EMBASE from1980 to June 2012 using the searchquery published elsewhere16). A flow-chart of the search (performed onJune 19, 2012) and the inclusion inUSE-IMT is displayed in eFigure 1(available at http://www.jama.com). Atpresent, 17 cohorts participate in USE-IMT of which 14 cohorts are includedin this analysis. One cohort was ex-cluded because only maximal com-mon CIMT values were measured.17

The individual information from 2 othercohorts was not available yet.18,19 Thecohorts were required to have avail-able baseline data on age, sex, ciga-rette smoking status, antihypertensivemedication use, blood pressure, cho-lesterol fractions, CIMT measure-

ments, history of cardiovascular dis-ease and diabetes mellitus, andfollow-up information on occurrenceof cardiovascular events. Individual datafrom cohorts were collected and har-monized for the statistical analysesusing SPSS version 17 (SPSS).

Study Population

Of the 63 514 individuals included inUSE-IMT, we selected 45 828 individu-als to whom the cardiovascular riskscores like Framingham Risk Score ap-ply (aged 45-75 years, systolic bloodpressure �180 mm Hg, total choles-terol �300 mg/dL; no symptomatic car-diovascular disease at baseline). Usingthese criteria, the number of excludedindividuals was 6154 because of age,2977 for total cholesterol level, 1757 forsystolic pressure, and 7740 for previ-ous cardiovascular disease (not mutu-ally exclusive). Incomplete data on com-mon CIMT, cardiovascular risk factors,and (time to) events resulted in 2.2%missing data points, which were im-puted using single imputation for eachcohort separately (using the Multivari-ate Imputation by Chained Equationspackage of R). Predictors in our impu-tation model included all variables in ourdatabase including the outcome of in-terest, as recommended previously.20 Fora sensitivity analysis, we also per-formed a complete case analysis.

Common CIMTand Outcome Measure

Per cohort, we averaged all availablecommon CIMT measurements (from thenumber of angles; from either the farwall, near wall, or both; and from one orboth sides of the neck). This choice wasbased on the observation that the mag-nitude of the relation between commonCIMT and cardiovascular events risk donot differ greatly across various mea-sures.21 All CIMT values were used in theanalysis, including values larger than1mm, which are suggestive of plaque. Toaccount for differences in absolute CIMTlevels across cohorts because of differ-ences in methodology, we also calcu-lated cohort-specific z scores, which werecreated by subtracting the individual

CIMT values from the cohort meanCIMT. This value was then divided bythe cohort CIMT standard deviation.First-timemyocardial infarctionand first-time stroke were included as a com-bined end point. These included both fa-tal and nonfatal events.


The original variables of the 10-yearFramingham Risk Score2 (age, sex, ciga-rette smoking status, blood pressure, an-tihypertensive medication use, total cho-lesterol level, high-density lipoproteincholesterol level, and presence of diabe-tes mellitus) were first refit using mul-tivariable Cox proportional-hazardsmodel. This baseline model was then ex-tended by a log-transformed commonCIMT variable. Both models included co-hort as a random effect using the frailtymodel. Heterogeneity in CIMT andevents across cohorts was tested with alikelihood ratio test for interaction be-tween cohort and CIMT in the Cox pro-portional-hazards model. In addition, wealso tested for heterogeneity of the haz-ards ratios across cohorts using a ran-dom effects meta-analysis.

The improvement of addition of meancommon CIMT to the baseline modelwas tested with the Wald test and thelikelihood ratio test. The predictive per-formance of both models was assessedby comparing the predicted vs the 10-year observed risk, based on the Kaplan-Meier estimate (eFigure 2). The discrimi-native value of both models wasexpressed with Harrell C index.22 The 10-year absolute risk to develop a myocar-dial infarction or stroke was calculatedand was used to classify individuals intorisk categories of less than 5% (low risk),5% to less than 20% (intermediate risk),or 20% or greater (high risk) accordingto the risk classification of the Framing-ham Heart Study.12 The net reclassifica-tion improvement was calculated andquantifies the percentage of correctmovement across categories for thosewith and without events. Correct move-ment is upward classification by a newmarker in those with events and down-ward classification for those withoutevents. Our risk prediction model was

CAROTID INTIMA-MEDIA THICKNESS AND CARDIOVASCULAR RISK



based on time-to-event data, which con-tain not only events and nonevents butalso individuals who discontinue pre-maturely. Therefore, the number of in-dividuals reclassified due to a change inrisk category was then described usingthe net reclassification improvement tak-ing survival time into account.23 The cor-responding 95% confidence intervalswere obtained with bootstrapping.

We also calculated the net reclassifi-cation improvement for 4 risk catego-ries: less than 5% (low risk), 5% to lessthan 10% (low to intermediate risk),10% to less than 20% (intermediate tohigh risk), and 20% or greater (high risk)because the 4-level risk category ap-proach is still widely used outside theUnited States. In addition, we assessedimprovement without cutoff by risk cat-egories using the integrated discrimina-tion improvement, which can be seen asequal to differences in discriminationslopes.24 The relative integrated discrimi-nation improvement was calculated bydividing the integrated discriminationimprovement by the discrimination ofthe baseline model (based on the pre-dicted probabilities in those with eventsand those without events).24

Finally, the net reclassification im-provement and the (relative) integrateddiscrimination improvement were as-sessed separately in men and women.This sex-specific analysis was per-formed as the relation between CIMTand first-time myocardial infarction orstroke was different for men and women(interaction term in Cox proportionalhazards model, P=.017). In addition, wespecifically addressed individuals clas-sified in the intermediate-risk groups (ac-cording to the baseline model and de-fined as a 10-year absolute risk of 5% to20%). All analyses were performed in thestatistical environment R (version2.10.0). We did not validate our modelwith CIMT measurements because theaim was not to create and validate a newprediction rule, but to assess the actualimprovement in risk prediction. All sta-tistical testing was 2-sided and a P� .05was considered statistically significant.

RESULTSBaseline characteristics of the cohorts arepresented in TABLE 1. The majority of thestudiedpopulationwaswhite.Mean(SD)common CIMT in USE-IMT was 0.73(0.16) mm. Mean CIMT increased with

age in every cohort (eTable 1). The me-dian (SD) follow-up in USE-IMT was 11(3.7) years, during which 4007 first-time myocardial infarctions or first-time strokes occurred (TABLE 2).

Common CIMT and First-TimeMyocardial Infarction or Stroke

The risk factors included in the Framing-ham Risk Score and increased com-mon CIMT were all related to first-time myocardial infarction or stroke(eTable 2), and there was no evidencefor heterogeneity in the relation be-tween CIMT and outcome between stud-ies (likelihood ratio rest for interac-tion, P=.18). Adjusted common CIMTwas positively related to myocardial in-farction and stroke with a hazard ratioper 0.1-mm difference of common CIMTof 1.12 (95% CI, 1.09-1.14) for womenand 1.08 (95% CI, 1.05-1.11) for men.The hazard ratio per 0.1-mm differ-ence of common CIMT was 1.08 (95%CI, 1.05-1.10) for myocardial infarc-tion and 1.12 (95% CI, 1.10-1.15) forstroke. The study-specific hazard ra-tios for mean common CIMT and first-time myocardial infarction or stroke aredisplayed in FIGURE 1. Based on a ran-

Table 1. Baseline Characteristics of the Cohorts in USE-IMTa

SourceMen,No.

Women,No.

Age,Mean

(Range), y

SBP,mm Hg,

Mean (SD)

No. (%) Mean (SD)

Smoking DiabetesStatinUse

HypertensiveLowering

Medication

TotalCholesterol,

mg/dL

HDLCholesterol,

mg/dLCommonCIMT, mm

ARIC,25 1994 6219 8099 54 (45-64) 120 (17.4) 3691 (26) 1530 (11) 3351 (23) 3920 (27) 212 (38) 54 (17) 0.65 (0.146)

CAPS,26 2006 1889 2000 52 (35-75) 128 (16.1) 810 (21) 103 (3) 180 (5) 743 (19) 224 (37) 58 (17) 0.74 (0.142)

Charlottesville,27 2006 341 269 57 (35-75) 138 (17.2) 52 (9) 21 (3) 150 (25) 264 (43) 220 (40) 46 (15) 0.82 (0.171)

CHS,28 2007 1183 1942 70 (65-75) 133 (18.6) 439 (14) 413 (13) 158 (5) 1141 (37) 212 (36) 54 (16) 0.85 (0.155)

FATE,8 2011 1438 3 51 (35-75) 128 (16.4) 184 (13) 38 (3) 130 (9) 163 (11) 205 (35) 50 (11) 0.72 (0.176)

Hoorn Study,29 2003 122 126 67 (60-75) 137 (18.4) 40 (16) 43 (17) 19 (8) 61 (25) 224 (37) 54 (15) 0.83 (0.152)

KIHD,30 1991 879 51 (42-61) 132 (14.6) 338 (39) 30 (3) 3 (1) 98 (11) 220 (35) 50 (12) 0.75 (0.157)

Malmo,31 2000 1973 2794 57 (46-68) 140 (17.4) 1077 (23) 356 (8) 118 (2) 692 (15) 232 (36) 54 (14) 0.76 (0.149)

MESA,32 2007 2800 3095 60 (44-75) 124 (19.3) 832 (14) 708 (12) 988 (17) 2093 (36) 193 (34) 50 (15) 0.74 (0.165)

Nijmegen Study,33 2009 562 638 61 (50-72) 128 (14.9) 194 (16) 58 (5) 97 (8) 238 (20) 224 (36) 54 (14) 0.83 (0.108)

NOMAS,34 2007 458 633 65 (50-75) 137 (17.5) 186 (17) 158 (14) 150 (19) 530 (49) 201 (37) 46 (14) 0.71 (0.087)

OSACA2 Study,35 2007 199 204 63 (39-75) 136 (17.3) 89 (22) 50 (12) 106 (26) 220 (55) 212 (33) 58 (16) 0.85 (0.258)

Rotterdam Study,36 1997 1536 2189 65 (55-75) 134 (19.2) 901 (24) 108 (3) 75 (20) 314 (8) 247 (35) 54 (14) 0.75 (0.137)

Tromsø Study,37 2000 2129 2111 60 (35-75) 141 (18.0) 1378 (3) 108 (3) 40 (1) 356 (8) 247 (36) 62 (17) 0.78 (0.152)

USE-IMT (total) 21 730 24 098 58 (35-75) 129 (19.4) 10 211 (22) 5131 (11) 5565 (12) 10 833 (24) 220 (40) 54 (17) 0.73 (0.163)

Abbreviations: ARIC, Atherosclerosis Risk in Communities; CAPS, Carotid Atherosclerosis Progression Study; CHS, Cardiovascular Health Study; CIMT, carotid intima-media thick-ness; FATE, Firefighters and Their Endothelium Study; HDL, high-density lipoprotein; KIHD, Kuopio Ischaemic Heart Disease Risk Factor Study; Malmo, Malmo Diet and CancerStudy; MESA, Multi-Ethnic Study of Atherosclerosis; Nijmegen Study, Nijmegen Biomedical Study; NOMAS, Northern Manhattan Study; OSACA2, Osaka Follow-up Study forCarotid Atherosclerosis 2; SBP, systolic blood pressure; USE-IMT, USE Intima-Media Thickness collaboration.

SI conversion factors: To convert total and HDL cholesterol to mmol/L, multiply by 0.0259.aValues are based on the individuals in the cohorts after applying the inclusion criteria as described in the “Methods” section.




dom-effects meta-analysis on the study-specific hazard ratios, there was no evi-dence for heterogeneity in CIMT andoutcome between studies (Q test ofheterogeneity P value, 0.24; I2, 12.30%).

Calibration and Discrimination

The addition of mean common CIMTimproved the baseline model (Wald testand likelihood ratio test, both P� .001).For both models, the 10-year pre-dicted risk was closely in agreementwith the 10-year cardiovascular dis-ease risk as estimated with Kaplan-Meier (eFigure 2). Harrell C index forthe baseline model was 0.757 (95% CI,0.749-0.764) and 0.759 (95% CI, 0.752-0.766) with addition of common CIMT.

Net Reclassification

FIGURE 2A shows the distribution of thenumberof individualswithoutandwitheventsacrossriskcategoriesbasedontheFramingham Risk Score and the distri-bution of individuals after the additionof the common CIMT. More than 90%of the individuals remained in the samerisk category. The numbers of individu-als shifting downward or upward with-out and with events were similar.

Figure 2B shows the observed risksof all the individuals in the categories.The observed risks of the individualsthat remained in the same risk catego-ries corresponded well to their allo-cated risk categories. Individualsreclassified to a higher risk categoryindeed had a significantly higher

observed risk compared than thosenot reclassified. Also, individualsreclassified to a lower risk categoryindeed had a lower observed risk thanthose not reclassified. Yet the confi-dence intervals indicate some overlapin observed risk in categories of thosereclassified.

Figure 1. Relation of Common Carotid Intima-Media Thickness With First-Time MyocardialInfarction or Stroke Across Studies

Source

Contribution to TotalUSE-IMT Population,

% of TotalHazard Ratio

(95% CI) aARIC,25 1994 31 1.11 (1.08-1.14)CAPS,26 2006 8 1.10 (0.99-1.23)Charlottesville,27 2006 1 0.88 (0.56-1.36)

FATE,8 2011 3 1.20 (1.01-1.42)Hoorn Study,29 2003 1 1.07 (0.72-1.59)KIHD,30 1991 2 1.05 (0.96-1.16)Malmo,31 2000 10 1.10 (1.04-1.17)MESA,32 2007 13 0.98 (0.89-1.08)Nijmegen Study,33 2009 3 1.34 (0.94-1.90)NOMAS,34 2007 2 1.36 (0.99-1.85)OSACA2 Study,35 2007 1 1.09 (0.96-1.24)Rotterdam Study,36 1997 8 1.13 (1.06-1.20)Tromsø Study,37 2000 9 1.04 (0.98-1.10)

I2 = 12.30%; Q test for heterogeneity, P = .24 1.09 (1.07-1.12)

CHS,28 2007 7 1.11 (1.06-1.16)

2.01.00.5

Hazard Ratio (95% CI) a

Study-specific hazard ratios (HRs) and the pooled hazard ratio based on a random-effects meta-analysis. Errorbars indicate 95% CI; data marker sizes indicate the sample sizes of the cohorts.aHazard ratios are per 0.1-mm increase in common carotid intima-media thickness.

Table 2. Baseline Risk and Follow-up Characteristics of the Cohorts in USE-IMTa

Source

Absolute 10-y Risk to Develop CVDBased on Framingham Risk Score

Variables, % (SD)Follow up,

Median (IQR), y MI, No.Strokes,

No.First-Time MIor Stroke, No.Overall Men Women

ARIC,25 1994 5.9 (5.2) 8.0 (5.9) 4.3 (3.8) 13.1 (12.3-13.9) 774 494 1196

CAPS,26 2006 4.5 (4.5) 5.8 (5.1) 3.3 (3.4) 8.1 (7.4-9.1) 59 77 130

Charlottesville,27 2006 5.4 (4.2) 6.5 (4.6) 4.1 (3.1) 4.1 (2.7-5.0) 5 3 8

CHS,28 2007 16.8 (10.0) 22.0 (11.1) 13.6 (7.7) 13.3 (8.8-14.5) 393 393 713

FATE,8 2011 3.8 (3.8) 3.8 (3.8) 0.5 (0.4) 8.0 (6.9-8.9) 23 10 33

Hoorn Study,29 2003 8.2 (5.8) 9.6 (5.6) 6.9 (5.7) 7.7 (7.5-8.2) 8 4 12

KIHD,30 1991 9.6 (5.8) 9.6 (5.8) 14.2 (13.2-15.2) 108 54 152

Malmo,31 2000 6.0 (4.9) 8.3 (5.8) 4.4 (3.4) 10.9 (10.2-11.6) 162 168 315

MESA,32 2007 5.5 (5.0) 7.1 (5.6) 4.2 (3.8) 6.5 (6.2-6.7) 98 72 167

Nijmegen Study,33 2009 5.5 (4.0) 7.2 (4.5) 4.0 (2.7) 3.9 (3.0-4.4) 12 5 17

NOMAS,34 2007 6.9 (5.1) 9.0 (6.0) 5.4 (3.7) 8.7 (5.7-10.3) 29 31 57

OSACA2 Study,35 2007 11.3 (8.3) 14.9 (9.1) 7.8 (5.7) 4.6 (3.3-6.1) 2 17 19

Rotterdam Study,36 1997 11.0 (7.3) 14.3 (8.1) 8.6 (5.6) 13.9 (10.6-14.8) 245 415 630

Tromsø Study,37 2000 11.9 (9.0) 14.7 (9.9) 9.0 (6.8) 10.7 (10.4-11.0) 366 228 558

USE-IMT (total) 7.5 (7.0) 9.4 (7.9) 5.8 (5.5) 10.8 (6.9-13.2) 2284 1971 4007Abbreviations: ARIC, Atherosclerosis Risk in Communities; CAPS, Carotid Atherosclerosis Progression Study; CHS, Cardiovascular Health Study; CIMT, carotid intima-media thick-

ness; CVD, cardiovascular disease; FATE, Firefighters and Their Endothelium Study; HDL, high-density lipoprotein; IQR, interquartile range; KIHD, Kuopio Ischaemic Heart Dis-ease Risk Factor Study; Malmo, Malmo Diet and Cancer Study; MI, myocardial infarction; MESA, Multi-Ethnic Study of Atherosclerosis; Nijmegen Study, Nijmegen BiomedicalStudy; NOMAS, Northern Manhattan Study; OSACA2, Osaka Follow-up Study for Carotid Atherosclerosis 2; USE-IMT, USE Intima-Media Thickness collaboration.

aValues are based on the individuals in the cohorts after applying the inclusion criteria as described in the methods section.




The net reclassification improve-ment indicated that the added valueof mean common CIMT was 0.8%(95% CI, 0.1%-1.6%) with no differ-ences between men and women(TABLE 3). The sex-specific reclassifi-cation tables are displayed in eFigures

3 and 4. The integrated discrimina-t ion improvement was 0 .0024(Table 3). The discrimination of thebaseline model based on the predictedprobabilities in those with and with-out events was 0.067. Thus, the rela-tive integrated discrimination improve-

ment was 3.6% and similar in men andwomen (Table 3).

Of the individuals at intermediaterisk, 88% remained in the same risk cat-egory after addition of CIMT to theFramingham Risk Score (Figure 2A).The reclassification was slightly morefavorable than in the whole popula-tion with more individuals withoutevents reclassified to a lower risk cat-egory and more individuals with eventsreclassified to a higher risk category.

Individuals classified to a higher riskcategory by CIMT had an observed riskabove 20% and those classified to a lowerrisk category by CIMT had an observedrisk less than 5%. The net reclassifica-tion improvement for the intermediate-risk group was 3.6% (95% CI, 2.7%-4.6%) with no differences between menand women (Table 3). The relative in-tegrated discrimination improvement in-dicated that the improvement in the pre-diction model was 3.6% (Table 3).

The net reclassification improve-ments in all individuals for myocardialinfarction and stroke separately were0.6% and 1.0%, respectively. When 4 riskcategories were applied (�5%, 5%-�10%, 10%-�20%, �20%), the netreclassification improvement in the over-all population was 1.2% (95% CI, 0.1%-2.2%) with no differences between menand women (eFigure 5). In individualsat intermediate risk, the net reclassifica-tion improvement was 4.6% (95% CI,3.1%-6.1%) with no differences be-tween men (3.9%; 95% CI, 2.3%-5.9%)and women (5.5%; 95% CI, 3.0%-6.9%). Results from the complete caseanalysis and from the analysis with thecohort-specific z scores were similar tothe results presented here.

Figure 2. Reclassification With CIMT Added to Framingham Risk Score

A Distribution of 45 828 individuals without and with events in USE-IMT across risk categories

B Observed Kaplan-Meier estimates in risk categories

Fram

ingh

am R

isk

Without events

<5%

>20%

5-20%

20 27139 162 (93.6) No change

Up classification–

1115

<5%

867

17 280

1229 (2.9%)

Down classification1430 (3.4%)

315

5%-20%

362

1611

>20%

Framingham Risk With CIMT



Fram

ingh

am R

isk

With events

Total without events, No. (%)

Total with events, No. (%)

All individuals, No. (%)

<5%

>20%

5-20%

5373684 (91.9%) No change


69

<5%

67

2410

169 (4.2%)


85

5%-20%

102

737

>20%

Fram

ingh

am R

isk

<5%

>20%

5-20%

2.2 (2.0-2.4)

42 846 (93.5%) No change


4.6 (3.3-5.9)

<5%

6.2 (4.5-7.9)

10.4 (10.0-10.9)

1398 (3.1%)


19.0 (14.6-23.1)

5%-20%

20.6 (16.4-24.6)

28.7 (26.7-30.6)

>20%

A, Individuals without and with events classified according to their 10-year absolute risk to develop a myo-cardial infarction or stroke predicted with the Framingham Risk Score variables or classified according to their10-year absolute risk to develop a first-time myocardial infarction or stroke predicted with the FraminghamRisk Score and a common carotid intima-media thickness (CIMT) measurement. B, Observed Kaplan-Meierabsolute risk estimates for all individuals (with and without events). The observed risk in reclassified individualsis significantly different from the observed risk of the individuals in the gray cells.

Table 3. Summary of the Indices of Added Value in the Total USE-IMT Cohort and in the Intermediate-Risk Categories, by Sex

All Men Women

USE-IMTNRI, % (95% CI) 0.8 (0.1 to 1.6) 0.9 (−0.2 to 1.9) 0.8 (−0.2 to 1.6)

IDI (95% CI) 0.0024 (0.0012 to 0.0036) 0.0024 (0.0004 to 0.0041) 0.0025 (0.0009 to 0.0040)

Relative IDI, % 3.6 3.6 3.7

USE-IMT, Intermediate-Risk Group (5% to �20%)NRI, % (95% CI) 3.6 (2.7 to 4.6) 3.2 (2.3 to 4.4) 3.9 (2.7 to 4.9)

IDI (95% CI) 0.0024 (0.0012 to 0.0036) 0.0019 (0.0003 to 0.0034) 0.0031 (0.0013 to 0.0048)

Relative IDI, % 3.6 2.7 4.6Abbreviations: IDI, integrated discrimination improvement; NRI, net reclassification improvement; USE-IMT, USE Intima-Media Thickness collaboration.




COMMENTIn this meta-analysis based on partici-pant data of 45 828 individuals from 14cohort studies worldwide, the addedvalue of common CIMT measure-ments to the Framingham Risk Scorein the general population was small(0.8% were correctly reclassified). In in-dividuals at intermediate risk, the addedvalue was 3.2% in men and 3.9% inwomen. Our results suggest that com-mon CIMT measurements should notroutinely be performed in the generalpopulation because the overall addedvalue is small and unlikely to be of clini-cal importance.

Recently, conflicting results havebeen published on the added value ofCIMT measurements in cardiovascu-lar risk prediction. These differencesmay be attributed to differences acrossstudies in CIMT measurement (eg, ca-rotid segments [common, bifurcation,internal], including or excluding ca-rotid plaques), individuals’ character-istics, cutoff values for risk categories,number of events (small numbers, es-pecially in those that are shifting riskcategories), and end-point definition.Within USE-IMT, we were able to sum-marize the majority of the existingevidence using uniform definitions ofcommon CIMT, study population, riskcategories, and cardiovascular events.We used only data on common CIMTand included only individuals to whomthe risk scores apply. Also, as fataland nonfatal myocardial infarction andstroke compose the majority of the car-diovascular events, we used theseoutcomes, which were available in allcohorts in USE-IMT. We used state-of-the-art statistical methods such as thenet reclassification improvement, whichincorporates time to event by Kaplan-Meier estimates rather than only dis-tinguishing between events and non-events. In addition, because thepopulations in USE-IMT may be verydifferent from that in Framingham,38 werefitted the cardiovascular risk factorsand also fitted the common CIMT mea-surements, which may be the moststraightforward method to assess theadded value of common CIMT mea-

surements. Finally, to evaluate the ro-bustness of our results, we also per-formed a complete-case analysis andused cohort-specific z scores of CIMT.These results were not different fromour main analysis. Our results indi-cate no improvement in risk stratifica-tion through common CIMT measure-ments for the general population,neither for men nor for women.

We based our analysis on measure-ments of the mean common CIMT. Werestricted to common CIMT measure-ments because they were available in allstudies, they are generally feasible to usein routine clinical practice, and their usehas been recommended.15,39 Measure-ments of CIMT obtained from other ca-rotid segments and the inclusion of aseparate measure of carotid plaque maybe important in risk prediction. Re-cently, the Framingham investigatorsshowed that the maximal CIMT of theinternal carotid artery has added valuein risk prediction whereas the com-mon CIMT of the mean common ca-rotid artery did not.12

Our results are very similar to thoseof the Framingham cohort, a studythat was not included in this meta-analysis. A recent meta-analysis sug-gested that carotid plaque was betterthan CIMT in predicting coronaryevents.40 In several cohorts included inthat meta-analysis, plaque was de-fined based on a certain arbitrary CIMTcutoff, and results were not presentedfor different definitions of plaque. In ad-dition, others found the opposite forrisk of stroke.41 The ARIC investiga-tors reported that plaque information,in addition to CIMT, resulted in a netreclassification improvement of 9.9% inthe overall population.11 In our study,we included all the reported CIMT val-ues, even the thicker CIMT values sug-gestive of plaque. However, we did notseparate plaque analysis, because sepa-rate information on plaque presence orabsence was not available in USE-IMT. Furthermore, the reproducibil-ity of plaque assessment is far less thanthat of CIMT (� for plaques, 0.60-0.70, vs intraclass correlation coeffi-cients for CIMT, 0.90-0.95).42,43 The

added value of CIMT measurementsfrom other sites than the common ca-rotid segment (eg, maximal CIMT) ob-tainable by carotid ultrasound is yet tobe determined.

Our results suggest that commonCIMT measurements should not rou-tinely be performed in the generalpopulation, as the overall added valuemay be too limited to result in healthbenefits. In individuals classified asbeing at intermediate risk by theFramingham Risk Score, informationon the common CIMT measurementsshowed a slightly higher yield (net re-classification improvement of 3.2% inmen and 3.9% in women). Yet, as de-scribed by Cook and Paynter,44 the netreclassification improvement for use-less markers may not be zero in the in-termediate-risk group, and one shouldbe cautious in overinterpreting the netreclassification improvement in the in-termediate-risk group. Therefore, theadded value of mean common CIMT in10-year risk prediction for cardiovas-cular disease, even in the intermediate-risk category, is most likely too smallto result in health benefit. However, asthe interest in risk prediction is cur-rently shifting from a 10-year risk tolifetime risk, the added value of a CIMTmeasurement and its cost-effective-ness using a horizon of 20 to 30 yearsmay be worthwhile to explore.

Our study has several limitations.The cohorts included in USE-IMTshowed variation in statin use becausethey were studied across different de-cades. Yet there was no heterogeneityin the relation between common CIMTmeasurements and cardiovascularevents, suggesting that differences instatin use did not affect the relation-ship between CIMT and events. Thereare differences in the adjudication ofevents across studies. Although we donot think that these differences are re-lated to CIMT measurement (so non-differential misclassification), weincluded hard end points such as myo-cardial infarction and stroke as thesewere least likely to be affected. It is wellestablished that ethnicity is an impor-tant determinant of CIMT.45 Because




most individuals in USE-IMT were de-rived from a white population, our find-ings on the added value of CIMT in riskprediction may not necessarily apply toother ethnicities.

In conclusion, the added value of com-mon CIMT in 10-year risk prediction ofcardiovascular events, in addition to theFramingham Risk Score, was small andunlikely to be of clinical importance.Author Affiliations: Julius Center for Health Sciencesand Primary Care (Drs Den Ruijter, Peters, Eijkemans,Grobbee, Koffijberg, Moons, and Bots) and Depart-ment of Experimental Cardiology (Dr Den Ruijter),University Medical Center Utrecht, Utrecht, theNetherlands; Department of Cardiac Sciences andLibin Cardiovascular Institute of Alberta, University ofCalgary, Calgary, Alberta, Canada (Dr Anderson);Department of Epidemiology and Public Health, Uni-versity College London, London, United Kingdom (DrBritton); Institute for Health and Care Research, VUMedical Center, Amsterdam, the Netherlands (DrsDekker and Nijpels); Department of Clinical Sciencesin Malmo, Lund University, Skane University Hospi-tal, Malmo, Sweden (Drs Engstrom, Hedblad, andRosvall); Department of Biostatistical Sciences andNeurology, Wake Forest School of Medicine,Winston-Salem, North Carolina (Mr Evans); Depart-ment of General Internal Medicine, Division of Vas-cular Medicine, Nijmegen University Medical Centre,the Netherlands (Drs de Graaf and Holewijn); Univer-sity of Malaya Medical Center, Kuala Lumpur,Malaysia (Dr Grobbee); Department of Epidemiology,Erasmus Medical Center, Rotterdam, the Netherlands(Drs Hofman, Kavousi, and Witteman); Osaka Medi-cal Center for Health Science and Promotion, Osaka,Japan (Drs Ikeda and Kitamura); Stroke Center,Department of Neurology, Osaka University Gradu-ate School of Medicine, Osaka (Drs Kitagawa andOkazaki); Department of Medicine, Division of Cardi-ology and Population Health Research Institute,McMaster University, Hamilton, Ontario, Canada (DrLonn); Department of Neurology, University Hospi-tal, Goethe-University, Frankfurt am Main, Germany(Drs Lorenz and Sitzer); Brain and CirculationResearch Group, Institute of Clinical Medicine, Uni-versity of Tromsø, Tromsø, Norway (Dr Mathiesen);Department of Radiology, Tufts Medical Center, Bos-ton, Massachusetts (Drs O’Leary and Polak); Centrefor Population Health Sciences, University of Edin-burgh, Edinburgh, United Kingdom (Drs Price andRobertson); Cardiology Division, Department ofInternal Medicine, University of Virginia, Charlottes-ville (Dr Rembold); Department of Neurology, MillerSchool of Medicine, University of Miami, Miami,Florida (Dr Rundek); MAS-Metabolic Analytical Ser-vices Oy, Helsinki, Finland (Dr Salonen); Departmentof Neurology Klinikum Herford, Herford, Germany(Drs Sitzer); and Department of Internal Medicineand Cardiovascular Research Institute Maastricht,Maastricht University Medical Center, Maastricht, theNetherlands (Dr Stehouwer).Author Contributions: Dr Den Ruijter had full accessto all of the data in the study and takes responsibilityfor the integrity of the data and the accuracy of thedata analysis.Study concept and design: Den Ruijter, Peters, de Graaf,Grobbee, Hofman, Koffijberg, Lonn, O’Leary,Rembold, Rosvall, Rundek, Salonen, Sitzer, Moons,Bots.Acquisition of data: Den Ruijter, Peters, Anderson,Britton, Dekker, Engstrom, Evans, de Graaf, Grobbee,Hedblad, Holewijn, Ikeda, Kavousi, Kitagawa,Kitamura, Koffijberg, Lonn, Lorenz, Mathiesen, Nijpels,

Okazaki, O’Leary, Polak, Price, Robertson, Rembold,Rosvall, Salonen, Sitzer, Stehouwer, Moons, Bots.Analysis and interpretation of data: Den Ruijter, Peters,Eijkemans, Grobbee, Koffijberg, Lonn, Price, Rosvall,Rundek, Salonen, Sitzer, Stehouwer, Moons, Bots.Drafting of the manuscript: Den Ruijter, Peters, Britton,Grobbee, Kavousi, Koffijberg, Rosvall, Salonen, Sitzer,Moons, Bots.Critical revision of the manuscript for important in-tellectual content: Peters, Anderson, Dekker,Eijkemans, Engstrom, Evans, de Graaf, Grobbee,Hedblad, Hofman, Holewijn, Ikeda, Kavousi, Kitagawa,Kitamura, Koffijberg, Lonn, Lorenz, Mathiesen, Nijpels,Okazaki, O’Leary, Polak, Price, Robertson, Rembold,Rosvall, Rundek, Salonen, Sitzer, Stehouwer, Witteman,Moons, Bots.Statistical analysis: Den Ruijter, Peters, Eijkemans,Grobbee, Koffijberg, Rosvall, Salonen, Sitzer, Moons,Bots.Obtained funding: Peters, Anderson, Grobbee,Hofman, Koffijberg, O’Leary, Polak, Rosvall, Salonen,Sitzer, Stehouwer, Moons, Bots.Administrative, technical, or material support: DenRuijter, Peters, Anderson, Britton, Engstrom, Evans,Grobbee, Hedblad, Holewijn, Ikeda, Kavousi, Kitamura,Koffijberg, Mathiesen, Okazaki, O’Leary, Polak,Rembold, Rosvall, Rundek, Salonen, Sitzer, Moons.Study supervision: Den Ruijter, Peters, de Graaf,Grobbee, Hofman, Koffijberg, Lorenz, Polak, Rosvall,Rundek, Salonen, Sitzer, Stehouwer, Moons, Bots.Conflict of Interest Disclosures: All authors have com-pleted and submitted the ICMJE Form for Disclosure ofPotential Conflicts of Interest. Dr de Graaf reported hav-ing received a Dutch Heart Foundation grant to per-form the NBS2 study (Nijmegen Biomedical Study). DrEngstrom reported being employed as a senior epide-miologist by AstraZeneca R&D. Dr Kitagawa reportedbeing employed by Osaka University Hospital; havingreceived a grant from the Ministry of Education, Cul-ture, Sports, and Technology of Japan; and having re-ceived lecture fees from sanofi-aventis. Dr Lonn re-ported having been a consultant for Merck and HoffmanLaroche; having provided expert testimony for Merck;having received grants from AstraZeneca, sanofi-aventis, Novartis, and GlaxoSmithKline; and having re-ceived lecture fees from Merck and Novartis. Dr Ma-thiesen reported having received a grant from the NorthNorwegian Health Authorities. Dr O’Leary reported own-ing stock in Medpace. Dr Polak reported having re-ceived a grant from the National Heart, Lung, and BloodInstitute. Dr Price reported having received a grant fromthe British Heart Foundation. Dr Rundek reported hav-ing received grants from the National Institutes of Health.Dr Salonen reported having received a grant from theUniversity of Eastern Finland for the funding of the Kuo-pio Ischaemic Heart Disease Risk Factor study. Dr Grob-bee and Dr Bots reported having been a consultant forand having received grants and lecture fees fromAstraZeneca. No other disclosures were reported.Funding/Support: This project is supported by a grantfrom the Netherlands Organisation for Health Re-search and Development (ZonMw 200320003).Role of the Sponsor: The funding source had no rolein the design and conduct of the study; in the collec-tion, analysis, and interpretation of the data; or in thepreparation, review, or approval of the manuscript.Online-Only Material: The eTables and eFigures areavailable at http://www.jama.com.Additional Contributions: We thank Thomas Debray,MSc ( Julius Center for Health Sciences and PrimaryCare, University Medical Center Utrecht), for his helpin performing the literature searches. He did not re-ceive compensation for the contribution.

REFERENCES

1. Berger JS, Jordan CO, Lloyd-Jones D, BlumenthalRS. Screening for cardiovascular risk in asymptom-

atic patients. J Am Coll Cardiol. 2010;55(12):1169-1177.2. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. Gen-eral cardiovascular risk profile for use in primary care:the Framingham Heart Study. Circulation. 2008;117(6):743-753.3. Akosah KO, Schaper A, Cogbill C, Schoenfeld P.Preventing myocardial infarction in the young adultin the first place: how do the National Cholesterol Edu-cation Panel III guidelines perform? J Am Coll Cardiol.2003;41(9):1475-1479.4. Greenland P, Alpert JS, Beller GA, et al. 2010 ACCF/AHA guideline for assessment of cardiovascular riskin asymptomatic adults: executive summary: a reportof the American College of Cardiology Foundation/American Heart Association Task Force on PracticeGuidelines. Circulation. 2010;112(25):2748-2764.5. Libby P, Theroux P. Pathophysiology of coronaryartery disease. Circulation. 2005;111(25):3481-3488.6. Lorenz MW, Markus HS, Bots ML, Rosvall M, SitzerM. Prediction of clinical cardiovascular events with ca-rotid intima-media thickness: a systematic review andmeta-analysis. Circulation. 2007;115(4):459-467.7. Society of Atherosclerosis Imaging and Preven-tion Developed in collaboration with the Interna-tional Atherosclerosis Society. Appropriate use crite-ria for carotid intima media thickness testing.Atherosclerosis. 2011;214(1):43-46.8. Anderson TJ, Charbonneau F, Title LM, et al. Mi-crovascular function predicts cardiovascular events inprimary prevention: long-term results from the Fire-fighters and Their Endothelium (FATE) study.Circulation. 2011;123(2):163-169.9. Elias-Smale SE, Kavousi M, Verwoert GC, et al. Com-mon carotid intima-media thickness in cardiovascu-lar risk stratification of older people: the RotterdamStudy. Eur J Prev Cardiol. 2012;19(4):698-705.10. Lorenz MW, Schaefer C, Steinmetz H, Sitzer M.Is carotid intima media thickness useful for individualprediction of cardiovascular risk? ten-year results fromthe Carotid Atherosclerosis Progression Study (CAPS).Eur Heart J. 2010;31(16):2041-2048.11. Nambi V, Chambless L, Folsom AR, et al. Ca-rotid intima-media thickness and presence or ab-sence of plaque improves prediction of coronary heartdisease risk: the ARIC (Atherosclerosis Risk in Com-munities) study. J Am Coll Cardiol. 2010;55(15):1600-1607.12. Polak JF, Pencina MJ, Pencina KM, O’Donnell CJ,Wolf PA, D’Agostino RB Sr. Carotid-wall intima-media thickness and cardiovascular events. N Engl JMed. 2011;365(3):213-221.13. Salonen JT, Salonen R. Ultrasonographically as-sessed carotid morphology and the risk of coronaryheart disease. Arterioscler Thromb. 1991;11(5):1245-1249.14. Xie W, Wu Y, Wang W, et al. A longitudinal studyof carotid plaque and risk of ischemic cardiovasculardisease in the Chinese population. J Am SocEchocardiogr. 2011;24(7):729-737.15. Naghavi M, Falk E, Hecht HS, et al; SHAPE TaskForce. From vulnerable plaque to vulnerable patient:part III, executive summary of the Screening for HeartAttack Prevention and Education (SHAPE) Task Forcereport. Am J Cardiol. 2006;98(2A):2H-15H.16. Plantinga Y, Dogan S, Grobbee DE, Bots ML. Ca-rotid intima-media thickness measurement in cardio-vascular screening programmes. Eur J Cardiovasc PrevRehabil. 2009;16(6):639-644.17. Kitamura A, Iso H, Imano H, et al. Carotid intima-media thickness and plaque characteristics as a risk fac-tor for stroke in Japanese elderly men. Stroke. 2004;35(12):2788-2794.18. Singh-Manoux A, Britton A, Kivimaki M, GueguenA, Halcox J, Marmot M. Socioeconomic status mod-erates the association between carotid intima-mediathickness and cognition in midlife: evidence from the




Whitehall II study. Atherosclerosis. 2008;197(2):541-548.19. Allan PL, Mowbray PI, Lee AJ, Fowkes FG. Rela-tionship between carotid intima-media thickness andsymptomatic and asymptomatic peripheral arterial dis-ease: the Edinburgh Artery Study. Stroke. 1997;28(2):348-353.20. Moons KGM, Donders RART, Stijnen T, HarrellFE Jr. Using the outcome for imputation of missing pre-dictor values was preferred. J Clin Epidemiol. 2006;59(10):1092-1101.21. Bots ML, de Jong PTVM, Hofman A, GrobbeeDE. Left, right, near or far wall common carotid intima-media thickness measurements: associations with car-diovascular disease and lower extremity arterialatherosclerosis. J Clin Epidemiol. 1997;50(7):801-807.22. Harrell FE Jr, Lee KL, Califf RM, Pryor DB, RosatiRA. Regression modelling strategies for improved prog-nostic prediction. Stat Med. 1984;3(2):143-152.23. Pencina MJ, D’Agostino RB Sr, Steyerberg EW.Extensions of net reclassification improvement calcu-lations to measure usefulness of new biomarkers. StatMed. 2011;30(1):11-21.24. Pencina MJ, D’Agostino RB, Vasan RS. Statisti-cal methods for assessment of added usefulness of newbiomarkers. Clin Chem Lab Med. 2010;48(12):1703-1711.25. Li R, Duncan BB, Metcalf PA, et al; Atheroscle-rosis Risk in Communities (ARIC) Study Investigators.B-mode-detected carotid artery plaque in a generalpopulation. Stroke. 1994;25(12):2377-2383.26. Lorenz MW, von Kegler S, Steinmetz H, MarkusHS, Sitzer M. Carotid intima-media thickeningindicates a higher vascular risk across a wide age range:prospective data from the Carotid AtherosclerosisProgression Study (CAPS). Stroke. 2006;37(1):87-92.27. Ali YS, Rembold KE, Weaver B, et al. Predictionof major adverse cardiovascular events by age-normalized carotid intimal medial thickness.Atherosclerosis. 2006;187(1):186-190.28. Cao JJ, Arnold AM, Manolio TA, et al. Associa-

tion of carotid artery intima-media thickness, plaques,and C-reactive protein with future cardiovascular dis-ease and all-cause mortality: the Cardiovascular HealthStudy. Circulation. 2007;116(1):32-38.29. Henry RMA, Kostense PJ, Spijkerman AMW, et al;Hoorn Study. Arterial stiffness increases with deterio-rating glucose tolerance status: the Hoorn Study.Circulation. 2003;107(16):2089-2095.30. Salonen R, Salonen JT. Determinants of carotidintima-media thickness: a population-based ultraso-nography study in eastern Finnish men. J Intern Med.1991;229(3):225-231.31. Rosvall M, Ostergren PO, Hedblad B, IsacssonSO, Janzon L, Berglund G. Occupational status, edu-cational level, and the prevalence of carotid athero-sclerosis in a general population sample of middle-aged Swedish men and women: results from theMalmo Diet and Cancer Study. Am J Epidemiol. 2000;152(4):334-346.32. Mora S, Szklo M, Otvos JD, et al. LDL particle sub-classes, LDL particle size, and carotid atherosclerosisin the Multi-Ethnic Study of Atherosclerosis (MESA).Atherosclerosis. 2007;192(1):211-217.33. Holewijn S, den Heijer M, Swinkels DW, StalenhoefAFH, de Graaf J. The metabolic syndrome and its traitsas risk factors for subclinical atherosclerosis. J Clin En-docrinol Metab. 2009;94(8):2893-2899.34. Prabhakaran S, Singh R, Zhou X, Ramas R, SaccoRL, Rundek T. Presence of calcified carotid plaque pre-dicts vascular events: the Northern Manhattan Study.Atherosclerosis. 2007;195(1):e197-e201.35. Kitagawa K, Hougaku H, Yamagami H, et al;OSACA2 Study Group. Carotid intima-media thick-ness and risk of cardiovascular events in high-risk pa-tients: results of the Osaka Follow-Up Study for Ca-rotid Atherosclerosis 2. Cerebrovasc Dis. 2007;24(1):35-42.36. Hofman A, van Duijn CM, Franco OH, et al. TheRotterdam Study: 2012 objectives and design update.Eur J Epidemiol. 2011;26(8):657-686.37. Stensland-Bugge E, Bønaa KH, Joakimsen O,Njølstad I. Sex differences in the relationship of riskfactors to subclinical carotid atherosclerosis mea-

sured 15 years later : the Tromsø study. Stroke. 2000;31(3):574-581.38. D’Agostino RB Sr, Grundy S, Sullivan LM, WilsonP; CHD Risk Prediction Group. Validation of theFramingham coronary heart disease prediction scores:results of a multiple ethnic groups investigation. JAMA.2001;286(2):180-187.39. Stein JH, Korcarz CE, Hurst RT, et al; AmericanSociety of Echocardiography Carotid Intima-MediaThickness Task Force; Endorsed by the Society for Vas-cular Medicine. Use of carotid ultrasound to identifysubclinical vascular disease and evaluate cardiovas-cular disease risk: a consensus statement from theAmerican Society of Echocardiography Carotid Intima-Media Thickness Task Force. J Am Soc Echocardiogr.2008;21(2):93-111.40. Inaba Y, Chen JA, Bergmann SR. Carotid plaque,compared with carotid intima-media thickness, moreaccurately predicts coronary artery disease events: ameta-analysis. Atherosclerosis. 2012;220(1):128-133.41. Hollander M, Hak AE, Koudstaal PJ, et al. Com-parison between measures of atherosclerosis and riskof stroke: the Rotterdam Study. Stroke. 2003;34(10):2367-2372.42. Peters SA, Dogan S, Meijer R, et al. The use ofplaque score measurements to assess changes in ath-erosclerotic plaque burden induced by lipid-loweringtherapy over time: the METEOR study. J AtherosclerThromb. 2011;18(9):784-795.43. Bots ML, Hofman A, De Jong PTVM, GrobbeeDE. Common carotid intima-media thickness as an in-dicator of atherosclerosis at other sites of the carotidartery: the Rotterdam Study. Ann Epidemiol. 1996;6(2):147-153.44. Cook NR, Paynter NP. Performance of reclassi-fication statistics in comparing risk prediction models.Biom J. 2011;53(2):237-258.45. Whincup PH, Nightingale CM, Owen CG, et al.Ethnic differences in carotid intima-media thickness be-tween UK children of black African-Caribbean andwhite European origin. Stroke. 2012;43(7):1747-1754.




CLINICIAN’S CORNERCLINICAL REVIEW

Lipid-Modifying Therapies and Riskof PancreatitisA Meta-analysisDavid Preiss, MD, PhDMatti J. Tikkanen, MD, PhDPaul Welsh, PhDIan Ford, PhDLaura C. Lovato, MSMarshall B. Elam, MD, PhDJohn C. LaRosa, MDDavid A. DeMicco, DPharmHelen M. Colhoun, MDIlan Goldenberg, MDMichael J. Murphy, MD, FRCPThomas M. MacDonald, MD, FRCPTerje R. Pedersen, MD, PhDAnthony C. Keech, MD, PhDPaul M Ridker, MDJohn Kjekshus, MD, PhDNaveed Sattar, MD, PhDJohn J. V. McMurray, MD

PANCREATITIS HAS A CLINICAL

spectrum ranging from a mild,self-limiting episode to a severeor fatal event. Case reports and

pharmacoepidemiologic studies haveclaimed that statins may cause pancre-atitis,1-4 although few of these studiescomprehensively considered confound-ing factors. Very few large randomizedtrials of statin therapy have publisheddata on incident pancreatitis. Recently re-ported data from the Study of Heart andRenal Protection (SHARP), a trial com-paring combination therapy of simva-

statin and ezetimibe with placebo on car-diovascular events in patients withchronic kidney disease, demonstrated areduction in pancreatitis cases in pa-tients receiving simvastatin and ezeti-mibe, suggesting a possible protective as-sociation.5 In addition, statins reduce bile

CME available online atwww.jamaarchivescme.comand questions on p 821.

Author Affiliations are listed at the end of this article.Corresponding Author: David Preiss, MD, PhD, BHFGlasgow Cardiovascular Research Centre, Universityof Glasgow, 126 University Pl, Glasgow G12 8TA,United Kingdom ([email protected]).Clinical Review Section Editor: Mary McGraeMcDermott, MD, Contributing Editor. We encour-age authors to submit papers for consideration as aClinical Review. Please contact Mary McGraeMcDermott, MD, at [email protected].

Context Statin therapy has been associated with pancreatitis in observational stud-ies. Although lipid guidelines recommend fibrate therapy to reduce pancreatitis risk inpersons with hypertriglyceridemia, fibrates may lead to the development of gall-stones, a risk factor for pancreatitis.

Objective To investigate associations between statin or fibrate therapy and inci-dent pancreatitis in large randomized trials.

Data Sources Relevant trials were identified in literature searches of MEDLINE,EMBASE, and Web of Science (January 1, 1994, for statin trials and January 1, 1972,for fibrate trials, through June 9, 2012). Published pancreatitis data were tabulatedwhere available (6 trials). Unpublished data were obtained from investigators (22trials).

Study Selection We included randomized controlled cardiovascular end-point trialsinvestigating effects of statin therapy or fibrate therapy. Studies with more than 1000participants followed up for more than 1 year were included.

Data Extraction Trial-specific data described numbers of participants developingpancreatitis and change in triglyceride levels at 1 year. Trial-specific risk ratios (RRs)were calculated and combined using random-effects model meta-analysis. Between-study heterogeneity was assessed using the I2 statistic.

Results In 16 placebo- and standard care–controlled statin trials with 113 800 par-ticipants conducted over a weighted mean follow-up of 4.1 (SD, 1.5) years, 309 par-ticipants developed pancreatitis (134 assigned to statin, 175 assigned to control) (RR,0.77 [95% CI, 0.62-0.97; P=.03; I2=0%]). In 5 dose-comparison statin trials with39 614 participants conducted over 4.8 (SD, 1.7) years, 156 participants developedpancreatitis (70 assigned to intensive dose, 86 assigned to moderate dose) (RR, 0.82[95% CI, 0.59-1.12; P=.21; I2=0%]). Combined results for all 21 statin trials pro-vided RR 0.79 (95% CI, 0.65-0.95; P=.01; I2=0%). In 7 fibrate trials with 40 162participants conducted over 5.3 (SD, 0.5) years, 144 participants developed pancre-atitis (84 assigned to fibrate therapy, 60 assigned to placebo) (RR, 1.39 [95% CI,1.00-1.95; P=.053; I2=0%]).

Conclusion In a pooled analysis of randomized trial data, use of statin therapy wasassociated with a lower risk of pancreatitis in patients with normal or mildly elevatedtriglyceride levels.JAMA. 2012;308(8):804-811 www.jama.com



cholesterol content,6 which may theo-retically reduce the risk of developinggallstones, a risk factor for pancreatitis.

Hypertriglyceridemia has been re-ported to be the third most commoncause of pancreatitis.7 This has led to ma-jor guidelines for lipid-modifying thera-pies, including advice to commence tri-glyceride-lowering therapy, usuallyfibrates, in persons with moderate andsevere hypertriglyceridemia (above 400to 500 mg/dL [to convert to mmol/L,multiply by 0.0113]).8,9 However, high-quality evidence for this approach is lack-ing, and only observational data ex-ist.10,11 Indeed, there is concern thatfibrates might increase the risk of pan-creatitis in individuals with triglyceridelevels lower than those mentioned inguidelines.12 Fibrates increase the cho-lesterol concentration in bile and mayincrease the risk of gallstones.13,14 How-ever, few large randomized placebo-controlled trials of fibrate therapy havepublished data on pancreatitis.

Consequently,theassociationsbetweenbothtypesoflipid-modifyingtherapyandtheriskofpancreatitis areuncertain.Wetherefore examined the associations be-tween use of a statin or a fibrate and theincidence of pancreatitis by conductingcollaborativemeta-analysesofpublishedand unpublished data from the relevantlarge randomized clinical trials.

METHODSWegathereddata fromlargerandomizedend-point trials primarily designed toassess the effects of statin therapy (in-cluding both placebo- and standardcare–controlledtrialsplusintensive-dose/moderate-dose trials) or fibrate therapyon cardiovascular events. Inclusion cri-teria were trials with 1000 or more par-ticipantsexposedtorandomizedtherapywithaminimummeanfollow-upof1year,as inprevious largemeta-analysesofstat-in trials.15 We excluded trials conductedinpatientswithpreviousorgantransplan-tationor those receivinghemodialysis aswell as trials comparing combinationtherapy with placebo.

WesearchedMEDLINE,EMBASE,andWebofSciencedatabasesusingthetermsstatin, HMG CoA reductase inhibitor, and

fibrate and also names of individual stat-ins (atorvastatin, fluvastatin, lovastatin,pitavastatin,pravastatin,rosuvastatin,sim-vastatin) and fibrates (bezafibrate, cipro-fibrate,clofibrate, fenofibrate,gemfibrozil)as title words and keywords, limited tostudiesdefinedasrandomizedcontrolledtrials, to identify relevant studies per-formed in adult patients (initial searchon October 28, 2011; search updatedJune 9, 2012) and published from Janu-ary 1, 1972 (fibrate trials), or January 1,1994 (statin trials), until June 9, 2012(FIGURE 1), without language restric-tions.Reference lists for thestudies iden-tified in the literature search weresearched for additional studies. The USFood and Drug Administration websitewas also searched for trial reports con-taining relevant data. Abstracts, manu-scripts, and reports were reviewed in-dependently by 2 readers (D.P., P.W.)inanunblindedfashion.Athirdreviewer(N.S.) settleddiscrepancies. In thesmallnumberof trials inwhichpublisheddataregarding incident pancreatitis andchange in triglyceride levels were avail-able, these data were tabulated. In themajority of trials in which no relevantdata were available, trial investigatorswerecontactedwitharequest toprovidethe required information.

Afterthefullarticleswerereviewedanddatawerereceivedfromcollaborators,21statin trials5,16-36 (TABLE 1) and 7 fibratetrials12,37-43 (TABLE2)wereincludedintheanalyses.BecauseunpublisheddataweremadeavailableforboththeHelsinkiHeartStudy40 and its smaller ancillary study44

conductedinsimilargroupsofparticipantsrandomizedtothesametherapiesoverthesame follow-uptimes, these resultswerecombined as a single overall study.

Data Sources

Published data for incident pancreatitiswereavailablefrom2statintrials5,22,36 and4 fibrate trials.12,37-39,41 Unpublisheddatawerecollectedfrom19statintrials16-21,23-35

and 3 fibrate trials.40,42,43 To examinewhethertherewasarelationshipbetweentheextentoftriglycerideloweringbetweenactive and control therapy groups in thetrialsandriskofpancreatitis,wecollecteddataonaveragechangeintriglyceridelev-

elsat1year.APRISMAchecklistwaspro-vided to the journal at the time of manu-script submission.45

Quality Assessment

Two authors (D.P., P.W.) used an estab-lished tool, the Jadad score, to indepen-dentlyevaluate thequalityofeach trial.46

TheJadadscoreisdesignedtoassesstrialswithregardtomethodofrandomization,whether the trial is double-blinded, andwhether withdrawals/dropouts are de-scribed, resulting in a score of up to 5points.A thirdreviewer(N.S.)wasavail-able to resolveanydisagreementbycon-sensus and discussion.

End Points

A patient was considered to have devel-oped pancreatitis during the trial if this

Figure 1. Literature Search

28 Trials included in meta-analysis21 Statin7 Fibrate

39 Full-text articles reviewed fortrial suitability

4082 Records screened

4426 Records identified4416 From databases

10 From other sources

33 Trials identified as suitable6 Had published data regarding

incident pancreatitis27 Unpublished data regarding

incident pancreatitis requested

5 Trials excluded (no data availableor provided)

344 Excluded (duplicates)

6 Excluded3 Had <1000 participants1 Follow-up <1 y1 Surrogate end point1 Unsuccessful randomization

4043 Excluded a

1650 Surrogate marker endpoint, <1000 participants,or follow-up <1 y

1388 Not randomized (baselinepaper, post hoc analysis,or review)

590 Nonstatin or fibrateintervention

415 Other

aMost records excluded for more than 1 reason, withonly the strongest reason recorded.

LIPID-MODIFYING THERAPIES AND RISK OF PANCREATITIS



was recorded as an adverse event or se-riousadverseevent.This informationwasidentified using different approaches

across the trials,1 namely text wordsearches of adverse event reports, in-cludingself-reportedhospitalizationdata,

for pancreatitis2; Medical Dictionary forRegulatory Activities event classifica-tion3; and International Classification of

Table 1. Baseline Data From 21 Large Statin Trials

Source

No.

Treatment,Active/Control

Follow-up, y

Trial Population(Triglyceride

Inclusion Criteria) Age, y

Triglycerides

Statin Control

Baseline,Mean (SD),

mg/dLDifference

at 1 y, %

Placebo- and Standard Care–Controlled Trials4S,16 1994 2223 2221 Simvastatin (10-40 mg)/

placebo5.4a Angina or previous MI

(triglycerides �222 mg/dL)134 (45) 18

WOSCOPS,17 1995 3302 3293 Pravastatin (40 mg)/placebo 4.9 Male, hypercholesterolemia,no history of MI (NR)

55 164 (69) 15

CARE,18 1996 2081 2078 Pravastatin (40 mg)/placebo 5.0a MI in previous 3-20 mo(triglycerides �350 mg/dL)

59 156 (61) 14b

AFCAPS/TexCAPS,19 1998 3304 3301 Lovastatin (20-40 mg)/placebo 5.2 Average cholesterol levels,no CVD (triglycerides�400 mg/dL)

58 181 (75) 14

LIPID,20 1998 4512 4502 Pravastatin (40 mg)/placebo 6.1 Hospitalization for unstableangina or previous MI(triglycerides �445 mg/dL)

62a 140a 11b

GISSI Prevenzione,21 2000 2138 2133 Pravastatin (20 mg)/standard care

2.0a Recent MI (NR) 166 (89) −4

HPS,5,22 2002 10 269 10 267 Simvastatin (40 mg)/placebo 5.4 CVD or diabetes (NR) 65 187 (125) 19PROSPER,23 2002 2891 2913 Pravastatin (40 mg)/placebo 3.3 Age 70-82 y with CVD or risk

factors (triglycerides�534 mg/dL)

75 138 (62) 17

GREACE,24 2002 800 800 Atorvastatin (to achieveLDL-C �100 mg/dL)/standard care

3.0 CHD (triglycerides �400 mg/dL) 59 181 28

ASCOT-LLA,25 2003 5168 5137 Atorvastatin (10 mg)/placebo 3.3a Hypertension, no CHD(triglycerides �400 mg/dL)

63 147 (80) 23

CARDS,26 2004 1428 1410 Atorvastatin (10 mg)/placebo 3.9a Type 2 diabetes mellitus,no CVD (triglycerides�603 mg/dL)

62 173 (97) 21

ASPEN,27 2006 1211 1199 Atorvastatin (10 mg)/placebo 4.0 Diabetes mellitus(triglycerides �600 mg/dL)

61 146a 14c

MEGA,28 2006 3866 3966 Pravastatin (10-20 mg)/no treatment

5.3 Hypercholesterolemia,no previous CHDor stroke (NR)

58 148 (83) 6

CORONA,29 2007 2514 2497 Rosuvastatin (10 mg)/placebo 2.7a Systolic heart failure (NR) 73 178 (114) 24d

JUPITER,30 2008 8901 8901 Rosuvastatin (20 mg)/placebo 1.9a No CVD, no diabetes,hsCRP �2.0 mg/L(triglycerides �500 mg/dL)

66a 118 (86-169)a 17

GISSI-HF,31 2008 2285 2289 Rosuvastatin (10 mg)/placebo 3.9a Chronic heart failure (NR) 68 NA NAIntensive- vs Moderate-Dose Trials

PROVE-IT TIMI 22,32 2004 2099 2063 Pravastatin (40 mg)/atorvastatin (80 mg)

2.0 Recent hospitalizationfor ACS (NR)

58 156a 21a

A to Z,33 2004 2265 2234 Placebo � simvastatin (20 mg)/simvastatin (40-80 mg)

2.0a Recent hospitalizationfor ACS (NR)

61a 149 (116-199)a 6

TNT,34 2005 4995 5006 Atorvastatin (80 mg)/atorvastatin (10 mg)

4.9a Stable CHD (triglycerides�600 mg/dL)

61 151 (71) NA

IDEAL,35 2005 4439 4449 Atorvastatin (80 mg)/simvastatin (20-40 mg)

4.8a Previous MI (triglycerides�600 mg/dL)

62 149 23

SEARCH,5,36 2010 6031 6033 Simvastatin (80 mg)/simvastatin (20 mg)

6.7 Previous MI (NR) 64 169 (107) 9

Total 76 722 76 692 4.3 (1.6)

Abbreviations: ACS, acute coronary syndromes; AFCAPS/TexCAPS, Air Force/Texas Coronary Atherosclerosis Prevention Study; ASCOT-LLA, Anglo-Scandinavian Cardiac OutcomesTrial–Lipid Lowering Arm; ASPEN, Atorvastatin Study for Prevention of Coronary Heart Disease Endpoints in Non–Insulin-Dependent Diabetes Mellitus; A to Z, Aggrastat to Zocor;CARDS, Collaborative Atorvastatin Diabetes Study; CHD, coronary heart disease; CORONA, Controlled Rosuvastatin Multinational Trial in Heart Failure; CVD, cardiovascular disease;GISSI-HF, SEARCH Study of the Effectiveness of Additional Reductions in Cholesterol and Homocysteine; GISSI Prevenzione, Gruppo Italiano per lo Studio della Sopravvivenza nell’Insufficienzacardiaca Prevenzione; GREACE, Greek Atorvastatin and Coronary Heart Disease Evaluation; HPS, Heart Protection Study; hsCRP, high-sensitivity C-reactive protein; IDEAL, Incre-mental Decrease in Events Through Aggressive Lipid Lowering; JUPITER, Justification for the Use of Statins in Prevention: an Intervention Trial Evaluating Rosuvastatin; LDL-C, low-density lipoprotein cholesterol; LIPID, Long-term Intervention With Pravastatin in Ischaemic Disease; MEGA, Management of Elevated Cholesterol in the Primary Prevention Group ofAdult Japanese Study Group; MI, myocardial infarction; NA, not available; NR, not reported (no triglycerides inclusion or exclusion criteria specified); PROSPER, Prospective Study ofPravastatin in the Elderly at Risk; PROVE-IT TIMI 22, Pravastatin or Atorvastatin Evaluation and Infection Therapy; TNT, Treating to New Targets; WOSCOPS, West of Scotland Coro-nary Prevention Study; 4S, Scandinavian Simvastatin Survival Study.

SI conversion factors:To convert triglyceride values mmol/L, multiply by 0.0113; to convert hsCRP values to nmol/L, multiply by 9.524.aMedian or median (interquartile range).bAverage difference over 5 years.cDifference at end of trial.dDifference at 3 months.




Diseases classifications (10th revision:K85, K86.0, K86.1; ninth revision: 577.0,577.1), according to the preference ofeach trial’s investigators. All reports ofpancreatitis were included, regardless ofsuggested etiology (information regard-ing alcohol intake was not available) orwhether the condition was described asacute, chronic, or neither, based on therationale that such additional data mayhave been largely absent or variably re-ported across trials.


Toidentifypotentialassociationsof lipid-modifying therapies with the risk of de-veloping pancreatitis, we calculated riskratios(RRs)as theratioofcumulative in-cidence and 95% CIs from the availabledata for all trial participants at baselineand for those who developed pancreati-tis during trial follow-up. Study-specificRRs were pooled using a random-effectsmodelmeta-analysisas thepreferableap-proachtomanagepotentialbetween-studyheterogeneity that may have been intro-ducedby thedifferingmethods for iden-tifyingparticipantswithincidentpancre-atitis available in the trials and differenttrialpopulations.Fortrialswithnoevents

with randomized or control therapy, anominal amount (0.5 cases) was addedto the results for both trial groups.

Statistical heterogeneity across stud-ies was quantified using both the �2 (orCochran Q statistic) and I2 statistics, withP� .10 considered statistically nonsig-nificant. The I2 statistic is derived fromthe Q statistic ([Q−df/Q]�100) and pro-vides a measure of the proportion of theoverall variation attributable to between-study heterogeneity.47

Placebo- and standard care–con-trolled statin trials plus intensive-dose/moderate-dose statin trials were ana-lyzed both separately (with comparisonof analyses by fixed-effect inverse-variance method) and in a combinedanalysis. In sensitivity analyses, onlytrials with previously published pancre-atitis data were examined; fixed-effectsmodel meta-analyses were also per-formed. We assessed the potential forpublication bias through formal statis-tical testing, namely, funnel plots andEgger tests. To evaluate the potential re-lationship between the associations oflipid-modifying agents with incidentpancreatitis and relative reductions in tri-glyceride levels achieved at 1 year using

statins and fibrates, respectively, ran-dom-effects meta-regression analyseswere performed.

All P values were 2-sided, and P� .05was considered statistically significant forthe meta-analyses and meta-regressionanalyses. Analyses were conducted usingStata version 10.1 (StataCorp).

RESULTSStatin Therapy and Pancreatitis

Twenty-onerandomizedclinical trialsofstatin therapy, 2 with published data re-gardingincidentpancreatitisand19withunpublished data, provided data on153414participantsoveraweightedmeanfollow-up period of 4.3 (SD, 1.6) years.Baselineaverage triglyceride levels in thetrials varied from 118 mg/dL to 187 mg/dL.Trialswereofhighquality,withame-dian Jadad score of 5 (range, 3-5) and100% agreement between reviewers.

In 16 placebo- and standard care–controlled statin trialswith113 800par-ticipants conducted over 4.1 (SD, 1.5)years,309participants(0.27%)developedpancreatitis (134 assigned to statin, 175assigned to control) (RR, 0.77 [95% CI,0.62-0.97; P=.03]) (Table 1, FIGURE 2).Thisrepresentsanumberneededtotreat

Table 2. Baseline Data From Trials Comparing Fibrate Therapy With Placebo

Source

No.

Treatment, Active/ControlFollow-

up, yTrial Population (Triglyceride

Inclusion Criteria) Age, y

Triglycerides

Fibrate Control

Baseline,Mean (SD),

mg/dLDifference

at 1 y, %Coronary Drug Project,37,38

1975b1103 2789 Clofibrate/placebo 6.2 Male, previous MI (NR) 184 25

WHO-COOP,39 1978b,d 5331 5296 Clofibrate/placebo 5.3 Male, upper third of cholesterol range(NR)

46 NA NA

HHS,40,44 1987c 2362 2347 Gemfibrozil/placebo 5.0 Male, no CHD or possible symptomsof CHD (NR)

47 177 (119) 35

VA-HIT,41 1999b 1264 1267 Gemfibrozil/placebo 5.1a Male, CHD (triglycerides �300mg/dL)

64 161 (68) 31

BIP,42 2000 1548 1542 Bezafibrate/placebo 6.2 Previous MI or stable angina(triglycerides �300 mg/dL)

60 145 (51) 21e

FIELD,12 2005 4895 4900 Fenofibrate/placebo 5.0a Diabetes mellitus, not taking statin(triglycerides 89-445 mg/dL)

62 174 (78) 30

ACCORD Lipid,43 2010 2765 2753 Simvastatin � fenofibrate/simvastatin � placebo

4.7 Diabetes mellitus, CVD or risk factors(triglycerides �750 mg/dL withno lipid-lowering therapy; �400mg/dL with therapy)

62 162 (113-229)a 20

Total 19 268 20 894 5.3 (0.5)

Abbreviations: ACCORD, Action to Control Cardiovascular Risk in Diabetes; BIP, Bezafibrate Infarction Prevention; CHD, coronary heart disease; CVD, cardiovascular disease; FIELD, Feno-fibrate Intervention and Event Lowering in Diabetes; HHS, Helsinki Heart Study; MI, myocardial infarction; NA, not available; NR, not reported (no triglycerides inclusion or exclusion criteriaspecified); VA-HIT, Veterans Affairs Cooperative Studies Program High-Density Lipoprotein Cholesterol Intervention Trial; WHO-COOP, World Health Organization Co-operative Trial.

SI conversion factor: To convert triglyceride values to mmol/L, multiply by 0.0113.aMedian or median (interquartile range).bOnly fatal cases of pancreatitis available.c Includes cases from both the HHS and its ancillary study (age, baseline triglyceride levels, and % difference in triglyceride levels are weighted means).d Includes cases during the trial and during first year after the trial.eAverage difference during trial.




of1175(95%CI,693-9195)over5years.Therewaslimitedheterogeneitybetweentrials for incident pancreatitis (�2=9.11;I2=0%).

In5dose-comparisonstatintrialswith39 614 participants conducted over 4.8(SD,1.7)years,156participants(0.39%)developedpancreatitis(70assignedtoin-tensive dose, 86 assigned to moderatedose) (RR, 0.82 [95% CI, 0.59-1.12;P=.21]) (Table 1, Figure 2). There wasagainlimitedheterogeneitybetweenthesetrials for incident pancreatitis (�2=1.29;I2=0%).

There was no evidence of statisticalheterogeneity between the analyses ofplacebo-controlled trials and intensive-dose/moderate-dose trials (P=.79 forinteraction).

In the combined data set of 21 trials,465participants(0.30%)developedpan-creatitis (of whom 204 were assigned tostatin therapy or intensive-dose statin

therapyand261wereassignedtoplacebo,standard care, or moderate-dose statintherapy,respectively)(RR,0.79[95%CI,0.65-0.95; P=.01; �2=10.48; I2=0%])(Table1,Figure2).Thisrepresentsanum-berneededtotreatof1187(95%CI,731-4768)over5years.Therewasnoevidenceof publication bias (P=.83) (eFigure 1A,available at http://jama.com). Meta-regression analysis found no relation-ship across the trials between risk of pan-creatitis and reduction in triglyceridelevels at 1 year, although this analysis wasof limited value given the limited statis-tical heterogeneity between trial-specific RRs (P=.23) (eFigure 2A).

Using a fixed-effects model ap-proach produced results (RR, 0.79 [95%CI 0.65-0.95; P=.01]) identical to thoseof the random-effects model. In a sen-sitivity analysis of only the 2 trials withpublished data,22,36 122 participants(0.37%) developed pancreatitis (52/

16 300 assigned to statin therapy or in-tensive-dose therapy, 70/16 300 as-signed to placebo or moderate-dosestatin therapy) (RR, 0.74 [95% CI, 0.52-1.07; P=.11; �2=0.30; I2=0%]).

Fibrate Therapy and Pancreatitis

Sevenrandomizedclinical trialsof fibratetherapy(4withpublisheddataand3withunpublisheddataregardingincidentpan-creatitis)provideddataon40 162partici-pantsoveraweightedmeanfollow-uppe-riodof5.3(SD,0.5)years.Baselineaveragetriglyceride levels inthetrialsvariedfrom145 mg/dL to 184 mg/dL. Trials were ofhigh quality, with a median Jadad scoreof5(range,5-5)and100%agreementbe-tween reviewers. During this time, 144participants (0.36%) developed pancre-atitis (84 assigned to fibrate therapy, 60assigned to placebo) (RR, 1.39 [95% CI,1.00-1.95;P=.053])(Table2,FIGURE3).Thisrepresentsanumberneededtoharm

Figure 2. Meta-analysis of Incident Pancreatitis in 21 Large Statin Trials

Control

Favors Statin Favors Control

0.01 101.0 1000.1

Risk Ratio (95% CI)

Cases Total

Statin

Cases TotalSourcePlacebo- and standard care–controlled

Risk Ratio(95% CI)Weight, %

7 22215 22234S,16 1994 0.71 (0.23-2.25)2.605 32932 3302WOSCOPS,17 1995 0.40 (0.08-2.06)1.28

17 207815 2081CARE,18 1996 0.88 (0.44-1.77)7.0710 33017 3304AFCAPS/TexCAPS,19 1998 0.70 (0.27-1.84)3.6723 450212 4512LIPID,20 1998 0.52 (0.26-1.05)7.02

2 21330 2138GISSI Prevenzione,21 2000 0.20 (0.01-4.16)0.3741 10 26733 10 269HPS,22 2002 0.80 (0.51-1.27)16.2911 29135 2891PROSPER,23 2002 0.46 (0.16-1.32)3.07

0 8000 800GREACE,24 2002 1.00 (0.02-50.46)0.2216 51378 5168ASCOT-LLA,25 2003 0.50 (0.21-1.16)4.76

4 14105 1428CARDS,26 2004 1.23 (0.33-4.61)1.985 11993 1211ASPEN,27 2006 0.59 (0.14-2.49)1.673 39663 3866MEGA,28 2006 1.03 (0.21-5.09)1.347 249712 2514CORONA,29 2007 1.70 (0.67-4.33)3.94

17 890117 8901JUPITER,30 2008 1.00 (0.51-1.96)7.587 22897 2285GISSI-HF,31 2008 1.00 (0.35-2.86)3.12

Subtotal: I2 = 0.0%, P = .87 0.77 (0.62-0.97)65.98

Control

Favors Statin Favors Control

0.01 101.0 1000.1

Risk Ratio (95% CI)

Cases Total

Statin

Cases TotalSourceIntensive vs moderate dose

Risk Ratio(95% CI)Weight, %

1 20631 2099PROVE-IT TIMI 22,32 2004 0.98 (0.06-15.72)0.452 22343 2265A to Z,33 2004 1.48 (0.25-8.86)

40 500633 4995TNT,34 2005 0.83 (0.52-1.31)16.0514 444914 4439IDEAL,35 2005 1.00 (0.48-2.10)6.2429 603319 6031SEARCH,36 2010 0.66 (0.37-1.17)10.22

Subtotal: I2 = 0.0%, P = .86 0.82 (0.59-1.12)34.02

Overall: I2 = 0.0%, P = .96 0.79 (0.65-0.95)100.00

For abbreviations, see Table 1. Size of data markers indicates relative weight of the study (from random-effects analysis).




of 935 (95% CI, 388 to �50 000) over5years.Therewas limitedheterogeneitybetween trials for incident pancreatitis(�2=4.48; I2=0%). Likewise, there wasno evidence of publication bias (P=.59)(eFigure 1B). Meta-regression analysisfound no relationship across the trialsbetween risk of pancreatitis and reduc-tion in triglyceride levels at 1 year acrossthe trials (P= .81) (eFigure 2B), al-though this analysis was of limited valuegiven the limited statistical heteroge-neity between trial-specific RRs and thesimilar relative reductions in triglycer-ide levels achieved across the trials.

Using a fixed-effects model approachproduced results identical to thoseachievedusingtherandom-effectsmodel(RR,1.39[95%CI,1.00-1.95;P=.053]).Inasensitivityanalysisofonly the4trialswithpublisheddata,12,37,39,4169participants(0.26%)developedpancreatitis(44/12593assignedto fibrate therapy,25/14 252as-signed to placebo) (RR, 1.75 [95% CI,1.07-2.86; P=.03; �2=1.19; I2=0%]).

COMMENTThis report of pooled randomized trialdata demonstrates that use of statintherapy was associated with a reduc-tion in the number of patients devel-oping pancreatitis. Broadly similar re-sults were obtained for statin comparedwith placebo as well as for intensive-dose statin therapy compared withmoderate-dose therapy, in keeping witha dose-dependent association. How-ever, we did not demonstrate an asso-ciation between use of fibrate therapyand risk of pancreatitis.

Previously published case reports andobservational pharmacoepidemiologicstudies have demonstrated an associa-tion between statin therapy and in-creased risk of pancreatitis.1-4 However,such analyses are susceptible to bias byunmeasured confounders and to con-founding by indication. The presentanalysis, however, indicates that statintherapy may be associated with a re-duced risk of pancreatitis overall. Al-though we cannot completely excludethe possibility that statin therapy maylead toveryoccasional idiosyncratic casesof pancreatitis, the randomized trial dataappear reassuring. Unlike fibrates, stat-ins are not known to increase the risk ofdeveloping gallstones.48 Studies show-ing both a reduction in bile cholesterollevels and an association with reducedrisk of gallstones with statin therapy sug-gest the possibility of a protective ef-fect.6,49 Furthermore, studies con-ducted in animal models suggest thatstatin therapy may be beneficial in bothestablishedacutepancreatitis andchronicpancreatitis.50-52

Major guidelines of lipid-modifyingtherapysuchas theNationalCholesterolEducation Program Third Report of theExpert Panel on Detection, Evaluation,andTreatmentofHighBloodCholesterolin Adults (NCEP ATP III)8 and the Na-tionalInstituteforHealthandClinicalEx-cellence (NICE) Type 2 Diabetes guide-line9suggesttheadditionoffibratetherapyin patients with moderately elevated tri-glyceride levels andabove(�400mg/dLand �500 mg/dL, respectively). This isbased on the rationale that hypertriglyc-

eridemiaisawell-recognizedcauseofpan-creatitis andthat loweringof triglyceridelevels should be clinically beneficial.7

However, no convincing trial data existto support use of any agents for preven-tionofpancreatitisinthisclinicalsituation.ParticipantsintheCoronaryDrugProjectassignedtoclofibratewereat50%higherrisk of developing cholelithiasis or cho-lecystitis than those receiving placebo,13

andgallstonesareawell-recognizedcauseof pancreatitis. In addition, it has beendemonstratedinsmallclinicalstudiesthatboth fenofibrate—a fibrate thought lesslikelytocausegallstones—andbezafibrateincrease the cholesterol content of bile,thereby theoretically increasing the riskofdevelopinggallstones.14,53FollowingtheCoronaryDrugProject,otherlargefibratetrialsdidnot findasignificant increase inthe incidence of gallbladder disease, al-though the total number of cases wassmall.40,41,43 Ouranalysisdidnotdemon-strate an association between fibratetherapyandriskofpancreatitis,althoughthe analysis may have lacked statisticalpowertoshowanincreasedriskinpatientswith slightly elevated triglyceride levels(the range at baseline in the trials we ex-amined was 145-184 mg/dL). It remainspossible,however,thatfibratesmighthaveadifferentneteffectinpatientswithhighertriglyceride levels.

Although the present results for bothstatinsand fibrates shouldbeconsideredhypothesis-generatingandthenumberofpancreatitis cases was small in this trialpopulationat lowriskofpancreatitis, theanalysis raises questions regarding thechoice of lipid-modifying agents in pa-

Figure 3. Meta-analysis of Incident Pancreatitis in 7 Large Fibrate Trials

Control

Favors Fibrate Favors Control

0.01 101.0 1000.1

Risk Ratio (95% CI)

Cases Total

Fibrate

Cases TotalSourceRisk Ratio(95% CI)Weight, %

1 27890 1103CORONARY Drug Project,37 1975 1.110 52963 5331WHO-COOP,39 1978 1.296 23473 2362HHS,40 1987 5.891 12671 1264VA-HIT,41 1999 1.476 15426 1548BIP,42 2000 8.82

23 490040 4895FIELD,12 2005 42.8323 275331 2765ACCORD Lipid,43 2010 38.59

Overall: I2 = 0.0%, P = .61

0.84 (0.03-20.71)6.95 (0.36-134.66)0.50 (0.12-1.99)1.00 (0.06-16.04)1.00 (0.32-3.10)1.74 (1.04-2.91)1.34 (0.78-2.31)

1.39 (1.00-1.95)100.00

For abbreviations, see Table 2. Size of data markers indicates relative weight of the study (from random-effects analysis).




tientswithhypertriglyceridemia.Inthosewith slightly elevated triglyceride levels,statins appear better supported by theavailabledatathanfibrates forpreventingpancreatitis.Lifestylemodificationsalsoremain important to improve lipid pro-files in such individuals. Inpatientswithseverehypertriglyceridemia, a trial com-paring fibratesandstatins forpreventingpancreatitiswouldbeclinicallyvaluable.

Strengths of this meta-analysis arethat the analysis was conducted usingdata from randomized trials, whichavoids most of the potential bias of un-measured confounders encountered inobservational studies, and that we wereable to include data from almost all ofthe relevant trials, both published andunpublished, thereby maximizingpower and providing the best answerpossible with existing data.

Thismeta-analysisalsohasseverallimi-tations. First, pancreatitis was not a pre-specified end point in the trials, whichwere primarily designed to assess the ef-fectof lipid-modifyingtherapyoncardio-vascular events.However, limitedstatis-tical heterogeneity between trial resultsfor statins and fibrates, plus evidence ofadose-dependentassociation for statins,providesconfidence in the findings. Sec-ond, the occurrence of pancreatitis wasnot recorded inastandardizedway,withresultantvariationbetweentrials.There-fore these results, especially for fibratetherapy when there were relatively fewevents dominated by 2 trials,12,43 shouldbe interpreted with caution.

Third,because itwas feltunlikely thatthecauseofpancreatitiswouldhavebeenconsistently recorded inanaccuratewayacross trials, we were unable to examinespecific causes such as gallstones. Like-wise, we were unable to separate reportsof pancreatitis into acute and chroniccases. However, given that the majorityof trials used the presence of hepatobili-ary disease as an exclusion criterion, it ishighly likely that themajorityofcases in-cluded in this report represent de novoacute pancreatitis. This is supported byevidence from SHARP.5 Fourth, we didnothaveaccess to individual-participantdata, which may have reduced our abil-ity to identify any relationship with the

extentof triglyceride lowering.Fifth,be-cause the trials tendedtoexcludepartici-pantswithmarkedhypertriglyceridemia,thesefindingsmaynotnecessarilybegen-eralizabletothatspecificgroupofpatients.

In summary, pooled analyses of ran-domized trial data suggest that statintherapy is associated with a reduction intheriskofpancreatitisinpatientswithnor-malormildlyelevatedtriglyceride levels.Author Affiliations: BHF Glasgow Cardiovascular Re-search Centre (Drs Preiss, Welsh, Sattar, andMcMurray) and Robertson Centre for Biostatistics (DrFord), University of Glasgow, Glasgow, United King-dom; University of Helsinki and Division of Cardiol-ogy, Helsinki University Hospital, and Folkhalsan Re-search Center, Helsinki, Finland (Dr Tikkanen); Divisionof Public Health Sciences, Department of Biostatisti-cal Science, Wake Forest University School of Medi-cine, Winston-Salem, North Carolina (Ms Lovato);Memphis Veterans Affairs Medical Center, Mem-phis, Tennessee (Dr Elam); SUNY Health Science Cen-ter at Brooklyn, New York, New York (Dr LaRosa);Pfizer Global Pharmaceuticals, New York, New York(Dr DeMicco); Medical Research Institute (Dr Col-houn), Department of Biochemical Medicine (Dr Mur-phy), and Medicines Monitoring Unit, Division of Medi-cal Sciences (Dr MacDonald), University of Dundee,Dundee, United Kingdom; Cardiac Rehabilitation In-stitute and Israeli Society for the Prevention of HeartAttacks, Leviev Heart Center, Sheba Medical Center,Tel Hashomer, Israel (Dr Goldenberg); University ofOslo and Centre for Preventative Medicine, Oslo Uni-versity Hospital, Ulleval, Oslo, Norway (Dr Peder-sen); NHMRC Clinical Trials Centre, University of Syd-ney, Sydney, Australia (Dr Keech); Harvard MedicalSchool, Cardiovascular Medicine, Brigham and Wom-en’s Hospital, Boston, Massachusetts (Dr Ridker); andDepartment of Cardiology, Oslo University HospitalRikshospitalet, Oslo, Norway (Dr Kjekshus).Author Contributions: Dr Preiss had full access to all ofthe data in the study and takes responsibility for the in-tegrity of the data and the accuracy of the data analysis.Study concept and design: Preiss, Tikkanen, McMurray.Acquisition of data: Preiss, Tikkanen, Ford, Lovato,Elam, LaRosa, Demicco, Colhoun, Goldenberg,Pedersen, Keech, Ridker, Kjekshus, McMurray.Analysis and interpretation of data: Preiss, Tikkanen,Welsh, LaRosa, Demicco, Colhoun, Goldenberg,Murphy, MacDonald, Keech, Ridker, Sattar, McMurray.Drafting of the manuscript: Preiss, Elam, LaRosa, Sattar,McMurray.Critical revision of the manuscript for important in-tellectual content: Preiss, Tikkanen, Welsh, Ford,Lovato, Elam, LaRosa, Demicco, Colhoun, Goldenberg,Murphy, MacDonald, Pedersen, Keech, Ridker,Kjekshus, McMurray.Statistical analysis: Preiss.Obtained funding: Pedersen.Administrative, technical, or material support: Welsh,Ford, Demicco, Goldenberg, Keech, Ridker.Study supervision: Ford, LaRosa, Pedersen, Sattar,McMurray.Conflict of Interest Disclosures: All authors have com-pleted and submitted the ICMJE Form for Disclosureof Potential Conflicts of Interest. The majority of trialsdiscussed in this article were funded partly or whollyby industry, and Drs Tikkanen, Ford, Elam, LaRosa,DeMicco, Colhoun, Goldenberg, Pedersen, Keech, Rid-ker, Kjekshus, and McMurray and Ms Lovato each re-ported serving as an investigator in at least 1 of thetrials. Dr Tikkanen reported receiving honoraria fromPfizer and consultant fees from Amgen Inc. Dr Elamreporting serving as a consultant, speaker, or both for

Abbott/Solvay, Merck Schering Plough, and PfizerCanada. Dr LaRosa reported receiving consultancy feesfrom Pfizer and AstraZeneca and participating in clini-cal trials funded by Pfizer. Dr Colhoun reported re-ceiving honoraria for advisory board participation andspeaker fees from Pfizer. Dr Pedersen reported re-ceiving speakers honoraria, consulting fees, or re-search grants from Merck, AstraZeneca, AMGEN,Roche, and Novartis. Dr Keech reported receivinghonoraria and research or travel grants from Abbott,Merck Sharpe & Dohme, Bristol-Myers Squibb, Novar-tis, Eli Lilly, Pfizer, Roche Diagnostics, Solvay, and As-traZeneca. Dr Ridker reported receiving research grantsupport from AstraZeneca and Novartis; receiving con-sultancy fees from Merck, Genzyme, Vascular Bio-genics, ISIS, and Boston Diagnostics; and being listedas a co-inventor on patents held by the Brigham andWomen’s Hospital that relate to the use of inflamma-tory biomarkers in cardiovascular disease and diabe-tes that have been licensed to AstraZeneca and Sie-mens. Dr Sattar reported consulting for and receivinglecture fees from Merck, Pfizer, and AstraZeneca andreceiving research grant support from Pfizer. No otherauthors reported disclosures.Funding/Support: This project was not supported byexternal funding. Dr Welsh is supported by British HeartFoundation fellowship grant FS/10/005/28147.Role of the Sponsor: The British Heart Foundation hadno role in the design and conduct of the study; the col-lection, analysis, and interpretation of the data; or thepreparation, review, or approval of the manuscript.Online-Only Material: eFigures 1 and 2 are availableat http://www.jama.com.Additional Contributions: We are grateful to the in-vestigators from the following trials for providingunpublished data: Helsinki Heart Study, Bezafibrate In-farction Prevention Study, Action to Control Cardio-vascular Risk in Diabetes Lipid Study, Scandinavian Sim-vastatin Survival Study, West of Scotland CoronaryPrevention Study, Air Force/Texas Coronary Athero-sclerosis Prevention Study, Long-term Intervention WithPravastatin in Ischaemic Disease Study, Gruppo Ital-iano per lo Studio della Sopravvivenza nell’Insufficienzacardiaca (GISSI) Prevenzione, Prospective Study ofPravastatin in the Elderly at Risk, Greek Atorvastatin andCoronary Heart Disease Evaluation Study, Anglo-Scandinavian Cardiac Outcomes Trial–Lipid LoweringArm, Collaborative Atorvastatin Diabetes Study, Prava-statin or Atorvastatin Evaluation and Infection TherapyStudy, Aggrastat to Zocor Study, Treating to New Tar-gets Study, Incremental Decrease in Events Through Ag-gressive Lipid Lowering Study, Atorvastatin Study forPrevention of Coronary Heart Disease Endpoints in Non–Insulin-Dependent Diabetes Mellitus, Management ofElevated Cholesterol in the Primary Prevention Groupof Adult Japanese Study Group, Controlled Rosuvasta-tin Multinational Trial in Heart Failure, Justification forthe Use of Statins in Prevention: an Intervention TrialEvaluating Rosuvastatin, and GISSI-Heart Failure. Bristol-Myers Squibb provided data for the Cholesterol and Re-current Events Trial.

REFERENCES

1. Van Woerkom RC, Adler DG. Report of simultane-ous acute pancreatitis and acute hepatitis in a patienttaking ezetimibe/simvastatin. J Clin Lipidol. 2010;4(4):314-315.2. Badalov N, Baradarian R, Iswara K, Li J, SteinbergW, Tenner S. Drug-induced acute pancreatitis: an evi-dence-based review. Clin Gastroenterol Hepatol. 2007;5(6):648-661.3. Singh S, Loke YK. Statins and pancreatitis: a system-atic review of observational studies and spontaneous casereports. Drug Saf. 2006;29(12):1123-1132.4. Tsigrelis C, Pitchumoni CS. Pravastatin: a potentialcause for acute pancreatitis. World J Gastroenterol. 2006;12(43):7055-7057.




5. USFoodandDrugAdministration (FDA). Endocrino-logic andMetabolicDrugsAdvisoryCommitteeBriefingDocument:SHARP(StudyofHeartandRenalProtection).FDA website. http://www.fda.gov/downloads/AdvisoryCommittees/CommitteesMeetingMaterials/Drugs/EndocrinologicandMetabolicDrugsAdvisoryCommittee/UCM277652.pdf. 2011. Accessed June 6,2012.6. Duane WC, Hunninghake DB, Freeman ML, PoolerPA, Schlasner LA, Gebhard RL. Simvastatin, a competi-tive inhibitor of HMG-CoA reductase, lowers choles-terol saturation index of gallbladder bile. Hepatology.1988;8(5):1147-1150.7. Ewald N, Hardt PD, Kloer HU. Severe hypertriglyc-eridemia and pancreatit is: presentation andmanagement. Curr Opin Lipidol. 2009;20(6):497-504.8. National Cholesterol Education Program. Third Re-port of the Expert Panel on Detection, Evaluation, andTreatment of High Blood Cholesterol in Adults (AdultTreatment Panel III). http://www.nhlbi.nih.gov/guidelines/cholesterol/. 2004. Accessed June 6, 2012.9. National Institute for Health and Clinical Excellence.The management of type 2 diabetes. http://www.nice.org.uk/nicemedia/live/12165/44320/44320.pdf. 2009.Accessed June 6, 2012.10. Athyros VG, Giouleme OI, Nikolaidis NL, et al. Long-term follow-up of patients with acute hypertriglyceri-demia-induced pancreatitis. J Clin Gastroenterol. 2002;34(4):472-475.11. Sandhu S, Al-Sarraf A, Taraboanta C, Frohlich J,Francis GA. Incidence of pancreatitis, secondary causes,and treatment of patients referred to a specialty lipid clinicwith severe hypertriglyceridemia: a retrospective co-hort study. Lipids Health Dis. 2011;10:157.12. Keech A, Simes RJ, Barter P, et al; FIELD StudyInvestigators. Effects of long-term fenofibrate therapyon cardiovascular events in 9795 people with type 2 dia-betes mellitus (the FIELD study): randomised con-trolled trial. Lancet. 2005;366(9500):1849-1861.13. Experience in the Coronary Drug Project. Gallblad-der disease as a side effect of drugs influencing lipidmetabolism. N Engl J Med. 1977;296(21):1185-1190.14. Palmer RH. Effects of fenofibrate on bile lipidcomposition. Arteriosclerosis. 1985;5(6):631-638.15. Baigent C, Blackwell L, Emberson J, et al; Choles-terol Treatment Trialists’ (CTT) Collaboration. Efficacyand safety of more intensive lowering of LDL choles-terol: a meta-analysis of data from 170,000 partici-pants in 26 randomised trials. Lancet. 2010;376(9753):1670-1681.16. Randomised trial of cholesterol lowering in 4444patients with coronary heart disease: the ScandinavianSimvastatin Survival Study (4S). Lancet. 1994;344(8934):1383-1389.17. Shepherd J, Cobbe SM, Ford I, et al; West of Scot-land Coronary Prevention Study Group. Prevention ofcoronary heart disease with pravastatin in men withhypercholesterolemia. N Engl J Med. 1995;333(20):1301-1307.18. Sacks FM, Pfeffer MA, Moye LA, et al; Cholesteroland Recurrent Events Trial Investigators. The effect ofpravastatin on coronary events after myocardial infarc-tion in patients with average cholesterol levels. N EnglJ Med. 1996;335(14):1001-1009.19. Downs JR, Clearfield M, Weis S, et al; Air Force/Texas Coronary Atherosclerosis Prevention Study. Pri-mary prevention of acute coronary events with lovas-tatin in men and women with average cholesterol levels:results of AFCAPS/TexCAPS. JAMA. 1998;279(20):1615-1622.20. Long-Term Intervention with Pravastatin in Isch-aemic Disease (LIPID) Study Group. Prevention of car-diovascular events and death with pravastatin in pa-tients with coronary heart disease and a broad rangeof initial cholesterol levels. N Engl J Med. 1998;339(19):1349-1357.

21. GISSI Prevenzione Investigators (Gruppo Italiano perlo Studio della Sopravvivenza nell’Infarto Miocardico).Results of the low-dose (20 mg) pravastatin GISSI Pre-venzione trial in 4271 patients with recent myocardialinfarction: do stopped trials contribute to overallknowledge? Ital Heart J. 2000;1(12):810-820.22. Heart Protection Study Collaborative Group. MRC/BHF Heart Protection Study of cholesterol lowering withsimvastatin in 20,536 high-risk individuals: a ran-domised placebo-controlled trial. Lancet. 2002;360(9326):7-22.23. Shepherd J, Blauw GJ, Murphy MB, et al; PROSPERStudy Group. Pravastatin in elderly individuals at risk ofvascular disease (PROSPER): a randomised controlledtrial. Lancet. 2002;360(9346):1623-1630.24. Athyros VG, Papageorgiou AA, Mercouris BR, et al.Treatment with atorvastatin to the National Choles-terol Educational Program goal versus “usual” care insecondary coronary heart disease prevention: the GREekAtorvastatin and Coronary-heart-disease Evaluation(GREACE) study. Curr Med Res Opin. 2002;18(4):220-228.25. Sever PS, Dahlof B, Poulter NR, et al; ASCOTInvestigators. Prevention of coronary and stroke eventswith atorvastatin in hypertensive patients who have av-erage or lower-than-average cholesterol concentra-tions, in the Anglo-Scandinavian Cardiac OutcomesTrial—Lipid Lowering Arm (ASCOT-LLA): a multicen-tre randomised controlled trial. Lancet. 2003;361(9364):1149-1158.26. Colhoun HM, Betteridge DJ, Durrington PN, et al;CARDS Investigators. Primary prevention of cardiovas-cular disease with atorvastatin in type 2 diabetes in theCollaborative Atorvastatin Diabetes Study (CARDS): mul-ticentre randomised placebo-controlled trial. Lancet.2004;364(9435):685-696.27. Knopp RH, d’Emden M, Smilde JG, Pocock SJ. Ef-ficacyandsafetyofatorvastatin in thepreventionof car-diovascular end points in subjects with type 2 diabetes:the Atorvastatin Study for Prevention of coronary heartdiseaseEndpointsinNon-insulin-dependentdiabetesmelli-tus (ASPEN). Diabetes Care. 2006;29(7):1478-1485.28. Nakamura H, Arakawa K, Itakura H, et al; MEGAStudy Group. Primary prevention of cardiovascular dis-ease with pravastatin in Japan (MEGA Study): a pro-spective randomised controlled trial. Lancet. 2006;368(9542):1155-1163.29. Kjekshus J, Apetrei E, Barrios V, et al; CORONAGroup. Rosuvastatin in older patients with systolic heartfailure. N Engl J Med. 2007;357(22):2248-2261.30. Ridker PM, Danielson E, Fonseca FA, et al; JUPITERStudy Group. Rosuvastatin to prevent vascular eventsin men and women with elevated C-reactive protein.N Engl J Med. 2008;359(21):2195-2207.31. Tavazzi L, Maggioni AP, Marchioli R, et al; Gissi-HF Investigators. Effect of rosuvastatin in patients withchronic heart failure (the GISSI-HF trial): a ran-domised, double-blind, placebo-controlled trial. Lancet.2008;372(9645):1231-1239.32. Cannon CP, Braunwald E, McCabe CH, et al;Pravastatin or Atorvastatin Evaluation and InfectionTherapy−Thrombolysis in Myocardial Infarction 22Investigators. Intensive versus moderate lipid lower-ing with statins after acute coronary syndromes. N EnglJ Med. 2004;350(15):1495-1504.33. de Lemos JA, Blazing MA, Wiviott SD, et al; A toZ Investigators. Early intensive vs a delayed conser-vative simvastatin strategy in patients with acute coro-nary syndromes: phase Z of the A to Z trial. JAMA.2004;292(11):1307-1316.34. LaRosa JC, Grundy SM, Waters DD, et al; Treatingto New Targets (TNT) Investigators. Intensive lipid low-ering with atorvastatin in patients with stable coronarydisease. N Engl J Med. 2005;352(14):1425-1435.35. Pedersen TR, Faergeman O, Kastelein JJ, et al; In-cremental Decrease in End Points Through AggressiveLipid Lowering (IDEAL) Study Group. High-dose ator-vastatinvsusual-dosesimvastatin for secondarypreven-

tion after myocardial infarction: the IDEAL study: a ran-domized controlled trial. JAMA. 2005;294(19):2437-2445.36. Armitage J, Bowman L, Wallendszus K, et al; Studyof the Effectiveness of Additional Reductions in Choles-terolandHomocysteine (SEARCH)CollaborativeGroup.Intensive lowering of LDL cholesterol with 80 mg ver-sus 20 mg simvastatin daily in 12,064 survivors of myo-cardial infarction:adouble-blindrandomisedtrial.Lancet.2010;376(9753):1658-1669.37. Clofibrateandniacinincoronaryheartdisease.JAMA.1975;231(4):360-381.38. CannerPL,BergeKG,WengerNK,etal. Fifteenyearmortality in Coronary Drug Project patients: long-termbenefit with niacin. J Am Coll Cardiol. 1986;8(6):1245-1255.39. Report From the Committee of PrincipalInvestigators. A co-operative trial in the primary pre-vention of ischaemic heart disease using clofibrate.Br Heart J. 1978;40(10):1069-1118.40. FrickMH,EloO,HaapaK,etal.HelsinkiHeartStudy:primary-prevention trialwithgemfibrozil inmiddle-agedmen with dyslipidemia: safety of treatment, changes inrisk factors, and incidence of coronary heart disease.N Engl J Med. 1987;317(20):1237-1245.41. Rubins HB, Robins SJ, Collins D, et al; Veterans Af-fairs High-Density Lipoprotein Cholesterol InterventionTrialStudyGroup.Gemfibrozil for thesecondarypreven-tion of coronary heart disease in men with low levels ofhigh-density lipoproteincholesterol.NEngl JMed.1999;341(6):410-418.42. SecondarypreventionbyraisingHDLcholesterolandreducingtriglycerides inpatientswithcoronaryarterydis-ease: the Bezafibrate Infarction Prevention (BIP) study.Circulation. 2000;102(1):21-27.43. Ginsberg HN, Elam MB, Lovato LC, et al; ACCORDStudyGroup.Effectsofcombination lipid therapy in type2 diabetes mellitus. N Engl J Med. 2010;362(17):1563-1574.44. Frick MH, Heinonen OP, Huttunen JK, Koskinen P,Manttari M, Manninen V. Efficacy of gemfibrozil in dys-lipidaemic subjects with suspected heart disease: an an-cillarystudy intheHelsinkiHeartStudyframepopulation.Ann Med. 1993;25(1):41-45.45. MoherD,LiberatiA,Tetzlaff J,AltmanDG;PRISMAGroup. Preferred reporting items for systematic reviewsandmeta-analyses: thePRISMAstatement.BMJ. 2009;339:b2535.46. Jadad AR, Moore RA, Carroll D, et al. Assessing thequality of reports of randomized clinical trials: is blindingnecessary? Control Clin Trials. 1996;17(1):1-12.47. HigginsJP,ThompsonSG,DeeksJJ,AltmanDG.Mea-suring inconsistency in meta-analyses. BMJ. 2003;327(7414):557-560.48. Baigent C, Landray MJ, Reith C, et al; SHARPInvestigators.Theeffectsof loweringLDLcholesterolwithsimvastatin plus ezetimibe in patients with chronic kid-neydisease (StudyofHeartandRenalProtection): a ran-domised placebo-controlled trial. Lancet. 2011;377(9784):2181-2192.49. BodmerM,BrauchliYB,KrahenbuhlS, JickSS,MeierCR. Statin use and risk of gallstone disease followed bycholecystectomy. JAMA. 2009;302(18):2001-2007.50. Choi OS, Park SJ, Seo SW, Park CS, Cho JJ, Ahn HJ.The3-hydroxy-3-methylglutarylcoenzymeA(HMG-CoA)reductase inhibitor, lovastatin (statin) ameliorates CCK-inducedacutepancreatitis in rats.BiolPharmBull. 2005;28(8):1394-1397.51. Almeida JL, Sampietre SN, Mendonca Coelho AM,etal.Statinpretreatmentinexperimentalacutepancreatitis.JOP. 2008;9(4):431-439.52. Wei L, Yamamoto M, Harada M, Otsuki M. Treat-mentwithpravastatinattenuatesprogressionof chronicpancreatitis in rat. Lab Invest. 2011;91(6):872-884.53. vonBergmannK,LeissO.Effectof short-termtreat-mentwithbezafibrateandfenofibrateonbiliary lipidme-tabolisminpatientswithhyperlipoproteinaemia.EurJClinInvest. 1984;14(2):150-154.




CLINICIAN’S CORNERJAMA CLINICAL CHALLENGE

Cutaneous Nodule in a Young Man

Jianjun Qiao, MD, PhDHong Fang, MD

A 21-YEAR-OLD MALE STUDENT PRESENTS WITH A NODULE ON HIS CHEST THAT

has been progressively enlarging over 1 month. He also has a 1-monthhistory of upper abdominal pain associated with diarrhea but without

fever or chills. He was diagnosed with acute gastroenteritis and treated with oralomeprazole and ofloxacin without relief. He has lost 5 kg during the past halfyear. He denies any history of chronic illness or family history of cancer. Physi-cal examination reveals left supraclavicular and cervical lymphadenopathy. Asolitary asymptomatic, nontender, and indurated nodule is found on his chest(FIGURE 1). Gastrointestinal endoscopy shows multiple ulcers on the gastric fun-dus and gastric corpus.

What Would You Do Next?

A. Consult a surgeon for excision ofthe nodule

B. Do nothing; the nodule willresolve over time

C. Inject corticosteroids intra-lesionally

D. Obtain a biopsy of the nodule

See www.jama.com for online ClinicalChallenge.

Figure 1. Solitary nodule on the chest.

Author Affiliations: Department of Dermatology, The First Affiliated Hospital, College of Medicine, Zhejiang University, Hangzhou, China.Corresponding Author: Hong Fang, MD, Department of Dermatology, The First Affiliated Hospital, College of Medicine, Zhejiang University, No. 79, Qingchun Road,Hangzhou, 310003 Zhejiang Province, People’s Republic of China ([email protected]).JAMA Clinical Challenge Section Editor: Huan J. Chang, MD, Contributing Editor. We encourage authors to submit papers for consideration as a JAMA Clinical Chal-lenge. Please contact Dr Chang at [email protected].



DiagnosisCutaneous metastasis of gastric adeno-carcinoma

What to Do NextD. Obtain a biopsy of the nodule

The key clinical feature to recognizein this case is the rapidly growing asymp-tomatic nodule associated with gastro-intestinal symptoms. Although cutane-ous metastases of gastric adenomas areuncommon especially in younger per-sons, a rapidly growing asymptomaticskin nodule of undetermined causesshould be biopsied. If cutaneous metas-tasis is diagnosed, the origin of metas-tases should be found.

CommentVirtuallyanytumorof thevisceralorganscanmetastasize totheskin.However,cu-taneousmetastasesareuncommon,withareported frequencyof0.7%to10.4%.1-5

Metastases are usually a late manifesta-tion of widely disseminated disease, butoccasionally theymaybe the first indica-tion of an undiagnosed internal malig-nancy.2,3,6 Cutaneousmetastasesusuallyoriginatefromthebreast,lungs,colon,rec-tum, ovary, head and neck, or kidney.2,3

Malignant cutaneous lesions indica-tive of systemic cancer most commonlyoccur as asymptomatic multiple nod-ules (46.6%), single nodules (37.7%),plaques or erythematous patches (9.4%),and ulcers (6.5%).6 Metastases typicallypresent with relatively sudden onset andrapid growth.1,6,7 Skin metastases may beisolated or associated with metastases atmultiple sites.3,8

Gastric carcinoma is usually de-tected after the fifth decade of life9 andis rare in adults younger than 30 years.9

The most common metastatic sites fromgastric carcinoma are liver, intra-abdominal lymph nodes, ovary, andperitoneal cavity.3,6,7 Gastric carci-noma presents as skin metastasis 6.4%to 7.8% of the time.2,8 The lesions of cu-taneous metastases of gastric carci-noma are usually located on the ab-dominal wall or near the lymph nodes2,8

and are similar to cutaneous metasta-ses of other malignancies.

Lesions of cutaneous metastases maybe clinically indistinguishable from be-nign lesions, such as hemangiomas, epi-dermal cysts, pyogenic granulomas, orneurofibromas.1,2,7 Diagnosis depends onhistological evaluation combined withimmunohistochemical staining.8 Histo-logically, cutaneous metastases usuallyshow features consistent with the un-derlying primary malignancy. How-ever, metastases may exhibit less differ-entiation and be more anaplastic.2,7 Thesurvival rate of patients with cutaneousmetastases from internal malignancies isusually low.6,7 The median duration ofsurvival after cutaneous metastases fromgastric carcinoma is not known.

Patient OutcomeIn thispatient, abiopsyrevealed infiltrat-ingneoplasticcells scatteredorclusteredin the dermis. Some neoplastic cells ap-peared as signet ring cells (FIGURE 2A).Immature lumens of glands consistentwith neoplastic cells were found in thedeep dermis (Figure 2B). There was anarrow layer beneath the epidermis thatis not infiltratedbyatypical cellsdistrib-uted in the deep dermis. Immunohisto-chemically, the signet ring cells werepositive for carcinoembryonic antigen,epithelialmembraneantigen, andCK20and negative for markers of CD3, CD4,CD8,CD30,andgrosscysticdiseasefluidprotein 15 (GCDFP-15). Staining withKi-67 indicated prominent proliferativeactivity. Mucin stained positively in themetastatic cells with Alcian blue. Thesepathological studies indicated a diagno-sis of secondary signet ring cell adeno-carcinoma and suggested digestive tract

origin. Pathology of biopsy specimensfrom ulcers in the stomach was com-patible with a gastric signet ring celladenocarcinoma. Abdominal com-puted tomography detected many meta-static nodules in mesentery, hepatichilar region, and retroperitoneal lymphnodes. In view of the metastases, sur-gical intervention would be noncura-tive. The patient received chemo-therapy with oxaliplatin (day 1) and S-1(egafur, gimestat, and oteracil potas-sium) (days 1 to 7). His symptomsimproved and the size of the nodulereduced after 2 weeks of the chemo-therapy.Conflict of Interest Disclosures: Both authors have com-pleted and submitted the ICMJE Form for Disclosure ofPotential Conflicts of Interest and none were reported.Additional Contributions: We thank the patient forproviding permission to publish his information.

REFERENCES

1. Chopra R, Chhabra S, Samra SG, et al. Cutaneousmetastases of internal malignancies. Indian J DermatolVenereol Leprol. 2010;76(2):125-131.2. Lookingbill DP, Spangler N, Helm KF. Cutaneous me-tastases in patients with metastatic carcinoma: a retro-spective study of 4020 patients. J Am Acad Dermatol.1993;29(2 pt 1):228-236.3. Sariya D, Ruth K, Adams-McDonnell R, et al. Clini-copathologic correlation of cutaneous metastases. ArchDermatol. 2007;143(5):613-620.4. Nashan D, Muller ML, Braun-Falco M, et al. Cuta-neous metastases of visceral tumours: a review.J Cancer Res Clin Oncol. 2009;135(1):1-14.5. Hu SC, Chen GS, Wu CS, et al. Rates of cutaneousmetastases from different internal malignancies: expe-rience from a Taiwanese medical center. J Am AcadDermatol. 2009;60(3):379-387.6. Hu SC, Chen GS, Lu YW, et al. Cutaneous metas-tases from different internal malignancies. J Eur AcadDermatol Venereol. 2008;22(6):735-740.7. Nashan D, Meiss F, Braun-Falco M, ReichenbergerS. Cutaneous metastases from internal malignancies. Der-matol Ther. 2010;23(6):567-580.8. Saeed S, Keehn CA, Morgan MB. Cutaneous me-tastasis: a clinical, pathological, and immunohistochemi-cal appraisal. J Cutan Pathol. 2004;31(6):419-430.9. Torpy JM. Stomach cancer [JAMA Patient Page].JAMA. 2010;303(17):1771.

A B

Figure 2. Histology of the nodule. A, Some neoplastic cells appeared as signet ring cells. B,Immature lumens of glands (both panels: hematoxylin-eosin, original magnification �400).

JAMA CLINICAL CHALLENGE



EDITORIAL Editorials represent the opinionsof the authors and JAMA and

not those of the American Medical Association.

New-Generation Drug-Eluting Stentsfor Patients With Myocardial InfarctionSalvatore Cassese, MDAdnan Kastrati, MD

OVER THE PAST 2 DECADES, NUMEROUS RANDOM-ized trials have been performed to define the op-timal reperfusion therapy in patients with acuteST-segment elevation myocardial infarction

(STEMI). An analysis of 23 randomized trials has assessedthe relative merits of primary percutaneous coronary inter-vention (PCI) vs thrombolysis in approximately 8000 pa-tients with STEMI and demonstrated a clear advantage forPCI.1 Fewer trials were required to establish the implanta-tion of bare-metal stents (BMSs) compared with plainballoon-angioplasty as the first-choice PCI option in thesepatients.2 More recently, assessment of the role of drug-eluting stents (DESs) during primary PCI has drawn con-siderable attention with approximately 10 000 patients withacute myocardial infarction (AMI) enrolled in randomizedtrials comparing DESs with BMSs in the last decade alone.3,4

Indeed, in no other setting has DES use elicited as muchskepticism as in patients with STEMI.

The use of first-generation DESs has substantially reducedthe risk of restenosis compared with BMSs without any sig-nificant trade-off in safety.5 Initial randomized trials of pa-tients with STEMI demonstrated that DESs were better thanBMSs in outcomes at 1 year.6,7 However, although there is gen-eral agreement about the higher anti-restenotic efficacy of DESsin patients with STEMI, concerns exist regarding the long-term safety of these devices in this setting.8 Pathologic samplesof patients with stent thrombosis confirmed that first-generation DESs led to delayed arterial healing, incompleteendothelialization, and persistent fibrin deposition.9 More-over, patients receiving first-generation DESs for AMI, com-pared with patients receiving DESs for stable angina, had evi-dence of less neointimal thickness at the stented site as wellas more uncovered stent struts and inflammation.10 The dif-ferential arterial response to DESs depending on plaque mor-phology of the culprit lesions might increase the intrinsicthrombotic risk of patients presenting with STEMI.11 Stent strutpenetration of a necrotic core may inhibit neointimal growth,resulting in uncovered stent struts, especially with DESs.10 Stentundersizing due to poor epicardial flow, vasoconstriction, or

thrombus dissolution or embolization may lead to incom-plete stent apposition. In fact, the incidence of positive arte-rial remodelling and late-acquired incomplete stent apposi-tion was significantly higher in patients with STEMI.12

Although concerns generated by preclinical and patho-logical investigations have not always been validated by clini-cal studies, long-term analyses at 3 to 5 years after the pro-cedure suggest that compared with use of BMSs the use offirst-generation DESs in STEMI is associated with an excessof very late thrombotic complications. A meta-analysis of 15trials, including 7867 patients randomly assigned to first-generation DESs or BMSs in STEMI, unravelled the signifi-cant interaction existing between DESs and time of observa-tion regarding thrombotic events.4 Although the overall riskof definite and definite or probable stent thrombosis was simi-lar for DESs and BMSs, very late (after the first year) throm-botic events were more likely to occur in the DES group.

Second-generation DESs may overcome some of the limi-tations of first-generation DESs. New alloys, improved bio-compatibility of durable polymer coatings, biodegradablepolymer coatings, and reduced toxicity of the antiprolifera-tive drugs are increasing the safety of current DES technol-ogy while maintaining or even improving its efficacy. Thesecond-generation permanent polymer stents eluting evero-limus significantly reduced the risk of stent thrombosis com-pared with paclitaxel- and sirolimus-eluting stents.13 Newer-generation DESs eluting limus drugs from a biodegradablepolymer coating are emerging as an effective and safe treat-ment option over the long term. A pooled patient-level meta-analysis of 3 randomized trials (4062 patients) found notonly increased efficacy but also a significant reduction inthe risk of stent thrombosis over 4 years with biodegrad-able polymer DES vs permanent polymer sirolimus-elutingstent.14 Patients with AMI obtained the major benefit fromthese DESs.15 Two opposite, time-dependent effects have beendescribed for polymers used in DES technology—an earlyprotective effect against stent thrombosis16 and a late pro-inflammatory and prothrombotic effect.17 By their very na-ture, the biodegradable polymers used in newer-generation DESs might offer the early advantages of polymerswhile avoiding the very late hazards, which may turn outto be especially useful in patients with STEMI.

See also p 777.AuthorAffiliations:DeutschesHerzzentrum,TechnischeUniversitat,Munich,Germany.Corresponding Author: Adnan Kastrati, MD, Deutsches Herzzentrum, Tech-nische Universitat, Lazaretstr 36, Munich, Germany ([email protected]).



Inthis issueof JAMA,Raberandcolleagues18 report theresultsofa randomizedcontrolled trial comparingbiodegradablepoly-mer biolimus-eluting stent vs its BMS counterpart in 1161patients with STEMI treated with primary PCI (Comparisonof Biolimus Eluted From an Erodible Stent Coating With BareMetal Stents in Acute ST-Elevation Myocardial Infarction[COMFORTABLE AMI] trial). At 1-year follow-up, biode-gradable polymer DES reduced major adverse cardiac events(cardiacdeath, targetvessel–relatedreinfarction,and ischemia-driven target-lesion revascularization) mostly due to a signifi-cant reduction in the rates of reinfarction and reintervention.Fewer cases of definite stent thrombosis were observed in thegroup of biodegradable polymer DES, with most of the reduc-tion occurring in the immediate peri-interventional period. Aper-protocol comprehensive imaging substudy in 100 patients(withboth intravascularultrasound-virtualhistologyandopti-cal coherence tomography) is also planned in this trial19 andwill provide valuable information about the vessel wall reac-tion to BMSs and biodegradable polymer DESs in patients withSTEMI.Additional intracoronary imaging inpatientswhosub-sequentlypresentwithangiographicallyconfirmedstent throm-bosis may have given further insights into the pathophysiol-ogy of thrombotic events in terms of strut coverage, stentmalapposition, and positive vessel wall remodelling.

The COMFORTABLE AMI trial is a well-done trial with con-vincing results regarding its primary end point, which showedthat among patients with STEMI, compared with the BMS, thebiodegradable polymer biolimus-eluting stent had lower ratesof the composite outcome of major adverse cardiovascularevents during 1 year of follow-up. The trial has, however, nei-ther the required sample size nor the sufficient length of fol-low-up to provide the definitive answer about the long-termsafety of the new biodegradable DES in this setting. The trialoffers positive signals that, along with previous findings in-dicating the excellent 4-year safety of biodegradable polymerDESs, form a good basis to believe that these new DESs maybelong to the “first-choice” devices in patients with STEMI.Bioabsorption of the drug carrier and even of the entire stentplatform20 may be particularly useful in this setting.

The findings from the COMFORTABLE AMI trial18 andfrom the series of previous trials on DESs in patients withAMI provide several important lessons. First, the efficacyof DESs vs BMSs in STEMI is already established and, there-fore, further studies comparing these interventions mightnot be needed. Second, concerns about a possible very latesafety issue with DESs are apparently DES-type specific,mostly related to first-generation DESs and less justified withnewer DESs. Larger randomized trials with longer fol-low-up and head-to-head comparisons of the available DEStechnologies are, however, required to completely elimi-nate these concerns. These studies should also take broaderadvantage of intravascular imaging technologies to pro-vide mechanistic insights into the clinical findings. Third,although there is almost no rationale for performing DESvs BMS studies anymore, it might be conceivable to expect

studies that test the hypothesis of noninferiority of new, im-proved BMSs to available DESs. Until then, recent studiessuch as the COMFORTABLE AMI trial should make cardi-ologists feel more comfortable with the use of new-generation DESs in patients with STEMI.

Conflict of Interest Disclosures: Both authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Kastrati holds apatent related to polymer-free sirolimus and probucol coating and reported re-ceiving honoraria from Abbott, Biosensors, Cordis, and Medtronic. Dr Cassese re-ports no conflicts.Funding/Support: This work was supported by the European Commission (underthe Seventh Framework Programme [Acronym: PRESTIGE, Project No. 260309]).Role of the Sponsor: The European Commission had no role in the preparation,review, or approval of the manuscript.

REFERENCES

1. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous throm-bolytic therapy for acute myocardial infarction. Lancet. 2003;361(9351):13-20.2. De Luca G, Suryapranata H, Stone GW, et al. Coronary stenting versus balloonangioplasty for acute myocardial infarction. Int J Cardiol. 2008;126(1):37-44.3. Dibra A, Tiroch K, Schulz S, et al. Drug-eluting stents in acute myocardial infarction.Clin Res Cardiol. 2010;99(6):345-357.4. Kalesan B, Pilgrim T, Heinimann K, et al. Comparison of drug-eluting stentswith bare metal stents in patients with ST-segment elevation myocardial infarction.Eur Heart J. 2012;33(8):977-987.5. Kastrati A, Dibra A, Spaulding C, et al. Meta-analysis of randomized trials ondrug-eluting stents vs bare-metal stents in patients with acute myocardial infarction.Eur Heart J. 2007;28(22):2706-2713.6. Spaulding C, Henry P, Teiger E, et al. Sirolimus-eluting versus uncoated stentsin acute myocardial infarction. N Engl J Med. 2006;355(11):1093-1104.7. Laarman GJ, Suttorp MJ, Dirksen MT, et al. Paclitaxel-eluting versus uncoatedstents in primary percutaneous coronary intervention. N Engl J Med. 2006;355(11):1105-1113.8. Joner M, Finn AV, Farb A, et al. Pathology of drug-eluting stents in humans.J Am Coll Cardiol. 2006;48(1):193-202.9. Luscher TF, Steffel J, Eberli FR, et al. Drug-eluting stent and coronary thrombosis.Circulation. 2007;115(8):1051-1058.10. Nakazawa G, Finn AV, Joner M, et al. Delayed arterial healing and increasedlate stent thrombosis at culprit sites after drug-eluting stent placement for acutemyocardial infarction patients. Circulation. 2008;118(11):1138-1145.11. Dangas GD, Caixeta A, Mehran R, et al. Frequency and predictors of stentthrombosis after percutaneous coronary intervention in acute myocardial infarction.Circulation. 2011;123(16):1745-1756.12. Hassan AK, Bergheanu SC, Stijnen T, et al. Late stent malapposition risk is higherafter drug-eluting stent compared with bare-metal stent implantation and asso-ciates with late stent thrombosis. Eur Heart J. 2010;31(10):1172-1180.13. Baber U, Mehran R, Sharma SK, et al. Impact of the everolimus-eluting stenton stent thrombosis. J Am Coll Cardiol. 2011;58(15):1569-1577.14. Stefanini GG, Byrne RA, Serruys PW, et al. Biodegradable polymer drug-eluting stents reduce the risk of stent thrombosis at 4 years in patients undergo-ing percutaneous coronary intervention. Eur Heart J. 2012;33(10):1214-1222.15. Stefanini GG, Kalesan B, Serruys PW, et al. Long-term clinical outcomes ofbiodegradable polymer biolimus-eluting stents versus durable polymer sirolimus-eluting stents in patients with coronary artery disease (LEADERS). Lancet. 2011;378(9807):1940-1948.16. Kolandaivelu K, Swaminathan R, Gibson WJ, et al. Stent thrombogenicity earlyin high-risk interventional settings is driven by stent design and deployment andprotected by polymer-drug coatings. Circulation. 2011;123(13):1400-1409.17. Koppara T, Joner M, Bayer G, Steigerwald K, Diener T, Wittchow E. Histo-pathological comparison of biodegradable polymer and permanent polymer basedsirolimus eluting stents in a porcine model of coronary stent implantation. ThrombHaemost. 2012;107(6):1161-1171.18. Raber L, Kelbæk H, Ostoijc M, et al. Effect of biolimus-eluting stents with bio-degradable polymer vs bare-metal stents on cardiovascular events among pa-tients with acute myocardial infarction: the COMFORTABLE AMI randomized trial.JAMA. 2012;308(8):777-787.19. Raber L, Kelbæk H, Ostoijc M, et al. Comparison of biolimus eluted from anerodible stent coating with bare metal stents in acute ST-elevation myocardial in-farction (COMFORTABLE AMI trial). EuroIntervention. 2012;7(12):1435-1443.20. Serruys PW, Onuma Y, Dudek D, et al. Evaluation of the second generationof a bioresorbable everolimus-eluting vascular scaffold for the treatment of de novocoronary artery stenosis. J Am Coll Cardiol. 2011;58(15):1578-1588.

EDITORIAL



EDITORIAL Editorials represent the opinionsof the authors and JAMA and

not those of the American Medical Association.

Cardiovascular Risk Assessmentin the 21st CenturyJ. Michael Gaziano, MD, MPHPeter W. F. Wilson, MD

IN THE 1960S, THE FRAMINGHAM HEART STUDY (FHS) IN-vestigators coined the term “factors of risk” for cardio-vascular disease1; today risk factors are used routinelyin daily practice to assist in risk prediction and to tar-

get prevention strategies for cardiovascular disease as wellas for other diseases. The FHS investigators subsequentlyused risk factors to develop a multivariable model to pre-dict the likelihood of a future coronary heart disease event.The model included several major cardiovascular risk fac-tors that estimate the probability of developing an initial coro-nary disease event over the next 10 years and result in a riskscore.2 This basic model is used around the globe.

Many guidelines now suggest this type of risk estima-tion may be useful as part of a general cardiovascular dis-ease prevention evaluation before embarking on various pre-vention strategies. Researchers have been trying to improveon this concept by adding novel cardiovascular markers andtests to the basic model.3 This discussion is timely becausethe National Cholesterol Education Program fourth AdultTreatment Panel guidelines may be released this year.

What makes a good prediction rule? First, it has to pro-vide information that can help with prognostication in or-der to make a better informed decision concerning care. Sec-ond, the rule should be easy to use, and the risk factors usedshould be easily obtainable at the time the information isneeded by the clinician. Third, the risk factor should in-volve minimal risk, such as smoking status. Risk factors thatexpose individuals to greater risk such as ionizing radia-tion or uncomfortable procedures, should be minimized.Fourth, research findings should support that a risk scoresuch as Framingham or other formulations can be appro-priately applied to the patient. Fifth, the added informa-tion should be worth the cost of acquiring the risk factordata. And sixth, it would be helpful if the tool could be usedto monitor changes in risk over time.4,5

The Framingham Risk Score (FRS) sets a high bar be-cause it predicts fairly well using several simple risk factorsthat are inexpensive and easy to obtain, such as sex, smok-ing status, blood pressure, and lipid values.2,6 The FRS ex-

plains a fair amount of the variability in risk of a popula-tion. Adding some conventional risk factors does not addmaterially to the model largely because of the high corre-lation with factors already in the model. For example, obe-sity and body mass index are clearly associated with risk ofcardiovascular events, but they are correlated with other fac-tors such as lipid levels and blood pressure that are alreadyin the model, so the basic FRS is not improved when thesefactors are added.6,7

In this issue of JAMA, 2 articles address the utility of ad-ditional techniques designed to enhance the predictive valueof the FRS. In the report by Den Ruijter and colleagues,8 apooled analysis of a number of studies demonstrates thatcarotid intima-media thickness (CIMT) does not add clini-cally meaningful information to the standard prediction mo-dalities, (ie, the FRS), and the authors rightly conclude thatthere is not much utility using CIMT measurement amongpatients who are at intermediate risk. In the report by Yeboahand colleagues,9 the authors extensively evaluated a num-ber of laboratory and imaging ancillary risk markers in onelarge cohort and demonstrate that among several risk mark-ers, coronary artery calcium (CAC) score was associated withsubstantially improved risk estimation among patients atmodest risk.

These reports illustrate the means by which the enhance-ment of a given model such as FRS with additional data isevaluated. Area under the receiver operating characteristiccurve has been used for many diagnostic tests. A clinicallyuseful approach describes the proportion of individuals whoare reclassified, known as the net reclassification index (NRI).However, not all reclassifications are created equally. Re-fining risk estimation at the low or high end of the risk spec-trum is not helpful because it will not likely alter manage-ment. On the other hand, reclassification can assist in decisionmaking for patients who are near clinical decision bound-aries. For this reason, both studies provided information onthe reclassification that occurred among those at 5% to 20%risk. In this intermediate range, risk estimates inform clini-cal decisions about certain interventions.

See also pp 788 and 796.

Author Affiliations: Boston VA Healthcare System, Brigham and Women’s Hos-pital, Harvard Medical School, Boston, Massachusetts (Dr Gaziano); and AtlantaVA Medical Center and Emory Clinical Cardiovascular Research Institute, Atlanta,Georgia (Dr Wilson). Dr Gaziano is also Contributing Editor, JAMA.Corresponding Author: J. Michael Gaziano, MD, MPH, Boston VA Healthcare Sys-tem, Brigham and Women’s Hospital, Harvard Medical School, 1620 Tremont St,Boston, MA 02120 ([email protected]).



Although there has been a great deal of work on the im-provement in prediction modeling, less work has been donein 2 areas: the cost and risk in the screened population andrisk prediction over time. There seems to be little informa-tion gained from CIMT that would alter current decisionmaking, and it appears that the extra cost is not justified.On the other hand, the CAC score, which did improve theNRI substantially, requires additional assessment to deter-mine whether the added information from this testing is trulyuseful. The CAC imaging also involves radiation exposureto the patient, which raises some concerns and limitsrescanning. The costs of the scan and the costs of the ad-ditional interventions must also be considered.

Much of the research in this area considers the addedutility of these adjunctive methods at a single point in time.This is the assumption for NRI. Not applying the newmarker could result in the misclassification of a patient andsubjecting the individual to overtreatment or undertreat-ment. However, this is not how medicine is practiced,because in most cases, the physician will likely see thepatient again.

For an intermediate-risk patient who is near a decisionboundary for lipid-lowering therapy, the physician couldplan to see that patient in a few months and reassess vas-cular disease risk. Repeating tests may improve accuracyand reveal trends. Knowing whether a patient’s vasculardisease risk is increasing or decreasing may inform clinicaldecisions, especially if the risk estimate is near a decisionboundary. For some patients, a CAC scan may help guidea clinical decision, but radiation exposure and costs areimportant considerations. Coronary artery calcium find-ings also are somewhat resistant to change even in the faceof improvement in risk factors and may be useful as asingle measure for assessment, especially when refinementof a risk estimate is important, but might not be useful fortracking risk over time.

In summary, although many guidelines stress the impor-tance of understanding overall vascular disease risk, a fewsimple questions can provide substantial information. First,does the patient have known cardiovascular disease? If theanswer is yes, there is not really any need to use a risk cal-culator because these individuals are at sufficiently highenough risk to warrant aggressive management. If the an-swer is no, a risk calculator is useful. Second, does the pa-

tient have diabetes mellitus? If the answer is yes, more ag-gressive management of cardiovascular disease risk factorsshould be considered, especially because the long-term bur-den of cardiovascular disease in diabetic patients is so high.10

Third, if the simple risk calculator, such as FRS or AdultTreatment Panel calculators, identifies patients at low risk(�5% over 10 years), or high risk (�20% over 10 years),then risk refinement is not as informative. For patients atintermediate risk (ie, 5% to �20% estimated risk over 10years), clinicians may be in a position to refine the risk es-timate with a CAC scan. Another possibility is repeating arisk assessment in a few months, and that strategy deservesfurther study. Coronary artery calcium score appears to aug-ment the risk assessment process but may have limited util-ity in terms of tracking an individual patient’s vascular riskbecause it is unlikely that CAC imaging will be tracked overtime. This limitation might change in the future as non-ionizing imaging methods are developed to quantify the ath-erosclerotic burden.Conflict of Interest Disclosures: Both authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest and none were re-ported.

REFERENCES

1. Kannel WB, Dawber TR, Kagan A, Revotskie N, Stokes J III. Factors of risk inthe development of coronary heart disease—six-year follow-up experience: theFramingham Study. Ann Intern Med. 1961;55(1):33-50.2. Wilson PW, D’Agostino RB, Levy D, Belanger AM, Silbershatz H, Kannel WB.Prediction of coronary heart disease using risk factor categories. Circulation. 1998;97(18):1837-1847.3. Ridker PM, Paynter NP, Rifai N, Gaziano JM, Cook NR. C-reactive protein andparental history improve global cardiovascular risk prediction: the Reynolds RiskScore for men. Circulation. 2008;118(22):2243-2251.4. Wilson PW. Challenges to improve coronary heart disease risk assessment. JAMA.2009;302(21):2369-2370.5. Lloyd-Jones DM. Cardiovascular risk prediction: basic concepts, current status,and future directions. Circulation. 2010;121(15):1768-1777.6. D’Agostino RB Sr, Vasan RS, Pencina MJ, et al. General cardiovascular risk pro-file for use in primary care: the Framingham Heart Study. Circulation. 2008;117(6):743-753.7. D’Agostino RB Sr, Grundy S, Sullivan LM, Wilson P; CHD Risk Prediction Group.Validation of the Framingham coronary heart disease prediction scores: results ofa multiple ethnic groups investigation. JAMA. 2001;286(2):180-187.8. Den Ruijter HM, Peters SAE, Anderson TJ, et al. Common carotid intima-media thickness measurements in cardiovascular risk prediction: a meta-analysis.JAMA. 2012;308(8):796-803.9. Yeboah J, McClelland RL, Polonsky TS, et al. Comparison of novel risk markersfor improvement in cardiovascular risk assessment in intermediate-risk individuals.JAMA. 2012;308(8):788-795.10. Fox CS, Pencina MJ, Wilson PW, Paynter NP, Vasan RS, D’Agostino RB Sr.Lifetime risk of cardiovascular disease among individuals with and without diabe-tes stratified by obesity status in the Framingham Heart Study. Diabetes Care. 2008;31(8):1582-1584.

EDITORIAL



LETTERS

Cardiac Device Infective Endocarditisand Patient Survival

To the Editor: The prospective cohort study on cardiac de-vice infective endocarditis (CDIE) found that early deviceremoval was associated with improved survival.1 It is re-grettable that no data regarding the means of removal (suchas percutaneous or surgical) were available. During the pastdecades, percutaneous lead removal has been widely usedfor infected leads as it is less invasive than cardiac surgery.Future studies should include the removal procedure in theinvestigation.

In addition, the timing and types of previous electro-therapeutic interventions, number of leads and their char-acteristics, location of leads (including the coronary sinusand cardiac veins), devices and techniques used, and the du-ration of the procedure should all be taken into consider-ation because they are known to increase the risk of com-plications.2,3

Another factor that may have influenced the outcome of de-vice removal is patient volume among the different centers inthe study. Overall, the in-hospital mortality was high at 12.8%and a mean of fewer than 3 patients was collected from eachcenter. High procedural success rate with low major compli-cations, no procedural deaths, and low postoperative 30-daymortality, which were not related directly to the extractionprocedure, have been achieved in centers with high vol-umes.4 Operator experience may play a crucial role, as it hasbeen reported that both the risk of failed extraction and com-plications increase with less experienced physicians.2,3

We agree with the Heart Rhythm Society guidelines5 thatcomplete extraction of implanted devices should be con-sidered standard therapy for most patients with CDIE, es-pecially when experienced surgeons are available.

Xingli Wu, MDDingyou Yang, MDAuthor Affiliations: Institute of Geriatric Cardiology, China PLA General Hospital,Beijing, China ([email protected]).Conflict of Interest Disclosures: Both authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest and none were re-ported.

1. Athan E, Chu VH, Tattevin P, et al; ICE-PCS Investigators. Clinical character-istics and outcome of infective endocarditis involving implantable cardiac devices.JAMA. 2012;307(16):1727-1735.2. Da Costa A, Kirkorian G, Isaaz K, Touboul P. Secondary infections after pace-maker implantation. Rev Med Interne. 2000;21(3):256-265.3. Byrd CL, Wilkoff BL, Love CJ, et al. Intravascular extraction of problematic orinfected permanent pacemaker leads: 1994-1996: US Extraction Database, MEDInstitute. Pacing Clin Electrophysiol. 1999;22(9):1348-1357.4. Arujuna A, Williams S, Whittaker J, et al. Trends, indications and outcomes ofcardiac implantable device system extraction: a single UK centre experience overthe last decade. Int J Clin Pract. 2012;66(2):218-225.5. Wilkoff BL, Love CJ, Byrd CL, et al; Heart Rhythm Society; American HeartAssociation. Transvenous lead extraction: Heart Rhythm Society expert consen-sus on facilities, training, indications, and patient management: this document wasendorsed by the American Heart Association (AHA). Heart Rhythm. 2009;6(7):1085-1104.

To the Editor: The study by Dr Athan and colleagues1 aboutthe clinical characteristics and outcomes of CDIE evalu-ated the association between device removal and clinical out-comes. We would like to raise a few points about the analy-sis that may be pertinent.

First, cardiac resynchronization therapy is another pre-dictor of long-term mortality after infection of cardiac im-plantable devices.2 It would be interesting to know if theauthors investigated differences in the rate of infection andmortality in patients with dual-chamber pacemakers vs ven-tricular pacemakers.

Second, removal of the cardiac device was not associ-ated with lower in-hospital mortality, but there was im-proved survival at 1 year. Patients in whom the device wasremoved also had an increased incidence of complicationssuch as valve infection. Acute physiological severity and em-bolic events are early predictors of in-hospital death in in-fective endocarditis.3 Did the investigators adjust for thesecomplications while analyzing in-hospital mortality?

Third, it would be interesting to know if there was anychange in mortality after removal and reimplantation of newcardiac devices following CDIE when compared with un-infected cardiac devices. If so, then the findings could beuseful in identifying high-risk patients, and a more aggres-sive approach should be adopted in such patients.

In addition, the investigators defined CDIE as valvularor lead vegetations detected by echocardiography or meet-ing Duke’s criteria.4 Four patients receiving hemodialysishave been described with superior vena cava and right atrialwall infective endocarditis.5 Did the authors include pa-tients with superior vena cava or right atrial wall endocar-ditis?

Abhishek Sharma, MDAjay Vallakati, MDAuthor Affiliations: Maimonides Medical Center, Brooklyn, New York([email protected]).

GUIDELINES FOR LETTERS. Letters discussing a recent JAMA article shouldbe submitted within 4 weeks of the article’s publication in print. Letters receivedafter 4 weeks will rarely be considered. Letters should not exceed 400 words oftext and 5 references and may have no more than 3 authors. Letters reportingoriginal research should not exceed 600 words of text and 6 references and mayhave no more than 5 authors. They may include up to 2 tables or figures but on-line supplementary material is not allowed. All letters should include a word count.Letters must not duplicate other material published or submitted for publication.Letters not meeting these specifications are generally not considered. Letters willbe published at the discretion of the editors and are subject to abridgementand editing. Further instructions can be found at http://jama.com/public/InstructionsForAuthors.aspx. A signed statement for authorship criteria and re-sponsibility, financial disclosure, copyright transfer, and acknowledgment and theICMJE Form for Disclosure of Potential Conflicts of Interest are required beforepublication. Letters should be submitted via the JAMA online submission and re-view system at http://manuscripts.jama.com. For technical assistance, please con-tact [email protected].

Letters Section Editor: Jody W. Zylke, MD, Senior Editor.



Conflict of Interest Disclosures: Both authors have completed and submitted the ICMJEForm for Disclosure of Potential Conflicts of Interest and none were reported.

1. Athan E, Chu VH, Tattevin P, et al; ICE-PCS Investigators. Clinical character-istics and outcome of infective endocarditis involving implantable cardiac devices.JAMA. 2012;307(16):1727-1735.2. Deharo JC, Quatre A, Mancini J, et al. Long-term outcomes following infec-tion of cardiac implantable electronic devices: a prospective matched cohort study.Heart. 2012;98(9):724-731.3. Chu VH, Cabell CH, Benjamin DK Jr, et al. Early predictors of in-hospital deathin infective endocarditis. Circulation. 2004;109(14):1745-1749.4. Durack DT, Lukes AS, Bright DK; Duke Endocarditis Service. New criteria fordiagnosis of infective endocarditis: utilization of specific echocardiographic findings.Am J Med. 1994;96(3):200-209.5. Thakar S, Janga KC, Tolchinsky T, et al. Superior vena cava and right atriumwall infective endocarditis in patients receiving hemodialysis. Heart Lung. 2012;41(3):301-307.

In Reply: Drs Wu and Yang have raised the question ofmeans of device removal in the setting of CDIE, in addi-tion to the results presented regarding any deviceremoval vs no removal. Unfortunately, our case reportforms did not collect this information from site investiga-tors, although ICE-PCS (International Collaboration onEndocarditis-Prospective Cohort Study) sites were pre-dominantly referral centers.1 Similarly, procedural com-plications directly associated with device removal,including mortality, were not collected. However, in ourstudy, in-hospital mortality was all cause, including bothprocedure-related mortality as well as other causes ofdeath. Specifically, death related to sepsis or comorbidmedical conditions in this population of patients withadvanced age and multiple medical problems may beindependent from the device removal procedure itself. Asshown in a recent study of consecutive device removalprocedures,2 patients with systemic infection as the indi-cation for device removal had significantly higher mortal-ity in both short-term and longer-term follow-up. Weagree that operator and institutional experience and vol-ume in device and lead removal is critical for optimaloutcome in these cases.

Although patients with permanent pacemakers andimplanted cardioverter-defibrillators (ICDs) had similarin-hospital mortality rates in our study, we agree withDrs Sharma and Vallakati that the type of cardiac devicein the setting of CDIE may have prognostic implications.Data on the type of ICD, specifically possible biventricu-lar ICDs, were not collected in the study. Similarly,embolic events and acute physiology may influence mor-tality as shown in previous studies, but the small numberof in-hospital deaths limited the validity of multivariableanalysis to determine variables independently associatedwith mortality. Device reimplantation, its timing relativeto infected device removal, and CDIE relapse were notvariables collected prospectively in our cohort. In addi-tion, regarding locations of vegetations, we are not awareof any cases of CDIE in our study with vegetationslocated on the right atrial or superior vena cava surface inthe absence of device lead or valve vegetation, but agreethat this should be considered as CDIE.

Eugene Athan, MDAndrew Wang, MD

Author Affiliations: Department of Infectious Diseases, Barwon Health, Geelong,Australia (Dr Athan); and Duke University Medical Center, Durham, North Caro-lina (Dr Wang; [email protected]).Conflict of Interest Disclosures: The authors have completed and submittedthe ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Wangreported providing expert testimony for trial defense; receiving grants fromGilead Sciences, Edwards Lifesciences, Boehringer-Ingleheim, American HeartAssociation Mid-Atlantic Grant in Aid, and Abbott Vascular; receiving paymentfor lectures from American Physician; receiving royalties for serving as an editorfrom Springer; and receiving payment for development of educational presen-tations from the American College of Cardiology Foundation. Dr Athanreported no disclosures.

1. Cabell CH, Abrutyn E. Progress toward a global understanding of infective en-docarditis: early lessons from the International Collaboration on Endocarditisinvestigation. Infect Dis Clin North Am. 2002;16(2):255-272.2. Maytin M, Jones SO, Epstein LM. Long-term mortality after transvenous leadextraction. Circ Arrhythm Electrophysiol. 2012;5(2):252-257.

Neonatal Abstinence Syndrome

To the Editor: The article by Dr Patrick and colleagues1 pro-vides important data on US trends in maternal opioid use,neonatal abstinence syndrome (NAS), and NAS-associatedhospital costs. These findings should spark forward-thinking research. However, certain cautions regarding theresults must be considered.

Future research should focus on the relationshipbetween diagnostic groups (eg, mothers with long-termmethadone use) and neonatal outcomes, including butnot limited to NAS. Different outcomes—with quite dif-ferent costs—may occur for neonates of women who areuntreated or misusing opioids vs those maintained onopioid agonists or medicated for pain. While many of theclinical conditions in Table 1 of the article may occur ininfants born to methadone-stabilized women, their fre-quency in infants born to methadone-stabilized women islower2 than the frequency reported for all infants bornwith NAS.

The marked change in NAS incidence over the 10-yearperiod examined is troubling1 and may be due to both in-creased awareness of the issue and an increase inadult licit and illicit opioid use. One possible approach toprevention is universal urine screening of pregnant womento identify those using opioids; this approach may seem ap-pealing but has notable drawbacks.3 Future research needsto determine not only the most accurate screening meth-ods but also the most efficacious prevention and treatmentinterventions to reduce the incidence and prevalence of opi-oid dependence among women.

No single standardized NAS measure exists. Thus, fu-ture research should not simply focus on complementarytools4 for the Finnegan measures. Rather, tools that bettermeasure NAS variants (eg, opioid-associated NAS vs ben-zodiazepine-associated NAS) may yield more effective treat-ments and attendant cost reductions.

No first-line treatment for newborns with NAS exists.1 Ran-domized controlled trials examining different NAS treat-

LETTERS



LETTERS


















ment protocols are needed to determine the most effectiveand cost-effective initiation, treatment, and weaning pro-tocols to optimize outcomes.

Future research must acknowledge that opioid useoccurs within a larger biopsychosocial context that con-founds the assessment of both the short- and longer-termeffects of in utero exposure to opioids. Poor neonatal out-comes often result from the concurrence of variousmaternal life experiences, including poverty and depriva-tion, physical, emotional, and sexual abuse, inadequatenutrition, and psychiatric comorbidity.5 Although illicitopioid addiction is one of the leading proximal causes ofNAS, meaningful reductions in incidence will likely onlyoccur when effective comprehensive woman-centeredtreatment programs for opioid-addicted pregnant womenare widely implemented.

Hendree E. Jones, PhDKarol Kaltenbach, PhD

Author Affiliations: RTI International, Research Triangle Park, North Carolina (DrJones; [email protected]); and Department of Pediatrics, Thomas Jefferson Univer-sity, Philadelphia, Pennsylvania (Dr Kaltenbach).Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Jones reported re-ceiving reimbursement for time and travel from Reckitt Benckiser Inc. Dr Kalten-bach reported no disclosures.

1. Patrick SW, Schumacher RE, Benneyworth BD, Krans EE, McAllister JM, DavisMM. Neonatal abstinence syndrome and associated health care expenditures: UnitedStates, 2000-2009. JAMA. 2012;307(18):1934-1940.2. Jones HE, Kaltenbach K, Heil SH, et al. Neonatal abstinence syndrome aftermethadone or buprenorphine exposure. N Engl J Med. 2010;363(24):2320-2331.3. American College of Obstetricians and Gynecologists. ACOG Committee Opin-ion No. 422: at-risk drinking and illicit drug use: ethical issues in obstetric and gy-necologic practice. Obstet Gynecol. 2008;112(6):1449-1460.4. Hayes MJ, Brown MS. Epidemic of prescription opiate abuse and neonatalabstinence. JAMA. 2012;307(18):1974-1975.5. Robins LN, Mills JL. Effects of in utero exposure to street drugs. Am J PublicHealth. 1993;83(suppl):1-32.

In Reply: Drs Jones and Kaltenbach comment on keyaspects of NAS and related research that may shape thenext decade of inquiry about this condition and its pre-vention and therapy. They suggest that the recent 3-foldincrease in NAS may be explained by heightened clini-cian awareness of the condition. We agree in principleand cited this as a potential limitation in the article. How-ever, the increasing trend in maternal use and abuse ofopiates also described in the study makes “heightenedawareness” less likely as a major factor in the increasingtrend. In addition, both maternal and newborn trendsappear to correlate with the general population trendtoward more opiate use and abuse.1

Jones and Kaltenbach correctly describe both mothersand newborns in our sample as heterogeneous popula-tions. Our study was limited by the inability to distin-guish the type and use of opiates among mothers. Asresearch in NAS continues to move forward, it is vitalthat investigators analyze the effects of different drugexposures on mothers and newborns. As an example,there is a paucity of data related to use of opiate pain

relievers during pregnancy and both short- and long-termnewborn outcomes.2

They also mention the need to improve methods to iden-tify newborns with NAS. In addition to the limitations theydescribe (ie, identifying benzodiazepine withdrawal vs opi-ate withdrawal), current measures are also generally lim-ited to term or near-term infants.3 To advance the field, fu-ture research must not only include identification of infantswith antenatal exposure to a broad spectrum of medica-tions but also more varied populations (eg, preterm in-fants).

As Jones and Kaltenbauch suggest, there should berenewed attention to promote clinical advances andpolicy interventions aimed at limiting antenatal opiateexposure and improving care for newborns with NAS.Recently, Senators Casey (D, Pennsylvania) and Alexan-der (R, Tennessee) wrote the Centers for Medicare &Medicaid Services Acting Administrator MarilynTavenner to “reach out to state Medicaid direc-tors . . . and begin to formulate a plan to stop NAS.”4 In aseparate communication, Senator Schumer (D, NewYork) encouraged the Substance Abuse and MentalHealth Services Administration to work toward educatingclinicians to properly identify and treat newborns withNAS and also encouraged the Centers for Disease Controland Prevention and the National Institutes of Health to“ramp up” research efforts on NAS.5

Ultimately, we agree with Jones and Kaltenbach: futurework on NAS should involve approaches from multipledisciplines, designed to minimize and manage opiateexposure before pregnancy, treat mothers with evidence-based practice during pregnancy, and identify and pro-vide high-quality care for newborns with drug with-drawal.

Stephen W. Patrick, MD, MPH, MSRobert E. Schumacher, MDMatthew M. Davis, MD, MAPP

Author Affiliations: Division of Neonatal-Perinatal Medicine (Drs Patrick andSchumacher) ([email protected]), Robert Wood Johnson Foundation Clini-cal Scholars Program (Dr Davis), University of Michigan Health System, Ann Ar-bor.Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Patrick reportedreceiving grant and travel support from the Robert Wood Johnson Foundation Clini-cal Scholars Program. Drs Schumacher and Davis reported no disclosures.

1. Signs V. Overdoses of Prescription Opioid Pain Relievers—United States,1999–2008. MMWR Morb Mortal Wkly Rep. 2011;60(43):1481-1487.2. Kellogg A, Rose CH, Harms RH, Watson WJ. Current trends in narcotic use inpregnancy and neonatal outcomes. Am J Obstet Gynecol. 2011;204(3):259, e1-e4.3. Finnegan LP, Kron RE, Connaughton JF, Emich JP. Assessment and treatmentof abstinence in the infant of the drug-dependent mother. Int J Clin PharmacolBiopharm. 1975;12(1-2):19-32.4. Casey RP. Senators Casey and Alexander urge administrat ionto protect infants from harm of prescription drug abuse [press re-lease]. http://www.casey.senate.gov/newsroom/press/release/?id=5fe46461-015f-41de-b7da-68506670ea34. Accessibility verified July 5,2012.5. Schumer CE. Schumer pushes plan to battle huge spike in infants born ad-dicted to prescription drugs in Buffalo hospitals [press release]. http://www.schumer.senate.gov/Newsroom/record.cfm?id=336982. Accessibility verified July 5, 2012.

LETTERS
















LETTERS



ment protocols are needed to determine the most effectiveand cost-effective initiation, treatment, and weaning pro-tocols to optimize outcomes.

Future research must acknowledge that opioid useoccurs within a larger biopsychosocial context that con-founds the assessment of both the short- and longer-termeffects of in utero exposure to opioids. Poor neonatal out-comes often result from the concurrence of variousmaternal life experiences, including poverty and depriva-tion, physical, emotional, and sexual abuse, inadequatenutrition, and psychiatric comorbidity.5 Although illicitopioid addiction is one of the leading proximal causes ofNAS, meaningful reductions in incidence will likely onlyoccur when effective comprehensive woman-centeredtreatment programs for opioid-addicted pregnant womenare widely implemented.

Hendree E. Jones, PhDKarol Kaltenbach, PhD

Author Affiliations: RTI International, Research Triangle Park, North Carolina (DrJones; [email protected]); and Department of Pediatrics, Thomas Jefferson Univer-sity, Philadelphia, Pennsylvania (Dr Kaltenbach).Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Jones reported re-ceiving reimbursement for time and travel from Reckitt Benckiser Inc. Dr Kalten-bach reported no disclosures.

1. Patrick SW, Schumacher RE, Benneyworth BD, Krans EE, McAllister JM, DavisMM. Neonatal abstinence syndrome and associated health care expenditures: UnitedStates, 2000-2009. JAMA. 2012;307(18):1934-1940.2. Jones HE, Kaltenbach K, Heil SH, et al. Neonatal abstinence syndrome aftermethadone or buprenorphine exposure. N Engl J Med. 2010;363(24):2320-2331.3. American College of Obstetricians and Gynecologists. ACOG Committee Opin-ion No. 422: at-risk drinking and illicit drug use: ethical issues in obstetric and gy-necologic practice. Obstet Gynecol. 2008;112(6):1449-1460.4. Hayes MJ, Brown MS. Epidemic of prescription opiate abuse and neonatalabstinence. JAMA. 2012;307(18):1974-1975.5. Robins LN, Mills JL. Effects of in utero exposure to street drugs. Am J PublicHealth. 1993;83(suppl):1-32.

In Reply: Drs Jones and Kaltenbach comment on keyaspects of NAS and related research that may shape thenext decade of inquiry about this condition and its pre-vention and therapy. They suggest that the recent 3-foldincrease in NAS may be explained by heightened clini-cian awareness of the condition. We agree in principleand cited this as a potential limitation in the article. How-ever, the increasing trend in maternal use and abuse ofopiates also described in the study makes “heightenedawareness” less likely as a major factor in the increasingtrend. In addition, both maternal and newborn trendsappear to correlate with the general population trendtoward more opiate use and abuse.1

Jones and Kaltenbach correctly describe both mothersand newborns in our sample as heterogeneous popula-tions. Our study was limited by the inability to distin-guish the type and use of opiates among mothers. Asresearch in NAS continues to move forward, it is vitalthat investigators analyze the effects of different drugexposures on mothers and newborns. As an example,there is a paucity of data related to use of opiate pain

relievers during pregnancy and both short- and long-termnewborn outcomes.2

They also mention the need to improve methods to iden-tify newborns with NAS. In addition to the limitations theydescribe (ie, identifying benzodiazepine withdrawal vs opi-ate withdrawal), current measures are also generally lim-ited to term or near-term infants.3 To advance the field, fu-ture research must not only include identification of infantswith antenatal exposure to a broad spectrum of medica-tions but also more varied populations (eg, preterm in-fants).

As Jones and Kaltenbauch suggest, there should berenewed attention to promote clinical advances andpolicy interventions aimed at limiting antenatal opiateexposure and improving care for newborns with NAS.Recently, Senators Casey (D, Pennsylvania) and Alexan-der (R, Tennessee) wrote the Centers for Medicare &Medicaid Services Acting Administrator MarilynTavenner to “reach out to state Medicaid direc-tors . . . and begin to formulate a plan to stop NAS.”4 In aseparate communication, Senator Schumer (D, NewYork) encouraged the Substance Abuse and MentalHealth Services Administration to work toward educatingclinicians to properly identify and treat newborns withNAS and also encouraged the Centers for Disease Controland Prevention and the National Institutes of Health to“ramp up” research efforts on NAS.5

Ultimately, we agree with Jones and Kaltenbach: futurework on NAS should involve approaches from multipledisciplines, designed to minimize and manage opiateexposure before pregnancy, treat mothers with evidence-based practice during pregnancy, and identify and pro-vide high-quality care for newborns with drug with-drawal.

Stephen W. Patrick, MD, MPH, MSRobert E. Schumacher, MDMatthew M. Davis, MD, MAPP

Author Affiliations: Division of Neonatal-Perinatal Medicine (Drs Patrick andSchumacher) ([email protected]), Robert Wood Johnson Foundation Clini-cal Scholars Program (Dr Davis), University of Michigan Health System, Ann Ar-bor.Conflict of Interest Disclosures: The authors have completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Patrick reportedreceiving grant and travel support from the Robert Wood Johnson Foundation Clini-cal Scholars Program. Drs Schumacher and Davis reported no disclosures.

1. Signs V. Overdoses of Prescription Opioid Pain Relievers—United States,1999–2008. MMWR Morb Mortal Wkly Rep. 2011;60(43):1481-1487.2. Kellogg A, Rose CH, Harms RH, Watson WJ. Current trends in narcotic use inpregnancy and neonatal outcomes. Am J Obstet Gynecol. 2011;204(3):259, e1-e4.3. Finnegan LP, Kron RE, Connaughton JF, Emich JP. Assessment and treatmentof abstinence in the infant of the drug-dependent mother. Int J Clin PharmacolBiopharm. 1975;12(1-2):19-32.4. Casey RP. Senators Casey and Alexander urge administrat ionto protect infants from harm of prescription drug abuse [press re-lease]. http://www.casey.senate.gov/newsroom/press/release/?id=5fe46461-015f-41de-b7da-68506670ea34. Accessibility verified July 5,2012.5. Schumer CE. Schumer pushes plan to battle huge spike in infants born ad-dicted to prescription drugs in Buffalo hospitals [press release]. http://www.schumer.senate.gov/Newsroom/record.cfm?id=336982. Accessibility verified July 5, 2012.

LETTERS



ONLINE FIRST

RESEARCH LETTER

Competitive Sports Participation in AthletesWith Congenital Long QT Syndrome

To the Editor: Competitive sports participation for ath-letes with long QT syndrome (LQTS) is guided by the36th Bethesda Conference, which recommends thatpatients with either (1) symptoms, (2) a corrected QTinterval (QTc) greater than 470 milliseconds (males) or480 milliseconds (females), or (3) an implantablecardioverter-defibrillator (ICD) not participate in mostsports.1 The European Society of Cardiology guidelinesare more restrictive, disqualifying athletes from all sportsbased solely on a stringent QTc cutoff (�440 millisec-onds in males, �460 milliseconds in females).2 Wesought to determine the outcomes of patients with LQTSwho chose to remain athletes against guideline recom-mendations.

Methods. In this institutional review board–approvedstudy (with waiver of consent), we reviewed records forpatients with LQT1-3 genotypes, aged 6 to 40 years, whowere first evaluated in the Mayo Clinic LQTS Clinicbetween July 2000 and November 2010. Records werereviewed for athletic participation after LQTS diagnosis andLQTS-related events during a mean (SD) follow-up of 5.1(2.9) years. No patients were lost to follow-up. All werereevaluated or contacted by phone after July 1, 2011. A“competitive athlete” was defined as one participating inorganized competitive sports at the little league, middle orhigh school, collegiate, or professional level.

The approach in the Mayo Clinic LQTS Clinic is toprovide the athlete and their family with sufficient infor-mation to enable an informed decision regarding sportscontinuation. All patients received a comprehensive 2- to3-day clinical and genetic evaluation, including a 1- to2-hour consultation with an LQTS specialist (M.J.A.) andadditional consultations as needed.3-5 Extensive counsel-ing was provided to discuss individual prognosis and ath-letic participation guidelines. If a minor, the athlete andboth parents had to agree to sports continuation. Tailoredtherapy included �-blockers, left cardiac sympatheticdenervation, an ICD, or a combination. QT drug avoid-ance, electrolyte and hydration replenishment, and mini-mization of core body temperature elevations wereadvised. Each athlete obtained an automatic externaldefibrillator as part of the sports gear, and relevant schoolofficials and coaches were informed.

Statistical analysis was performed using JMP version 8.0 (SASInstitute). A 2-tailed P� .05 was considered significant.

Results. Of 353 LQT1-3 patients (199 females; mean[SD] age, 17 [11] years; mean [SD] QTc, 472 [42] milli-seconds), the majority (223, 63%) either were not

involved in sports (196, 88%) or chose to discontinuesports (27, 12%) following evaluation. Overall, 130patients (37%, 60 females; age, 11 [7] years; QTc, 471[46] milliseconds) remained in competitive athletics,including 20 with ICDs. There were no significant differ-ences between the total cohort and athletes except thatthe nonathletes were older (TABLE). The 130 athletescompeted in a variety of competitive sports (FIGURE),6

and 49 of 130 (38%) participated in more than 1 sport.There were 32 athletes (25%) competing in high schooland 8 (6%) competing at the college, university, or pro-fessional level.

Seventy athletes (54%) were competing contrary toEuropean guidelines but within Bethesda guidelines.None had a sport-related event. Of the 60 LQTS athletes(46%) continuing in sports contrary to both guidelines,only 1 experienced sporting-related events: a 9-year-oldboy with LQT1, extreme QT prolongation (QTc �550milliseconds), and a history of aborted cardiac arrest. Hereceived 2 appropriate ventricular fibrillation-terminatingICD shocks, both while warming up before games. Eachepisode occurred in the setting of admitted �-blockernonadherence.

The overall rate of events per athlete-year was 0.003(1 event in 331 athlete-years; 95% CI, 1 in 92 to 1 in 2763athlete-years).

Comment. Although many individuals with LQTS willchoose to remain in competitive sports, athletes and theirfamilies are capable of self-disqualification. With more than650 athlete-years of follow-up, we report a low rate of LQTS-triggered cardiac events during sports. Limitations includethe small sample size, limited length of follow-up, and un-known generalizability.

Jonathan N. Johnson, MDMichael J. Ackerman, MD, PhDAuthor Affiliations: Division of Pediatric Cardiology (Dr Johnson), Division of Car-diovascular Diseases (Dr Ackerman), Mayo Clinic, Rochester, Minnesota ([email protected]).Published Online: July 21, 2012. doi:10.1001/jama.2012.9334Author Contributions: Dr Ackerman had full access to all of the data in the studyand takes responsibility for the integrity of the data and the accuracy of the dataanalysis.Study concept and design: Johnson, Ackerman.Acquisition of data: Johnson.Analysis and interpretation of data: Johnson, Ackerman.Drafting of the manuscript: Johnson, Ackerman.Critical revision of the manuscript for important intellectual content: Johnson,Ackerman.Statistical analysis: Johnson.Obtained funding: Ackerman.Administrative, technical, or material support: Johnson, Ackerman.Study supervision: Ackerman.Conflict of Interest Disclosures: Both authors have completed and submittedthe ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Ackerman isa consultant for Biotronik, Boston Scientific, Medtronic, St Jude Medical, andTransgenomic. Intellectual property derived from Dr Ackerman’s research pro-gram resulted in license agreements in 2004 between Mayo Clinic Health Solu-tions (formerly Mayo Medical Ventures) and PGxHealth (formerly GenaissancePharmaceuticals and now Transgenomic). The Mayo Foundation for MedicalEducation and Research receives royalties from Transgenomic for the intellec-tual property used in its FAMILION LQTS genetic test. No other disclosureswere reported.Funding/Support: Dr Ackerman’s research program and this study were sup-

LETTERS



ported by the Mayo Clinic Windland Smith Rice Comprehensive Sudden CardiacDeath Program.Role of the Sponsor: The funding source had no role in the design and conduct ofthe study; in the collection, analysis, and interpretation of the data; or in the prepa-ration, review, or approval of the manuscript.Additional Contributions: We thank the patients who sought evaluation at Mayo’sLQTS Clinic, as well as Katy Harris, MS, RN (Mayo Clinic); Heidi J. Owen, RN (MayoClinic); and Carla M. Haglund (Mayo Clinic) for their assistance with patient follow-up. None of these individuals received compensation.

1. Zipes DP, Ackerman MJ, Estes NA III, Grant AO, Myerburg RJ, Van Hare G.Task Force 7: arrhythmias. J Am Coll Cardiol. 2005;45(8):1354-1363.2. Pelliccia A, Fagard R, Bjørnstad HH, et al; Study Group of Sports Cardiology ofthe Working Group of Cardiac Rehabilitation and Exercise Physiology; WorkingGroup of Myocardial and Pericardial Diseases of the European Society of Cardiology.Recommendations for competitive sports participation in athletes with cardiovas-cular disease: a consensus document from the Study Group of Sports Cardiologyof the Working Group of Cardiac Rehabilitation and Exercise Physiology and theWorking Group of Myocardial and Pericardial Diseases of the European Society ofCardiology. Eur Heart J. 2005;26(14):1422-1445.3. Ackerman MJ. Cardiac channelopathies: it’s in the genes. Nat Med. 2004;10(5):463-464.4. Goldenberg I, Moss AJ, Peterson DR, et al. Risk factors for aborted cardiac ar-rest and sudden cardiac death in children with the congenital long-QT syndrome.Circulation. 2008;117(17):2184-2191.5. Liu JF, Jons C, Moss AJ, et al; International Long QT Syndrome Registry. Riskfactors for recurrent syncope and subsequent fatal or near-fatal events in childrenand adolescents with long QT syndrome. J Am Coll Cardiol. 2011;57(8):941-950.6. Mitchell JH, Haskell W, Snell P, Van Camp SP. Task Force 8: classification ofsports. J Am Coll Cardiol. 2005;45(8):1364-1367.

CORRECTION

Incorrect Wording in Patient Page: In the Patient Page entitled “NeonatalHyperbilirubinemia,” published in the May 16, 2012, issue of JAMA (2012;307[19]:2115), incorrect wording was used. The lighting in the Figure shouldbe labeled “Special light source (blue-green spectrum),” and in the “Photo-therapy” section under “Treatment,” the words “ultraviolet (blue) lights”should read “a special light source (blue-green spectrum).” This article was cor-rected online.

Figure. Primary Sport Classification (According to the 2005 BethesdaConference Guidelines6) for the Competitive Athletes With Long QTSyndrome

Sta

tic C

ompo

nent

, Pea

k %

Max

imal

Volu

ntar

y Le

ft Ve

ntric

ular

Con

trac

tion

No. of Patients

Dynamic Component, Peak % Maximal Oxygen Uptake

IIIHigh>50

IIModerate

20-50

ILow<20

ALow<40

25

0

3

CHigh>70

0

34

28

BModerate

40-70

4

14

22

Table. Demographics of Competitive Athletes With Long QT Syndromea

Total Cohort Athletes

Within Bethesdabut Contrary toESC Guidelines

Contrary to BothGuidelines

PValueb

No. of patients 353 130 70 60

Age at diagnosis, mean (SD), y 17 (11) 11 (7) 11 (7) 12 (6) .91

Sex .21Male 154 70 41 29

Female 199 60 29 31

Mean (SD) QTc, ms 472 (42) 471 (46) 444 (23) 501 (46) �.001

Symptoms, No. (%) 111 (31) 29 (22) 1 (1) 28 (47) �.001

�-Blockers, No. (%) 280 (79) 112 (87) 55 (79) 57 (95) �.008

ICD, No. (%) 78 (22) 20 (15) 0 20 (33) �.001

Follow-up available, mean (SD), y 5.5 (3.4) 5.1 (2.9) 5.1 (2.9) 5.0 (3.0) .98

Genotype, No. (%)LQT1 182 (52) 74 (57) 41 (59) 33 (55) .65

LQT2 130 (37) 41 (32) 20 (29) 12 (35) .47

LQT3 37 (10) 11 (8) 8 (11) 3 (5) .20

Multiplec 4 (1) 4 (3) 1 (1) 3 (5) .19Abbreviations: ESC, European Society of Cardiology; ICD, implantable cardioverter-defibrillator; LQTS, long QT syndrome; QTc, corrected QT interval.aDemographics of the overall cohort of eligible LQT1-3 patients 6-40 years old (N=353), which includes the 130 LQTS athletes including those participating against ESC guidelines

but within Bethesda guidelines (n=70) and those athletes participating against both guidelines (n=60).bComparing athletes who are participating within Bethesda Conference guidelines but against ESC guidelines vs those participating against both guidelines.cPatients with �1 LQTS-causing mutation.

LETTERS



Methadone Overdose Deaths RiseWith Increased Prescribing for PainBridget M. Kuehn

METHADONE WAS INVOLVED IN

more than 30% of overdosedeaths linked to the use of

prescription painkillers in 2009, de-spite the drug making up only about 2%of painkiller prescriptions that year, ac-cording to a report from the US Cen-ters for Disease Control and Preven-tion (CDC).

Methadone, which has been usedsuccessfully for more than 40 years asa treatment for heroin addiction, hasbeen widely prescribed over the past de-cade for the treatment of pain. In 2009,more than 4 million prescriptions formethadone were written for pain pa-tients, according to the CDC. As thenumber of methadone prescriptions forpain has increased, methadone over-dose deaths have increased, noted CDCdirector Thomas R. Frieden, MD, MPH,in a press briefing. There are 5000 suchoverdose fatalities each year, more thanthe number of heroin and cocaine over-doses combined, he said.

“All the data suggest the increase isrelated to the use of methadone to treatpain,” said Frieden.

CHEAPER, GREATER RISK

Increasing numbers of methadonedeaths are part of a larger trend of grow-ing opioid use, misuse, and abuse thathas shadowed an effort by clinicians tomore aggressively manage pain. Twocharacteristics of methadone have con-tributed to the drug’s disproportionaterole in painkiller-related overdosedeaths: it is cheaper, and it carries greaterrisks than other drugs in this class.

The lower cost of methadone has ledstates and insurance companies to list

it as the preferred opioid medication intheir formularies, according to Frie-den. The drug also has a long and some-times hard-to-predict half-life, which canlead to toxic levels of the drug buildingup in patients and causing respiratorydepression, according to the CDC re-port (http://tinyurl.com/7lgxzez). Ad-ditionally, the drug may interact with an-tianxiety medications, which are alsooften prescribed to patients with painand also frequently abused.

Lewis Nelson, MD, an emergencyphysician and medical toxicologist atNew York University School of Medi-cine, explained in an interview thatmethadone is a difficult drug to use.Drug treatment programs, which op-erate under strict federal regulations,carefully control methadone induc-tion for the first 1 to 2 weeks that a pa-tient is in the program. Physicians usingthe drug to treat pain may know less

about the drug’s risks, Nelson said, andmay give doses too rapidly, causing re-spiratory depression and death amongsome patients.

“Many physicians, who are wellmeaning, don’t know how to use metha-done,” he said.

SOCIETAL COSTS

In Washington state, a Pulitzer Prize–winning investigative series docu-mented how the state’s decision to listmethadone as a preferred painkiller tocut costs contributed to increasing num-bers of overdoses among patients cov-ered by Medicaid (http://tinyurl.com/c7gbb4b).

During the briefing, Frieden de-scribed the use of methadone as a cost-effective way to treat pain as “penny-wise and pound-foolish.” He explainedthat although the drug itself is rela-tively inexpensive, it generates sub-

Methadone use for pain (kg/100 000 persons)Methadone-related overdose deaths per 100 000 personsMethadone prescriptions for pain per 100 persons

Rat

e

Year

3.0

2.5

2.0

1.5

1.0

0.5

0

Rates of methadone distribution for pain, methadone-relatedoverdose deaths, and methadone prescriptions for pain—

United States, 1999–2010

1999 2001 2003 2005 2007 2009

Source: Morbidity and Mortality Weekly Report, Centers for Disease Control and Prevention. http://www.cdc.gov/mmwr/PDF/wk/mm6126.pdf. Accessed July 10, 2012.

An increase inmethadone overdosedeaths has followedan increase inprescribing ofmethadone for pain,according to the USCenters for DiseaseControl andPrevention.

MEDICAL NEWS& PERSPECTIVES



stantial societal costs from increasednumbers of methadone-related emer-gency department visits and other con-sequences of overdoses.

Frieden said physicians should avoidusing methadone to treat patients withacute pain or chronic noncancer pain,noting that there is limited evidencesupporting such use. Instead, he said,physicians treating a patient with painnot caused by cancer should considerthe full range of medications and otherinterventions. “Use opiates only whennecessary, and use safer opioids,” hesaid.

Nelson agreed that there should bemore limited use of opioids for

chronic noncancer pain. He explainedthat while the risks of such use maybe tolerable in the short-term, therisks to individuals taking these medi-cations over the long-term are sub-stantial, with minimal evidence forbenefit.

“I don’t think methadone should beused for chronic [noncancer] pain,” hesaid. “It’s the least safe option.”

Physicians and regulators may needto change their approach to long-acting opioid medications, said Nel-son, who recently chaired the USFood and Drug Administration (FDA)Drug Safety and Risk ManagementAdvisory Committee. He explained

that the FDA cannot regulate thepractice of medicine and is so far rely-ing on pharmaceutical companies toeducate physicians about proper useof opioids. However, he said physi-cians also need other sources of edu-cation about these medications andsuggested that linking such educationto Drug Enforcement Agency (DEA)registration may be necessary. Duringa press briefing in July, FDA Commis-sioner Margaret Hamburg, MD, indi-cated that the FDA was supportingproposed legislation that would linksuch education with DEA registration.

“These drugs must be treated with ahealthy respect,” Nelson said. �

Experts Question Recommendationsfor Universal Lipid Screenings in ChildrenMike Mitka

RECOMMENDATIONS FOR UNIVER-sal lipid screening of children,presented late last year by a

National Heart, Lung, and BloodInstitute (NHLBI) panel, are drawingnew criticism. The aim of the screen-ing is to detect and treat abnormalcholesterol levels in childhood inhopes of reducing cardiovascular dis-ease risk much later in life.

The recommendations, publishedin December 2011 and endorsed bythe American Academy of Pediatrics,call for universal screening of 9- to11-year-old children with a nonfast-ing lipid panel, plus targeted screen-ing with 2 fasting lipid profilesof children ages 2 to 8 years and 12to 16 years (involving between 30%and 40% of all children). Previousrecommendations called for screen-ing only children considered highrisk using a nonfasting total choles-terol test.

Children who are identified ashaving abnormal lipid levels wouldthen be treated mostly through life-

style modification involving diet andexercise. Medical therapy, mostlywith statins, would be an optionfor children with severely abnormallipid levels and would affect less than1% of children, other research hassuggested.

CRITICS’ CONCERNS

The critics, from the University ofCalifornia, San Francisco (UCSF),argue the NHLBI screening recom-

mendations were made without pro-viding estimates of the health ben-efits, harms, and costs that mightresult from such screening. They alsoargue that the evidence used tojustify such screening is not as strongas implied by the NHLBI and thatthe recommendations are basedheavily on expert opinion. They alsoexpress concerns that the NHLBIpanel members had conflicts of inter-est with industry that could have

Lifestyle modificationthrough exercise anddiet remains thefront-line treatmentfor children withabnormal lipid levels.

MEDICAL NEWS & PERSPECTIVES



affected their ability to provide unbi-ased recommendations (Newman TBet al. Pediatrics. doi:10.1542/peds.2012-0481 [published online July 23,2012]).

In a published response, membersof the NHLBI said such universalscreening is justified, as it will helpidentify children with familial hyper-cholesterolemia, which affects about1 in 500 people and leads to earlierdevelopment of clinical cardiovascu-lar disease and increased risk formyocardial infarction, sometimesstarting before age 40 years. Theydefended their use of evidence, say-ing it is the best available for a popu-lation in which outcomes that willoccur decades into the future are dif-ficult to test using standard studymethods. They also stated that theNHLBI recommendation process wasevidence-based and specified before-hand and that all relationships withindustry were declared (McCrindleBW et al. Pediatrics. doi:10.1542/peds.2012-1137 [published online July 23,2012]).

LOW-GRADE EVIDENCE

One of the critics, Thomas B. New-man, MD, MPH, a professor in UCSF’sDivision of Epidemiology, said the mainconcerns for him and his colleagueswere the low grade of scientific evi-dence used as the basis of the recom-mendations and the panel members’ tiesto industry. “We disagreed with the pro-cess in which the NHLBI selected thepanel and subsequent weighing of theevidence,” said Newman in an inter-view. “When guidelines are based onexpert opinion, it is important to avoidconflict of interest.”

Newman also noted that the NHLBImade these recommendations eventhough the US Preventive ServicesTask Force has stated since 2007 thatthe evidence is insufficient to recom-mend for or against routine screeningfor lipid disorders in infants, children,adolescents, or young adults up to age20 years.

Stephen R. Daniels, MD, PhD, whochaired the NHLBI panel and is

chairman of the department of pedi-atrics at the University of ColoradoSchool of Medicine, said that univer-sal screening is needed, as the tradi-tional use of family history too oftenfails to identify children at risk, andthat studies of less-than-desired rigorwere used because of the long lagtimes between identifying a markerfor risk and seeing an actual adverseoutcome. “You are talking aboutdeveloping risk for outcomes that arelikely to occur in adulthood, so thereis this time between screening andusing potential interventions andchanges in outcomes, and that makesit very difficult for us to collect someof the types of evidence that couldbe helpful to those making guide-lines,” Daniels said. “If you look atthe kinds of evidence that haveevolved over the past 30 to 40 years,you see atherosclerosis begins inchildhood.”

Meanwhile, as the debate on thescreening guidelines continues,research has turned up good newsabout childhood lipids. An August 8,2012, study by the US Centers forDisease Control and Prevention foundthat from 1988-1994 through 2007-2010 in youths aged 6 to 19 years,mean total cholesterol decreased from165 mg/dL to 160 mg/dL, and theprevalence of elevated total choles-terol decreased from 11.3% to 8.1%(Kit BK et al. JAMA. 2012;308[6]:591-600).

Sarah D. de Ferranti, MD, MPH, anassistant professor of pediatrics atHarvard Medical School and BostonChildren’s Hospital, who wrote anEditorial accompanying the JAMAstudy, supports universal screening.“Guidelines are moving targets thatare developed in the setting of theavailable evidence,” she said inan interview. “This new study isinteresting and contributes to ourunderstanding of cholesterol in chil-dren, but I would say we have notsolved the problem of childhoodcholesterol and adult heart disease,and there are still many children atrisk.” �

news@JAMAFrom JAMA’s Daily News Site

Pregnant Women Are DrinkingAbout 1 in 13 pregnant women drinks alco-hol during pregnancy and 1 in 70 binge drink.

An analysis of 2006-2010 Behavioral RiskFactor Surveillance System data estimated that7.6% of pregnant women had at least 1 alco-holic beverage in the 30 days prior to beingsurveyed, and 1.4% of pregnant women hadat least 1 binge drinking session (4 or moredrinks) during that time frame.

http://tinyurl.com/7bv3ujs

Housing Woes Harm ChildrenThe flurry of delin-quent mortgages andhome foreclosuresfollowing the 2008recession has beenlinked by research-ers with an increasedrisk for child abuse.

From 2000 to2009, admissions ofchildren and infants in 38 hospitals for physi-cal abuse and high-risk traumatic brain in-jury—an indicator of abuse—increased eachyear while all-cause injury rates declined. Ad-missions were linked with a community’s cur-rent mortgage delinquency rate and withchanges in delinquency and foreclosure ratesfrom the previous year.

http://tinyurl.com/7meeccg

No Benefit From Milk ThistleSilymarin, an extract of milk thistle that’spopular among patients with hepatitis C–re-lated liver conditions, doesn’t appear to im-prove clinical outcomes or quality of life.

In a recent study, 154 patients with chronichepatitis C virus infection who did not ben-efit from standard interferon therapy re-ceived 420-mg silymarin, 700-mg silyma-rin, or a placebo 3 times daily for 24 weeks.Results showed that silymarin didn’t reduceviral RNA levels, improve liver function, orenhance quality of life.

http://tinyurl.com/7ohr9kp

Babesiosis WarningPeople in the northeastern or upper mid-western United States have more reason tobe vigilant about preventing tick bites. Sur-veillance reports show that tick-borne babe-siosis is common in the United States.

Reports from 18 states and New York Cityidentified 1124 confirmed and probable ba-besiosis cases, including 10 associated withblood transfusion. Most were reported inConnecticut, Massachusetts, Minnesota, NewJersey, New York, Rhode Island, and Wis-consin. Symptom onset usually was duringsummer months.

http://tinyurl.com/6vz3qhs

For more on these stories and other medicalnews, visit http://newsatjama.jama.com/.

Reicaden/iStockphoto.com

MEDICAL NEWS & PERSPECTIVES



Source: Kaiser Family Foundation (www.kff.org) analysis. Original data and detailed source information are available at http://facts.kff.org/jama_082212.*Produced by: Rachel Garfield, PhD, Robin Rudowitz, MPA, Barbara Lyons, PhD, Anne Jankiewicz, and David Rousseau, MPH

VISUALIZING HEALTH POLICY


Kaiser Family Foundation*


Military PTSD CareFewer than half of the individuals whoscreened positive for posttraumatic stressdisorder (PTSD) after serving with theUS military in Iraq or Afghanistan havebeen referred for further evaluation ortreatment, according to a report (http://tinyurl.com/bn9dwlk) from the Insti-tute of Medicine on the handling of PTSDby the Departments of Defense (DOD)and Veterans Affairs (VA).

According to the report, 13% to 26%of the 2.6 million service members whohave served in these war theaters since2001mayhavePTSD.Butonly40%ofin-dividuals who screen positive for PTSDarereferred for furthercare, andof these,only 65% receive care. The report notesthatsomepatientsmayforgocarebecausethey are concerned that the stigma asso-ciatedwithapsychiatricdisorderwillside-track their career. Individuals may alsostruggle to access care because of logis-ticaldifficulties incombatsettingsorgeo-graphic and other barriers to care oncethey have returned to the United States.

Thereport recommendsthatDODcli-nicians caring for military personnel an-nuallyassesspatients forPTSDasVAcli-nicianscaringforveteransdoandthatbothdepartments better assess the care theircliniciansprovide topatientswithPTSD.

Egg Allergy TreatmentExposure to tiny amounts of pow-dered egg white over many months mayhelp resolve egg allergies for many chil-dren, according to results of a studyfunded by the National Institute of Al-lergy and Infectious Diseases.

Ateamofresearchersconductedaclini-cal trial of 55 children aged 5 to 11 yearswithaneggallergy todeterminewhetherexposuretotinyamountsofeggovertimemight desensitize children to egg. Thisstrategyhaspreviouslyshownpromiseforindividuals with peanut allergy (BurkesAW et al. N Engl J Med. doi:10.1056/NEJMoa1200435 [published online July

19, 2012]). The children were random-ized to receive a placebo or an oral doseof powdered egg white daily. At 10months, more than half of the treatedchildren were desensitized, and at 22months, 75% of the treated children were

desensitized. None of the children re-ceiving placebo were desensitized. Af-ter 22 months, desensitized children dis-continued therapy and avoided egg for4 to 6 weeks; of these individuals, 28%remained desensitized when retested.

Smoking Cessation RemindersFewer patients reported being advised toquit smoking by their physicians in 2010than in 2005, according to a study ledby scientists from the US Centers for Dis-ease Control and Prevention (CDC).

Advice fromaphysician toquit smok-ing has been shown to have a positive ef-fect on patients’ cessation rates, but fewstudies have examined how often physi-ciansoffersuchadvice.Toassessthisphy-sician behavior, a team of scientists ana-lyzeddata fromthe2000,2005,and2010Cancer Control Supplement of the Na-tional Health Interview Survey to deter-minehowmany individuals reportedre-ceivingcessationadvice fromphysicians(Kruger J et al. Prev Chronic Dis. 2012;9:

E130).Theyfoundthat thepercentageofindividuals who reported that they re-ceived cessation advice increased from53.3%in2000to58.9%in2005but thendeclinedto50.7%in2010.Womenweremorelikelythanmentobeadvisedbytheirphysiciansaboutsmokingcessation.Olderagewasassociatedwithanincreasedlike-lihoodof receivingcessationadvice.His-panicindividualswerelesslikelytoreceivecessation advice than members of otherethnic groups.

Fetal Alcohol ExposureChildren whose mothers engaged inheavy alcohol consumption during preg-nancy may have cognitive deficits, evenif they do not have the abnormal facialfeatures associated with fetal alcohol syn-drome, according to a study funded bythe Eunice Kennedy Shriver NationalInstitute of Child Health and HumanDevelopment.

To determine the prevalence of indi-vidual symptoms of fetal alcohol syn-drome in children whose mothers con-sumed large amounts of alcohol duringpregnancy, the scientists compared dataon 101 pregnant mothers in Chile whodrank at least 4 drinks per day with amatched group of 101 mothers who didnot drink alcohol during pregnancy(Kuehn D et al. Alcohol Clin Exp Res. doi:10.1111/j.1530-0277.2012.01794.x.[published online July 23, 2012]). Theyfound that 44% of the children born toheavy maternal drinkers had func-tional central nervous system abnor-malities, compared with 13.6% ofchildren who were not exposed to alco-hol. Additionally, 27.2% of the ex-posed children had growth restriction,compared with 12.5% of the controlgroup, and 17.3% of the exposed chil-dren had facial abnormalities, com-pared with 1.1% of the control group.However, one limitation of the study wasthat data to assess abnormalities wereavailable for only about half of the ex-posed children.—Bridget M. Kuehn

Exposure to tiny amounts of egg over manymonths may desensitize children with eggallergies, a study has found.

Yin

Yan

g/iS

tock

phot

o.co

m

HEALTH AGENCIES UPDATE



UnexplainedRespiratory DiseaseOutbreak WorkingGroup Activities—Worldwide, March2007–September 2011

MMWR. 2012;26:480-483

2 tables omitted.

THE UNEXPLAINED RESPIRATORY DIS-ease Outbreak (URDO) working groupis a multidisciplinary team composedof approximately 40 scientists fromacross CDC with expertise in infec-tious and noninfectious respiratory dis-eases. The URDO working group wasformed in 2004 to streamline CDC re-sponse efforts to assist local, state, andinternational public health officials ininvestigations of unexplained respira-tory disease outbreaks. This report sum-marizes URDO working group activi-t ies from March 2007 throughSeptember 2011. During that period,the URDO working group was noti-fied of 57 investigations and facili-tated consultations with subject mat-ter experts (in all 57 investigations),laboratory testing at CDC (in 42 inves-tigations), and on-site field investiga-tion support (in eight investigations).Of these 57 investigations, 41 oc-curred domestically, and 16 occurredinternationally. An etiology was iden-tified in 29 (51%) investigations; amongthese, the most commonly implicatedpathogens were noninfluenza respira-tory viruses (41%), influenza viruses(17%), Mycoplasma pneumoniae (14%),and Bordetella pertussis (14%). Notifi-cation occurred a median of 33 days af-ter illness onset in the first case, whichmight have limited the ability to col-lect early laboratory specimens or epi-demiologic data. Reducing delays insample collection, epidemiologic in-

vestigations, and consultation with theURDO working group might increasethe ability to identify etiologies and leadto more rapid control of these unex-plained respiratory disease outbreaks.

The objectives of this analysis wereto describe the investigations re-ported to the URDO working groupfrom March 2007 through September2011 and identify opportunities to im-prove the URDO working group’spublic health response. The URDOworking group might be notified of un-explained respiratory disease out-breaks before the initiation of an in-vestigation or during an ongoinginvestigation if uncertainty exists re-garding the etiology or co-etiologies.Notifications were tracked beginning inMarch 2007, and correspondence fromURDO working group investigators wascollected. Because these materials of-ten contained preliminary informa-tion from evolving investigations, thisinformation was supplemented withfield investigation final reports and pub-lications, when available.1-6 Investiga-tions were classified by the level of sup-port provided, affected age group(children, adults, or both), setting, andwhether the etiology was determined.Medians and ranges for the number ofdays from illness onset in the first caseto notification were calculated, as werenumbers of cases, hospitalizations, anddeaths.

During March 2007–September2011, the URDO working group wasnotified by state or international pub-lic health officials of 57 investiga-tions. Of these, 41 occurred in theUnited States and its territories, and 16occurred internationally (FIGURE). No-tifications per year ranged from eightto 15. The median time from illness on-set in the first case to notification (avail-able for 37 investigations) was 33 days(range: 4–218 days). For all investiga-tions, the URDO working group pro-vided input through either telephoneor e-mail consultation. Depending on

the specific needs of the requestor, ad-ditional assistance ranged from advicegiven via conference calls with theworking group (in 70% of investiga-tions), to laboratory testing at CDC(74%), to on-site epidemiologic inves-tigation assistance (14%). Once anetiology was identified or highly sus-pected, laboratory testing and investi-gation assistance were provided by CDCdivisions with relevant subject matterexpertise. Laboratory testing includedmolecular diagnostics for respiratorypathogens using polymerase chain re-action (PCR), serology, culture, histo-pathology, immunohistochemistry, andurine antigen testing. Additionally, in13 investigations, the URDO working

What is already knownon this topic?

Respiratory disease outbreaks canpresent investigation challenges, es-pecially because of the many poten-tial etiologies with overlapping clini-cal presentations.

What is added by this report?The Unexplained Respiratory Dis-ease Outbreak (URDO) workinggroup provided varying forms of sup-port, ranging from telephone con-sultation to laboratory testing to on-site field investigation assistance for57 domestic and international inves-tigations during March 2007–September 2011. A cause was foundfor 51% of the outbreaks investi-gated. The most common causeswere influenza viruses, Mycoplasmapneumoniae, Bordetella pertussis, andnoninfluenza respiratory viruses.

What are the implicationsfor public health practice?Reducing delays in sample collec-tion, epidemiologic investigations, andconsultation with the URDO work-ing group might increase the ability toidentify etiologies and lead to morerapid control of these unexplained re-spiratory disease outbreaks.

Morbidity and Mortality Weekly Report

FROM THE CENTERSFOR DISEASE CONTROLAND PREVENTION



group provided laboratory support withTaqMan Array Cards (Life Technolo-gies, Grand Island, New York), a PCR-based technology that tests simultane-ously for approximately 20 respiratorypathogens.7

Fifty-one investigations involved twoor more cases. Of the other six investi-gations, one involved a pseudo-outbreak caused by clinical specimencontamination,4 and five involved singlecase consultations. The exact case countwas available for 49 investigations; themedian case number was 15 (range: oneto 409 cases). The number of hospital-izations was available in 32 (56%) in-vestigations (median: three; range: zeroto 18). Among 26 (46%) investigationswith two or more cases and numbers ofcases and hospitalizations reported, themedian percentage of cases resulting inhospitalization was 23% (range: zero to100%). A total of 36 (63%) investiga-tions had the number of deaths re-ported (median: zero, range: zero to 12,with one outlier from a wild-type polio-virus outbreak with 169 deathsreported).3 The URDO working groupwas notified of this outbreak because itappeared initially to have a respiratorycomponent. Age ranges of affectedpersons were reported in 52 (91%) in-vestigations; adults were affected mostcommonly. Communities (i.e., nonin-stitutional settings) and long-term–care facilities were the most common set-tings for outbreaks.

The etiology was determined in 29(51%) investigations, based on the in-terpretation of laboratory results withclinical and epidemiologic informa-tion. The most commonly identified eti-ologies were influenza viruses, Myco-plasma pneumoniae, Bordetella pertussis,and noninfluenza respiratory viruses(e.g., respiratory syncytial virus, ad-enovirus, and parainfluenza virus). Five(9%) investigations involved multipleetiologies. CDC provided laboratorysupport for 24 (83%) of 29 investiga-tions with confirmed etiologies and for18 (64%) of 28 that remained unex-plained. Among 13 investigations in-volving TaqMan Array Cards, the eti-ology was identified for six (e.g.,

parainfluenza virus 3, Streptococcuspneumoniae, Chlamydophila pneumo-niae, human parechovirus, humanmetapneumovirus, rhinovirus, and hu-man enterovirus 68).6 Of the seven in-vestigations involving TaqMan ArrayCards for which the etiology re-mained unclear, two or more patho-gens were identified in five.

Reported by: Paul R. Cieslak, MD, Oregon Health Au-thority. April S. Britt, PharmD, Lauri Hicks, DO, LauraConklin, MD, Chris Van Beneden, MD, Laurel E. Gar-rison, MPH, Jonas Winchell, PhD, Respiratory Dis-eases Br, Div of Bacterial Diseases; Eileen Schneider,MD, Dean Erdman, DrPH, Div of Viral Diseases; Ali-cia Fry, MD, Seema Jain, MD, Tim Uyeki, MD, LynFinelli, DrPH, Steve Lindstrom, PhD, Influenza Div;Thomas A. Clark, MD, Maria-Lucia Tondella, PhD,Meningitis and Vaccine Preventable Diseases Br, Divof Bacterial Diseases, National Center for Immuniza-tions and Respiratory Diseases; Wun-Ju Shieh, MD,Sherif Zaki, MD, PhD, Div of High-Consequence Patho-gens and Pathology, National Center for Emerging andZoonotic Infectious Diseases, for the Unexplained Re-spiratory Disease Outbreak Working Group; Kather-ine E. Fleming-Dutra, MD, EIS Officer, CDC. Corre-sponding contributor: Katherine E. Fleming-Dutra, [email protected], 404-639-4243.

Editorial Note: Respiratory disease out-break investigations present several chal-lenges. Clinical presentation alone usu-ally is insufficiently distinct to permitidentification of an etiology. Respira-

tory specimens of good quality can be dif-ficult to obtain in a timely manner, andlocal laboratory capability to test for mul-tiple potential causes often is limited.Furthermore, rapid identification of theetiology is important for timely imple-mentation of control measures (e.g., ap-propriate infection control, vaccina-tion, and chemoprophylaxis). Toenhance public health responses to un-explained respiratory outbreaks,CDCin-tegrated epidemiologic and laboratoryexpertise from across the agency to formthe URDO working group to advise, con-duct laboratory testing, and facilitateCDC assistance for these outbreaks. Dur-ing the evaluation period (March 2007–September 2011), the URDO workinggroup was notified of 57 investigations,29 (51%) of which resulted in the iden-tification of the etiology. Common re-spiratory pathogens, such as influenza vi-ruses,Mycoplasmapneumoniae,Bordetellapertussis, and noninfluenza respiratoryviruses, were identified as the etiology forseveral outbreaks. Additionally, theURDO working group assisted with in-vestigations of illnesses with respira-tory symptoms that were determined to

FROM THE CENTERS FOR DISEASE CONTROL AND PREVENTION



be primarily nonrespiratory (e.g., Rick-ettsia rickettsii infection and polio). TheURDO working group has capacity toprovide rapid, multipathogen testingwith TaqMan Array Cards for investiga-tions in which several etiologies are un-der consideration.

Information is available to guide in-vestigations of unexplained respira-tory disease outbreaks. These materi-als include documents that can assistpublic health officials with establish-ing case definitions, forming line lists,and generating epidemic curves, as wellas a sample respiratory illness ques-tionnaire and instructions for speci-men collection.8 Of particular impor-tance are disease outbreaks that 1)might be interrupted by timely vacci-nation,5 environmental interventions,or other control methods; 2) occur ininstitutional settings or among vulner-able populations; 3) might involve bio-terrorism agents; 4) are severe, large,or rapidly progressive; or 5) cause pub-lic concern.8,9 Such outbreaks mightwarrant notification of the URDO work-ing group and further investigation.

The findings in this report are sub-ject to at least three limitations. First,the fraction of unexplained respira-tory outbreaks that are reported is notknown; reported outbreaks might notrepresent the distribution of etiolo-gies among all unexplained respira-tory outbreaks. Second, the URDOworking group is provided informa-tion while investigations are under waybut does not systematically collect fi-nal reports from local public health of-ficials. Therefore, the data in this re-port likely underestimate case counts,hospitalizations, and deaths. Finally, in-formation was not collected about un-explained respiratory disease investi-gations occurring before March 2007,so determining whether the URDOworking group has improved identifi-cation of etiologies or outbreak miti-gation is not possible.

An etiology was identified in onlyabout half of all investigations by theURDO working group. This might re-sult, in part, from delays in clinicalspecimen collection because many

pathogens (e.g., influenza) only can bedetected for a short time after illness on-set.10 The URDO working group wasnotified of each outbreak a median of33 days after illness onset in the firstcase. These delays might have re-sulted from delayed outbreak recogni-tion or because local investigatorswaited until local testing failed to iden-tify the etiology. Reducing the time be-tween outbreak recognition and noti-fication might increase the likelihoodthat optimal respiratory specimens arecollected. However, even when timelyspecimens are obtained, sensitive andspecific laboratory diagnostic testsmight not be available locally or at CDCfor some known pathogens and for newpathogens. Finally, gaps in availableclinical and epidemiologic informa-tion might decrease the URDO work-ing group’s ability to determine plau-sible etiologies and to recommendappropriate diagnostics.

Health-care providers and facilitiesare encouraged to report suspected out-breaks early to local public health of-ficials, and health officials are invitedto consult the URDO working groupearly in the course of any unexplainedrespiratory disease investigation. Ad-ditionally, CDC recommends that pub-lic health officials collect and store clini-cal specimens as an investigationevolves for potential future testing. TheURDO working group provides an ex-ample of a successful, interdisciplin-ary approach to providing assistance topublic health professionals in theUnited States and abroad. The URDOworking group can be contactedthrough CDC’s Emergency Opera-tions Center by telephone, at 770-488-7100. Additional information and re-sources are available at http://emergency.cdc.gov/urdo.

AcknowledgmentsMembers of the URDO working group: Sasha Kli-mov, Mike Jhung, Ashley Fowlkes, Greg Armstrong,Larry Anderson, David Shay, Gayle Fischer Langley,Nicole Alexander, Barry Fields, Marc Fischer.

REFERENCES

10 Available.

Work-RelatedAsthma—38 States andDistrict of Columbia,2006–2009MMWR. 2012;61:375-377.

1 Table omitted.

WORK-RELATED ASTHMA (WRA) IN-cludes work-exacerbated asthma (pre-existing or concurrent asthma wors-ened by factors related to the workplaceenvironment) and occupational asthma(new onset asthma attributed to theworkplace environment).1,2 WRA is apreventable occupational lung diseaseassociated with serious adverse healthand socioeconomic outcomes.1,2 Amongworkers with similar occupational ex-posures, WRA diagnosis offers uniqueopportunities for prevention.2,3 TheAmerican Thoracic Society estimatedthat 15% of U.S. adults with asthmahave asthma attributable to occupa-tional factors.3 State-level informationon the proportion of asthma that isWRA is limited but could be useful toprioritize and guide investigations andinterventions. To estimate currentasthma prevalence and the proportionof asthma that is WRA, CDC analyzeddata from the 2006–2009 BehavioralRisk Factor Surveillance System(BRFSS) from 38 states and the Dis-trict of Columbia (DC). This reportsummarizes the results of that analy-sis, which indicated that among ever-employed adults with current asthma,the overall proportion of current asthmathat is WRA was 9.0%. State-specificproportions of asthma that are WRAranged from 4.8% to 14.1%. Propor-tions of WRA were highest among per-sons aged 45–64 years (12.7%), blacks(12.5%), and persons of other races(11.8%). These findings provide a base-line that state and national health agen-cies can use to monitor the proportionof WRA among persons with currentasthma. Enhancing WRA surveillancethrough routine collection of industry




and occupation information will greatlyincrease understanding of WRA.

BRFSS is a state-based, random-digit–dialed telephone survey of the nonin-stitutionalized U.S. civilian popula-tion aged �18 years. The survey collectsinformation on health risk behaviors,preventive health practices, health-care access, and disease status.* In 2005,the Asthma Call-Back Survey (ACBS)†was pilot tested in three states and hasbeen conducted every year since. ACBScollects detailed information on asthma,including data on asthma symptoms,health-care utilization, medication use,knowledge of asthma, cost of asthmacare, work-related asthma, comorbidconditions, and complementary and al-ternative medicine use for asthma.BRFSS respondents are eligible to par-ticipate in ACBS if they answer “yes”to the question, “Have you ever beentold by a doctor, nurse, or other healthprofessional that you had asthma?”Those who agree are contacted to par-ticipate in ACBS within 2 weeks of theBRFSS completion date. Data fromBRFSS and ACBS for 2006–2009 from38 states and DC are included in thisanalysis. The Council of American Sur-vey and Research Organizations me-dian response rates among the 38 statesand DC ranged from 47.5% in 2007 to51.4% in 2009 for BRFSS and from47.2% in 2009 to 54.3% in 2007 forACBS.

For this analysis, participants inBRFSS and ACBS who responded “yes”to the questions, “Have you ever beentold by a doctor, nurse, or other healthprofessional that you had asthma?” and“Do you still have asthma?” were listedas having current asthma. ACBS par-ticipants were considered to be ever-employed if they indicated that theycurrently were “employed full-time” or“employed part-time” or that they hadever been employed outside the home.Ever-employed adults with currentasthma who responded “yes” to thequestion, “Were you ever told by a doc-tor or other health professional thatyour asthma was related to any job youever had?” were classified as havingWRA.

Combined data for 2006–2009were weighted to account for unequalprobability of sample selection andnonresponse differences in thesample.§ For states with multipleyears of data, annual weights wereproportionately adjusted based on thenumber of years and the sample sizein each year. Statistical software wasused to calculate estimates and 95%confidence intervals (CIs), accountingfor the complex survey design. Statis-tically significant differences in distri-bution were determined using theRao-Scott chi-square test of indepen-dence (p�0.05).

During 2006–2009, in the 38 statesand DC included in the analysis,1,082,135 adults participated in BRFSS(representing an estimated annual av-erage of 198 million adults), and 56,097adults participated in ACBS (represent-ing an estimated annual average of 26million adults). During this period, anestimated 8.4% of adults had currentasthma. The prevalence of currentasthma significantly differed by age, sex,and race/ethnicity.¶ Prevalence waslowest among persons aged �65 years(7.6%), men (6.3%), and Hispanics(6.3%). State-specific estimates of theprevalence of current asthma rangedfrom 6.3% to 10.4%.

A total of 38,306 adults whoparticipated in ACBS were ever-employed and had current asthma,representing an estimated 16 millionadults in the 38 states and DC. Ofthese, the estimated proportion whohad WRA was 9.0% (representing anestimated annual average of 1.4 mil-lion adults). Distributions of the pro-portion of WRA differed significantlyby age and race/ethnicity and werehighest among persons aged 45–64years (12.7%), blacks (12.5%), andpersons of other races (11.8%). Thees t imated propor t ion of ever -employed adults with current asthmawho had WRA was similar amongmen (9.1%) and women (8.9%). Bystate, the estimated proportions ofever-employed adults with currentasthma who reported WRA rangedfrom 4.8% to 14.1%.

Reported by: Gretchen E. Knoeller, MPH, Jacek M. Ma-zurek, MD, Div of Respiratory Disease Studies, Na-tional Institute for Occupational Safety and Health;Jeanne E. Moorman, MS, Div of Environmental Haz-ards and Health Effects, National Center for Environ-mental Health, CDC. Corresponding contributor:Gretchen E. Knoeller, [email protected], 304-285-5838.

Editorial Note: The results of this analy-sis indicate that exposures in the work-place continue to contribute to asthmamorbidity among adults in the UnitedStates and that blacks with asthma ap-pear to be affected disproportionatelyby occupational conditions. Amongadults who have ever been employed,an estimated annual average of 1.4 mil-lion WRA cases could have been pre-vented. These findings are consistentwith the estimated proportion of adultasthma that is WRA reported from the2005 ACBS in Michigan (7.6%[CI=4.9%–10.3%]), Minnesota (5.6%

What is already knownon this topic?

Work-related asthma, one ofthe most common occupationallung diseases, is preventable but of-ten undiagnosed.

What is added by this report?These results indicate that an esti-mated annual average of 1.4 millioncases of adult asthma (9.0% of cur-rent asthma cases among ever-employed adults) could have beenprevented and that ever-employedblacks with current asthma are dis-proportionately affected by work-related asthma.

What are the implicationsfor public health practice?Enhancing surveillance for work-related asthma through routine col-lection of data on industry, occupa-tion, and workplace exposures couldgreatly expand understanding of po-tential causes and triggers. Such in-formation could be useful to state andlocal health departments to guide in-vestigation and prevention efforts,such as the use of engineering andadministrative controls to diminishthe current burden of work-relatedasthma.




[CI=2.9%–8.2%]), and Oregon (9.0%[CI=6.7%–11.4%]).4

Strategies to reduce or eliminateworkplace exposures for persons withWRA range from substitution of chemi-cals to engineering and administrativecontrols and will aid in the preventionof new cases and slow the progressionof subclinical cases in the same work-place.2,5 For example, in the early 1990s,health-care workers and other work-ers exposed to powdered, non–rubberlatex gloves experienced high inci-dence of WRA. After recommenda-tions were made to change the type ofglove used and to reduce the powderand non–rubber latex protein contentof the gloves if they needed to be used,considerable reductions in the occur-rence of WRA were observed in thehealth-care industry.5 Another ex-ample is the substantial reduction inWRA prevalence among workers inthe detergent industry after detergentenzymes were encapsulated duringthe production process to reduceexposure.5

Continued administration of ACBSwill allow state asthma programs tomonitor the proportion of asthma thatis WRA. Information on WRA respon-dents’ industry and occupation is nec-essary to guide the development of suc-cessful intervention strategies. WRAmanagement and prevention includesa public health aspect (i.e., work-places suspected to pose a high risk fordevelopment of WRA should be inves-tigated, and appropriate exposure con-trol measures should be implementedto prevent WRA).1

The findings in this report are sub-ject to at least six limitations. First, re-sults likely are underestimates of the ac-tual proportion of WRA because WRAis underdiagnosed in the UnitedStates.6,7 Second, ACBS might be sub-ject to selection bias because BRFSS re-spondents with asthma were asked ifthey agreed to be called back for ACBS.Those who agreed to participate inACBS might have more severe asthmaor might be more likely to attributeasthma to their work.8 No informa-tion on asthma symptoms or work-

relatedness was available in BRFSS forthose who refused to participate. Third,BRFSS was not designed to allow as-sessment of the prevalence of currentasthma among ever-employed adults.Therefore, findings on the prevalenceof current asthma and the proportionof current asthma that is WRA weredetermined based on different denomi-nator populations and should be inter-preted with caution. Fourth, no infor-mation on industry and occupation wasavailable for these participants. Infor-mation on industry and occupation forWRA cases is limited because CDC’ssentinel-event surveillance currently isconducted only in selected states.9 Fifth,exclusive use of landline telephones insome years might mean some groupsare underrepresented in the sample.10

Finally, because ACBS had low re-sponse rates and data are limited to the38 states and DC that conducted thesurvey, these estimates are not gener-alizable to the entire U.S. populationand do not represent the populationsof nonparticipating states.

Currently, CDC provides technicaland financial assistance to five states(California, Massachusetts, Michigan,New Jersey, and New York) to con-duct expanded WRA surveillance.**These systems collect in-depth, case-based information on WRA cases, in-cluding workplace exposure and em-ployment information, but do not allowassessment of WRA burden in thepopulation. For many states, ACBS pro-vides the only state-based estimates ofWRA, and some states already have ini-tiated the collection of information onindustry and occupation in BRFSS. In2013, CDC will sponsor a BRFSS op-tional module designed to collect re-spondents’ current industry and occu-pation information.

Expanding surveillance for WRA toinclude collection of information on in-dustry and occupation will increase un-derstanding of WRA epidemiology.These important additions will enablestates, other government agencies,health professionals, employers, work-ers, and worker representatives to tar-get intervention and prevention ef-

forts more effectively to reduce theburden of WRA.

AcknowledgmentsBRFSS state coordinators. Kitty H. Gelberg, PhD, NewYork State Dept of Health. Lee Petsonk, MD, Na-tional Institute for Occupational Safety and Health,CDC.

REFERENCES

1. Friedman-Jimenez G, Beckett WS, Szeinuk J, PetsonkEL. Clinical evaluation, management, and preventionof work-related asthma. Am J Ind Med. 2000;37(1):121-141.2. Tarlo SM, Balmes J, Balkissoon R, et al. Diagnosisand management of work-related asthma: AmericanCollege of Chest Physicians consensus statement.Chest. 2008;134(3)(Suppl):1S-41S.3. Balmes J, Becklake M, Blanc P, et al; Environmen-tal and Occupational Health Assembly, American Tho-racic Society. American Thoracic Society Statement:Occupational contribution to the burden of airwaydisease. Am J Respir Crit Care Med. 2003;167(5):787-797.4. Lutzker LA, Rafferty AP, Brunner WM, et al. Preva-lence of work-related asthma in Michigan, Minne-sota, and Oregon. J Asthma. 2010;47(2):156-161.5. Liss GM, Nordman H, Tarlo SM, Bernstein DI. Pre-vention and surveillance. In: Bernstein IL, Chan-Yeung M, Malo J, Bernstein DI, eds. Asthma in theworkplace. 3rd ed. New York, NY: Taylor & FrancisGroup; 2006:353-375.6. Henneberger PK, Kreiss K, Rosenman KD, ReillyMJ, Chang YF, Geidenberger CA. An evaluation of theincidence of work-related asthma in the United States.Int J Occup Environ Health. 1999;5(1):1-8.7. Sama SR, Milton DK, Hunt PR, Houseman EA,Henneberger PK, Rosiello RA. Case-by-case assess-ment of adult-onset asthma attributable to occupa-tional exposures among members of a health main-tenance organization. J Occup Environ Med. 2006;48(4):400-407.8. Menzies D, Nair A, Williamson PA, et al. Respira-tory symptoms, pulmonary function, and markers ofinflammation among bar workers before and after alegislative ban on smoking in public places. JAMA.2006;296(14):1742-1748.9. CDC. Work-related lung disease surveillance sys-tem (eWoRLD). Asthma. Cincinnati, OH: US Depart-ment of Health and Human Services, CDC, NationalInstitute for Occupational Safety and Health; 2008.Ava i l ab le a t h t tp : //www2a.cdc .gov/drds/worldreportdata/sectiondetails.asp?archiveid=1&sectiontitleid=9. Accessed May 21, 2012.10. Hu SS, Balluz L, Battaglia MP, Frankel MR. Im-proving public health surveillance using a dual-framesurvey of landline and cell phone numbers. Am JEpidemiol. 2011;173(6):703-711.

*Additional information and survey questions avail-able at http://www.cdc.gov/brfss.†Additional information and survey questions avail-able at http://www.cdc.gov/asthma/survey/brfss.html#callback and http://www.cdc.gov/brfss/acbs/index.htm.§Additional information available at http://www.cdc.gov/brfss/pdf/userguide.pdf.¶Persons identified as Hispanic might be of any race.Persons identified as white, black, or other race areall non-Hispanic.**Information on WRA surveillance programs fromCDC-funded states is available at http://www.cdc.gov/niosh/topics/surveillance/ords/statebasedsur-veillance/wra.html.




THE COVER

WHEN THE MIST RISES FROM THE SCIOTO AND OHIO

rivers, the fog rolls into Portsmouth and the streetsfill with the same miasma that Clarence Holbrook

Carter (1904-2000) knew when he was a boy in that unas-suming southern Ohio town. Although decades have passedsince Carter left Portsmouth and its industrial mien—the fac-tories and mines have now largely vacated the area—muchof the town remains clothed in the remnants of its history, andthe natural setting of the confluence of the rivers has not beenirrevocably altered. Carter retained the influences of, and in-terest in, his riparian birthplace, after his education at the Cleve-land School of Art and his travels in Italy and in France. HansHofmann, master teacher and abstract artist ( JAMA cover, June27, 2012), helped Carter find lodgings in Capri in 1927; Carterin turn accepted Hofmann’s gracious offer of art lessons—6weeks for the price of 4—despite Carter’s penchant for inde-pendent work, even while enrolled in formal study.

Carter’s early work, a mix of portraiture, still life, nude, land-scape, and architectural drawing, has been termed of the re-gionalist,Americanscene, andsuperrealist styles;his laterpaint-ings, in watercolor, acrylic, and oil, evolved into surrealismand abstraction. Port Huron (cover) dates from 1936, whenCarter lived in Cleveland and held a multifunctional positionat the Cleveland Museum of Art. Carter scholars maintain thatCarter had not reached his mature style until after he ac-cepted a faculty appointment in 1938 at the Carnegie Insti-tute of Technology in Pittsburgh, yet Port Huron melds ab-stract elements with aspects of regionalist painting. Abuttingthe border with Canada, Port Huron, Michigan, sits at thesouthern-most tip of Lake Huron. Its location places it in themiddle of the Rust Belt, and Carter’s selection of subject paystribute to the industrial heritage of the lakefront.

Regionalistartemphasizedthespecialqualitiesofareas,suchas theMidwestor thedeepSouth, intheUnitedStates; thisstylereached its peak during and after the Great Depression. PortHuron can certainly be classified in this way, its storage facil-ity a symbol of the intersection between the Great Lakes andtheroadsoftheheartland,awaystationbetweenshipsandtrains,the point of process of crude or raw materials into functionalproducts.SomeresemblancetoPrecisionism,suchaspracticedbyCharlesSheeler( JAMAcover,October13,2010)andCharlesDemuth( JAMAcovers,January26,2011,andFebruary9,2011),can be sensed in Port Huron, but Carter did not subscribe tothePrecisionist aesthetic.Hiscomposition iscompelling,withthe placement of the vats in lines, arranged with puzzling per-spective: theirtopsblendseamlesslyintotheabstractlydesignedclouds. Carter raises the question, almost tongue-in-cheek,whether the sky is formed of those clouds, or if the heavensare blocked by industrial effluent.

With his transition into less representational art, Carter usedas his signature device an ovoid—usually transparent, in vari-ous positions within the composition. Sometimes singular,sometimes multiple, the ovoid appears as a character, recur-

ring throughout the story of Carter’s oeuvre; at times the ovoidbecomes the star, at other times it shows its “face” like a childpeering around the edges of a posed photograph. It is diffi-cult not to interpret the ovoid as a symbol of fertility, of birth,of life, reflecting the classical tradition. However, sources sug-gest that Carter was obsessed with death, the infinite natureof the universe, and the concept of man as consciousness—and the ovoid represented, in its simple shape and interest-ing treatments, all of those themes.

Carter has been described as an oneiric artist, one who paintsdreams, visions, or hallucinations. Port Huron, executed whenCarter was 32 years old, does not appear to have elements ofa dream, nor does its subject seem a likely product of a hal-lucination. When the progress of Carter’s work into abstrac-tion, and then into surrealism, is examined, it becomes clearthat the surrealist pillar of oneirism supports much of Cart-er’s later vision. Another surrealist, Roberto Matta (JAMA cover,October 27, 2010), expressed this same theory with his sig-nature style, psychological morphology, and his favorite re-curring characters, the vitreurs.

In 1938, when Carter painted murals in the post office inPortsmouth, he remained true to his ties to the river commu-nityand its surrounds.HedepictedPortsmouth’s too-frequentflooding in the pre–flood control days, before levies and damsrestrained the deluges; he also portrayed workers and inhab-itants of both the Ohio and Kentucky sides of the river, fran-tically tryingtoholdbackthesurgingwaterswithwallsofsand-bags. These murals, along with those in the Ravenna, Ohio,postoffice, adhere toCarter’s adaptationof regionalism;paint-ings like Port Huron, Coal Docks at Superior (1939), Fairport(1934), and Wheeling Steel Mill, Portsmouth, Ohio (1955) areall excellent examples of Carter’s contributions to the region-alist and American scene styles. Carter died when he was 96years old, at his home in Holland Township, New Jersey. Hiscreativity spanned more than 80 years and reached across—likethebridgesofPortsmouthcrossingovertheOhioandSciotorivers—artificial stylistic boundaries, leaving a large body ofthought-provoking and visually stimulating work.

Janet M. Torpy, MD

Courtesy of the Cleveland Museum of Art (http://www.clevelandart.org/), Cleve-land, Ohio; Hinman B. Hurlbut Collection; 2698.1936.

Clarence Holbrook Carter(1904-2000), Port Huron,1936, American. Oil oncanvas. 68.9 � 81.6 cm.



POETRY AND MEDICINE

Torsades de Pointes

torsades de pointeslike most French wordsthey sound refinedcultivated

torsades de pointesunlike vtachprolonged QTarrhythmia

torsades de pointeslike the classicsa painting byMonet Degas

torsades de pointesunlike scribblesoutside the linesmy kid could do

torsades de pointeslike a fine foodfoie gras soufflequiche escargot

torsades de pointesunlike fast foodcanned beans burnt toastcold leftovers

torsades de pointeslike love and lifecrème de la crèmea tour de force

torsades de pointesunlike the deaththe coup de graceit really is

torsades de pointeslike a ballettwisting aroundthe true meaning

torsades de pointesunlike two padsa high-pitched whinetwo hundred joules

Adam Possner, MD

Author Affiliation: Medical Faculty Associates, George Washington University, Washington, DC ([email protected]).Poetry and Medicine Section Editor: Charlene Breedlove, Associate Editor. Poems may be submitted [email protected].

Philip Greenland, MD

September 19, 20122 to 3 PM ET

Novel Markers forCardiovascular Risk

Assessment inIntermediate-Risk

Individuals

Discuss the article'simplications for clinical

practice with the author at the next teleconference

http://jama.jamanetwork.com/multimedia.aspx#Teleconferences

JAMA, August 22/29, 2012—Vol 308, No. 8 745

©2012 American Medical Association. All rights reserved.


JAMA 100 YEARS AGOAUGUST 17, 1912

Current Comment

THE PRACTICAL VALUE OF DIRECT TRANSFUSIONOF BLOOD

The time is not yet at hand when final statements as to thedefinite indications for direct transfusion of blood can be made.In secondary anemia, such as results from traumatic hemor-rhage, from post-partum hemorrhage, and from other condi-tions, if would often be desirable to transfuse after the bleed-ing has been controlled, if the technic were as simple as, forinstance, the intravenous injection of salt solution. Unfortu-nately, we are not yet in possession of a technic for transfu-sion which is easily mastered; the operation is not without con-siderable difficulty—even some danger to the patient. It shouldbe remembered always that the selection of a donor is a mostimportant matter, requiring special care. All this does not meanthat recent accomplishments in this field are insignificant, butrather that the ultimate goal desired is not attained. With thepresent technic, probably only three classes of patients withsecondary anemia should be considered as fit subjects for trans-fusion: those who have bled so profusely that their life is injeopardy; those who have lost much blood and are sufferingfrom uncontrollable hemorrhage, and those who require someoperation or other while in a state of very grave anemia. Thepractice of transfusion of patients suffering from moderatelysevere hemorrhage such as may be associated with extrauter-ine pregnancy is not generally commendable because it leadsto a routine performance of transfusion, when in reality con-trol of the bleeding is all that is necessary. When transfusionbecomes a simpler matter than now, many patients no doubtmay be transfused with benefit, but at present the operationwith the selection of the donor is so complicated as to be em-ployed only after careful deliberation. In other kinds of cases,such as pernicious anemia, Hodgkin’s disease, and chronic in-fections with a hopeless outlook, e. g., tuberculosis with ad-vanced anemia, evidence is still wanting that transfusion is ofany real benefit. At the most, the improvement resulting hasbeen simply temporary. Perhaps the outlook in tuberculosisis a little more favorable; here the content of the blood incomplement is low and as transfusion tends to increase comple-ment, theoretically it would be beneficial. The outcome of re-peated transfusions in tuberculosis and other similar dis-eases may be awaited with interest.

THE THERAPEUTIC USE OF BLOODAn interesting development in minor surgery is the use

of blood to stop hemorrhage. For this purpose blood maybe drawn from a normal person and injected subcutane-ously at once or, if preferred, a sterile flask with a coiled

wire may be used to receive the blood, which is then defi-brinated so as to be ready for injection. If haste is not re-quired a simple procedure is to allow the blood to stand inthe ice-box in a sterile bottle, the serum being used for in-jection as required. From 10 to 20 c.c. may be injected, andthe dose repeated several times if desirable. As substitutesfor human serum, beef-serum or that of the horse or rabbitmay be used. The possibility of anaphylactic reactions fromthe use of foreign serums makes the latter somewhat lessdesirable. Blood-serum is also efficacious when adminis-tered locally, and applied to oozing surfaces it possessesmarked hemostatic properties. Good results follow injec-tions of blood or serum in hemorrhage of the new-born, ex-cept when life is endangered from exsanguination; in thesecases direct transfusion is indicated. In hereditary hemo-philia injection of serum is useful, probably because it causesa stimulation of the bone-marrow, resulting in an increaseof thrombokinase, which both Sahli and Morawitz find islacking in hemophilics. The treatment of the toxemias ofpregnancy with serum of normal pregnant women also prom-ises well, and even now many cases have been recorded inwhich the treatment has given satisfactory results. The modeof action in these cases is unknown.

WHOLESALE GEOGRAPHIC SURGERYWhat might be called a capital operation is the pro-

posed cutting of a canal from the Mediterranean to the Des-ert of Sahara, converting a considerable portion of the lat-ter into an inland sea. Now that the Panama Canal is almostcompleted and the edge of its novelty has been somewhatdulled in the public thought, France steps forward to pro-pose a new, stupendous engineering feat.1 There are two factsthat make this a feasible proposition. The first is that partof the desert is below the sea-level and the second that thecanal to admit the waters of the Mediterranean would haveto be but fifty miles long. The results need but to be hintedat: an inland sea for large ships, tropical verdure to replacebarren desert, agricultural possibilities to support nationsof people and great modification of the climate. Since theparts of the Sahara which are below the sea-level are bro-ken up by higher spots, this inland sea would be filled withislands and promontories. The proposal is an ingenious one,full of great possibilities for benefit to Africa.

1. Thompson, G. A.: A Plan for Converting the Sahara Desert Into a Sea, Scient.Am., Aug 12, 1912, p. 114.

JAMA. 1912;59(7) 548-550

Editor’s Note: JAMA 100 Years Ago is transcribed verbatim from articles pub-lished a century ago, unless otherwise noted.

JAMA 100 Years Ago Section Editor: Jennifer Reiling, Assistant Editor.



BOOK AND MEDIA REVIEWS

MADE IN INDIA: A FILM ABOUT SURROGACY

DVDBy Rebecca Haimowitz and Vaishali Sinha97 min, color, Hindi/English (subtitled), pricing available fromthe distributorNew York, NY, Women Make Movies, 2010Available for purchase at http://www.wmm.com/filmcatalog/pages/c805.shtml

WHAT IS NOW CALLED “MEDICAL TOURISM” WAS ONCE LARGELY

a privilege of the elite. England’s Bath, Germany’s BadenBaden, and Belgium’s Spa all achieved fame as places for Eu-rope’s wealthy to “take the waters” as cures for various healthailments.

Modern medical tourism is a much more diverse phe-nomenon and has a different clientele. Many of today’s medi-cal tourists are patients of modest means who travel to lessdeveloped nations for treatment that is unaffordable at homeand much cheaper abroad. Their trips may include sight-seeing but are motivated primarily by financial consider-ations.

This new category of medical tourism is the subject ofMade in India: A Film About Surrogacy. The film illustratesnumerous clinical, legal, and ethical issues raised bymedical tourism, both in general and in the context ofsurrogacy.

In the film, a childless couple from San Antonio, Texas,risks their savings to use a California-based medical tour-ism company to arrange for a surrogate pregnancy with anIndian surrogate mother. The couple lives in what appearsto be, by US standards, a middle-income neighborhood. Thesurrogate mother, who has 3 children of her own, is illiter-ate and lives in Mumbai, where the clinic responsible forthe assisted reproductive services is located. The couple fliesto Mumbai, and after additional fertility treatment there, thegenetic mother conceives. Several of her embryos are im-planted in the surrogate.

In month 7 of a mostly uneventful pregnancy, the surro-gate moves from her family’s apartment to a surrogacyhome. Shortly thereafter she experiences prepartum hem-orrhage, at which point she is transferred to a largeMumbai hospital, where she gives birth to healthy twinsby cesarean delivery. The twins are monitored in neonatalintensive care.

The surrogate mother remains in the hospital at least an-other 11 days, during which time the genetic mother (joinedlater by the father) has again arrived from the United States.In the interval, a dispute concerning the legal status of thetwins has arisen among the genetic parents, the surrogate,

and the hospital and clinic, requiring intervention by theUS consulate. After a near stalemate, it is resolved in favorof having the birth certificate issued in the name of the ge-netic mother. The twins are granted US citizenship, and thecouple returns to the United States.

At the end of the film, one of the twins is reported to havedied from sudden infant death syndrome. The other re-mains healthy and is reported to be normal. Although thesurrogate was paid much less than she was promised, andfar less than the Texas couple was initially told she wouldbe, she says that she plans to undertake another surrogacyin the future.

Made in India refrains from making explicit judgmentsabout the issues it raises, but it highlights these issues in away that constitutes a message. Both the genetic parents andthe surrogate appear satisfied with their choices. Things mighthave been different, however, if the outcomes had been lessfavorable, and it is evident from the film that this could haveoccurred at any point in the process. Conception or im-plantation might not have succeeded. The surrogate mothercould have died as a result of hemorrhage or surgery. Thelegal status of the twins might have been resolved in favorof the surrogate mother.

The film also suggests that the lines of responsibility amongthe medical tourism company in California and the clini-cians in India were poorly drawn. This, combined with thedistance involved and the disparate cultural and legal set-tings, appears to have impaired their ability to address un-expected problems as they arose. For example, the stale-mate about the children’s legal status could easily haveremained a stalemate.

Similar difficulties would have confronted any effort toobtain recourse for adverse outcomes through litigationafter the fact. Malpractice litigation is cumbersomeenough in the United States. A lawsuit involving partieshalfway around the world, in very different cultures, isorders of magnitude more challenging. Aside from spe-cifically legal questions such as where a suit arising fromthese circumstances would be heard (Texas, India, orboth) and which jurisdiction would provide the relevantlegal standards, the practical costs of obtaining evidenceand securing witnesses at such a distance would almostsurely be prohibitive.

The ethics of surrogate motherhood have been exten-sively debated, but medical tourism casts a different lighton these issues. For example, the surrogate motherreceived a little more than $100 per month for running

Book and Media Reviews Section Editor: John L. Zeller, MD, PhD, Contributing Editor.



the risks and bearing the burdens of pregnancy, raisingquestions of exploitation that have not been fully exam-ined in this context. The extent to which the surrogatewas able to provide informed consent to the procedure isalso questionable, especially given her level of education.The film does not indicate, for example, whether she wasadvised of the possibilities of multiparity, hemorrhage, orsurgery.

The medical tourism industry is expanding. As healthcare costs increase in developed countries it will almostsurely continue to do so. Made in India tells the story ofonly a single case, and there is no reason to believe that itis representative. Other arrangements for medical tour-ism may be, and likely are, more reassuring than thosedepicted in this film.

Still, even a single case such as this points out the needfor a far more coherent and serviceable internationalregulatory effort to govern this area of health care. It isunlikely that such an effort will soon be forthcoming. Inthe meantime, Made in India delivers an important cau-tionary message to patients considering medical tourismfor themselves and to professionals advising them aboutthis option.

Samuel Yancey Sessions, MD, JD

Author Affiliation: Department of Psychiatry and Biobehavioral Sciences, Harbor-UCLA Medical Center, Torrance, California ([email protected])Conflict of Interest Disclosures: The author has completed and submitted the ICMJEForm for Disclosure of Potential Conflicts of Interest and reported receiving a grantfrom the Institute of Medicine to author a commissioned paper for and serve as aconsultant to the institute’s Committee on Public Health Strategies to Improve Health.

INTERNATIONAL HEALTH AND AID POLICIES:THE NEED FOR ALTERNATIVES

By Jean-Pierre Unger, Pierre De Paepe, Kasturi Sen,and Werner Soors314 pp, $52Cambridge, United Kingdom, Cambridge University Press, 2010ISBN-13: 978-0-521-17426-8

IF ANYTHING GOOD HAS COME OUT OF THE ONGOING CRISIS

in the world economy that emerged in 2008, it is the real-ization in many quarters that old solutions do not work.The neoliberal trend that sought to replace the dominanceof government in welfare systems has come under criticalscrutiny. The focus on regulated competition, contractingout, and consumer choice, although not devoid of con-structive lessons about policy, has left policy makers, aca-demics, and citizens a bit disillusioned. It is clear thatthere are no panaceas, and after every major swing of thependulum from a socialist to a capitalist orientation andback, there is a need to pick up the pieces and fine-tunethe system.

Nowhere is this better illustrated than in health systems.Western European health systems have been experiment-

ing with various forms of regulated competition in healthcare, and, although there have been some positive out-comes in terms of responsiveness and perhaps even effi-ciency, costs continue to increase and seem to be driven upby market reforms. For Central and Eastern European healthsystems, the rapid move from a Soviet-style system to a promi-nent role for private enterprise has resulted in mixed out-comes that in many cases include greater health inequali-ties. For developing countries, the relevance of market modelsfor health care has been questioned from the outset, and theconnection of these types of reforms to the population healthchallenges faced by these countries has been difficult to ar-ticulate.

The experience in developing countries receiving con-siderable international aid to their health systems is thesubject of International Health and Aid Policies: The Needfor Alternatives. The purpose of the book is to demon-strate that neoliberal aid policies have not led to improve-ments in health status in developing countries but haveactually exacerbated the situation. The opening chapterspresent the conceptual model that underpins the book.The model posits that health policy in developing coun-tries has become dominated by commodification ofhealth care and the separation of health care from diseasecontrol and that a new form of comprehensive healthcare provision is needed.

The book is based on numerous research projects andarticles that have been produced by the Public Policy andManagement Unit at the Department of Public Health ofthe Institute of Tropical Medicine in Antwerp. After a sec-tion that lays out the existing paradigms of internationalpolicies, such as those led by the World Bank and theEuropean Union, other sections present studies of theimplementation and results of international health and aidpolicies in countries that span Africa, Latin America, andAsia. Although the data presented support the underlyingthesis of the collection, the research is rigorous and wellthought out and paints a picture that, whatever the ideo-logical or academic orientation of the reader, elicits seriousconcern. Numerous examples are provided of how inter-national market–oriented policies tend to separate provi-sion of health care from disease control, to the detrimentof both. From malaria control in Mali, to contracting outin Colombia, to the role of cooperative medical systems inChina, the authors provide substantial evidence support-ing the case that the drive to introduce market mecha-nisms and privatize health systems has been associatedwith fragmentation of health care provision from diseasecontrol, again to the detriment of both. Abandonment ofcooperation based on social capital, as in the case of thereplacement of Community Health Systems, first withcommercialization and then with the top down NewCooperative Medical System, entailed loss of the commu-nity role without commensurate efficiencies in terms ofhealth costs and outcomes.




The final sections offer a set of principles for publicly ori-ented health policies and a strategic toolkit for implemen-tation of these principles. Policies should be mission ori-ented, not profit oriented. Services, such as health care anddisease control, should be integrated. To increase motiva-tion, staff should be involved in organizational design of thenew systems. Communities, and the social solidarity inher-ent in them, should be activated in restructuring the healthsystem. Many if not all of these proposals sound utopian,but when presented in the wake of the above-mentionedanalysis of current international health and aid policies, theyseem more realistic and the simple need to implement themis made clearer. It should be pointed out that the authorsdo not suggest eliminating the activities of private nongov-ernmental organizations or even private for-profit health ser-vices. Instead, they put forth an approach that capitalizeson the best features of private ownership, in an attempt tobring these types of entities into the integrated system theyenvision.

International Health and Aid Policies makes a strong casethat current neoliberal-oriented health and international aid

policies are not succeeding. The compartmentalizing of healthcare as separate from disease control redounds negativelyto both. Market- and product-oriented interventions haveled to fragmentation of health systems. The solutions putforward are idealistic and perhaps unrealistic, but their chanceof being adopted seem stronger when placed in the contextof the current disarray in world economies and social poli-cies, some of which are compellingly described in this book.What is perhaps most striking is that, although the bookfocuses on low- and middle-income countries, its propos-als seem relevant to many developed countries, perhaps no-tably the United States, which is embarking on a restruc-turing of its health system that begs for a new paradigm andthat could benefit from many of the insights and policy di-rections provided.

David Chinitz, PhD

Author Affiliation: Department of Health Policy and Management, Hebrew Uni-versity–Hadassah, Jerusalem, Israel ([email protected]).Conflict of Interest Disclosures: The author has completed and submitted theICMJE Form for Disclosure of Potential Conflicts of Interest and none werereported.

Books are to be call’d for, and supplied, on the assumptionthat the process of reading is not a half sleep, but, in high-est sense, an exercise, a gymnast’s struggle; that the readeris to do something for himself, must be on the alert, musthimself or herself construct indeed the poem, argument, his-tory, metaphysical essay—the next furnishing the hints, theclue, the start of frame-work. Not the book needs so muchto be the complete thing, but the reader of the book does.That were to make a nation of supple and athletic minds,well-train’d, intuitive, used to depend on themselves, andnot on a few coteries of writers.

—Walt Whitman (1819-1892)




ported by the Mayo Clinic Windland Smith Rice Comprehensive Sudden CardiacDeath Program.Role of the Sponsor: The funding source had no role in the design and conduct ofthe study; in the collection, analysis, and interpretation of the data; or in the prepa-ration, review, or approval of the manuscript.Additional Contributions: We thank the patients who sought evaluation at Mayo’sLQTS Clinic, as well as Katy Harris, MS, RN (Mayo Clinic); Heidi J. Owen, RN (MayoClinic); and Carla M. Haglund (Mayo Clinic) for their assistance with patient follow-up. None of these individuals received compensation.

1. Zipes DP, Ackerman MJ, Estes NA III, Grant AO, Myerburg RJ, Van Hare G.Task Force 7: arrhythmias. J Am Coll Cardiol. 2005;45(8):1354-1363.2. Pelliccia A, Fagard R, Bjørnstad HH, et al; Study Group of Sports Cardiology ofthe Working Group of Cardiac Rehabilitation and Exercise Physiology; WorkingGroup of Myocardial and Pericardial Diseases of the European Society of Cardiology.Recommendations for competitive sports participation in athletes with cardiovas-cular disease: a consensus document from the Study Group of Sports Cardiologyof the Working Group of Cardiac Rehabilitation and Exercise Physiology and theWorking Group of Myocardial and Pericardial Diseases of the European Society ofCardiology. Eur Heart J. 2005;26(14):1422-1445.3. Ackerman MJ. Cardiac channelopathies: it’s in the genes. Nat Med. 2004;10(5):463-464.4. Goldenberg I, Moss AJ, Peterson DR, et al. Risk factors for aborted cardiac ar-rest and sudden cardiac death in children with the congenital long-QT syndrome.Circulation. 2008;117(17):2184-2191.5. Liu JF, Jons C, Moss AJ, et al; International Long QT Syndrome Registry. Riskfactors for recurrent syncope and subsequent fatal or near-fatal events in childrenand adolescents with long QT syndrome. J Am Coll Cardiol. 2011;57(8):941-950.6. Mitchell JH, Haskell W, Snell P, Van Camp SP. Task Force 8: classification ofsports. J Am Coll Cardiol. 2005;45(8):1364-1367.

CORRECTION

Incorrect Wording in Patient Page: In the Patient Page entitled “NeonatalHyperbilirubinemia,” published in the May 16, 2012, issue of JAMA (2012;307[19]:2115), incorrect wording was used. The lighting in the Figure shouldbe labeled “Special light source (blue-green spectrum),” and in the “Photo-therapy” section under “Treatment,” the words “ultraviolet (blue) lights”should read “a special light source (blue-green spectrum).” This article was cor-rected online.

Figure. Primary Sport Classification (According to the 2005 BethesdaConference Guidelines6) for the Competitive Athletes With Long QTSyndrome

Sta

tic C

ompo

nent

, Pea

k %

Max

imal

Volu

ntar

y Le

ft Ve

ntric

ular

Con

trac

tion

No. of Patients

Dynamic Component, Peak % Maximal Oxygen Uptake

IIIHigh>50

IIModerate

20-50

ILow<20

ALow<40

25

0

3

CHigh>70

0

34

28

BModerate

40-70

4

14

22

Table. Demographics of Competitive Athletes With Long QT Syndromea

Total Cohort Athletes

Within Bethesdabut Contrary toESC Guidelines

Contrary to BothGuidelines

PValueb

No. of patients 353 130 70 60

Age at diagnosis, mean (SD), y 17 (11) 11 (7) 11 (7) 12 (6) .91

Sex .21Male 154 70 41 29

Female 199 60 29 31

Mean (SD) QTc, ms 472 (42) 471 (46) 444 (23) 501 (46) �.001

Symptoms, No. (%) 111 (31) 29 (22) 1 (1) 28 (47) �.001

�-Blockers, No. (%) 280 (79) 112 (87) 55 (79) 57 (95) �.008

ICD, No. (%) 78 (22) 20 (15) 0 20 (33) �.001

Follow-up available, mean (SD), y 5.5 (3.4) 5.1 (2.9) 5.1 (2.9) 5.0 (3.0) .98

Genotype, No. (%)LQT1 182 (52) 74 (57) 41 (59) 33 (55) .65

LQT2 130 (37) 41 (32) 20 (29) 12 (35) .47

LQT3 37 (10) 11 (8) 8 (11) 3 (5) .20

Multiplec 4 (1) 4 (3) 1 (1) 3 (5) .19Abbreviations: ESC, European Society of Cardiology; ICD, implantable cardioverter-defibrillator; LQTS, long QT syndrome; QTc, corrected QT interval.aDemographics of the overall cohort of eligible LQT1-3 patients 6-40 years old (N=353), which includes the 130 LQTS athletes including those participating against ESC guidelines

but within Bethesda guidelines (n=70) and those athletes participating against both guidelines (n=60).bComparing athletes who are participating within Bethesda Conference guidelines but against ESC guidelines vs those participating against both guidelines.cPatients with �1 LQTS-causing mutation.

LETTERS



JAMA PATIENT PAGE

Ectopic Pregnancy

Intrauterine pregnancy

Ovary

Implanted embryo

UTERUS

VAGINA

Ectopic pregnancy

Cervix

Cervical canal

Endometrium

IsthmusAmpulla

Fallopian tube

Embryo

Awoman’s reproductive system includes the uterus, the ovaries,and the fallopian tubes. In a healthy pregnancy, the fertilizedegg implants on the uterine wall. The growing fetus develops,

nourished by the placenta, for about 40 weeks until the fetus isready for delivery. When a fertilized egg does not make it all the wayto the uterus and stays in the fallopian tube (or rarely in the ovaryor in the abdominal cavity), an ectopic (out-of-place) pregnancyexists. The abnormally placed cells of an ectopic pregnancy cannotbecome normally placed inside the uterus or develop into anormal pregnancy.

RISK FACTORS

• Past ectopic pregnancies• Past operations on the fallopian tubes or ovaries• Scar tissue from other abdominal or pelvic surgery• History of pelvic infections• Endometriosis• Smoking• In vitro fertilization

SIGNS AND SYMPTOMS

• Symptoms similar to a healthy pregnancy, such as missed menstrual periods,breast tenderness, back pain, or nausea

• Abdominal, pelvic, or rectal pain• Vaginal bleeding• Dizziness, paleness, fainting, and even shock can happen if the ectopic pregnancy

ruptures (bursts open) and causes bleeding.

DIAGNOSISA medical history that is taken will include questions about sexual activity and thepossibility of pregnancy. Physical examination may reveal abdominal or pelvictenderness, especially during vaginal or rectal examination. A pregnancy test is done,including a blood test for levels of a hormone called beta human chorionicgonadotropin (beta-HCG), because a healthy pregnancy shows a different amount ofincreasing beta-HCG. Your doctor will look for signs of anemia. An abdominal andpelvic ultrasound is often performed, looking for whether a pregnancy can be seeninside the uterus and whether an ectopic pregnancy can be seen outside the uterus(for example, in the fallopian tube). This ultrasound also looks for blood in theabdominal cavity or in the pelvis (a result of rupture of the ectopic pregnancy).

TREATMENTIn some cases, if an ectopic pregnancy is discovered, a medication called methotrexatemay be used as treatment. In other cases, in particular when a ruptured ectopicpregnancy has occurred, a laparoscopy (using instruments to look into the abdomenand pelvis through small incisions) or a laparotomy (in which the abdomen is surgicallyopened) may be done to remove the ectopic pregnancy. Sometimes the fallopian tubemust be removed (salpingectomy).

Also, if there has been a significant amount of bleeding, support of the bloodcirculation may be required, including intravenous (through a vein) fluids andsometimes blood transfusion.

FOR MORE INFORMATION• US Department of Health

and Human Services Officeon Women’s Healthwww.womenshealth.gov

• American Congress of Obstetriciansand Gynecologistswww.acog.org

INFORM YOURSELF

To find this and previous JAMAPatient Pages, go to the PatientPage Index on JAMA’s website atwww.jama.com. Many are available inEnglish and Spanish. A Patient Pageon smoking and pregnancy waspublished in the March 9, 2005, issue;one on prenatal care was published inthe January 7, 2004, issue; and oneon miscarriage was published in theOctober 16, 2002, issue.Sources: US Department of Health and HumanServices Office on Women’s Health, AmericanCongress of Obstetricians and Gynecologists

Janet M. Torpy, MD, Writer

Alison E. Burke, MA, Illustrator

Robert M. Golub, Editor

The JAMA Patient Page is a public service of JAMA. The information and recommenda-tions appearing on this page are appropriate in most instances, but they are not a substi-tute for medical diagnosis. For specific information concerning your personal medicalcondition, JAMA suggests that you consult your physician. This page may be photocopiednoncommercially by physicians and other health care professionals to share with patients.To purchase bulk reprints, call 312/464-0776.

WO

MEN

’SH

EA

LTH

The Journal of the American Medical Association



EDITORIAL STAFF

EDITOR IN CHIEFHoward Bauchner, MD

Executive EditorPhil B. Fontanarosa, MD, MBA

DeputyEditors:DrummondRennie,MD,RonnaHenry,MD,RobertM.Golub,MD,Edward H. Livingston, MDManaging Deputy Editor: Annette FlanaginSenior Editor: Jody W. Zylke, MDAssociate Senior Editor: Roxanne K. YoungContributing Editors: Derek C. Angus, MD, MPH, Robert G. Badgett, MD,Anne Rentoumis Cappola, MD, ScM, Huan J. Chang, MD, MPH, Helene M.Cole, MD, Thomas B. Cole, MD, MPH, J. Michael Gaziano, MD, MPH, RichardM. Glass, MD, David H. Mark, MD, MPH, Mary McGrae McDermott, MD,Robert A. McNutt, MD, George T. O’Connor, MD, Boris Pasche, MD, PhD,Eric D. Peterson, MD, MPH, Jeanette M. Smith, MD, Janet M. Torpy, MD,Wolfgang C. Winkelmayer, MD, MPH, ScD, John L. Zeller, MD, PhDStatistical Editor: Naomi Vaisrub, PhDAssociate Editor: Charlene Breedlove

Fishbein Fellow: Denise M. Goodman, MD, MSContributing Writers: Robert H. Brook, MD, ScD, Christine K. Cassel, MD, NicholasA. Christakis, MD, PhD, MPH, Allan S. Detsky, MD, PhD, Mark Duggan, PhD,Ezekiel J. Emanuel, MD, PhD, Elliott S. Fisher, MD, MPH, Lawrence O. Gostin,JD, Kathy Hudson, PhD, John P. A. Ioannidis, MD, DSc, Ashish K. Jha, MD, MPH,David S. Ludwig, MD, PhD, Howard Markel, MD, PhD, Steven E. Nissen, MD,Bruce M. Psaty, MD, PhD, Joshua M. Sharfstein, MD, Stephen M. Shortell, PhD,MPH, MBA, Harold C. Sox, MD, Robert Steinbrook, MD, Mary E. Tinetti, MD,Abraham Verghese, MD, Jason Wang, MD, PhD, Steven H. Woolf, MD, MPHMedical News & Perspectives: Joan Stephenson, PhD (editor);Bridget M. Kuehn, Mike Mitka (senior staff writers); Rebecca Voelker (associatemanaging editor)Assistant Editors: Carrie Butt, Angela Grayson, Jennifer Reiling

Manuscript Editing: Stacy L. Christiansen (director); Emily A. Greenhow,Joy K. Jaeger, Phil Sefton, Heather A. Shebel, Kristine B. Simmons,Beverly Stewart (senior manuscript editors); Kim S. P. Campbell;Cara Wallace (freelance manuscript editor)

Editorial Graphics: Cassio Lynm, Alison E. Burke (senior medical illustrators);David Song (editorial assistant)

Electronic Media: Michelle Kurzynski (director, development and design);Christopher Hastings, Peter Kim, Craig McCaffrey (developers/designers);Reuben Rios (Web administrative assistant)

Administration: Marla A. Hall (director); Gloria Tate (supervisor);Deanna M. Willis (staff assistant)

Editorial Assistants: Mary Cannon, Susan Collister, Carmina Conley, Dawn Cortez,Cheryl Dixon, Georgia Eaton, Audrey Forman, Gwenn Gregg, Lisa Hardin,Rachell M. Lozano, Rosa E. Miranda, Lupe Morales

Workflow and Content Management Systems: Larry Bryant (director); Paul Frank(systems administrator); Monica Mungle (editorial systems manager);J.D. Neff (associate developer)

Editorial Systems and Administration: Elaine Williams (director); Gale Saulsberry(database specialist); Lenette Gardner-Gullens (electronic input supervisor);Delayna Brown, Fanny L. Brown (electronic input assistants)

Editorial Counsel: Joseph P. Thornton, JD

Survey Research Specialist: Joseph S. Wislar

JAMA/Archives Media Relations: Jann Ingmire (director);Jim Michalski, Deanna Bellandi, Cassie Brasseur

EDITORIAL BOARDDaniel M. Albert, MD, MSUniversity of Wisconsin Hospital

and Clinics, Madison, WisconsinEditor, Archives of Ophthalmology

Donald M. Berwick, MD, MPPNewton, Massachusetts

Philip Cagle, MDThe Methodist HospitalHouston, Texas

Jordan J. Cohen, MDThe George Washington UniversityWashington, DC

Joseph T. Coyle, MDHarvard Medical SchoolBoston, MassachusettsEditor, Archives of General Psychiatry

Ezekiel J. Emanuel, MD, PhDUniversity of PennsylvaniaPhiladelphia, Pennsylvania

Julie Freischlag, MDJohns Hopkins University

School of MedicineBaltimore, MarylandEditor, Archives of Surgery

Thomas R. Frieden, MD, MPHCenters for Disease Control

and PreventionAtlanta, Georgia

Michael M. E. Johns, MDRobert W. Woodruff Health Sciences

Center of Emory UniversityAtlanta, Georgia

Wayne F. Larrabee Jr, MDUniversity of WashingtonSeattle, WashingtonEditor, Archives of Facial

Plastic Surgery

Paul A. Levine, MDUniversity of VirginiaCharlottesville, VirginiaEditor, Archives of Otolaryngology–

Head & Neck Surgery

Donald A. B. Lindberg, MDNational Library of MedicineBethesda, Maryland

C. David Naylor, MD, DPhilUniversity of TorontoToronto, Ontario, Canada

Rita F. Redberg, MD, MScUniversity of California, San Francisco

School of MedicineSan Francisco, CaliforniaEditor, Archives of Internal Medicine

Uwe E. Reinhardt, PhDPrinceton UniversityPrinceton, New Jersey

Frederick P. Rivara, MD, MPHUniversity of WashingtonSeattle, WashingtonEditor, Archives of Pediatrics

& Adolescent Medicine

June K. Robinson, MDThe Feinberg School of Medicine,

Northwestern UniversityChicago, IllinoisEditor, Archives of Dermatology

Griffin Rodgers, MDNational Institute of Diabetes and

Digestive and Kidney DiseasesBethesda, Maryland

Roger N. Rosenberg, MDUniversity of Texas Southwestern

Medical Center, Dallas, TexasEditor, Archives of Neurology

Joshua M. Sharfstein, MDMaryland Department of Health

and HygieneBaltimore, Maryland

Harold C. Sox, MDDartmouth Medical SchoolHanover, New Hampshire

JOURNAL OVERSIGHT COMMITTEEEdward H. Shortliffe, MD, PhD (chair)Columbia UniversityNew York, New YorkArizona State UniversityScottsdale, Arizona

Kent R. AndersonJournal of Bone and Joint SurgeryNeedham, Massachusetts

Karen Antman, MDBoston University

School of MedicineBoston, Massachusetts

Steven L. Kanter, MDUniversity of Pittsburgh

School of MedicinePittsburgh, Pennsylvania

Raynard S. Kington, MD, PhDGrinnell CollegeGrinnell, Iowa

James L. Madara, MDAmerican Medical AssociationChicago, Illinois

Claire Pomeroy, MD, MBAUC Davis Health SystemUniversity of California, Davis

School of MedicineSacramento, California

FORMER EDITORSNathan S. Davis, MD (1883-1888)John B. Hamilton, MD

(1889, 1893-1898)John H. Hollister, MD (1889-1891)James C. Culbertson, MD (1891-1893)Truman W. Miller, MD (1899)George H. Simmons, MD (1899-1924)Morris Fishbein, MD (1924-1949)Austin Smith, MD (1949-1958)

Johnson F. Hammond, MD(1958-1959)

John H. Talbott, MD (1959-1969)Hugh H. Hussey, MD (1970-1973)Robert H. Moser, MD (1973-1975)William R. Barclay, MD (1975-1982)George D. Lundberg, MD

(1982-1999)Catherine D. DeAngelis, MD, MPH

(2000-2011), Editor in Chief Emerita

To Promote the Science and Art of Medicine and the Betterment of the Public HealthAugust 22/29, 2012, Vol 308, No. 8 Pages 731-830



PERIODICAL PUBLICATIONS STAFF

Senior Vice President and Interim Publisher, Periodical PublicationsRobert A. Musacchio, PhD

Staff: Monica Smith

ADVERTISING PRINT AND ONLINE SALES AND MARKETINGDirector: Jeffery J. Bonistalli

Phone: (312) [email protected]

Manager: Kathy Russell

Staff: Mirna Monal

Sales Representatives:Maureen A. Reichert

Phone: (973) 263-9191 ext [email protected]

Nancy E. SouzaPhone: (973) 263-9191 ext [email protected]

Advertising Sales, Marketing and Research

Director: Mark ThornburyPhone: (973) 263-9191 ext [email protected]

Reprint & ePrint Communications

Manager: Susan LovenbergPhone: (312) [email protected]

Reprint Sales:Reprints Desk, Inc

United States: +1 (800) 622-3920Worldwide: +1 (310) [email protected]

Classified Advertising/Circulation Promotion

Sales Manager: Thalia MossPhone: (312) [email protected]

Senior Sales Representative:Susan J. Henning


Sales Representative:Tem Harrier


CIRCULATION PRINT AND ONLINE SALES AND MARKETINGWorldwide Sales

Director: Vida DamijonaitisPhone: +1 (312) [email protected]

Regional Managers:Saskia Bolore, Americas

Phone: +1 (312) [email protected]

Gretchen Linder, Europe/Middle East/AfricaPhone: +1 (312) [email protected]

Brian McCartney, Asia PacificPhone: +1 (312) [email protected]

Sales Representatives:Rebecca Haas, Americas


Christine Hearne, AmericasPhone: +1 (312) [email protected]

Staff:Irma Millie Lara, Europe/Middle East/Africa


Audra LawlorPhone: +1 (312) [email protected]

Maliya Scott, Asia/PacificPhone: +1 (312) [email protected]

Marketing and Product Development

Directors:Matt Herron


Elizabeth SolaroPhone: +1 (312) [email protected]

Associate Publisher/Senior Product Development:Geoffrey Flick


Managers:Rick Bell


Paul GeePhone: +1 (312) [email protected]

Michael McGrawPhone: +1 (312) [email protected]

Stacy TuckerPhone: +1 (312) [email protected]

Staff:Millette Jackson-Bates


Sara NicholsonPhone: +1 (312) [email protected]

BUSINESS OPERATIONSDirectors: Karen Adams-Taylor, Nicole Bossert,Karl Elvin, Sue Sherrill, Brian Shields

Managers: David Antos, Chris Borden,Suzanne J. Bukovsky, Debbie Camp, Lydia Cruz,Diane Darnell, Colleen Kelly, Chris Meyer,Jaye Mize, Sean O’Donnell, Sean Ohlson,Phil O’Leary, Ryan Patella, Susan Price,Emily Mitchell, Denise Steinhauser

Advertising and Production Services:Bernadette Bauer, Kim Boler,Rhonda Bailey Brown, Timothy Burke,Susan Chiara, Daniel DeGroot,Jordynn Farrar, Euwanaha Joiner-Smith,Cindy Rendlen, Ellen Yoshihara

Circulation Services: Julie Burton,Dexter Golden, Erin Hinchliff,Noemi Rivera, Geneine Van Someren

Composition and New Media: Gail Barrett,Regina Brownlee, Brenda Chandler-Grayer, Scott Curl,Amy Evers, Teri Hutchison, Mary Ann Kuranda,Joshua Lampinen, Fred Marshall, Patricia E. Panek,La’son Diggs Sledge, George Thomson

Graphics: Joe Amft, Maria Duda, Fred Furtner,Carolyn Hall, Maria Kowalkowski

Proofreading: Carol Joy Farrell, Daniel James,Judith A. Literskis, Teresa H. Omiotek

AMA EXECUTIVES AND TRUSTEESExecutive Vice President, Chief Executive Officer:James L. Madara, MD

Chief Operating Officer: Bernard L. Hengesbaugh

President: Jeremy A. Lazarus, MD

President-Elect: Ardis D. Hoven, MD

Immediate Past President: Peter W. Carmel, MD

Secretary: Stephen R. Permut, MD, JD

Speaker, House of Delegates:Andrew W. Gurman, MD

Vice-Speaker, House of Delegates:Susan R. Bailey, MD

Trustees: Joseph P. Annis, MD; David O. Barbe,MD, MHA (chair-elect); Peter W. Carmel, MD;

Malini P. Daniel; Alexander Ding, MD, MS;Julie K. Goonewardene; Patrice A. Harris, MD;Ardis D. Hoven, MD; William E. Kobler, MD;Jeremy A. Lazarus, MD; Barbara L. McAneny, MD;Mary Anne McCaffree, MD; Albert J. Osbahr III,MD; Stephen R. Permut, MD, JD; Carl A. Sirio, MD;Steven J. Stack, MD (chair); Georgia A. Tuttle, MD;Robert M. Wah, MD (immediate past chair);Monica C. Wehby, MD

The JAMA Network is a consortium of peer-reviewed print and online medical publications that includes JAMA and 9 specialty Archives Journals.

© 2012 American Medical Association. All rights reserved. Reproduction without permission is prohibited.

To Promote the Science and Art of Medicine and the Betterment of the Public HealthAugust 22/29, 2012, Vol 308, No. 8 Pages 731-830



jama issue august 22 (2012)

Documents

receptor blockers

david geffen

refused definite

acute lung

community

percutaneous

dual antiplatelet

affordable