key articles and guidelines for the prevention of venous thromboembolism

P R A C T I C E I N S I G H T S

Key Articles and Guidelines for the Prevention ofVenous Thromboembolism

Paul P. Dobesh, Pharm.D., FCCP, Ann K. Wittkowsky, Pharm.D., FASHP, FCCP,Zachary Stacy, Pharm.D., William E. Dager, Pharm.D., FCSHP, Stuart T. Haines, Pharm.D., FCCP,

Larry M. Lopez, Pharm.D., FCCP, Edith Nutescu, Pharm.D., FCCP, Katherine W. Phillips, Pharm.D.,Toby C. Trujillo, Pharm.D., and Tom Vondracek, Pharm.D.

Venous thromboembolism (VTE) is an important medical problem that affectsmillions of patients each year. With appropriate prophylaxis, many of thesethromboembolic events can be prevented. Although strong evidencesupporting VTE prophylaxis spans several decades, several large Americanand global registries have documented very poor use of appropriateprophylaxis. Because of increasing regulatory requirements, hospitalsnationwide are in the process of developing documentation of appropriateVTE prophylaxis programs for both surgical and medical patients. A widerange of clinicians must understand what constitutes appropriate VTEprophylaxis in various patient populations. With the existence of numerouspharmacologic agents, abundance of data from major clinical trials, andseveral nationally recognized clinical guidelines, compiling the neededreference material to make evidence-based decisions on appropriate VTEprophylaxis can be difficult for clinicians. Therefore, we provide abibliography of key articles and guidelines related to the prevention of VTE invarious patient groups. We hope this compilation will serve as a resource forpharmacists, physicians, nurses, residents, and students responsible for thecare of patients who may be at risk for VTE.Key Words: anticoagulation, evidence-based medicine, pharmacy practice.(Pharmacotherapy 2009;29(4):410–458)

OUTLINE

Clinical GuidelinesClinical Predictors, Incidence, Prevalence, and Risk

StratificationGeneral SurgeryOrthopedic SurgeryCentral Nervous System Procedures and InjuriesTraumaIntensive Care SettingMedical IllnessesCancerPregnancyObesity and Bariatric SurgeryRenal Insufficiency and DialysisLong-Distance TravelMechanical Prophylaxis

Extended ProphylaxisQuality Improvement Initiatives

Venous thromboembolism (VTE), whichencompasses both deep vein thrombosis (DVT)and pulmonary embolism, is a significant healthcare problem, causing considerable morbidity,mortality, and resource utilization.1 Every year inthe United States, there are more than 1 millionDVT events and more than 100,000 deaths frompulmonary embolism. These events occur in awide range of patients in both surgical andmedical populations. Among patients dischargedfrom U.S. hospitals, VTE was the second leadingcause of medical complications and prolongedlength of stay, and third leading cause of

KEY ARTICLES FOR THE PREVENTION OF VENOUS THROMBOEMBOLISM Dobesh et al

preventable mortality and excessive hospitalcharges.2 Clinicians must realize that manythromboembolic events can be prevented withappropriate VTE prophylaxis. Despite more than30 years of demonstrated effectiveness and safety,VTE prophylaxis is substantially underutilized.This underutilization has led to the recentinvolvement of government and other regulatoryagencies in an attempt to improve VTEprophylaxis in U.S. hospitals.

The Cardiology Practice and Research Network(PRN) of the American College of ClinicalPharmacy has taken the initiative to compile listsof key articles and guidelines in major focus areasof cardiology. From 2004–2006, five collectionsof annotative bibliographies were published onthe topics of acute coronary syndromes,arrhythmias, hypertension, systolic heart failure,and dyslipidemias.3–7 These documents are beingupdated and published in Pharmacotherapy.8

Since the prevention of VTE is not only acardiology issue, the Cardiology PRN joined withthe Internal Medicine PRN and the AmbulatoryCare PRN to compile this document that focuseson key articles and guidelines in the preventionof VTE. We collected guidelines and significant

articles published in the area of VTE preventionand provide a summary of the results of theclinical trials, as well as clinical insights on theimplications for clinical practice and research.This document will serve as an excellent reviewand resource for pharmacists, physicians, nurses,residents, and students, especially in this time ofincreased attention on VTE prevention.

Clinical Guidelines

Geerts WH, Bergqvist D, Pineo GF, et al.Prevention of venous thromboembolism:American College of Chest Physicians evidence-based clinical practice guidelines, 8th ed. Chest2008;133(suppl 6):S381–453.

The 8th edition of the American College ofChest Physician (ACCP) guidelines, published inthe summer of 2008, provides updatedrecommendations on the prevention of VTE in awide variety of patient populations. In relationto other clinical practice guidelines on VTEprevention, the ACCP guidelines are the mostcomprehensive in terms of patient populationsdiscussed, and often are the primary guidelinesource that most clinicians turn to regarding notonly recommendations for the prevention ofVTE, but also the appropriate use of antithrom-botic agents in general. General recommen-dations include that every hospital shoulddevelop and implement a formal, institution-wide strategy to assess each patient for risk ofVTE. The guidelines also encourage specificresources such as computer decision support,preprinted orders, and periodic audit to facilitateappropriate use of VTE prophylaxis. In a changefrom the 7th edition published in 2004, the 8thedition outlines a risk assessment scheme withthree levels of risk, as opposed to four levels inthe previous version. Low-risk patients includethose undergoing minor surgery or patients whoare fully mobile with medical conditions.Patients at moderate risk include most patientsundergoing general surgery, open gynecologicsurgery, or urologic surgery, as well as patientswho are confined to bed rest with medicalconditions. High-risk patients include thoseundergoing major orthopedic surgery and thosewith major trauma or spinal cord injury (SCI).General recommendations for the prevention ofVTE depending on risk level are as follows: lowrisk—early ambulation, no specific preventiveinterventions needed; moderate risk—low-molecular-weight heparin (LMWH) at recom-mended doses, low-dose unfractionated heparin

411

From the Department of Pharmacy Practice, College ofPharmacy, University of Nebraska Medical Center, Omaha,Nebraska (Dr. Dobesh); the University of WashingtonSchool of Pharmacy, and Anticoagulation Services,University of Washington Medical Center, SeattleWashington (Dr. Wittkowsky); the Division of PharmacyPractice, St. Louis College of Pharmacy, St. Louis, Missouri,and the Department of Pharmacy, St. Luke’s Hospital,Chesterfield, Missouri (Dr. Stacy); the University ofCalifornia School of Pharmacy, San Francisco, California,and the University of California–Davis Medical Center andSchool of Medicine, Sacramento, California (Dr. Dager); theUniversity of Maryland School of Pharmacy, Baltimore,Maryland, and the Department of Pharmacy, West PalmBeach Veterans Affairs Medical Center, West Palm Beach,Florida (Dr. Haines); the Department of Pharmacy Practice,University of Florida College of Pharmacy and Medicine,Gainesville, Florida (Dr. Lopez); the Department ofPharmacy Practice, College of Pharmacy, and the Center forPharmacoeconomic Research, University of Illinois atChicago, and the Antithrombosis Center, University ofIllinois Medical Center, Chicago, Illinois (Dr. Nutescu); theDepartment of Pharmacy, Boston Medical Center, Boston,Massachusetts (Dr. Phillips); the Department of ClinicalPharmacy, School of Pharmacy, University of ColoradoDenver, and University of Colorado Hospital, Denver,Colorado (Dr. Trujillo); and the Department of Pharmacy,Exempla Saint Joseph Hospital, Denver, Colorado (Dr.Vondracek).

Address reprint requests to Paul P. Dobesh, Pharm.D.,FCCP, BCPS, Department of Pharmacy Practice, College ofPharmacy, University of Nebraska Medical Center, 986045Nebraska Medical Center, Omaha, NE 68198-6045; e-mail:[email protected].

PHARMACOTHERAPY Volume 29, Number 4, 2009

(UFH) given 2 or 3 times/day, fondaparinux, ormechanical methods in patients at high risk forbleeding; high risk—LMWH at recommendeddoses, fondaparinux, and warfarin with dosagetitration to an international normalized ratio(INR) of 2.0–3.0. Low-dose UFH tends to be lesseffective than other pharmacologic options andshould be used in conjunction with mechanicalmethods in high-risk patients. Given thecomprehensive nature of the guidelines, thereader is referred to the parent document formore details regarding recommendations inspecific patient populations.

Other recommendations should be noted.First, aspirin alone should never be recom-mended for prophylaxis in any patient group.Second, careful attention to renal function shouldbe a factor in the selection of the specific agentfor prophylaxis. When considering the use ofagents that accumulate in the urine, health carepractitioners should avoid the use of theseagents, use a reduced dosage of the agent, ormonitor the drug level or anticoagulant effect ofthe agent if applicable. Finally, when mechanicalmethods of prophylaxis are selected, carefulattention should be paid to ensuring that they areused properly.

Committee on Practice Bulletins–Gynecology,American College of Obstetricians andGynecologists. ACOG practice bulletin no. 84:prevention of deep vein thrombosis andpulmonary embolism. Obstet Gynecol 2007;110(2 pt 1):429–40.

Noting that the risk of developing VTE eventscan be significant in patients undergoing majorgynecologic surgery (estimated rates of 15–40%in the absence of prophylaxis), the AmericanCollege of Obstetricians and Gynecologistsdeveloped guideline to address recommendationsfor prevention of DVT and pulmonary embolism.In terms of defining the risk of VTE ingynecologic surgery, the authors make use of thefour-level risk scheme that was found in the 7thedition of the ACCP guidelines published in2004. Additional sections of the guidelinediscuss the available mechanical and pharma-cologic options for prophylaxis, as well as therole of hypercoagulable states in the developmentof VTE. Recommendations for VTE prophylaxisin gynecologic surgery based on risk are asfollows: low risk—no specific prophylaxisintervention is recommended other than earlyambulation; moderate risk—low-dose UFH 5000U subcutaneously twice/day, dalteparin 2500 IU

subcutaneously once/day, enoxaparin 40 mgsubcutaneously once/day, graduated compressionstockings (GCS), or intermittent pneumaticcompression (IPC) devices are all equivalentoptions; high risk—low-dose UFH 5000 Usubcutaneously 3 times/day, dalteparin 5000 IUsubcutaneously once/day, enoxaparin 40 mgsubcutaneously once/day, or IPC devices are allequivalent options; highest risk—low-dose UFH5000 U subcutaneously 3 times/day, dalteparin5000 IU subcutaneously once/day, enoxaparin 40mg subcutaneously once/day, GCS, or IPC deviceplus recommended doses of low-dose UFH orLMWH are all equivalent options. Recommen-dations are based on available literature in thegynecologic surgery arena for prevention of VTE.

Other recommendations should be noted.First, clinicians must appropriately consider thetiming of dose administration and placement orremoval of epidural anesthesia to minimize therisk of spinal hematoma. Second, if mechanicalmethods are to be used, they should be started atthe beginning of surgery. Third, the authors donot recommend the discontinuation of hormonereplacement therapy or contraceptives beforegynecologic surgery because of the lack ofevidence that discontinuation of these agentsreduces the risk of postoperative VTE. Finally, itis reasonable for patients with a history of VTE toundergo testing for hypercoagulable conditions.

Lyman GH, Khorana AA, Falanga A, et al.American Society of Clinical Oncology guideline:recommendations for venous thromboembolismprophylaxis and treatment in patients withcancer. J Clin Oncol 2007;25:5490–505.

Recognizing that there exists a close relation-ship between cancer and the occurrence of VTE,and that there are some distinct aspects inprevention or treatment of VTE in patients beingtreated for malignancy, the American Society ofClinical Oncology developed this guideline,which encompasses recommendations for bothprevention and treatment of VTE in patients withcancer. In addition, as a rationale for developingthis document, the authors cite the relative lackof focus on patients with cancer in the 2004version of the ACCP guidelines. Althoughpatients with cancer share many of the same riskfactors for VTE as the general population,specific risk factors in patients with malignancyinclude the following: time since cancerdiagnosis, with the highest risk in the first 3–6months; current metastatic disease; primary siteof malignancy; active chemotherapy; and current

412


or recent antiangiogenic therapy with thalidomide,lenalidomide, or bevacizumab.

The guideline was developed to address fiveprimary questions on the prevention or treatmentof VTE in patients with cancer. Only thequestions related to prevention of VTE will beaddressed here. First, hospitalized patients withcancer should be considered candidates toreceive prophylaxis with low-dose UFH, LMWH,or fondaparinux in the absence of contraindi-cations; this recommendation is based on trials ofmedically ill patients that contained subsets ofpatients with cancer. Second, routine prophylaxisfor ambulatory patients without VTE andreceiving chemotherapy is not indicated. Theonly exception is patients with myeloma who arereceiving thalidomide or lenalidomide pluschemotherapy or dexamethasone. In those cases,an LMWH or warfarin (INR target 1.5) should beused for prevention of VTE. Finally, all patientsundergoing major surgical interventions formalignant disease should receive prophylaxiswith low-dose UFH, LMWH, or fondaparinux inthe absence of contraindications. Prophylaxisshould extend for at least 7–10 days. Mechanicalmethods of prophylaxis such as IPC may beadded to pharmacologic agents but should not beused alone unless that patient is at high-risk forbleeding. In addition, it is reasonable to useLMWH for prophylaxis for up to 4 weeks aftersurgery in patients thought to be at very highrisk, defined as major abdominal or pelvicsurgery with residual malignant disease, obesity,or previous history of VTE.

Horlocker TT, Wedel D, Benzon H, et al.Regional anesthesia in the anticoagulated patient:defining the risks—the second ASRA consensusconference on neuraxial anesthesia and anti-coagulation. Reg Anesth Pain Med 2003;28:172–97.

Despite being more than 5 years old, theserecommendations remain the primary source forhow to handle anticoagulation in patients alsoreceiving neuraxial anesthesia and analgesia inthe perioperative period. This issue first came tothe forefront in the 1990s with the release of anumber of case reports in which patients haddeveloped spinal hematomas associated with theuse of neuraxial anesthesia. Although LMWHswere used in many of these cases, it is importantto note that many patients had hemostaticabnormalities and had received other agents thatmay affect hemostasis (aspirin, nonsteroidal

antiinflammatory drugs, UFH). Common themesof all the recommendations include considerationfor the timing of needle placement, epiduralcatheter removal, and drug administration. Inpatients receiving recent fibrinolytic therapy, theuse of spinal or epidural anesthesia shouldgenerally be avoided. In addition, generalguidelines for fibrinolytics suggest avoiding theiruse in patients with puncture of noncompressiblevessels in the past 10 days. In patients receivinglow-dose UFH subcutaneously, there is nocontraindication to neuraxial techniques, as therisk for developing spinal hematoma appears tobe minimal. In patients who are to receiveintraoperative anticoagulation with UFH, theUFH infusion should be started at least 1 hourafter needle placement, and indwelling cathetersshould be removed 2–4 hours after discontin-uation of the UFH infusion and only after thepatient’s coagulation status has been assessed.

In patients for whom LMWH is being used forVTE prophylaxis, concomitant use of other oralantiplatelet and anticoagulant drugs should beavoided. Also, needle placement should bedelayed until 10–12 hours after the dose ofLMWH, when LMWH is being started preoper-atively. If LMWH has been administered intreatment doses, needle placement shouldcommence at least 24 hours after the last dose.Finally, postoperative LMWH prophylaxis shouldbegin no sooner than 2 hours after needle orindwelling catheter removal. In patients whohad been receiving long-term warfarin therapy orare receiving warfarin therapy for VTEprophylaxis, long-term warfarin therapy ideallyshould be held for a minimum of 4–5 days beforesurgery, and assessment of the INR should beundertaken before needle insertion. In patientsreceiving preoperative warfarin for prophylaxis,an INR should be checked before needleplacement if it has been over 24 hours since thefirst dose, or if two or more doses of warfarinhave been given. For patients with an indwellingcatheter who are receiving warfarin, the cathetershould be removed when the INR is less than 1.5.The use of fondaparinux and neuraxial anesthesiaor analgesia should follow the conditions used inclinical trials as closely as possible. Additionalrecommendations are provided regarding the useof antiplatelet agents and herbal therapy. Ineither case, significant risk for spinal hematomais noted only with the use of adenosinediphosphate antagonists, such as ticlopidine andclopidogrel, and glycoprotein IIb-IIIa antagonists.

413


American Academy of Orthopedic Surgeons.Clinical guideline on prevention of symptomaticpulmonary embolism in patients undergoing totalhip or knee arthroplasty. Available from http://www.aaos.org/Research/guidelines/PE_guideline.pdf. Accessed September 27, 2008.

The guidelines from the American Academy ofOrthopedic Surgeons (AAOS) have generated ahigh level of controversy and debate. Althoughthe guidelines have not been peer reviewed, norhave they been published in a peer-reviewedjournal, the AAOS thoroughly documents thesystematic process used to produce theirconclusions. Therefore, readers can make theirown interpretations and conclusions regardingthe methods used as if the document were apublished article. Important differences guidethe development of AAOS recommendations ascompared with other guidelines that address VTEprevention. First, the authors of the AAOSguidelines question the relationship between theprevention of DVT and a subsequent decrease inpulmonary embolism. Whereas most consensusguidelines and systemic reviews take theapproach that interventions that have supportingevidence for lowering the risk of DVT will alsolower the risk of pulmonary embolism, theauthors of AAOS guidelines state that there islittle compelling literature validating thisrelationship. Second, as such, the authors of theAAOS guidelines by default place little emphasison the complications of DVT development suchas venous insufficiency, postthrombotic syndrome,and recurrent events. Finally, although bleedingis discussed and emphasized in other clinicalguidelines on VTE prevention, the AAOS authorsplace a higher degree of emphasis on the risk ofbleeding in the development of their recom-mendations, citing that bleeding may result inchronic joint stiffness, infection, and potentially areturn to surgery for evacuation of hematomas.The authors also speculate that the reported rateof bleeding in clinical trials is an underestimateof real-world experience. Consequently, therecommendations for VTE prevention in thisclinical guideline focus solely on an intervention’scapacity to reduce the frequency of clinicalpulmonary embolism, as well as to cause majorbleeding.

The availability of extensive literature on theprevention of DVT by various pharmacologicmodalities is given minimal consideration. Theconsensus document provides recommendationsfor the preoperative, intraoperative, andpostoperative periods. The recommendations for

prevention of symptomatic pulmonary embolismin patients at standard risk for both pulmonaryembolism and major bleeding include aspirin 325mg/day for up to 6 weeks, LMWH dosed accordingto the package insert for 7–12 days, fondaparinuxdosed according to the package insert for 7–12days, and warfarin with goal INR less than 2.0 for2–6 weeks. In patients with an elevated risk forpulmonary embolism but standard risk for majorbleeding, only LMWH, fondaparinux, or warfarinis recommended. In patients at high risk forbleeding, regardless of the risk of pulmonaryembolism, only aspirin and warfarin arerecommended as options. It is important to notethat while aspirin has been shown to be effectiveat preventing DVT compared with placebo, it hasconsistently been shown to be inferior toanticoagulant options at preventing DVT. Hence,the recommendation for the use of aspirin forprophylaxis is not found in other clinicalguidelines and often is actively discouraged.Resolving the discrepancy over recommendationsfor aspirin use for prophylaxis in various clinicalguidelines continues to be one of the challengesin improving the rates of VTE prophylaxis forpatients requiring acute care.

Kaufman JA, Kinney TB, Streiff MB, et al.Guidelines for the use of retrievable and con-vertible vena cava filters: report from the Societyof Interventional Radiology multidisciplinaryconsensus conference. Surg Obes Relat Dis2006;2:200–12.

The goal of this consensus document was toprovide clinical guidance to physicians regardingthe appropriate use of vena cava filters, especiallythose that may be removed after a period of time.In 2003, the United States Food and DrugAdministration approved changes to instructionsfor several available vena cava filters such thatthey could be used on a temporary basis.Subsequently, the overall use of vena cava filtersincreased, especially with nonpermanent devices.Although physicians now had technicalinstructions on how to place and removenonpermanent filters, no information wasavailable regarding which patients could becandidates for a removable filter or what specificconsiderations to evaluate before placing aremovable filter. To further complicate matters,available literature addressing the use of venacava filters, either permanent or removable, issparse at best, with most reports being in theform of observational studies.

In 2005, the Society of International Radiology

414


convened a multidisciplinary consensusconference to develop guidance for the use ofnonpermanent filters. The consensus groupadvocates several general principles. Theprimary means of prophylaxis and treatment ofVTE are pharmacologic in nature. The soleindication for use of a vena cava filter is theprevention of pulmonary embolism in patientswho cannot receive pharmacologic prophylaxisor treatment, or who despite pharmacologicprophylaxis or treatment remain at unacceptablyhigh risk for pulmonary embolism. Theindications for use of a nonpermanent vena cavafilter are no different from those for the use ofpermanent filters. The decision to use anonpermanent filter should be based on theanticipated required duration of protection frompulmonary embolism, as well as the time periodin which pharmacologic therapy may becontraindicated—when either or both of theseare transient, the use of a retrievable filter can bejustified. Regardless of whether a filter is inplace, pharmacologic prophylaxis or treatmentshould be started as soon as it is determined safeto do so. And finally, whereas the placement ofvena cava filters to prevent pulmonary embolismin patients who have no objectively confirmedacute DVT events is controversial and has littlesupport in the literature, there are patientpopulations in which some support is availableand use seems reasonable. These include thosewho sustained major trauma, critically ill patientswith a history of VTE and contraindications toanticoagulation, and those undergoing bariatricsurgery. The guideline contains additionalinformation regarding the removal of nonper-manent vena cava filters. Important managementpoints include the absence of a need todiscontinue therapeutic anticoagulation duringthe removal of a temporary filter, and, after filterretrieval, patients should be treated according totheir VTE status according to standards of care asdelineated in national guidelines.

Clinical Predictors, Incidence, Prevalence, andRisk Stratification

Silverstein MD, Heit JA, Mohr DN, PettersonTM, O’Fallon WM, Melton LJ III. Trends in theincidence of deep vein thrombosis andpulmonary embolism: a 25-year population-based study. Arch Intern Med 1998;158:585–93.

This landmark study is the first in a seriesconducted by this research group examining theincidence of and risk factors for VTE. The

purpose of this initial investigation was tocharacterize trends in the incidence of DVT andpulmonary embolism over a prolonged period oftime, specifically 1966–1990. The researchersused a retrospective, population-based studydesign whereby they identified all known cases(2218 cases) of first-lifetime DVT or pulmonaryembolism from medical records, death certificates,and autopsy reports from residents of OlmstedCounty, Minnesota. In all cases included in theanalysis, the diagnosis of DVT or pulmonaryembolism was determined by objective methods.The mean age of patients in these first-time VTEcases was approximately 62 years. Forty-twopercent had DVT, 44% had pulmonary embolism,and 14% had both DVT and pulmonaryembolism at the time of diagnosis. The averageannual incidence of VTE over the 25-year periodwas 117/100,000 patients. The average annualincidence of pulmonary embolism with orwithout DVT (69/100,000 patients) wassomewhat higher than the annual incidence ofDVT alone (48/100,000 patients). Among adultsyounger than 45 years, women were more likelyto experience a VTE than men, but the reversewas true after the age of 45 years. The overalllifetime incidence of VTE was slightly greater inmen (male:female ratio 1.2:1). As expected, theincidence of VTE increased markedly with age—nearly doubling with each decade after the age of40 years. The annual incidence of VTE over timewas not constant. There was a notable 35%decline in the average annual incidence ofpulmonary embolism between 1977 and 1979,which may have been due to changes indiagnostic testing patterns, improvements in thecare of patients with DVT, and/or a decline in therate of autopsy. Regardless of the explanation forthis unexpected observation, this study clearlydemonstrated that VTE is a common disease andits occurrence among the elderly is quite high.Perhaps the most surprising finding of this studywas the relatively high incidence of pulmonaryembolism (with or without DVT). Indeed, manycases of pulmonary embolism were identified byautopsy alone—suggesting that many patients donot present with classic VTE symptoms and/orthey die suddenly. In either case, pulmonaryembolism is far more common than previouslyrealized, and as the population ages and high-risksurgeries are performed more frequently, theproblem is likely to escalate.

Levitan N, Dowlati A, Remick SC, et al. Rates ofinitial and recurrent thromboembolic disease

415


among patients with malignancy versus thosewithout malignancy: risk analysis using Medicareclaims data. Medicine 1999;78:285–91.

The association between cancer and thrombosisis well established and was first characterized byTrousseau in 1865. Patients with an activemalignancy are significantly more likely todevelop VTE, and likewise, patients with anidiopathic or unprovoked VTE are far more likelyto receive a diagnosis of malignancy. Given thatcancer is a heterogeneous group of diseases, thisstudy attempted to quantify the incidence of DVTand pulmonary embolism associated with specifictypes of malignancy. Further, the researcherswished to determine if the rate of recurrent VTEwas greater in patients with malignancy than inpatients who have VTE unrelated to malignancy.The investigators constructed a retrospectiveanalysis by using Medicare Provider Analysis andReview Record (MEDPAR) data from 1988–1990.The MEDPAR database includes data regardingnearly every hospital admission, includingprimary and secondary diagnosis codes, forMedicare recipients. The incidence of initial VTEand recurrence over 183 days after the initialevent was determined in four mutually exclusivegroups based on initial hospitalization dischargediagnosis: patients with both DVT or pulmonaryembolism and a malignant disease; patients withDVT or pulmonary embolism without malignantdisease; patients with malignant disease withoutDVT or pulmonary embolism; and patients withnonmalignant disease without DVT or pulmonaryembolism. Although the cumulative probabilityof hospitalization for VTE in patients withmalignancy (0.6%) over the 3 years was onlyslightly greater than that in patients withoutmalignancy (0.57%, p=0.001), the rates of VTEamong different types of cancers were strikinglydifferent. The rates of VTE among patients withovarian (120/10,000 patients), brain (117/10,000patients), and pancreatic (110/10,000 patients)cancers were 5–8 times greater than the ratesobserved in those with head and neck (16/10,000patients), bladder (22/10,000 patients), andbreast (22/10,000 patients) cancers. Othercommon malignancies such as prostate, colon,and lung cancers had rates of VTE of 61–79/10,000 patients. The investigators also foundthat the cumulative probability of re-admissionwith a VTE over the next 183 days wassignificantly greater in patients with malignancyand VTE (22%) when compared to patients withVTE without malignancy (6.5%), those withmalignancy without VTE (13.5%), or those with

nonmalignant disease without VTE (8%).Likewise, the probability of death was signifi-cantly higher in those patients with a dischargediagnosis of malignancy and VTE (94%)compared to those with malignancy alone (42%),no malignancy (29%), and VTE withoutmalignancy (26%).

Whereas the generalizability of these data issomewhat limited due to the older age of thepopulation sample, the findings clearly show thatthe risk of VTE is significantly different amongthe various types of malignancy and that the riskof recurrent VTE and death is very high amongthose patients with the two diagnoses. Indeed, inthe absence of autopsy data, it seems likely that asubstantial proportion of the deaths in the VTEwith malignancy group were attributable to fatalpulmonary embolism or to bleeding compli-cations related to anticoagulation therapy.

Heit JA, Silverstein MD, Mohr DN, et al. Riskfactors for deep vein thrombosis and pulmonaryembolism: a population-based case-control study.Arch Intern Med 2000;160:809–15.

The primary objectives of this case-controlstudy were to identify independent risk factorsfor VTE and to estimate the magnitude of riskimparted by each. Using data from the RochesterEpidemiology Project, the researchers identifiedall known patients from Olmsted County,Minnesota, with a first-lifetime VTE eventbetween 1976 and 2000 (625 patients), andcompared them with 625 age-, sex-, andcalendar-year–matched controls. The diagnosisof VTE among case patients was confirmed byobjective means. Over 500 different clinicalparameters were collected from the medicalrecords of the case patients and control subjectsby trained abstractors. Step-wise conditionallogistic regression analysis was performed on thedata to create a final model, and a bootstrapmethod was used to validate the variablesselected. The factors most powerfully associatedwith a first-time VTE were surgery (odds ratio[OR] 21.7), trauma (OR 12.7), and recenthospitalization or nursing home confinement(OR 8.0). Other factors that conferred consid-erable risk were malignancy with (OR 6.5) orwithout (OR 4.5) chemotherapy, central venouscatheter (CVC) or pacemaker placement (OR5.6), and neurologic disease with extremityparesis (OR 3.0). A history of superficial veinthrombosis (OR 4.3) also appeared to increasethe likelihood of VTE, as did varicose veinsbefore the age of 45 years (OR 4.2) but not later

416


in life. Among female patients, the postpartumperiod (OR 19) and gynecologic surgery (OR 11)were also found to increase the risk of VTE. Notsurprisingly, serious liver disease (OR 0.1)protected against VTE.

This study confirmed that surgery, trauma, andhospitalization were all strong and independentrisk factors for the development of VTE.However, given the relatively small number ofcases in this cohort study, the risk associated withless frequently encountered factors such asinflammatory bowel disease, chronic kidneydisease, and oral contraceptive use could not beaccurately estimated. Moreover, informationregarding inherited thrombophilias was notavailable. Finally, the study population wasrestricted to residents of Olmsted County,Minnesota, which limits the generalizability ofthe findings.

Crowther MA, Kelton JG. Congenital throm-bophilic states associated with venousthrombosis: a qualitative overview and proposedclassification system. Ann Intern Med 2003;138:128–34.

The authors of this review article regardinginherited prothrombotic states give readers aqualitative overview of the subject and propose asimplified classification scheme. The authorssegment the major hereditary thrombophiliasinto two major groups: group 1 being disorderscaused by a deficiency of coagulation factorinhibitors, and group 2 being disorders caused byincreased levels or function of coagulationfactors. Group 1 disorders include deficiencies ofantithrombin, protein C, or protein S. Group 2disorders include activated protein C resistance(or factor V Leiden); factor II (prothrombin)G20210A mutation; elevated levels of factorsVIII, IX, and XI; evaluated levels of lipopro-tein(a); and dysfibrinogenemia. Disorders ingroup 1 are relatively uncommon, with aprevalence in the general population estimated tobe less than 0.5%. However, among patients withan unprovoked or idiopathic VTE, as many as7.5% have a group 1 deficiency. Group 1deficiencies generally impart a greater lifetimerisk of VTE when compared with group 2disorders. Some of the group 2 disorders arerelatively common in the general population.Upward of 5–6% of healthy Northern Europeanshave factor V Leiden. Similarly, the prothrombingene mutation has been found in 4% of thegeneral population in some parts of the world.Although group 2 disorders clearly increase the

lifetime risk of VTE, it is uncommon for them tobe the primary cause. In other words, group 2disorders are a contributing factor that increasesthe likelihood of VTE in patients with otherunderlying risks such as surgery, trauma,estrogen use, or acute medical illness requiringhospitalization. Given the prevalence of group 2disorders in the population, co-inheritance oftwo prothrombotic states is not uncommon, andthe risk imparted by each disorder appears to beadditive. Whether clinicians should routinelyobtain a “hypercoagulable work-up” in patientswho have an idiopathic or seemingly unprovokedVTE at presentation is controversial. It remainsunclear whether this information would alter theintensity or duration of antithrombotictreatment. Nevertheless, a working knowledge ofthe most common thrombophilias, their relativeprevalence, and the relative risk (RR) for VTEthat each imparts will help clinicians interpretthe literature and understand the context offuture clinical trials.

Anderson FA Jr, Spencer FA. Risk factors forvenous thromboembolism. Circulation 2003;107:I9–16.

Written by two of the most frequently citedVTE epidemiology researchers, this compre-hensive review examines the evidence impli-cating a breadth of risk factors associated withVTE. The authors stratified these risk factorsinto two broad categories: individual risk factorsthat are sufficient to justify the use of antithrom-botic drugs for prophylaxis, and factors thatincrease VTE risk but are not sufficientindividually to justify pharmacologic approachesfor prevention. The strongest risk factors forVTE (OR > 10) include hip or leg fracture, hip orknee replacement, major general surgery, majortrauma, and SCI. The use of antithromboticdrugs is clearly warranted in these populations.Factors that place patients at moderate risk (OR2–9) include previous VTE, arthroscopic kneesurgery, malignancy, CVCs, cancer chemotherapy,heart failure, respiratory failure, pregnancy orpostpartum period, oral contraceptives, hormonereplacement therapy, and thrombophilia. Amongthese moderate risk factors, the authorssuggested that malignancy, heart failure, andrespiratory failure pose sufficient risk to warrantpharmacologic prophylaxis during hospital-ization—a premise that has been confirmed byrandomized clinical trials. It is important to notethat thrombophilias, such as antithrombindeficiency, protein C or S deficiency, activated

417


protein C resistance or factor V Leiden, and thefactor II G20210A mutation, are not, in and ofthemselves, considered sufficiently strong riskfactors to warrant primary prophylaxis withantithrombotic drugs. Factors that are relativelyweak contributors to VTE risk (OR < 2) includedbed rest for more than 3 days, prolonged sitting(e.g., due to car or air travel), increasing age,obesity, and varicose veins. These risk factors areadditive, and most patients who develop VTEwill have two or more identifiable risk factors.

Heit JA, Kobbervig CE, James AH, PettersonTM, Bailey KR, Melton LJ III. Trends in theincidence of venous thromboembolism duringpregnancy or postpartum: a 30-year population-based study. Ann Intern Med 2005;143:697–706.

This research report is an extension of theauthors’ data regarding VTE risks previouslyreported in the Archives of Internal Medicine in2000. The purpose of this study was tospecifically examine the risk for VTE duringpregnancy and the postpartum period, as well astrends over time. Using data from the RochesterEpidemiology Project, the researchers identified100 women known to have had a DVT orpulmonary embolism during pregnancy or thepostpartum period (e.g., up to 3 mo after thetermination of pregnancy) in Olmsted County,Minnesota, between 1966 and 1995. Theabsolute risk of VTE during pregnancy andpostpartum period combined was approximately200/100,000 woman-years. This represents anRR for VTE of 4.2 among pregnant andpostpartum women compared with the expectedage-specific rate of VTE among women living inOlmstead County. The annual incidence of VTEwas significantly greater among postpartumwomen (511/100,000 patient-yrs) whencompared with pregnant women (96/100,000patient-yrs). Moreover, the annual incidence ofpulmonary embolism was substantially greater inpostpartum women (160/100,000 patient-yrs)when compared with pregnant women(11/100,000 patient-yrs). Younger women (aged15–19 yrs) were more likely to experience a DVTduring pregnancy than older women, but theincidence of VTE, particularly pulmonaryembolism, increased with age during thepostpartum period. Trend analysis over 30 yearsrevealed that the incidence of VTE duringpregnancy remained essentially unchanged, butthe incidence of pulmonary embolism during thepostpartum period decreased by greater than 2-fold with a notable decrease observed from

1985–1995. The authors postulate that thisfinding may be due to shorter hospital stays andearlier mobilization of women after delivery.Although the generalizability of this study islimited because of the narrow population studied(98% Caucasian), the findings are a reminderthat the postpartum period places women atgreater risk for VTE. Whether universallyscreening women during the first trimester ofpregnancy and using appropriate interventionalstrategies (e.g., GCS) for those at most riskwould lower the frequency of VTE duringpregnancy is unknown.

Caprini JA. Thrombosis risk assessment as aguide to quality patient care. Dis Mon 2005;51:70–8.

This concise review by a longstanding andpassionate proponent for universal risk assess-ment is an excellent guide for clinicians andclinical administrators; it uses concrete examplesto improve the quality of patient care. Includedin the article is Evanston NorthwesternHealthcare’s Thrombosis Risk Factor Assessmentscore sheet that the author and his colleaguesdeveloped in the late 1980s. They have used andrefined this instrument over many years based on“science, logic, emotion, and experience.” TheThrombosis Risk Factor Assessment instrumentis a practical tool that can be completed by anyhealth care practitioner based on easily obtainedinformation gathered during a patient interviewsupplemented by the medical record. The toolhelps practitioners weigh the relative strength ofeach risk factor by scoring them from 1–5 points.Risk factors awarded only 1 point are consideredrelatively minor contributing causes, whereasthose risk factors awarded higher points are ofrelatively greater strength and importance interms of VTE risk. The cumulative total riskfactor score is determined by simple addition.Patients with a Thrombosis Risk Factor Assessmentscore of greater than 2 should receive VTEprophylaxis, using either a mechanical or phar-macologic method. Patients with a ThrombosisRisk Factor Assessment score of 5 or moreshould receive pharmacologic prophylaxis aloneor in combination with a mechanical method.The tool also provides the clinician guidanceregarding potential safety issues that must beconsidered before beginning pharmacologicstrategies. For practitioners and institutions thatare struggling to create a VTE risk assessmentinstrument, the simple but effective tooldescribed in this review is extremely helpful.

418


Anderson FA Jr, Zayaruzny M, Heit JA, Fidan D,Cohen AT. Estimated annual numbers of U.S.acute-care hospital patients at risk for venousthromboembolism. Am J Hematol 2007;82:777–82.

As the title implies, the objective of this studywas to determine the total number of patientsadmitted to U.S. hospitals in 2003 who were atrisk for VTE. The investigators used the 2003Nationwide Inpatient Sample from the HealthcareCost and Utilizations Project to determine thetotal number of patients at risk. The NationwideInpatient Sample is a robust all-payer inpatientdatabase with information regarding approxi-mately 8 million hospital stays from nearly 90%of hospitals in the United States. Adults (age >18 yrs) who were admitted to the hospital for atleast 2 days were considered potentially eligiblefor VTE prophylaxis and, therefore, wereincluded in the analysis. Of the more than 38million patients who met these criteria, nearly7.8 million were admitted for a major or minorsurgical procedure and 15.2 million wereadmitted for a medical indication. Not sur-prisingly, among surgical patients, a largeproportion (56% or 4.3 million) were eligible forthromboprophylaxis based on the 2001 ACCPguidelines. Among those patients who wereadmitted to a hospital for a medical indication,over 7.7 million (51%) were at risk for VTEbased on the 2001 ACCP guidelines. In total,nearly 12.1 million hospitalized patients were atsubstantial risk for VTE and should have receivedVTE prophylaxis according to the 2001 ACCPguidelines. Given that nearly 1 in 3 hospitalizedpatients are at risk for VTE, this represents a verysignificant public health threat. Although thesedata were derived from a 2003 population sampleand the VTE risk assessment and prophylaxisrecommendations are based on 2001 guidelines,it is likely that the number of patients who are atrisk for VTE and should receive a recommendedprophylaxis strategy will continue to increase inthe years to come due to the aging of thepopulation, increasing numbers of orthopedicprocedures performed each year, and increasingprevalence of heart failure, stroke, and otheracute medical illnesses that place patients at riskfor VTE. These data reinforce the importance ofscreening all patients for VTE risk factors at thetime of hospital admission.

Spencer FA, Lessard D, Emery C, Reed G,Goldberg RJ. Venous thromboembolism in theoutpatient setting. Arch Intern Med 2007;167:1471–5.

This study, conducted by the Worcester VenousThromboembolism Study Group, examined VTEfrom an outpatient perspective. Specifically, theinvestigators attempted to quantify the proportionof patients who first experience VTE symptomsin the outpatient setting, determine theprevalence of well-documented risk factors forVTE in the outpatient setting, and ascertainwhether the use (or nonuse) of VTE prophylaxisduring a recent hospitalization modifies VTErisk. This was a retrospective study in whichtrained abstractors reviewed the medical recordsof all residents of Worcester, Massachusetts, whohad a health care encounter (inpatient,outpatient, emergency department, or laboratory)coded with any of 34 VTE diagnostic codes in1999, 2001, and 2003. Information gatheredincluded patient demographics, history ofprevious VTE, results of clinical laboratory anddiagnostic tests, surgeries, hospitalizations, anduse of VTE prophylaxis during previoushospitalizations.

From 7222 potential cases identified in thestudy periods, 1897 Worcester residents had anindependently validated DVT, pulmonaryembolism, or both. Of these, 1399 (73.7% of allVTE) were considered cases that occurred in theoutpatient setting. More than 40% of the patientshad been hospitalized (36.8%) or had majorsurgery (23.1%) in the 3 months before the VTE.Furthermore, a substantial percentage had adiagnosis of active malignancy (29%) or a historyof VTE (19.9%). In those cases in which theVTE was preceded by a hospitalization, morethan two thirds presented within 30 days ofhospital discharge and more than 40% receivedno VTE prophylaxis (pharmacologic ormechanical) during the previous hospital stay.The lack of VTE prophylaxis during the previoushospital stay was equally striking among patientswith well-known and powerful risk factors forVTE such as surgery (38%), malignancy (36%),or history of VTE (26%).

These data confirm that most DVTs andpulmonary embolisms develop in patients whoare not hospitalized, but these events are oftentemporally related to a hospitalization. Moreover,patients with malignancy or previous VTE aredisproportionately represented in this population;therefore, an effective strategy to lower the risk ofVTE in the outpatient setting must be targeted tothese populations. A significant percentage ofpatients with strong VTE risk factors did notreceive any form of prophylaxis during theirhospital stays, and this may have led to

419


potentially preventable cases of VTE. However,most patients who were previously hospitalizeddid receive VTE prophylaxis—an observationthat suggests that extended prophylaxis afterhospital discharge might be a worthwhile strategyin more patient populations.

General Surgery

Mismetti P, Laporte S, Darmon J-Y, et al. Meta-analysis of low-molecular-weight heparin in theprevention of venous thromboembolism ingeneral surgery. Br J Surg 2001;88:913–30.

This meta-analysis sought to expand earlierobservations that showed that prophylaxis withLMWH compared with placebo was associatedwith a reduction in risk of symptomatic DVT,pulmonary embolism, and mortality. The statedobjectives of this meta-analysis were to confirmor refute these past findings, but also to evaluatecomparative effects of LMWH and UFH inpatients undergoing general surgery or surgeryfor cancer. In addition, the efficacy and safety oflow and high prophylactic doses of LMWH wereevaluated. Studies included in this analysis wereall randomized comparisons of a prophylacticregimen of an LMWH with any other prophy-lactic strategy in patients undergoing generalsurgery (defined as abdominal, thoracic,gynecologic, or surgery for malignant disease).Primary and secondary end points werefrequency of objectively detected DVT andsymptomatic pulmonary embolism, symptomaticVTE, death, major hemorrhage, wound hematoma,other hemorrhage, and postoperative transfusion.Both the number of studies (59) and patients(54,144) included in this meta-analysis wereconsiderable. Unfortunately, the number ofdifferent LMWH compounds was also large(nadroparin, certoparin, dalteparin, enoxaparin,parnaparin, and tinzaparin), and as a conse-quence, it was not possible to compare theirrelative effects. Low-dose LMWH was defined as3400 or fewer antifactor Xa U/day, whichcorresponds to an enoxaparin dose of approxi-mately 35 mg/day.

Compared with placebo or no treatment, use ofLMWH was associated with an overall 72%relative risk reduction (RRR) for DVT, a 75%RRR for pulmonary embolism, a 71% RRR forVTE, and a 46% RRR in mortality. All of thesedifferences were statistically significant (p valuerange 0.001–0.018) except for overall mortality(p=0.09). Failure to achieve statistical signifi-cance for mortality was most likely due to the

small number of studies (eight) included and theresultant large confidence interval (CI). Whencompared with UFH, LMWH did not result instatistically significant reductions in DVT,pulmonary embolism, or death (p value range0.10–0.63), although reduction in VTE did reachmarginal significance (p=0.049). Also, whenonly double-blind studies were considered, noneof these achieved statistical significance. In thesestudies, the dose of UFH was either 5000 Utwice/day (30 studies), 5000 U 3 times/day (20studies), or 2500 U/day (one study). In spite ofthe substantial number of studies included in thisportion of the analysis, no attempt was made tocompare the different dosages of UFH. Whenonly cancer surgery was considered, nosignificant difference could be detected betweenLMWH and UFH. Finally, no significantdifference was observed in the occurrence ofclinical events when low-dose or high-doseLMWH was compared with UFH. However, useof low-dose LMWH was associated with asignificant 24% RRR of major hemorrhage.Overall, this meta-analysis confirmed earlierobservations that prophylactic use of LMWH isclearly effective at preventing clinically importantVTE events.

Koch A, Bouges S, Ziegler S, et al. Low-molecular-weight heparin in thrombosisprophylaxis after major surgical intervention:update of previous meta-analysis. Br J Surg1997;84:750–9.

This meta-analysis may be considered supple-mentary to the one just discussed; although, itassessed different end points and includedpatients who had undergone orthopedicsurgeries. It analyzed randomized, double-blindclinical trials comparing thromboprophylacticefficacy of LMWH and UFH after general andorthopedic surgeries. The primary end pointswere objectively determined DVT and woundhematoma. Frequency of pulmonary embolismand other categories of major bleeding wereincluded under additional analyses. A total of 36trials involving 16,583 patients were included inthe final analysis. Although there were fewertrials and patients included in this meta-analysiscompared with the meta-analysis discussed above(51 and 48,624), observations regardingcomparative frequency of DVT between LMWHand UFH were similar in both analyses.Specifically, no significant difference was observedin either analysis in comparative frequency of DVT.In this analysis, however, there was a statistically

420


significant 39% reduction in frequency ofpulmonary embolism in patients receiving LMWH,whereas in the former analysis only a 12%reduction was observed and it was not statisticallysignificant. Also in this analysis, overall frequencyof wound hematoma and other indexes of majorhemorrhage were similar in the two treatmentgroups, observations also noted in the formermeta-analysis. Finally, both analyses were unableto demonstrate therapeutic superiority of high-dose LMWH (defined as > 3400 antifactor XaU/day) over either its low-dose counterpart or UFHwith respect to DVT. In both studies, however,there was an advantage in using low-dose LMWHwith respect to occurrence of bleeding.

Bergqvist D, Burmark US, Frisell J, et al. Low-molecular-weight heparin once daily comparedwith conventional low-dose heparin twice daily:a prospective double-blind multicentre trial onprevention of post-operative thrombosis. Br JSurg 1986;73:204–8.

This was one of the very early studies thatdocumented the usefulness of subcutaneousLMWH compared with low-dose UFH 5000 Usubcutaneously twice/day for prevention ofpostoperative DVT. As such, it is primarily ofhistorical importance. In this double-blind study,432 consecutive patients from four centers inSweden were randomly assigned in blocks of 10to receive either subcutaneous LMWH, describedas “heparin fragment,” or UFH for thrombo-prophylaxis. To preserve the double-blinddesign, patients assigned to the LMWH groupreceived a subcutaneous placebo injection astheir evening dose. Each patient wasadministered the initial dose 2 hours beforescheduled surgery and every 12 hours thereafterfor 5–7 days. Preparation of the LMWH wasdescribed, and the resultant compound wasreported to have a 4:1 ratio of antifactorXa:antifactor IIa, which is comparable to today’senoxaparin. The dosage of LMWH administeredwas reported as 5000 antifactor Xa U/day, adosage greater than that used today, which istypically 75–3500 U/day. The dosage of UFHadministered was the classic 5000 U twice/day.After 30 days of follow-up, the frequency ofobjectively detected DVT in the LMWH and UFHgroups was 4.3% and 6.4%, respectively, (p=NS),and no significant difference was observed inoverall mortality (data and p value not reported).However, the overall frequency of hemorrhagicevents was greater in the LMWH group than theUFH group (11.6% vs 4.6%, p=0.007), an

observation most likely due to the relatively highdose of LMWH used in this study. There was noassessment of the occurrence of any other VTEevents such as pulmonary embolism. Theequivalent effect on mortality, however, suggeststhat pulmonary embolism occurred equallyfrequently in the two study groups. Sincepublication of this trial, subsequent comparisonsof prophylactic doses of UFH and LMWH haverevealed minimal to no differences in thefrequency of hemorrhagic complications inpatients undergoing general surgery.

Leonardi MJ, McGory ML, Ko CY. The rate ofbleeding complications after pharmacologic deepvein thrombosis prophylaxis: a systematic reviewof 33 randomized control trials. Arch Surg2006;141:790–9.

Observations from this well-conducted meta-analysis have been used as part of the rationalefor identifying DVT prophylaxis as a Medicarequality measure. One strength of this article isthat the frequency of specified bleeding compli-cations was analyzed based on the administereddose of LMWH or UFH, and both were comparedwith placebo. The bleeding complicationsassessed were injection site bruising, woundhematoma, drain site bleeding, hematuria,gastrointestinal tract bleeding, retroperitonealbleeding, discontinuation of prophylaxis, andsurgery to correct bleeding. Thirty-threerandomized control trials involving 33,813patients were included in this analysis. High-dose LMWH was defined as more than 3400antifactor Xa U/day, and high-dose UFH wasdefined as 5000 U 3 times/day.

The two most common bleeding complicationswere injection site bruising (6.9%) and woundhematomas (5.7%). Not unexpectedly, thesecomplications were observed less frequently inpatients receiving placebo (2.8% and 0.8%,respectively). Of note, these bleeding complica-tions were observed more frequently in the low-dose LMWH group than in the high-dosecounterparts (bruising 6.8% vs 3.4%, p=0.04, andhematoma 6.6% vs 4.0%, p<0.001). Drain sitebleeding, hematuria, and discontinuation of DVTprophylaxis were observed less frequently (2.0%,1.6%, and 2.0%, respectively). Of note,hematuria was observed considerably morefrequently in both high-dose LMWH and UFHgroups compared with the low-dose group (5.8%vs 0.4% and 4.7% vs 0.2%, respectively, p<0.001for both). Finally, rates of the major bleedingcomplications—gastrointestinal bleeding,

421


retroperitoneal bleeding, and surgery forbleeding—were all very low (0.2%, 0.08%, and0.7%, respectively). Based on these observations,it was concluded that pharmacologic thrombo-prophylaxis was safe for patients undergoinggeneral surgery who are at moderate or high riskfor developing DVT. Unfortunately, thisconclusion is incompletely supported by the datasince there was no assessment of level of patientrisk in the 33 studies included in this analysis.

Haas S, Wolf H, Kakkar AK, et al. Prevention offatal pulmonary embolism and mortality insurgical patients: a randomized double-blindcomparison of LMWH and unfractionatedheparin. Thromb Haemost 2005;94:814–19.

Before publication of this article, efficacy ofpharmacologic prophylaxis had been wellestablished for DVT, but mortality, particularlyfatal pulmonary embolism, had not beenconsidered fully. The purpose of this well-designed, large clinical trial was to assess thefrequency of overall mortality and fatalpulmonary embolism associated with thrombo-prophylaxis with an LMWH or UFH. In thisdouble-blind study, 23,078 patients enrolled at 67centers in Germany, Austria, and the CzechRepublic were randomly assigned in a double-blind manner to receive either certoparin 3000antifactor Xa U once/day or UFH 5000 U 3times/day, both administered subcutaneously.Patients assigned to the certoparin group alsoreceived two placebo injections/day in order topreserve the double-blind design of the study.The first dose was given 2 hours before surgeryand continued for 5–20 days thereafter. Theprimary end point was autopsy-confirmedpulmonary embolism occurring during therapyand up to 14 days thereafter. All-cause mortalitywas the secondary end point and was recordedfor the same time period. Data were presentedand analyzed in an intent-to-treat manner only.

All-cause mortality was reported as 1.66% inthe LMWH group and 1.46% in the UFH group(p=0.28). Autopsy was performed in 70.2% ofthese patients, and pulmonary embolism wasfound in 0.15% of patients of both groups(p=0.87). If these latter data were to beextrapolated to the entire cohort of patients, itwas estimated that the final rate of fatalpulmonary embolism would have been 0.2%.

Overall, observations of this study revealedthat LMWH and UFH were equivalent inprevention of fatal pulmonary embolism and all-cause mortality. The certoparin dosage of 3000

antifactor Xa U/day was somewhat less than thetraditionally used LMWH dosage of 3400antifactor Xa U/day. Such a difference likely hadonly minimal impact on the outcomes, however,as previous observations with a higher dose ofcertoparin (5000 antifactor Xa U/day) revealed“no differential efficacy” when compared withthe lower dose.

Of note, the highest rates of fatal pulmonaryembolism and all-cause mortality were observedin patients who underwent surgery for hipfracture. The authors commented that thebenefits of thromboprophylaxis appeared to beblunted in this group of patients.

Agnelli G, Berqvist D, Cohen AT, et al, onbehalf of the PEGASUS Investigators.Randomized clinical trial of postoperativefondaparinux versus perioperative dalteparin forprevention of venous thromboembolism in high-risk abdominal surgery. Br J Surg 2005;92:1212–20.

Before publication of this trial, there had beena number of clinical trials comparing thrombo-prophylactic efficacy and safety of fondaparinuxand enoxaparin after orthopedic surgeries. Thiswas the first trial, to our knowledge, to similarlyevaluate fondaparinux and dalteparin afterabdominal surgery. In this trial, 2927 patientsfrom 131 hospitals in 22 European counties wererandomly assigned in a double-blind fashion toreceive either fondaparinux 2.5 mg/day ordalteparin 5000 IU/day, both administeredsubcutaneously. First doses of fondaparinux anddalteparin were administered 6 hours aftersurgical closure and 2 hours before surgery,respectively, and double-blind therapy continuedfor 5–9 days. The primary efficacy outcome wasobjectively detected VTE (asymptomatic and/orsymptomatic DVT and/or pulmonary embolism).The primary safety outcome was major bleedingdefined as fatal, retroperitoneal, intracranial,intraspinal, or involving any other critical organ,and bleeding leading to surgery or otherinterventions such as need for transfusion.

Notably, the trial was originally intended todemonstrate superiority of fondaparinux overdalteparin. The overall rate of VTE, however,was less than anticipated, and a noninferiorityanalysis was used instead. A VTE was detectedin 4.6% and 6.1% of patients in the fondaparinuxand dalteparin groups, respectively; a statisticallynonsignificant difference that met prespecifiednoninferiority criteria. Likewise, DVT wasobserved in 4.2% and 5.8%, respectively

422


(p=0.10), and frequency of pulmonary embolismwas nearly equal in both groups (0.3% and 0.2%,respectively, statistical analysis not reported).Major bleeding was noted in 3.4% and 2.4% ofpatients, respectively (p=0.122). Results ofprevious studies with fondaparinux revealed aclear superiority when compared with enoxaparinfor prevention of postoperative VTE in patientswho underwent orthopedic surgeries. In thisstudy, fondaparinux was essentially equivalent todalteparin for the prevention of VTE after majorabdominal surgery. It is noteworthy that thiseffective equivalence was achieved by using apostoperative administration strategy forfondaparinux, which is preferred by manysurgeons.

Orthopedic Surgery

Quinlan DJ, Eikelboom JW, Dahl OE, et al.Association between asymptomatic deep veinthrombosis detected by venography andsymptomatic venous thromboembolism inpatients undergoing elective hip or knee surgery.J Thromb Haemost 2007;5:1438–43.

Although contrast dye–enhanced venographyis commonly used to compare the efficacy ofvarious thromboprophylactic options in patientsundergoing major orthopedic surgery, the truerelationship between asymptomatic DVT andsymptomatic VTE is unclear. This articledescribes the relationship between asymptomaticDVT and symptomatic VTE in patients undergoingelective hip or knee surgery who were treatedwith enoxaparin 30 mg twice/day or 40 mgonce/day. The frequency of asymptomatic DVTin 10 studies in which venography was performed(5796 patients) was compared with the frequencyof symptomatic VTE in two studies in whichvenography was not performed (3500 patients).The frequency of asymptomatic DVT after totalhip replacement (THR) was 13.2% (95% CI12.2–14.2%) and after total knee replacement(TKR) was 38.1% (95% CI 35.5–40.8%). The 90-day rate of symptomatic VTE after THR was 2.7%(95% CI 2.1–3.4%) and after TKR was 1.8% (95%CI 0.9–2.7%). The asymptomatic venographicDVT:symptomatic VTE ratio was 5:1 in patientsundergoing THR and 21:1 for patientsundergoing TKR. Despite some differences in therelationship due to the type of surgery and thevenogram reading committee, a consistentrelationship was found between asymptomaticDVT and symptomatic VTE. The results of thisstudy further strengthen the evidence that DVT

detected by means of venography is a valid endpoint and surrogate for symptomatic VTE.

Pulmonary Embolism Prevention (PEP) TrialCollaborative Group. Prevention of pulmonaryembolism and deep vein thrombosis with low-dose aspirin: pulmonary embolism prevention(PEP) trial. Lancet 2000;355:1295–302.

Although antiplatelet therapy has been shownto be effective in reducing the risk of arterialthrombotic events, controversy exists regardingthe role and efficacy of aspirin as a primarythromboprophylactic agent against VTE inpatients undergoing major orthopedic surgery.The Pulmonary Embolism Prevention (PEP) trialrandomly assigned 13,356 patients undergoinghip fracture surgery and 4088 patients under-going elective THR or TKR to receive aspirin 160mg/day or placebo, which was started preoper-atively and continued for 35 days. In the hipfracture surgery group, aspirin-treated patientsshowed a 43% proportional reduction inpulmonary embolism (95% CI 18–60%, p=0.002)and a 29% reduction in symptomatic DVT (95%CI 3–48%, p=0.03). In the elective arthroplastygroup, rates of VTE were lower (1.1% in theaspirin group vs 1.4% in the placebo group,hazard ratio (HR) 0.81, 95% CI 0.47–1.42), butthe proportional effects of aspirin were similar tothose of the hip fracture surgery group. Amongall 17,444 randomized patients, the risk ofpulmonary embolism or DVT was decreased by34% (95% CI 17–47%, p=0.0003). Deaths due tobleeding were higher in the aspirin groupcompared with the placebo group (p=0.04).

This is one of the largest trials that support arole for aspirin as a thromboprophylactic optionagainst VTE in major orthopedic surgery.Although the results suggest that low-doseaspirin is beneficial in reducing the risk ofpulmonary embolism and symptomatic DVT inpatients undergoing hip fracture surgery and inelective hip or knee arthroplasty, data fromprevious and often methodologically limitedstudies are inconclusive and controversial. Asmore effective methods of prophylaxis areavailable in patients undergoing major orthopedicsurgery, the role of aspirin remains highlycontroversial. Because of the availability of moreeffective therapies, the ACCP guidelinesrecommend against the use of aspirin as a solethromboprophylactic agent in this patient group.

Leclerc JR, Geerts WH, Desjardins L, et al.Prevention of venous thromboembolism afterknee arthroplasty: a randomized, double-blind

423


trial comparing enoxaparin with warfarin. AnnIntern Med 1996;124:619–26.

Preventing VTE in patients undergoing TKR isdifficult due to the bleeding risk associated withthe procedure and the relative resistance of thisprocedure to the effect of most traditionalantithrombotic agents. This double-blind,randomized, controlled study added to thealready existing body of evidence supporting theefficacy and safety of LMWH when comparedwith warfarin in patients undergoing TKRsurgery. Six hundred seventy patientsundergoing TKR were randomly assigned toreceive enoxaparin 30 mg subcutaneously every12 hours or adjusted-dose warfarin (INR 2–3),both of which were started after surgery. The rateof DVT in patients with adequate bilateralvenograms was the primary end point, andbleeding was the secondary end point.

In the 417 patients with adequate venograms,the rate of DVT was lower in the enoxaparingroup compared with the warfarin group (36.9%vs 51.7%, p=0.003), with an absolute riskdifference of 14.8% in favor of enoxaparin (95%CI 5.3–24.1%). The rates of proximal DVT andmajor bleeding were not significantly differentbetween the two groups (p>0.2 for both). Thisstudy further confirmed that prophylaxis with anLMWH after TKR resulted in better efficacyoutcomes compared with traditional agents suchas warfarin and strengthened the role of LMWHas a preferred prophylactic option in thesepatients. Despite the efficacy benefit thatenoxaparin conferred over warfarin, the rate ofresidual thrombosis was still fairly high in bothtreatment groups, leaving the quest open for evenmore effective therapeutic options. The currentACCP guidelines recommend either of thesestrategies, or fondaparinux, for VTE preventionafter TKR surgery.

Colwell CW Jr, Collis DK, Paulson R, et al.Comparison of enoxaparin and warfarin for theprevention of venous thromboembolic diseaseafter total hip arthroplasty: evaluation duringhospitalization and three months after discharge.J Bone Joint Surg Am 1999;81:932–40.

Both warfarin and LMWH are now part of theaccepted drugs available for prophylaxis of VTEin patients undergoing THR. In this randomized,open-label, parallel group study, the efficacy andsafety of LMWH (enoxaparin) and warfarin inthe prevention of clinically symptomatic VTEwas evaluated in patients undergoing THR. Thestudy was divided into two phases: up to 14 days

during hospitalization, and a follow-up period of3 months after discharge. Three thousand elevenpatients were randomly assigned to in-hospitaltreatment with enoxaparin 30 mg subcutaneouslyevery 12 hours to begin within 24 hours ofsurgery, or adjusted-dose warfarin to begin within48 hours before surgery to 24 hours after. A totalof 1516 patients received enoxaparin, and 1495received warfarin. Mean duration of treatmentwith either drug was 7.3 days.

During the entire study period, clinicallyimportant VTE occurred in 3.6% (55 patients) ofthe enoxaparin group and 3.7% (56 patients) ofthe warfarin group (p=0.8). During hospital-ization, four enoxaparin-treated patients (0.3%)and 17 warfarin-treated patients (1.1%)experienced a VTE (p=0.0083). This benefit inthe enoxaparin group was lost after therapy wasstopped, with no significant difference in VTEbetween the enoxaparin and warfarin groups at 3months after discharge (3.4% vs 2.6%, p=0.22).Major or minor bleeding occurred in 152 (10%)enoxaparin-treated patients and in 110 (7.4%)warfarin-treated patients (p value not reported).There was a notable relationship between thetiming of the initial enoxaparin dose and majorbleeding, as 14 of 18 enoxaparin-treated patientswho had a major bleed received the first dose lessthan 12 hours after surgery. Although this studyfavors the use of LMWH over warfarin in patientsundergoing THR at least for the short term (up to14 days), it also raises questions about theduration of prophylaxis in these patients.

Handoll HH, Farrar MJ, McBirnie J, et al.Heparin, low-molecular-weight heparin andphysical methods for preventing deep veinthrombosis and pulmonary embolism followingsurgery for hip fractures. Cochrane Database SystRev 2000;(4):CD000305.

Similar to patients who undergo TKR and THR,patients undergoing hip fracture surgery are atvery high risk for postoperative VTE. Thissystematic review evaluated the efficacy of UFH,LMWH, and physical methods of prophylaxis inpatients undergoing hip fracture surgery. A totalof 31 trials involving 2958 patients wereincluded. Compared with the control, UFH andLMWH resulted in a lower rate of DVT (26% vs42%, RR 0.60, 95% CI 0.5–0.71). Overallmortality was not significantly different betweenthe UFH-LMWH and control groups. Data wereinsufficient to determine if there was an efficacyor safety difference between UFH and LMWH.Mechanical foot pumping devices offered benefit

424


over control (7% vs 22%, RR 0.31, 95% CI0.19–0.51), but the studies evaluating them werelimited by methodologic flaws, thus these resultshave to be interpreted with caution. The influ-ence of various treatments on fatal pulmonaryembolism and mortality could not be determineddue to insufficient data. This analysis confirmsthat mechanical foot pumping devices, UFH, andLMWH are effective in the prophylaxis of DVTafter hip fracture surgery, but mechanicalmethods are limited by compliance. Whetherthere is an efficacy or safety benefit of LMWHover UFH in this patient population is stillunclear based on the results of this systematicreview. The LMWHs are generally preferred inpractice mainly because of convenience.

Mismetti P, Laporte S, Zufferey P, et al.Prevention of venous thromboembolism inorthopedic surgery with vitamin K antagonists: ameta-analysis. J Thromb Haemost 2004;2:1058–70.

This meta-analysis of randomized, controlledstudies was conducted to better assess the placein therapy of adjusted-dose vitamin K antagonistsamong the other available VTE thrombo-prophylactic strategies in patients undergoingmajor orthopedic surgery. The use of vitamin Kantagonists was found to be more effective thanplacebo or no treatment in reducing DVT (567patients, RR 0.56, 95% CI 0.37–0.84, p<0.01)and clinical pulmonary embolism (651 patients,RR 0.23, 95% CI 0.09–0.59, p<0.01) but wasassociated with a higher rate of woundhematoma (162 patients, RR 2.91, 95% CI1.09–7.75, p=0.03). Vitamin K antagonists werealso more effective than IPC in preventingproximal DVT (534 patients, RR 0.46, 95% CI0.25–0.82, p=0.009). The benefit:risk ratio ofvitamin K antagonists was comparable to that ofUFH and antiplatelet agents, but the number ofstudies comparing these options was low.Compared with LMWH, vitamin K antagonistswere less effective in preventing total DVT (9822patients, RR 1.51, 95% CI 1.27–1.79, p<0.001)and proximal DVT (6131 patients, RR 1.51, 95%CI 1.04–2.17, p=0.028). No significantdifferences in clinical pulmonary embolism,major bleeding, and wound hematoma werefound between LMWH and vitamin Kantagonists. The type of surgery (THR vs TKR)or the timing of administration of LMWH relativeto surgery (pre- vs postoperative) did not have asignificant effect on the results. As both vitaminK antagonists and LMWHs are common VTE

prophylactic options in patients undergoingmajor orthopedic surgery, the data from thisanalysis tips the balance in favor of LMWHs inthis population.

Menzin J, Colditz GA, Regan MM, et al. Cost-effectiveness of enoxaparin vs low-dose warfarinin the prevention of deep vein thrombosis aftertotal hip replacement surgery. Arch Intern Med1995;155:757–64.

The ACCP guidelines support both LMWH andwarfarin as appropriate prophylactic optionsagainst VTE in patients undergoing THR. Inaddition to efficacy and safety considerations,cost-effectiveness of available therapies is often afactor in agent selection. This article provides adecision-analysis model in a hypothetical cohortof 10,000 patients undergoing THR in a studythat compared a prophylactic strategy ofwarfarin, enoxaparin, and no prophylaxis. Foreach strategy, estimates were derived from data inthe literature on expected cases of DVT,pulmonary embolism, and deaths, as well as costsof VTE care. Compared with no prophylaxis,warfarin was estimated to reduce DVTs from1000 to 420/10,000 patients and deaths due toVTE from 250 to 110/10,000 patients. The costof care for DVT was estimated to be reduced from$530 to $330/patient with warfarin. Prophylaxiswith enoxaparin was estimated to further reducethe rate of DVT and related mortality (250 and70/10,000 patients, respectively), but increasecosts by $50/patient treated. Still, enoxaparinwould offer an overall cost benefit over warfarinby $12,000/death avoided. Although theacquisition cost of enoxaparin is higher than thatof warfarin, it appears to be cost-effective whenused for prophylaxis in patients undergoing THR,at least based on the estimates used in thisanalysis.

White RH, Romano PS, Zhou H, et al. Incidenceand time course of thromboembolic outcomesfollowing total hip or knee arthroplasty. ArchIntern Med 1998;158:1525–31.

This is one of the landmark studies describingthe frequency and time course of clinical VTEafter THR and TKR. Patients with a diagnosis ofDVT or pulmonary embolism within 3 months ofa THR or TKR surgery were identified from aState of California–linked hospital dischargedatabase. The cumulative frequency of DVT orpulmonary embolism within 3 months of surgerywas 2.8% after THR (556/19,586 cases) and 2.1%after TKR (508/24,059 cases, difference 0.7%,

425


95% CI 0.4–1.0%). These rates are lower thanthe 30–45% rate of VTE reported in studies thatused screening venography and indicate that inmost cases VTE resolves without causingsymptoms that patients would perceive assignificant enough to seek further care. Themajor finding of the study was that VTE wasdiagnosed after hospital discharge in 76% ofpatients who underwent THR and 47% of thosewho had TKR (difference 29%, 95% CI 23–34%,p<0.001), raising the question of needing toextend prophylaxis beyond hospital discharge.The median time to diagnosis of VTE was longerin the THR than the TKR group (17 vs 7 days,p<0.001). The type and duration of thrombo-prophylaxis used was identical in the THR andTKR groups, with 95% of cases receivingprophylaxis. Based on these findings, a shorterbut more aggressive course of prophylaxis shouldbe used in patients undergoing TKR, whereasthose undergoing THR need a longer course oftherapy. The findings of this study are significantin that they allow clinicians to further define theappropriate duration of prophylaxis in patientsundergoing THR or TKR surgery.

Eikelboom JW, Quinlan DJ, Douketis JD.Extended-duration prophylaxis against venousthromboembolism after total hip or kneereplacement: a meta-analysis of the randomisedtrials. Lancet 2001;358:9–15.

This meta-analysis evaluated the efficacy ofextended-duration out-of-hospital prophylaxis onsymptomatic VTE after THR and TKR surgery.Nine studies (3999 patients) were included:eight used LMWH and one used UFH. Thefrequency of symptomatic VTE was significantlyreduced in patients receiving extendedprophylaxis for 30–42 days versus those whoseprophylaxis was stopped at hospital discharge(1.3% vs 3.3%, OR 0.38, 95% CI 0.24–0.61). Thereduction in VTE was greater in the THR group(1.4% vs 4.3%, OR 0.33, 95% CI 0.19–0.56)compared with those who underwent TKR (1.0%vs 1.4%, OR 0.74, 95% CI 0.26–2.15). The rateof major bleeding was not affected, but there wasan increase in the rate of minor bleedingassociated with extended prophylaxis (3.7% vs2.5%, OR 1.56, 95% CI 1.08–2.26).

The results of this study support that byextending prophylaxis with LMWH or UFHbeyond hospital discharge in patients undergoingTHR and TKR surgery, the significant reductionin symptomatic VTE events is equivalent to a RRof 20 events/1000 patients treated. Consistent

with findings from previous studies, this meta-analysis also suggests that an extended durationof prophylaxis may be less effective in patientsundergoing TKR compared with thoseundergoing THR. Unlike previous studies thatraised questions about the clinical relevance ofasymptomatic DVT and its link to symptomaticevents, this analysis showed that a reduction insymptomatic thrombosis was also paralleled by asimilar decrease in asymptomatic events. Thesefindings help support the link betweenasymptomatic and symptomatic DVT and thatasymptomatic clots detected with venographycan be used as a surrogate marker for sympto-matic events.

Hull RD, Brant RF, Pineo GF, et al. Preoperativevs postoperative initiation of low-molecular-weight heparin prophylaxis against venousthromboembolism in patients undergoingelective hip replacement. Arch Intern Med1999;159:137–41.

Clinical practice differences exist betweenEuropean and North American countries withregard to timing of initiation of LMWH preoper-atively or postoperatively in patients undergoingmajor orthopedic surgery. Preoperative initiationof prophylaxis is based on the premise that DVTstarts during surgery and the aim is to optimizeanticoagulant efficacy, whereas postoperativeinitiation is usually performed 12–24 hours aftersurgery with the aim of minimizing the risk ofbleeding. This meta-analysis sought to analyzeefficacy and safety outcomes between preoperativeand postoperative initiation of LMWH (enoxaparin)prophylaxis in patients undergoing THR surgery.Included trials were selected by strict criteria,including double-blind design, DVT documentationby contrast-enhanced venography performedbefore or at the time of hospital discharge, andthe use of the enoxaparin started either before orafter surgery in doses previously shown to beeffective. The authors found that preoperativeinitiation of enoxaparin was more effective thanpostoperative initiation, with a DVT frequency of10% versus 15.3% (p=0.02). Major bleeding wasalso less frequent in patients in whom enoxaparinwas started preoperatively compared with thosein whom therapy was started postoperatively(0.9% vs 3.5%, p=0.01).

The findings of this meta-analysis support theview that optimal protection against DVTinduced by surgery is attained by preoperativeinitiation of prophylaxis. These findings wererather surprising, as typical clinical practice in

426


the United States was to begin thromboprophy-laxis with LMWH postoperatively in patientsundergoing major orthopedic surgery. Theresults of this meta-analysis were viewed ashypothesis generating and led to the evaluationof this same clinical question in subsequentrandomized clinical trials.

Hull RD, Pineo GF, Francis C, et al, for theNorth American Fragmin Trial Investigators.Low-molecular-weight heparin prophylaxis usingdalteparin in close proximity to surgery vswarfarin in hip arthroplasty patients: a double-blind, randomized comparison. Arch Intern Med2000;160:2199–207.

As clinical practice patterns differ with regardto the timing of initiation of LMWH prophylaxisin patients undergoing major orthopedic surgery,this study sought to elucidate the timing for startof prophylaxis that will maximize treatmentefficacy. A “just-in-time” concept was appliedthat would clarify the right balance betweenimmediate pre- or postoperative treatmentinitiation and bleeding. This was a double-blind,randomized trial that evaluated three differenttreatment approaches: dalteparin 2500 IU startedwithin 2 hours before surgery and a second dosegiven at least 4 hours after surgery, then 5000IU/day; dalteparin 2500 IU given at least 4 hoursafter surgery, then 5000 IU/day; or warfarinonce/day (goal INR 2–3) started on the eveningafter surgery. A total of 1472 patients undergoingTHR were enrolled. The primary end point wasDVT detected by contrast-enhanced venographyafter surgery (mean 5.7 days). The rates of DVTin the groups receiving preoperative or post-operative dalteparin or warfarin for all DVT were10.7%, 13.1%, and 24%, respectively (p<0.001for both pre- and postoperative dalteparin vswarfarin). The rates of proximal DVT were 0.8%,0.8%, and 3%, respectively (p=0.04 and p=0.03for pre- and postoperative dalteparin vswarfarin). The RRR ranged from 45–72% infavor of the dalteparin groups compared withwarfarin. The frequency of major bleeding washigher in the preoperative dalteparin groupcompared with warfarin (p=0.01).

This study demonstrates that dalteparin whenstarted in close proximity to surgery either beforeor after surgery results in better efficacyoutcomes than warfarin in patients undergoingTHR surgery. As both warfarin and LMWH arerecommended prophylactic options according tothe ACCP guidelines, this study demonstratedthat further efficacy benefit can be gained with

LMWH compared with warfarin if the timing ofLMWH therapy initiation is in close proximity tosurgery (mean ± SD 6.6 ± 2.4 hrs). In addition,this was the first major study to demonstrateimproved efficacy in the prevention of proximalDVT with an LMWH compared with warfarin.This finding was in contrast to results of studiesthat used 12–24-hour postoperative initiation ofLMWH and that failed to show a significantbenefit compared with oral anticoagulants. Theresults of this study confirmed the role of thejust-in-time or close-to-surgery initiation ofanticoagulant agents, and all major studiesevaluating new anticoagulant agents are nowfollowing this strategy to maximize anticoagulantefficacy.

Hull RD, Pineo GF, Stein PD, et al. Timing ofinitial administration of low-molecular-weightheparin prophylaxis against deep vein thrombosisin patients following elective hip arthroplasty: asystematic review. Arch Intern Med 2001;161:1952–60.

Despite the new knowledge gained regardingthe appropriate timing of LMWH after THR, thetranslation of the just-in-time or close-to-surgeryinitiation of prophylaxis to clinical practice isstill lagging behind in the United States. At thetime this report was published in 2001, standardpractice was still a delayed initiation of LMWH at12–24 hours after surgery. This systematicreview was conducted to assess the efficacy andsafety of LMWH started at different times inrelation to surgery compared with oral anticoag-ulation. Four trials met the predefined inclusioncriteria. The results showed that LMWH startedin close proximity to surgery resulted in anabsolute RR of 11–13%, and an RRR of 43–55%for DVT, compared with oral anticoagulation.Initiation of LMWH 12 hours before surgery or12–24 hours after surgery was not more effectivethan oral anticoagulation. When LMWH wasstarted at half the usual dose in close proximityto surgery, the rate of clinical or major bleedingwas not significantly affected (p=0.16). Theresults of this systematic review confirmed thatappropriate timing of LMWH in relation to majororthopedic surgery is key in maximizinganticoagulant efficacy. Close-to-surgery initiationof LMWH (4–6 hrs after surgery) is nowsupported by the ACCP guidelines. Delayinginitiation of LMWH to 12–24 hours after majororthopedic surgery results in suboptimal efficacywithout gaining a major safety advantage, andthis practice should be discouraged.

427


Bauer KA, Eriksson BI, Lassen MR, et al.Fondaparinux compared with enoxaparin for theprevention of venous thromboembolism afterelective major knee surgery. N Engl J Med2001;345:1305–10.

Because of a high residual VTE rate after TKR,the quest for new and more effective anticoagu-lants continues. Fondaparinux, an indirectsynthetic factor Xa inhibitor, was evaluated inthis study. A total of 1049 patients undergoingTKR surgery were randomly assigned in adouble-blind fashion to receive fondaparinux 2.5mg/day subcutaneously or enoxaparin 30 mgsubcutaneously twice/day. Both treatments werestarted postoperatively, with enoxaparin started12–24 hours after surgery and fondaparinuxstarted a mean ± SD of 6 ± 2 hours after surgery.The primary efficacy outcome was VTE (DVTdetected with venography, documentedsymptomatic DVT, or documented symptomaticpulmonary embolism) up to postoperative day11, and the primary safety outcome was majorbleeding. The frequency of VTE was significantlylower in the fondaparinux group (12.5%, 45/361)compared with the enoxaparin group (27.8%,101/363), with an RRR of 55.2% (95% CI36.2–70.2, p<0.001). The frequency ofsymptomatic VTE was not significantly differentbetween the two groups. Major bleeding wasmore common in the fondaparinux group(p=0.006). In this study, once-daily fondaparinuxthat was started in close proximity after surgerywas more effective than enoxaparin started 12–24hours after surgery, but at a cost of higher majorbleeding complications. Criticisms of this studyincluded the differences in timing of initiation ofthe two agents after surgery, with fondaparinuxbeing started closer to surgery.

Eriksson BI, Bauer KA, Lassen MR, et al.Fondaparinux compared with enoxaparin for theprevention of venous thromboembolism afterhip-fracture surgery. N Engl J Med 2001;345:1298–304.

This study is the second in a series of studiesthat evaluated the efficacy of fondaparinux inpatients undergoing major orthopedic surgery, inthis case hip fracture surgery. At the time of thisstudy, data with other antithrombotics in hipfracture surgery were limited and recommen-dations were based mainly on expert opinion.This was a multicenter, double-blind, randomizedstudy to compare the efficacy and safety offondaparinux with that of enoxaparin in patientsundergoing hip fracture surgery. A total of 1711

patients were randomly assigned to receivefondaparinux 2.5 mg/day subcutaneously startedpostoperatively or enoxaparin 40 mg/daysubcutaneously started preoperatively for at least5 days. The primary efficacy outcome was VTE(DVT detected with venography, documentedsymptomatic DVT, or documented symptomaticpulmonary embolism) up to postoperative day11, and the primary safety outcomes were majorbleeding and mortality from all causes. Theduration of follow-up was 6 weeks.

The frequency of VTE was significantly lowerin the fondaparinux group (8.3%, 52/626)compared with the enoxaparin group (19.1%,119/624), with an RRR of 56.4% (95% CI39.0–70.3%, p<0.001). The frequency ofsymptomatic VTE was not significantly differentbetween the two groups, nor was the frequencyof death or clinically relevant bleeding. Becauseof planned regional anesthesia, only 25.6% ofenoxaparin-treated patients received thepreoperative dose, indicating the difficulty ofgiving an LMWH dose before surgery in patientsundergoing hip fracture surgery. Fondaparinuxwas found to be more effective than enoxaparinin hip fracture surgery and without an increase inmajor bleeding. The results of this studycontributed to the ACCP guidelines recognizingand categorizing fondaparinux as the preferredagent in hip fracture surgery (grade 1A recom-mendation), as data with warfarin and LMWHare limited.

Turpie AGG, Bauer KA, Eriksson BI, et al.Postoperative fondaparinux vs postoperativeenoxaparin for prevention of venous thrombo-embolism after elective hip-replacement surgery:a randomised double-blind trial. Lancet2002;359:1721–6.

This is the third in a series of four studies thatevaluated the efficacy and safety of fondaparinuxin major orthopedic surgery, in this case THRsurgery. A total of 2275 patients were randomlyassigned in a double-blind fashion to receivefondaparinux 2.5 mg/day subcutaneously orenoxaparin 30 mg subcutaneously twice/day.Both treatments were started postoperatively,with enoxaparin started at 12–24 hours aftersurgery and fondaparinux started a mean ± SD of6 ± 2 hours after surgery. The primary efficacyoutcome was VTE (DVT detected withvenography, documented symptomatic DVT, ordocumented symptomatic pulmonary embolism)up to postoperative day 11, with the primarysafety outcome being a composite of major

428


bleeding and death.The rate of VTE was 6% in the fondaparinux

group (48/787) compared with 8% in theenoxaparin group (66/797) with an RRR of 26.3%(95% CI -10.8–52.8%, p=0.099). The frequencyof symptomatic VTE was lower in the enoxaparingroup compared with that in the fondaparinuxgroup (0.1% vs 1%, 95% CI -0.0–1.9%, p=0.006).The frequency of major bleeding or death was notsignificantly different between the two groups. Inthis study, once-daily fondaparinux started inclose proximity after surgery was not significantlymore effective than enoxaparin started 12–24hours after surgery. Although the reduction inrisk of VTE achieved by fondaparinux in thisstudy was not statistically significant, the resultsof this study are consistent with those of the otherthree studies in this series, which showed greaterefficacy of fondaparinux compared withenoxaparin in TKR, hip fracture, and THRsurgeries.

Turpie AGG, Bauer KA, Eriksson BI, et al.Fondaparinux vs enoxaparin for the preventionof venous thromboembolism in major orthopedicsurgery. Arch Intern Med 2002;162:1833–40.

This meta-analysis of four major randomized,double-blind trials in 7344 patients undergoingTKR, THR, or hip fracture surgery was conductedto evaluate whether the efficacy and safety offondaparinux 2.5 mg/day subcutaneously starting6 hours after surgery was more effective than andas safe as the approved enoxaparin regimens,started either before or after surgery, inpreventing VTE. The primary efficacy outcomewas VTE (DVT detected with venography,documented symptomatic DVT, or documentedsymptomatic pulmonary embolism) up topostoperative day 11, and the primary safetyoutcome was major bleeding.

The rate of VTE was 6.8% in the fondaparinuxgroup (182/2682) compared with 13.7% in theenoxaparin group (371/2703), with a commonodds reduction of 55.2% (95% CI 45.8–63.1%,p<0.001). This effect was consistent across alltypes of surgery and all subgroups. Theoccurrence of major bleeding was more frequentin the fondaparinux group (p=0.008), but that ofclinically relevant bleeding was not significantlydifferent. Fondaparinux 2.5 mg/day subcu-taneously started 6 hours after surgery was foundto be more effective than approved doses ofenoxaparin in patients undergoing majororthopedic surgery, with an overall risk reductionof greater than 50% but with an increase in major

bleeding. These findings have led to fondaparinuxbeing included in current national guidelinerecommendations as a grade 1A therapeuticoption along with LMWH and warfarin inpatients undergoing major orthopedic surgery.

Central Nervous System Procedures andInjuries

Agnelli G, Piovella F, Buoncristiani P, et al.Enoxaparin plus compression stockings comparedwith compression stockings alone in theprevention of venous thromboembolism afterelective neurosurgery. N Engl J Med 1998;339:80–5.

When this article was published, little wasknow about the most appropriate method ofprophylaxis in patients undergoing neurosurgery.Mechanical methods were often primarily usedbecause of perceived established efficacy as wellas the lack of bleeding risk. In fact, pharmacologicprophylaxis was often avoided for fear of thepotential for intracranial hemorrhage. This studyassessed the safety of enoxaparin when added tomechanical methods of VTE prophylaxis forpatients undergoing neurosurgery. The authorsconducted a multicenter, randomized, double-blind trial in which all patients received GCS. Inaddition to GCS, patients received eitherenoxaparin 40 mg/day for at least 7 days ormatching placebo injection. Enoxaparin wasstarted at least 24 hours after the end of thesurgical procedure. The primary end point wasthe occurrence of VTE by day 8. Venousthromboembolism was identified either throughsymptoms and confirmed with venography, orthrough bilateral venography to identifyasymptomatic clots.

Of the 307 patients assigned to treatmentgroups, 129 (84%) of the 154 patients receivingplacebo and 130 (85%) of the 153 patientsreceiving enoxaparin had venographic studiesadequate for analysis. The rate of VTE was lowerin patients receiving enoxaparin (17%) comparedwith those receiving placebo (32%). The RRRwith enoxaparin added to GCS was 0.52 (95% CI0.33–0.82, p=0.004). The rate of proximal DVTwas also lower in patients receiving enoxaparinversus placebo (5% vs 13%, p=0.04). Clinicalsymptomatic VTE events occurred in ninepatients receiving placebo and one patientreceiving enoxaparin. The frequency of majorbleeding did not differ significantly betweengroups, including the rates of intracranialhemorrhage. Death also occurred at a similar

429


rate during the 60-day study period. Ofimportance, this study established that not onlywas pharmacologic prophylaxis more effectivethan mechanical prophylaxis with GCS inpatients undergoing elective neurosurgery, butalso that enoxaparin 40 mg/day could be usedsafely in this patient population.

Nurmohamed MT, van Riel AM, Henkens CMA,et al. Low-molecular-weight heparin andcompression stockings in the prevention ofvenous thromboembolism in neurosurgery.Thromb Haemost 1996;75:233–8.

As in the study discussed above, the authorsnoted that although patients undergoingneurosurgery are at an elevated risk for VTE, theoptimal method of prophylaxis had yet to bedetermined. Whereas the preference of neuro-surgeons was to use mechanical methods ofprophylaxis as they did not evaluate the risk ofbleeding, a substantial risk for VTE remainedeven when these methods were used. Therefore,the authors sought to evaluate the effects ofadding an LMWH, in this case nadroparin, toGCS for VTE prophylaxis in patients undergoingneurosurgery. Patients older than 18 yearsundergoing craniotomy or spinal column surgerywere randomly assigned to receive bilateral GCSalone (control group), or the combination ofGCS and nadroparin 7500 antifactor Xa Uonce/day (nadroparin group). Compressionstockings were continued until hospitaldischarge, whereas nadroparin, which was started18–24 hours after surgery, was continued for 10days or until hospital discharge, whicheveroccurred first. Patients were assessed daily forsigns and symptoms of VTE and bleeding. Inaddition, patients underwent ultrasonographicsurveillance at 6, 8, and 10 days after surgery, aswell as bilateral venography for all patients whohad not developed VTE by day 10.

An initial 485 patients were enrolled, but only166 of the nadroparin group and 174 of thecontrol group had an adequate venogram; thus340 patients were randomized. A DVT occurredin 18.7% of patients in the nadroparin group and26.3% of the control group (p=0.047). ProximalDVT was also reduced in the LMWH groupcompared with the control group, although thisdid not reach statistical significance (p=0.065).Major bleeding occurred in six patients in thenadroparin group and in two patients in thecontrol group (p=0.087). The reductions in theoccurrence of VTE were maintained out to 8weeks. Of concern was that mortality was

increased in the 56-day study in patientsreceiving nadroparin (9.1% vs 4.1%), althoughnone of the deaths could be attributed to theoccurrence of bleeding. Despite the mortalityresults, this trial adds important information andconfirms the results from previous trials thatLMWH when used with GCS reduces theoccurrence of VTE and can be used safely inpatients undergoing neurosurgical procedures.

Iorio A, Agnelli G. Low-molecular-weight andunfractionated heparin for prevention of venousthromboembolism in neurosurgery: a meta-analysis. Arch Intern Med 2000;160:2327–32.

Although there had been some smallinvestigations regarding the use of low-dose UFHand LMWH for the prevention of VTE inneurosurgery, confidence in the overall safety andefficacy of these agents was still questioned byhealth care practitioners. The absolute magnitudeof benefit of pharmacologic compared withmechanical VTE prophylaxis in patientsundergoing neurosurgery was not clear, andconcerns remained regarding the risk of bleeding,especially intracranial hemorrhage. Hence, thismeta-analysis of controlled, randomized trials onthe efficacy and safety of heparins in theprophylaxis of VTE in patients undergoingneurosurgery was performed. Pertinent publica-tions that evaluated either low-dose UFH orLMWH were identified by searching MEDLINEand scanning meeting abstracts, as well as byreviewing references from available systematicreviews. This process identified four controlled,randomized studies, three of which involvedLMWH, that met criteria to be included in theanalysis. One hundred eighty-seven VTE events(22.6%) were recorded in 827 patients. Prophylaxiswith low-dose UFH or LMWH resulted in a 45%RRR in VTE events (OR 0.48, 95% CI 0.35–0.66,p<0.001). No fatal bleeding events were noted,but the use of heparins for prophylaxis did resultin a 71% RR increase of major bleeding (OR 1.72,95% CI 0.69–4.27, p=0.24). The authors notedthat the use of heparins for prophylaxis resultedin a number needed to treat of eight to preventone VTE event, and 16 to prevent one proximalDVT. The number needed to harm was 102.Overall, this meta-analysis confirmed the resultsof earlier smaller trials that the use of a heparin,primarily LMWH, reduces the occurrence of VTEin patients undergoing neurosurgery. Furthermore,the use of LMWH in this population does notappear to increase the risk of intracranialhemorrhage.

430


Norwood SH, McAuley CE, Berne JD, et al.Prospective evaluation of the safety of enoxaparinprophylaxis for venous thromboembolism inpatients with intracranial hemorrhagic injuries.Arch Surg 2002;137:696–702.

Before the publication of this article, littleinformation was available on whether LMWHcould be used safely to prevent VTE events inpatients experiencing intracranial hemorrhagesecondary to blunt-force trauma. In fact,previous investigations of VTE prophylaxis intrauma patients excluded those who hadexperienced intracranial hemorrhage. However,patients with trauma who also experienceintracranial hemorrhage are at a particularly highrisk for VTE, and the ability to use a convenientand effective LMWH regimen would beadvantageous. Therefore, the investigatorsconducted this study to evaluate the safety ofearly anticoagulation in patients with intracranialhemorrhagic injury. This was a prospective,single-center, observational study evaluating thesafety of administering enoxaparin 30 mgsubcutaneously twice/day, starting approximately24 hours after initial evaluation. Importantaspects of the protocol included the following:withholding enoxaparin for the first 72 hours inpatients with splenic injuries managed withoutsurgery, withholding enoxaparin for 24 hours inpatients who underwent craniotomy orcranioplasty, and use of IPC only before theinitiation of enoxaparin or during the 24 hourssurrounding craniotomy or cranioplasty whenenoxaparin was withheld. Enoxaparin wasadministered for the duration of hospitalizationor until a patient met the prespecified exclusioncriteria.

During an 18-month period at the study center,1428 patients were identified as experiencingblunt-force trauma, with 177 patients having adocumented intracranial hemorrhagic injury oninitial computed tomographic (CT) scan. Onehundred fifty of the 177 patients receivedenoxaparin beginning approximately 24 hoursafter hospital admission until discharge. Fourhundred sixty-eight CT scans were obtained inthe study population, with 34 showingprogression of intracranial hemorrhagic injury.Twenty-eight of those occurred before initiationof enoxaparin with no further worsening of theinjury occurring after initiation of enoxaparin.Six patients showed progression of intracranialhemorrhagic injury after enoxaparin initiation,with the LMWH subsequently being discon-tinued. All six patients survived hospitalization.

No significant differences were noted withrespect to bleeding complications betweenpatients who underwent surgery (2/24 [8%]) andthose who did not (4/126 [3%]). A DVT wasidentified in two patients (2%). The authorsnoted that the rate of bleeding complications wassimilar to previously reported findings on therate of bleeding in patients undergoing electiveneurosurgery with low-dose UFH as the methodof VTE prevention.

This trial added important informationregarding whether the use of enoxaparin is safein patients with intracranial hemorrhagic injurysecondary to blunt-force trauma. Many patientsundergo the placement of a Greenfield filter toprevent fatal pulmonary embolism when traumais complicated by an intracranial hemorrhagicinjury because of the fear of worsening thehemorrhage. The LMWHs are considered themost effective means of VTE prophylaxis fortrauma patients, and the ability to use theseagents even in patient with intracranialhemorrhagic injury would likely decrease the riskof both VTE and complications from Greenfieldfilter placement.

Spinal Cord Injury ThromboprophylaxisInvestigators. Prevention of venous thrombo-embolism in the acute treatment phase afterspinal cord injury: a randomized, multicentertrial comparing low-dose heparin plusintermittent pneumatic compression withenoxaparin. J Trauma 2003;54:1116–24.

These authors, noting that patients with acuteSCI have one of the highest rates of VTE,conducted a prospective, multicenter studycomparing low-dose UFH 5000 U subcu-taneously every 8 hours plus IPC, withenoxaparin 30 mg subcutaneously twice/day withno IPC during the early phase (first 2 wks) ofSCI. Previous investigations had established thatmechanical methods were inadequate in thispopulation and that although combiningmechanical methods with low-dose UFHimproved efficacy, the combination could beconsidered impractical to implement. Based onprevious investigations in orthopedic surgery inwhich LMWH (enoxaparin) was demonstrated tobe superior to low-dose UFH, the authorsconducted this study to assess whether LMWHcould provide similar or better efficacy than thecombination of low-dose UFH plus IPC.

Four hundred seventy-six patients wererandomly assigned to receive at least one dose ofstudy drug. From this population, only 107

431


patients were evaluated for efficacy, as theycompleted the full course of prophylaxis andunderwent adequate proximal and distalvenography, had proximal DVT confirmed byduplex ultrasonography, or had clinical evidenceof pulmonary embolism. In this evaluationcohort, the frequency of VTE was 63.3% in theUFH plus IPC group versus 65.5% in theenoxaparin group (p=0.81). Rates for proximalpulmonary embolism in 181 patients whounderwent at least proximal venography werenot statistically significantly different betweengroups (UFH plus IPC 6.5% vs enoxaparin9.0%). The frequency of pulmonary embolismwas 18.4% in the UFH plus IPC group comparedwith 5.2% in the enoxaparin group (p=0.03).

All 476 randomized patients were evaluated forsafety, and the rate of major bleeding was 5.3%with UFH plus IPC compared with 2.6% in theenoxaparin group (p=0.14). Minor bleeding wasreported in 17.9% of the UFH plus IPC–treatedpatients and in 14.8% of the patients receivingenoxaparin. The authors concluded that in theearly treatment phase after SCI, safety andefficacy were generally similar with UFH plusIPC and enoxaparin.

Other relevant analyses from the study includethe compliance rate for mechanical devices, aswell as risk factors for the development of VTE inthis population. Overall, this study wassignificant in establishing a more feasibleregimen of LMWH alone as being as effective andsafe as UFH plus IPC in patients in the earlyphase of SCI. Of importance, although it isreasonable to think that the combination ofLMWH and IPC would produce additive effects,the hypothesis would need to be tested in anappropriately designed trial.

Spinal Cord Injury ThromboprophylaxisInvestigators. Prevention of venous thrombo-embolism in the rehabilitation phase after spinalcord injury: prophylaxis with low-dose heparinor enoxaparin. J Trauma 2003;54:1111–15.

Little information was available on the risk ofVTE in patients during the rehabilitation phaseafter SCI. To address this issue, a prospective,multicenter study compared low-dose UFH withenoxaparin for prophylaxis against VTE duringthe rehabilitation phase after SCI (wks 3–8 afterinjury). This study was a continuation of theearly-phase investigation comparing low-doseUFH 5000 U every 8 hours plus IPC versusenoxaparin. Patients were eligible for therehabilitation-phase portion if they had

completed the early-phase study withoutobjective evidence of VTE. Patients in therehabilitation phase received the sameprophylaxis regimen to which they wererandomly assigned in the early phase, except thatIPC was discontinued in patients assigned toreceive low-dose UFH, and the dosage ofenoxaparin was reduced from 30 mg twice/day to40 mg once/day. Patients who had no clinicalsymptoms of VTE during the rehabilitation phasesubsequently underwent repeat bilateral lowerextremity duplex ultrasonography at the end ofthe study period.

A total of 172 patients completed the early-treatment phase period. Of these, 119 patientscompleted the rehabilitation phase and hadadequate imaging. New VTE was demonstratedin 13 of 60 patients treated with low-dose UFHversus 5 of 59 enoxaparin-treated patients(21.7% vs 8.5%, p=0.052). Patients who receivedat least one dose of study drug were included inthe safety analysis. No major bleeding eventsoccurred in the enoxaparin group and oneoccurred in the low-dose UFH group. Only onepatient from each group was discontinued fromthe study because of bleeding. Significantobservations from this study include theconfirmation that patients in the rehabilitationphase of SCI remain at high risk for VTE events,as well as that enoxaparin 40 mg/day appears tobe more effective that low-dose UFH 5000 Uevery 8 hours at preventing VTE.

Hebbeler SL, Marciniak CM, Crandall S, ChenD, Nussbaum S, Mendelewski S. Daily vs twicedaily enoxaparin in the prevention of venousthromboembolic disorders during rehabilitationfollowing acute spinal cord injury. J Spinal CordMed 2004;27:236–40.

Previous investigations of the value of LMWHfor VTE prophylaxis in patients with acute SCIused enoxaparin 30 mg subcutaneouslytwice/day. This study was undertaken to evaluatethe safety and efficacy of enoxaparin 40 mgsubcutaneously once/day compared with the 30-mg twice-daily regimen. Patients receiving eitherenoxaparin prophylactic regimen were identifiedthrough a retrospective chart review over a 2-yearperiod at a single rehabilitation facility. A total of129 patients with acute SCI received eitherenoxaparin 40 mg once/day (80 patients) orenoxaparin 30 mg twice/day (49 patients) as theprimary mode of VTE prophylaxis. Enoxaparinwas started an average of 23.6 days and 20.6 daysafter SCI for the once- and twice-daily regimens,

432


respectively. The mean length of prophylaxiswith enoxaparin was 42.7 days in the twice-dailyregimen group and 39.5 days in the once-dailyregimen group. Despite a lack of randomization,the patient groups were similar with respect togeneral demographics, the level of SCI, and othercomorbidities.

Over the duration of prophylaxis, one patientin each group experienced a DVT. Onepulmonary embolism occurred during theevaluation period in the twice-daily regimengroup. Bleeding complications were numericallyhigher in the once-daily regimen group (20 vs 5complications), with the difference not achievingstatistical significance. Limitations of the studyinclude the retrospective nature along with thelack of randomization. Although the two groupsof patients were well matched with majorvariables that may serve as confounders, not allconfounders can be identified and eliminated instudies of this type. Nevertheless, the resultswould suggest that enoxaparin 40 mg once/day isas safe and efficacious as a regimen of 30 mgtwice/day in patients with SCI. However, giventhe number of patients assessed in the study, itmay have been difficult to detect smalldifferences between the regimens. Advantageswith using the once-daily regimen includedecreased cost and ease of use in a patientpopulation that may receive prophylaxis for anextended duration. Although the study mayprovide enough evidence for some to useenoxaparin 40 mg once/day for prophylaxis inpatients with SCI, confirmation of these results ina larger randomized trial would be optimal beforeuniversal implementation.

Slavik RS, Chan E, Gorman SK, et al.Dalteparin versus enoxaparin for venousthromboembolism prophylaxis in acute spinalcord injury and major orthopedic traumapatients: DETECT trial. J Trauma 2007;62:1075–81.

This was a retrospective case-control analysisthat evaluated the efficacy and safety ofdalteparin 5000 IU subcutaneously once/daycompared with enoxaparin 30 mgsubcutaneously twice/day in patients with eitheracute SCI or major lower extremity orthopedictrauma. In discussing the rationale for the study,the authors succinctly reviewed literature thatdescribed a very elevated risk of proximal DVTand pulmonary embolism in this patientpopulation, and that stated that LMWH issuperior to other forms of pharmacologic and

mechanical prophylaxis. Most of the publishedliterature discussed enoxaparin. Citing somesmall published work with dalteparin in thesepatient populations, the authors implemented aninstitutional interchange of dalteparin forenoxaparin as the primary mode of prophylaxis.Presumably this switch was made on the basis ofequivalent efficacy with reduced cost. The studyis an evaluation of the implemented interchange.Despite being a nonrandomized comparison, thegroups were well matched with respect to patientcharacteristics, as well as in hospital events orprocedures.

In their analysis, the authors noted thatpatients had a risk for proximal DVT pluspulmonary embolism of 1.6% with enoxaparinand 9.7% with dalteparin (p=0.103). Theseresults failed to meet the prespecified criteria fornoninferiority with the CI crossing the upperboundary of an absolute risk increase of 5%.Bleeding rates were approximately the same,although an interesting finding is that theenoxaparin group had more missed doses,suggesting that once-daily regimens may beeasier to comply with.

There are many limitations to this analysisgiven the retrospective nature of data collection,as well as the absence of statistical methods suchas logistic regression to account for confoundersbetween the two groups. However, given theresults of this analysis when assessed within thecontext of other studies using enoxaparin in thispatient population, it would seem prudent tooperate with the assumption that in acute SCI,enoxaparin 30 mg subcutaneously twice/dayshould be the preferred strategy since it is wellsupported by the scientific literature. Given thedesign of the comparison, it would be misleadingto state that enoxaparin 30 mg twice/day issuperior to dalteparin 5000 IU once/day.However, given the lack of evidence fordalteparin in VTE prevention for patients withSCI, justifying utilization in this group may bedifficult without further investigation.

Trauma

Upchurch GR, Demling RH, Davies J, Gates JD,Knox JB. Efficacy of subcutaneous heparin inprevention of venous thromboembolic events intrauma patients. Am Surg 1995;61:749–55.

The investigators of this meta-analysisincluded five trials along with 66 patients fromtheir own data that evaluated the efficacy ofsubcutaneous UFH compared with no prophylaxis

433


for prevention of VTE in patients with trauma. Atotal of 1102 patients in this meta-analysis hadmajor trauma (injury severity score > 10) andunderwent prospective duplex ultrasonographyfor DVT. The frequency of VTE in patientsreceiving UFH was 9.9% compared with 7.2% inpatients receiving no prophylaxis (p=0.771). Therate of pulmonary embolism was 1.3% in patientsreceiving UFH compared with 2.2% in patientsreceiving no prophylaxis (p=0.157). The rate ofbleeding was not reported. Although this meta-analysis includes trials conducted about 15 yearsago, the results suggest that subcutaneous UFH isnot effective at reducing the risk of VTE inpatients with major trauma. The investigatorsprovide a good review of the patients andmethods of each of the trials included in themeta-analysis. The dosing of subcutaneous UFHused in the individual trials, however, is notmentioned.

Geerts WH, Jay RM, Code KI, et al. Acomparison of low-dose heparin with low-molecular-weight heparin as prophylaxis againstvenous thromboembolism after major trauma. NEngl J Med 1996;335:701–7.

Three hundred forty-four trauma patients inthis single-center trial were randomly assigned ina double-blinded fashion to receive UFH 5000 Utwice/day or enoxaparin 30 mg twice/day for upto 14 days. Patients in the trial had an injuryseverity score above 9 and received their firstdose of anticoagulant within 36 hours ofadmission. No additional mechanical prophylaxiswas permitted. The primary end point of thetrial was the occurrence of venographic DVT, orsymptomatic or fatal pulmonary embolism. Theuse of enoxaparin demonstrated a significant30% relative reduction in the rate of any DVT(31% vs 44%, p=0.014) and a significant 58%relative reduction in the rate of proximal DVT(6% vs 15%, p=0.012). Major bleeding in thesetrauma patients occurred rarely and was notsignificantly different between the groups (fiveevents with enoxaparin vs one event with UFH,p=0.12).

This trial demonstrated improved efficacy ofenoxaparin 30 mg twice/day compared with UFH5000 U twice/day, without an increase inbleeding. Therefore, based on these data, UFHtwice/day should not be used in trauma patients,and enoxaparin is preferred. It should be notedthat the dosage of enoxaparin is not the typical40 mg once/day used for VTE prophylaxis inother patient populations. One of the main

issues with using pharmacologic VTEprophylaxis in trauma patients is when to startprophylaxis in relation to the risk of bleeding andthe extent of the trauma. Patients in this trialhad a mean injury severity score of 23, 85%required surgery, and 38% required bloodtransfusion within the first 24 hours. In thatsetting, pharmacologic prophylaxis was able tobe started within 36 hours of injury, with asignificant reduction of DVT and few majorbleeding events when enoxaparin 30 mgtwice/day was used.

Cothren CC, Smith WR, Moore EE, Morgan SJUtility of once-daily dose of low-molecular-weight heparin to prevent venous thrombo-embolism in multisystem trauma patients. WorldJ Surg 2007;31:98–104.

In an attempt to simplify VTE prophylaxis intrauma patients with a once-daily regimen,investigators of this observational trial evaluated743 trauma patients receiving dalteparin 5000 IUonce/day. The patients had a mean injuryseverity score of 19.5 and length of stay of 14days; 174 patients had brain injury. On average,patients received 9.2 doses of dalteparin duringtheir hospital stay, but they did not receive theirfirst dose until 3.3 days after injury. The primaryend point evaluated was the frequency ofultrasonography-detected DVT. The initialultrasonographic examination was performedwithin 48 hours of injury, before prophylaxis wasstarted. Patients with an initial negativeultrasound scan were then included in the trialand reevaluated between days 7 and 10. Duringthe 2 years of study enrollment, 3.9% of patientsdeveloped DVT and 0.8% developed pulmonaryembolism with no cases of fatal pulmonaryembolism. Major bleeding occurred in 2.7% ofpatients, and 3% required transfusion.

The obvious limitations of this trial are theobservational design with no comparator group.Although the authors attempt to state that theirresults are similar to or even better than theresults from the single-center trial discussedabove, this conclusion is inappropriate. Thesingle-center trial used a randomized comparisontrial design and documented venographic DVT.The use of dalteparin once/day may be a usefuloption for VTE prophylaxis in trauma patients,but a higher quality trial will be needed toconfirm these initial results. The main benefit ofthis trial to the trauma literature is the findingthat pharmacologic VTE prophylaxis does notneed to be delayed while waiting for most

434


surgical procedures and does not need to be heldbefore most surgical procedures. Patients in thisobservational trial had on average five surgicalprocedures during their hospital stay, and LMWHtypically was not withheld.

Ginzburg E, Cohn SM, Lopez J, Jackowski J,Brown M, Hameed SM. Randomized clinical trialof intermittent pneumatic compression and low-molecular-weight heparin in trauma. Br J Surg2003;90:1338–44.

In this trial, 442 trauma patients were prospec-tively randomized to subcutaneous enoxaparin30 mg every 12 hours or IPC. Enoxaparin wasstarted within 24 hours of trauma, but the timingof the initiation of IPC was not provided. Theprimary end point was proximal DVT detectedwith duplex ultrasonography and objectivelyconfirmed symptomatic pulmonary embolism.The frequency of the primary end point was 3%with IPC and 0.5% with enoxaparin (p=0.122).Major bleeding was less than 2% in both groups.

Although it may seem that IPC is a reasonablealternative to LMWH for VTE prophylaxis intrauma patients, there are several issues to beconsidered. First, trauma patients in this studywere considered to be at moderate risk since theirinjury severity score was less than 19 (meanscore 17), which is less than that in most of theother trauma VTE prophylaxis trials. It is alsounlikely that the trial was powered to find adifference in ultrasonography-detected proximalDVT in moderate-risk patients. Therefore, the83% relative reduction in VTE in this trial mayrequire further evaluation. The fact that bleedingwas similar between the groups demonstrates thesafety of LMWH in trauma patients.

Nathens AB, McMurray MK, Cuschieri J, et al.The practice of venous thromboembolismprophylaxis in the major trauma patient. JTrauma 2007;62:557–63.

The investigators of this cohort trial attemptedto answer the important question about timing ofinitiation of VTE prophylaxis in patients withtrauma. Three hundred fifteen patients experi-enced hemorrhagic shock after their trauma and,therefore, were at high risk for both VTE andbleeding. Pharmacologic VTE prophylaxis wasstarted within 48 hours in 25% of patients andafter 7 days in another 25%. Early prophylaxiswas defined as starting 4 days or less after trauma(174 patients), and late prophylaxis was definedas starting more than 4 days after trauma (137patients). Four patients who experienced a DVT

within the first 48 hours after admission wereexcluded from the analysis. Early prophylaxiswas associated with a 5% rate of VTE comparedwith 15% for late prophylaxis (RR 3.0, 95% CI1.4–6.5). Early prophylaxis was actuallyassociated with significantly less bleeding andneed for blood transfusion. This may be becausethe trial was observational and patients whoreceived “late” prophylaxis were treated in thatmanner due to their increased risk of bleeding.Although the design of this trial is not optimal, itdoes provide some insight into the issue ofstarting pharmacologic prophylaxis as soon aspossible to provide maximum protection againstVTE.

Velmahos GC, Kern J, Chan LS, Oder D, MurrayJA, Shekelle P. Prevention of venous thrombo-embolism after injury: an evidence-based report.I. Analysis of risk factors and evaluation of therole of vena caval filters. J Trauma 2000;49:132–9.

Both parts I and II of this meta-analysis wereconducted by a panel of 17 experts on VTE whorepresented the academic, private, and managedcare sectors. The investigators set out to answerfour questions related to prevention of VTE.What is the role of different chemical ormechanical methods in preventing VTE? Whatare the factors placing trauma patients at highrisk for VTE? Which is the optimal method toscreen for DVT? What is the role of vena cavafilters in preventing pulmonary embolism? Thisarticle, which is part I, addressed the firstquestion. Three of the investigators screened2437 titles, which led to the screening of 225 fullarticles. Of these articles, 73 randomized,controlled trials were accepted for the meta-analysis.

The rate of DVT and pulmonary embolismvaried widely among the trials, with the pooledanalysis providing a DVT rate of 11.8% and apulmonary embolism rate of 1.5%. The dramaticdifference in screening methods and reporting inrandomized controlled trials made it difficult tohave confidence in the accuracy of these findings.Overall, the results of the meta-analysisdemonstrated no evidence of efficacy for the useof subcutaneous UFH (OR 0.965, 95% CI0.353–2.636) or mechanical methods (OR 0.769,95% CI 0.265–2.236) in the prevention of VTE intrauma patients. The lack of benefit ofsubcutaneous UFH and mechanical methods wasstill evident when trials that were not randomizedand controlled were added to the analysis of therandomized controlled trials. When evaluating

435


trials of LMWH, no significant difference wasnoted in the frequency of pulmonary embolismwith LMWH compared with subcutaneous UFH,but the authors mention that the CIs were verywide, and a significant difference cannot beexcluded. The use of LMWH was associated witha significant reduction in DVT when comparedwith subcutaneous UFH, as well as whencompared with mechanical prophylaxis, althoughthe data were not presented. The pooled rate ofbleeding was 3.6% with subcutaneous UFH and3.1% with LMWH. The pooled rate ofthrombocytopenia was 1.9% with subcutaneousUFH and 0.4% with LMWH. The investigatorsprovide an insightful discussion of the poorquality of randomized controlled trials thatevaluated different VTE prophylaxis methods intrauma patients. This meta-analysis representsthe best summary of VTE prophylaxis methods intrauma patients in the literature and is frequentlyreferenced in the ACCP guidelines.

Velmahos GC, Kern J, Chan LS, Oder D, MurrayJA, Shekelle P. Prevention of venous thrombo-embolism after injury: an evidence-based report.II. Analysis of risk factors and evaluation of therole of vena caval filters. J Trauma 2000;49:140–4.

This article is part II of the meta-analysis andaddresses two of the remaining questions posedby the investigators. The same methodsdescribed in part I were used in part II. The onlysignificant risk factors for DVT in traumapatients that were identified were spinal fracturesand SCI, which increased the risk by 2- and 3-fold, respectively. Patients with DVT were anaverage of 9 years older than patients withoutDVT, but no specific age cutoff point could bedetermined based on the quality of the trialsavailable. Patients with prophylactic vena cavafilters had a rate of pulmonary embolism of 0.2%,whereas the comparator patients without filtersin these trials had a rate of pulmonary embolismof 1.5%, and historical controls had a rate ofpulmonary embolism of 5.8%. Because of thepoor quality of the trials, no statistics wereconducted on these findings. These articlesprovide a broad overview of VTE prophylaxisdata in trauma patients and should probably bein the files of clinicians who care for traumapatients in their hospitals.

Intensive Care Setting

Geerts W, Selby R. Prevention of venousthromboembolism in the ICU. Chest 2003;124:S357–63.

This review, by the lead writer of the ACCPsection on prevention of VTE, summarizes therisk factors and prevalence of VTE in patients inthe intensive care unit (ICU). It builds on anearlier systematic review and examines the fewprophylaxis trials using routine, objectivescreening specifically in patients in the ICU, aswell as providing the reader with a simple,practical, individualized approach to prophylaxisin the ICU based on bleeding and thrombosisrisk. The review addresses the relevant literaturebut highlights the lack of robust data specificallyderived from the ICU population. Only fourprospective studies were identified that usedobjective testing and reported VTE event rates inintensive care patients not receiving prophylaxis.The overall DVT rate in these four untreatedcohorts ranged from 13–31%. One study was asmall, prospective, cohort study that usedfibrinogen leg scanning for diagnostic testing,whereas the other three were control arms of theonly randomized, comparative trials in this arena.In these three studies, one also used fibrinogenleg scanning, and the largest trial was onlyavailable in abstract form at the time of thiswriting. The RRRs with heparin 5000 Usubcutaneously twice/day in two of the studieswere 55% and 65%, respectively. Unfortunately,neither proximal DVT nor bleeding event rateswere reported in either study. The trial that usedLMWH showed a 45% RRR but used venographyand provided more details. A schematic withspecific pharmacologic or mechanical prophylaxisrecommendations based on bleeding andthrombosis risk is presented, although details tostratify the bleeding component are not provided.The authors advocate that this assessment occuron admission to the ICU and daily thereafter, andthat compliance is enhanced with use of pre-printed orders as well as the active involvementof a pharmacist on daily ICU rounds.

Haas CE, Nelsen JL, Raghavendran K, et al.Pharmacokinetics and pharmacodynamics ofenoxaparin in multiple trauma patients. J Trauma2005;59:1336–44.

Previous studies have shown efficacy ofstandard doses of LMWH in trauma patients andhave suggested that LMWHs are superior to othermethods of VTE prophylaxis in this population.However, it is not clear that standard LMWHdoses are optimal in patients in the ICU. Thiswell-written, prospective, observational study oftwo cohorts of trauma patients clearly illustratesthat plasma antifactor Xa concentrations are

436


highly variable after administration of enoxaparin30 mg subcutaneously every 12 hours andfrequently undetectable in one subset of patients.Strict inclusion and exclusion criteria were used,and the 21 patients evaluated were divided intoedema or no edema groups based on whetherthey had gained at least 10 kg since admissionand had peripheral edema at physicalexamination. No patients were receivingvasopressors at the time of serum antifactor Xasampling, which took place at multiple timepoints after at least four enoxaparin doses hadbeen administered to ensure steady state. Thecohort with edema, in addition to having a 12-hour area under the concentration-time curveless than half that of the no edema cohort(p=0.01), also had significantly less antithrombin(antifactor IIa) activity (82% vs 66%, p<0.01).This raises the possibility that the reportedantifactor Xa levels may overestimate the true invivo response to LMWH. What remainsunanswered is whether clinical outcomes in thisarena, or least in those intensive care patientswith significant edema, would be enhanced byinitial weight-based LMWH dosing and/oradjusting LMWH doses based on target antifactorXa levels. As more comparative data becomeavailable, this study will also serve to clarify ifsome of the observed variability is a property ofthe individual drugs studied or more a byproductof this patient population.

Khouli H, Shapiro J, Pham VP, et al. Efficacy ofdeep venous thrombosis prophylaxis in themedical intensive care unit. J Intensive Care Med2006;21:352–8.

This was a prospective cohort study of 141consecutive patients admitted to a medical ICUwho received DVT prophylaxis according to arisk-stratified protocol and had bilateral lowerextremity ultrasonographic examinationsperformed within 48 hours of admission. Anultrasonographic examination was alsoperformed weekly thereafter, or upon dischargeor clinical suspicion of an event. The 801patients were screened according to presetinclusion and exclusion criteria to arrive at thefinal cohort. Most (528 patients) were excludedbecause the anticipated length of medical ICUstay was not greater than 48 hours. The protocolused in this center evaluated for the presence orabsence of 10 predetermined risk factors. If 1–3risk factors were present, the patient wasconsidered at moderate risk and was givensubcutaneous UFH, or IPC if UFH was

contraindicated. Four or more risk factorsresulted in UFH and IPC being used together. AnLMWH was not used. The high rate of DVT(9.9%), despite prophylaxis, is fairly consistentwith treatment arms of the few previouscomparative trials of patients in an ICU. Of note,10 of the 14 events were proximal DVT, butoverall only two of the events were symptomatic.Since 38% of the cohort had contraindications toUFH by their criteria, prophylaxis was fairlyevenly distributed between IPC (38%), UFH(32%), and UFH plus IPC (30%). Thedistribution of DVT did not differ by the type ofprophylaxis used. A post hoc analysis of fiveprimary risk factors showed the odds of having aDVT were significantly increased if at least tworisk factors were present. Overall this studyadvances our understanding regardingprophylaxis in the intensive care patient. Weawait the results of the ongoing Prophylaxis ofThromboembolism in Critical Care Trial(PROTECT), which compares daily LMWH withtwice-daily UFH, to further clarify specificprophylaxis recommendations.

Limpus A, Chaboyer W, McDonald E, LukmanT. Mechanical thromboprophylaxis in critically illpatients: a systematic review and meta-analysis.Am J Crit Care 2006;15:402–12.

Five randomized trials, 13 observationalstudies, and three surveys were evaluated to formthe basis for the rather general recommendationsby the ACCP that for critical care patients withhigh bleeding risk, mechanical prophylaxisshould be used until the bleeding risk decreases.Thereafter, pharmacologic prophylaxis can besubstituted or added to the mechanical method(GCS and/or IPC). Combining data from the fiverandomized controlled trials was limited becauseof the heterogeneity of the patient populationsand study methodologies. All patient populationtypes (trauma vs surgical vs medical) wereconsidered but had to be specified for inclusion.Four of the studies involved trauma patients, andof those, only two compared mechanical withpharmacologic (enoxaparin at different doses)prophylaxis. The aggregate rate of DVT detectedwith ultrasonographic examination on admissionand weekly while in the ICU appeared to favorLMWH over IPC (1.4% vs 3.5%) but failed toachieve significance given that only 562 patientswere studied.

The authors compiled a complete summarytable of observational studies encompassing over3000 patients, but conclusions were limited.

437


Three studies evaluated some combination ofmechanical compared with pharmacologicprophylaxis and failed to detect a significantdifference between the two, but their samplesizes and power to detect a true difference werelimited. Survey data highlighted the frequent useof GCS and underscored that practices appearedto vary throughout different worldwidegeographic regions. This review is compre-hensive but underscores the need for furthercomparative risk-to-benefit trials involvingmechanical prophylaxis, especially in thispopulation often with high clot and highbleeding risks.

Medical Illnesses

Gårdlund B, for the Heparin Prophylaxis StudyGroup. Randomised, controlled trial of low-doseheparin for prevention of fatal pulmonaryembolism in patients with infectious disease.Lancet 1996;347:1357–61.

This study represents the largest trial everconducted in medically ill patients. It included11,693 patients with infectious diseases whowere randomly assigned in an unblinded fashionto receive UFH 5000 U every 12 hours untilhospital discharge (maximum 21 days) orcontrol. The primary end point of the trial wasautopsy-verified pulmonary embolism. Nosignificant difference was noted between thepatients receiving UFH twice/day compared withthe control group in overall mortality (5.3% vs5.6%, p=0.39) or autopsy-detected thrombo-embolic complications (49% vs 49.2%, p=NS).Also, no significant difference was noted inpulmonary embolism or DVT at autopsy.Although the end point of this trial was unique,the trial was adequately powered to find adifference if one existed. Unfractionated heparin5000 U every 12 hours is a commonly usedregimen for prevention of VTE. The results ofthis large trial suggest that there is no benefit tothe use of this regimen in the prevention of VTEin these medically ill patients.

Belch JJ, Lowe GDO, Ward AG, Forbes CD,Prentice CRM. Prevention of deep veinthrombosis in medical patients by low-doseheparin. Scott Med J 1981;26:115–17.

The 100 medically ill patients in this trial hadheart failure or chest infection and wererandomly assigned to receive UFH 5000 U every8 hours or control during their hospital stay. Theprimary end point was the occurrence of DVT

detected by fibrinogen scanning. The first scanwas obtained within 24 hours of admission, andscanning was repeated every second day for 14days or until hospital discharge. The rate of DVTwas significantly reduced from 26% in thecontrol group to 4% in UFH-treated patients(p<0.01). Hematoma occurred in 20% of patientsreceiving UFH, but bleeding was not significantlydifferent between groups. Although questionshave been raised about the efficacy of UFH 5000U twice/day for VTE prevention in medically illpatients, this trial demonstrated the efficacy ofUFH 5000 U 3 times/day.

King CS, Holley AB, Jackson JL, Shorr AF,Moores LK. Twice vs three times daily heparindosing for thromboembolism prophylaxis in thegeneral medical population: a meta-analysis.Chest 2007;131:507–16.

Although a twice-daily UFH regimen comparedwith a thrice-daily UFH regimen for VTEprophylaxis has never been evaluated in a head-to-head comparison trial, both regimens havebeen compared with another active therapy,placebo, or control. The investigators of thismeta-analysis included 12 trials (7978 patients)that compared UFH twice/day (6314 patients) or3 times/day (1664 patients) with placebo orcontrol for VTE prevention to determine if adifference exists between these two UFHregimens. The rate of any VTE was 5.4% withUFH twice/day and 3.5% with UFH 3 times/day(p=0.87). There was a trend toward lower ratesof pulmonary embolism (0.5% vs 1.5%, p=0.09)and lower proximal DVT and pulmonaryembolism (0.9% vs 2.3%, p=0.05) with the use ofUFH 3 times/daily versus twice/day. The rate ofmajor bleeding was significantly higher inpatients receiving UFH 3 times/day comparedwith twice/day (0.96% vs 0.35%, p<0.001), butthe overall risk was very low.

Although clinicians may want to use this meta-analysis as a justification to continue using UFHtwice/day, it should be noted that the numericdifferences in efficacy from the meta-analysis maybe clinically relevant, whereas the difference inmajor bleeding may not be. Also, there are somelimitations to this meta-analysis. First, themethod of detection of VTE varied greatly amongthese trials. The difference in the ability to detect(or prevent) VTE events by using autopsy,venography, ultrasonography, or symptomsmakes it difficult to compare these trials. Trialsincluded in this meta-analysis included not onlymedically ill patients but also critical care

438


patients, which alters their VTE and bleedingrisks compared with typical medically ill patients.Although the information provided by this meta-analysis is interesting, a standard head-to-headcomparison trial is needed to completely answerthis question.

Samama MM, Cohen AT, Darmon J-Y, et al, forthe MEDENOX Study Group. A comparison ofenoxaparin with placebo for the prevention ofvenous thromboembolism in acutely ill medicalpatients. N Engl J Med 1999;341:793–800.

The Prophylaxis in Medical Patients withEnoxaparin (MEDENOX) trial had two maingoals. The first was to evaluate whethermedically ill patients were truly at risk for VTE inthe current era of medical practice. A number ofolder trials had reported a VTE rate of about25%, but medical practice had changed from thetime of these older trials to the time of theMEDENOX trial. Therefore, there was a placebogroup in the MEDENOX trial. The second goalof the trial was to evaluate the efficacy and safetyof two different doses of enoxaparin for VTEprevention, if these patients were at risk. Toevaluate these goals, 1102 hospitalized patientsolder than 40 years were randomly assigned in adouble-blind fashion to placebo, enoxaparin 20mg subcutaneously once/day, or enoxaparin 40mg subcutaneously once/day. The primary endpoint of the study was the occurrence of VTE byday 14. Patients were evaluated for VTE byvenography if they had a symptomatic eventbefore they went home (average length of stay 7days) or by day 14, which was the end of the in-hospital follow-up period. Patients were alsoevaluated 3 months later for the accumulation ofany symptomatic events, but not with venography.

Most of the medical illness in the MEDENOXtrial consisted of heart failure (33%), acuterespiratory failure without a ventilator (53%),acute infectious disease (53%), and acuterheumatic disease (9%). The rate of VTE by day14 was 14.9% with placebo, 15.0% withenoxaparin 20 mg/day, and 5.5% with enoxaparin40 mg/day. Proximal DVT by day 14 was 4.9%with placebo, 4.5% with enoxaparin 20 mg/day,and 1.7% with enoxaparin 40 mg/day. Therefore,the MEDENOX trial demonstrated that medicallyill patients are at risk for VTE. The VTE eventrate of 14.9% with placebo would be defined as agroup in the moderate-risk category, but a groupof patients who should receive prophylaxis. Thetrial also demonstrated that enoxaparin 20mg/day is not effective for prevention of VTE in

these patients. Finally, the use of enoxaparin 40mg/day provides a significant 63% relativereduction in total VTE (p<0.001) and a 65%relative reduction in proximal DVT comparedwith placebo (p<0.0001 for both comparisons).These significant reductions were still evident atthe 3-month follow-up (p<0.001 for bothcomparisons). The benefits provided by the useof enoxaparin 40 mg/day did not produce anincrease in major or minor hemorrhage, but therewas a significant increase in the rate of hematoma(0% placebo vs 1.4% enoxaparin 40 mg, p=0.03).

Leizorovicz A, Cohen AT, Turpie AGG, OlssonC-G, Vaikus PT, Goldhaber SZ, for thePREVENT Study Group. Randomized, placebo-controlled trial of dalteparin for the prevention ofvenous thromboembolism in acutely ill medicalpatients. Circulation 2004;110:874–9.

Although the study design of ProspectiveEvaluation of Dalteparin Efficacy for Preventionfor VTE in Immobilized Patients Trial(PREVENT) was similar to that of theMEDENOX trial, the primary end point wasdifferent. Instead of evaluating for venographicevents, as done in MEDENOX, the PREVENTinvestigators wanted to evaluate what theydefined as “clinically relevant” VTE. Theprimary end point of the PREVENT trial was thecomposite of symptomatic distal or proximalDVT, symptomatic or fatal pulmonary embolism,and asymptomatic proximal DVT detected bymeans of compression ultrasonographic exami-nation at day 21. Since there is controversyabout the significance of preventing asympto-matic distal DVT, and the lack of use ofvenography in clinical practice, the investigators’definition of clinically relevant VTE seemsappropriate. The use of this more difficult endpoint explains the need for 3 times as manypatients in the PREVENT trial compared withthat in the MEDENOX trial. The 3706 patients(aged ≥ 40 yrs) in the PREVENT trial wererandomly assigned to placebo or dalteparin 5000IU subcutaneously once/day.

Medical illness in the PREVENT trial consistedof heart failure (52%), acute respiratory failure(30%), acute infectious disease (37%), andrheumatologic disease (11%). Dalteparinprovided a significant 45% relative reduction inthe primary end point at day 21 compared withplacebo (2.77% vs 4.96%, p=0.0015). ProximalDVT was also significantly reduced by 32% withthe use of dalteparin compared with placebo(1.79% vs 3.65%). No significant difference was

439


noted in the rate of hemorrhage between thegroups. Hematoma was not evaluated as anindependent safety outcome in the PREVENTtrial. Therefore, based on the results of thePREVENT trial, dalteparin 5000 IU subcu-taneously once/day is an effective and safe optionfor the prevention of VTE in medically illpatients.

Cohen AT, Davidson BL, Gallus AS, et al, forthe ARTEMIS Investigators. Efficacy and safetyof fondaparinux for the prevention of venousthromboembolism in older acute medicalpatients: randomised placebo controlled trial.BMJ 2006;332:325–9.

The Arixtra for Thromboembolism Preventionin a Medical Indications Study (ARTEMIS)investigators sought to evaluate the efficacy andsafety of fondaparinux in the prevention of VTEin medically ill patients. The 849 patients withgeneral medical illness who were aged 60 yearsor older were randomly assigned in a double-blinded fashion to placebo or fondaparinux 2.5mg subcutaneously once/day. The study designand end points of the trial were similar to thoseof the MEDENOX trial, with the primary endpoint being the occurrence of VTE by day 15detected with venography. Medical illnesses inthe ARTEMIS trial consisted of heart failure(25%), acute respiratory disease (20%), and acuteinfectious or inflammatory disease (25%).

Patients receiving fondaparinux had asignificant 47% reduction in VTE compared withpatients receiving placebo (5.6% vs 10.5%,p=0.029). The rate of hemorrhage was similarbetween the groups. As in the PREVENT trial,hematoma was not evaluated as an independentsafety outcome. There was an interesting trendtoward a reduction in mortality in the ARTEMIStrial (p=0.06). A trend was also suggested in theMEDENOX trial but was not as strong (p=0.31).A much larger trial would be needed to concludeif this possible effect on mortality is real.

The rate of VTE in patients with cancer was17% in fondaparinux-treated patients comparedwith 3.9% in placebo-treated patients (p=0.03).Although these data were not in the initial article,it clearly is an odd finding that requires furtherevaluation.

Kleber F-X, Witt C, Vogel G, Koppenhagen K,Schomaker U, Flosbach CW, for the THE-PRINCE Investigators. Randomized comparisonof enoxaparin with unfractionated heparin forthe prevention of venous thromboembolism inmedical patients with heart failure or severe

respiratory disease. Am Heart J 2003;145:614–21.Several trials have evaluated the efficacy of

different anticoagulants for VTE prophylaxiscompared with placebo or control in medically illpatients. There is probably no longer a need forplacebo-controlled trials for VTE prevention inmedically ill patients. These trials consistentlydemonstrate that medically ill patients are at riskand that providing them pharmacologicprophylaxis reduces that risk. There is now aneed for head-to-head comparison trials todetermine the optimal regimen for these patients.

The 665 patients in the Thromboembolism–Prevention in Cardiac or Respiratory Diseasewith Enoxaparin (THE-PRINCE) trial wererandomly assigned to either UFH 5000 U 3times/day or enoxaparin 40 mg once/day.Medically ill patients in THE-PRINCE trialincluded similar numbers of patients with heartfailure or severe respiratory disease. The primaryend point of the trial was the occurrence ofvenography-verified DVT and symptomaticpulmonary embolism up to 1 day after thetreatment period (maximum 10 days).

The primary end point occurred in 8.4% ofenoxaparin-treated patients compared with10.4% of UFH-treated patients (p=NS). The useof enoxaparin was associated with fewer adverseevents compared with UFH (45.8% vs 53.8%,p=0.44). The THE-PRINCE investigators thenseparately evaluated efficacy in patients withrespiratory disease or heart failure. Since patientswith medical illnesses represent a very hetero-geneous group, different levels of risk exist. Thefrequency of the primary end point in patientswith respiratory disease was 7.1% for enoxaparinand 5.9% for UFH (p=NS). In the higher riskpatients with heart failure, still no significantdifference was noted (9.7% enoxaparin vs 16.1%UFH), but clearly the separation of more than 6%requires further study. Therefore, THE-PRINCEinvestigators’ evaluation of different groups ofmedically ill patients was quite insightful and setup some of the future research in this area.

Wein L, Wein S, Haas SJ, Shaw J, Krum H.Pharmacological venous thromboembolismprophylaxis in hospitalized medical patients: ameta-analysis of randomized controlled trials.Arch Intern Med 2007;167:1476–86.

This meta-analysis included 36 trials thatevaluated the efficacy and safety of UFH, LMWH,or fondaparinux for the prevention of VTE inmedically ill patients. Fourteen of the trialsevaluated UFH versus control, 11 trials evaluated

440


LMWH versus control, 10 trials evaluated UFHversus LMWH, and one trial evaluatedfondaparinux versus placebo (ARTEMIS).Overall, the use of UFH provided a significant67% reduction in risk of DVT and a 44%reduction in risk of pulmonary embolism inmedically ill patients. The use of UFH 3times/day (83% risk reduction) was moreeffective than UFH twice/day (48% riskreduction) in preventing DVT. The use ofLMWH was associated with a significant 32%risk reduction of DVT when directly comparedwith UFH. Neither agent provided a reduction inmortality. Although the risk of major bleedingand thrombocytopenia were not significantlydifferent between UFH and LMWH, there was asignificant 53% reduction in risk of injection sitehematoma for patients receiving LMWHcompared with UFH. At least two additionalmeta-analyses have evaluated the efficacy andsafety of VTE prevention in medically ill patients.Although the methods of these other meta-analyses vary somewhat, the overall conclusionsof the benefits and safety of VTE prophylaxis inmedically ill patients are similar.

Sherman DG, Albers GW, Bladin C, et al, forthe PREVAIL Study Investigators. The efficacyand safety of enoxaparin versus unfractionatedheparin for the prevention of venous thrombo-embolism after acute ischaemic stroke (PREVAILstudy): an open-label randomised comparison.Lancet 2007;369:1347–55.

Patients with ischemic stroke represent notonly a group of medically ill patients at high riskfor VTE, but also a group of patients in whomconcern about bleeding is significant. The 1762patients with acute ischemic stroke in thePrevention of Venous Thromboembolism AfterAcute Ischaemic Stroke (PREVAIL) trial wererandomly assigned in an unblinded fashion toreceive enoxaparin 40 mg once/day or UFH 5000U twice/day for up to 10 days. The primary endpoint of the trial was venography-confirmedDVT, or symptomatic or fatal pulmonaryembolism.

Patients receiving enoxaparin demonstrated a43% relative reduction in VTE compared withUFH (10% vs 18%, p=0.0001). The rates of anybleeding (8% in both groups) and intracranialhemorrhage (1% in both groups) were notdifferent between the groups. The rate ofextracranial bleeding was higher with the use ofenoxaparin than with UFH, but the number ofoverall events was small (seven vs zero events,

p=0.015). Although smaller trials havedemonstrated a benefit of enoxaparin over UFH,these trials were not large enough to evaluatesafety. A limitation of the PREVAIL trial was theuse of UFH twice/day compared with 3 times/day.Smaller trials had already demonstrated a benefitof enoxaparin over UFH 3 times/day. Theinvestigators noted the concern of possibleincreased hemorrhagic transformation with UFH3 times/day. Therefore, the UFH twice-dailyregimen could explain the lower rate ofextracranial bleeding compared with that of theenoxaparin regimen, but it probably does notexplain the decreased efficacy.

Shorr AF, Jackson WL, Sherner JH, Moores LK.Differences between low-molecular-weight andunfractionated heparin for venous thrombo-embolism prevention following ischemic stroke:a meta-analysis. Chest 2008;133:149–55.

The investigators of this meta-analysisincluded three trials that evaluated an LMWHcompared with UFH for VTE prevention inpatients with ischemic stroke. Two of the trialsused enoxaparin, and one trial used certoparin.Two trials used UFH 3 times/day and one trialused UFH twice/day. Overall, compared withUFH, the use of LMWH was associated with asignificant reduction in total VTE by 46%(p<0.001), proximal VTE by 47% (p<0.001), andpulmonary embolism by 74% (p=0.042). Therates of bleeding, intracranial hemorrhage, andmortality were not significantly different betweenthe groups. When only the trials that usedenoxaparin were considered, the results wereconsistent with the overall findings. Unfortu-nately, no sensitivity analysis was conductedevaluating the outcomes based on twice-dailyversus thrice-daily dosing of UFH. As patientswith ischemic stroke have such a high risk forVTE, this evaluation would have beeninteresting. Also, this added evaluation wouldhave been beneficial since the only trial that usedUFH twice/day was the PREVAIL trial, but itcontributed more than 70% of the patients in themeta-analysis.

Creekmore FM, Oderda GM, Pendleton RC,Brixner DI. Incidence and economic implicationsof heparin-induced thrombocytopenia in medicalpatients receiving prophylaxis for venousthromboembolism. Pharmacotherapy 2006;26:1438–45.

To our knowledge, this study was the first toevaluate the overall cost to a health care system

441


when hospitalized medically ill patients whoreceive UFH or enoxaparin for the prevention ofVTE develop heparin-induced thrombocytopenia(HIT). This retrospective study used nested case-controls from a single hospital. Heparin-inducedthrombocytopenia was deemed to have occurredif heparin-dependent antibodies were detectedduring hospitalization. Each patient with HITwas matched with three control patients withoutHIT who were identified with prespecifiedcriteria. Hospital and physician charges for thepatients with HIT and control subjects wereextracted from the hospital’s database andcompared. All such charges were converted to2004 dollars to account for inflation.

During the more than 4 years of datacollection, 10,121 adult patients received eitherUFH 5000 U 2 or 3 times/day, or enoxaparin 30mg twice/day or 40 mg once/day. In patientsreceiving UFH, the rate of HIT was 0.51%,compared with 0.08% in patients who receivedenoxaparin (p=0.03). Compared with controlpatients, the patients who developed HITincurred an average of $82,266 more in charges.In addition, development of HIT was associatedwith an extra 12.5 days of hospitalization. Mostnotably, as a result of the markedly lower rate ofHIT in patients receiving enoxaparin comparedwith those receiving UFH, a total cost savings of$32,981 was observed in patients who receivedenoxaparin for VTE prophylaxis. This costsavings is equivalent to $13.88/patient, despitethe higher acquisition drug cost for enoxaparin.These observations suggest that use ofenoxaparin is preferred over UFH for preventionof VTE for both clinical and pharmacoeconomicreasons. It is possible that the savings calculatedfrom this study may actually represent aconservative estimate since the rate of detectedHIT was quite a bit lower than the 1–3% typicallyquoted.

Cancer

Chaukiyal P, Nautiyal A, Radhakrishnan S,Singh S, Navaneethan SD. Thromboprophylaxisin cancer patients with central venous catheters:a systematic review and meta-analysis. ThrombHaemost 2008;99:38–43.

Central venous catheters are frequentlyimplanted in patients with malignancy as ameans to deliver chemotherapy and otherintravenous drugs. These catheters have atendency to clot, which is typically preventedwith local anticoagulation by flushing the

catheter with UFH, saline, or other agents. In1990, a randomized clinical trial comparingwarfarin 1 mg/day with no treatment suggestedthat systemic anticoagulation was valuable inreducing the rate of central venous catheter–related thrombosis. Subsequent trials evaluatedsystemic anticoagulation with UFH and LMWHand also suggested a role for systemic anticoag-ulation. However, more recent, well-designed,double-blind, randomized, controlled trials thatevaluated warfarin and LMWHs in patients withcancer who have a central venous catheter havenot shown any benefit of using systemicanticoagulation to prevent catheter-associatedthrombosis. Thus, a meta-analysis was conductedto assess the available literature.

This meta-analysis, one of two published in2008, combined results from eight randomizedcontrolled trials involving 1428 patients assignedto receive warfarin, UFH, or LMWH, or toplacebo or no treatment. One trial comparednadroparin with warfarin 1 mg/day. Nostatistically significant difference in the rate ofcatheter-associated thrombosis was observed forwarfarin compared with placebo or no treatment(RR 0.75, p=0.63), for UFH or LMWH comparedwith placebo or no treatment (RR 0.46, p=0.06),or for any systemic anticoagulation (warfarin,UFH, or LMWH) versus placebo or no treatment(RR 0.59, p=0.11). The use of systemic anti-coagulation was not associated with anystatistically significant increase in the risk ofoverall bleeding (RR 1.25, p=0.28). The authorsconcluded that systemic anticoagulation forprevention of central venous catheter thrombosisprovides no benefit. An overall decline in theoccurrence of catheter-related thrombosis wasalso noted, likely associated with improvementsin catheter design and care, and the routine useof heparin or saline flushes.

Kakkar AK, Levine MN, Kadziola Z, et al. Low-molecular-weight heparin therapy withdalteparin and survival in advanced cancer: theFragmin advanced malignancy outcome study(FAMOUS). J Clin Oncol 2004;22:1944–8.

The effects of activated coagulation proteaseson tumor cell biology and the observation thatLMWH may prolong survival in patients withcancer led to speculation that anticoagulationmight have benefits beyond the prevention andtreatment of thrombosis in patients with cancer.The Fragmin Advanced Malignancy OutcomeStudy (FAMOUS), a randomized, double-blind,placebo-controlled trial, was designed to

442


investigate whether routine anticoagulation withan LMWH might improve cancer survival inpatients with malignancy but without thrombosis.A total of 385 patients were randomly assigned toreceive long-term dalteparin 5000 IU subcu-taneously once/day or placebo. The primary goalwas to assess mortality after 1 year of therapy,with secondary outcomes that included VTE andbleeding complications. The patients hadadvanced stage III or IV cancer of the breast,lung, gastrointestinal tract, ovary, or uterus.

At 1 year, survival estimates for patients treatedwith dalteparin and placebo were similar (46% vs41%, p=0.19). However, a subgroup analysis of102 patients with a better prognosis and whosurvived beyond 17 months found a significantsurvival benefit (p=0.03) for dalteparin at 2 yearsand at 3 years. The rate of symptomatic VTE waslow during the study period (dalteparin 2.4% vsplacebo 3.3%), as was the rate of major andminor bleeding (dalteparin 4.7% vs placebo2.7%). The high early mortality rate in theoverall population may have influenced theresults. Nevertheless, the trial demonstrated thefeasibility, safety, and effectiveness of long-termLMWH in patients with advanced cancer.

Lazo-Langner A, Goss GD, Spaans JN, et al. Theeffect of low-molecular-weight heparin on cancersurvival: a systematic review and meta-analysis ofrandomized trials. J Thromb Haemost 2007;5:729–37.

After FAMOUS, three additional randomizedcontrolled trials evaluated the mortality benefit ofLMWH in patients with malignancy but withoutthrombosis. Data from all four trials weresubsequently combined in this meta-analysis,which compared the effect of LMWH and placeboor no treatment on cancer survival. In total, thestudies evaluated 448 patients with solid tumorswho were treated with dalteparin 5000 IUsubcutaneously once/day or weight-adjustednadroparin 3600–7200 antifactor Xa U twice/day,for various durations depending on the studyinvolved. An additional 450 patients receivedplacebo or no treatment during the study. Withuse of OR analysis, the pooled results of thestudies showed that LMWH was associated witha 30% reduction in the odds of death at 1 yearand a 43% reduction in the odds of death at 2years. The RR analysis found that LMWH wasassociated with a 13% reduction in the risk ofdeath at 1 year, and a 10% reduction in the riskof death at 2 years. In contrast to the FAMOUStrial, this meta-analysis found that the survival

benefit of LMWH extended to patients withadvanced disease. When patients with stage I orII malignancy were excluded, the OR of death at1 year was 0.75, and the RR of death at 1 yearwas 0.89. At 2 years, these figures were 0.90 and0.92, respectively. Despite the results of thismeta-analysis, routine use of LMWH in patientswith cancer has not become the standard of care.Even primary VTE prophylaxis in patients withcancer is currently not recommended. Futuretrials will need to determine the effectiveness andcost:benefit ratio of both of these potentialstrategies.

The ENOXACAN Study Group. Efficacy andsafety of enoxaparin versus unfractionatedheparin for prevention of deep vein thrombosisin elective cancer surgery: a double-blindrandomized multicentre trial with venographicassessment. Br J Surg 1997;84:1099–103.

The Enoxaparin and Cancer (ENOXACAN)trial is a landmark trial that was among the firstto establish the effectiveness and safety of anLMWH compared with UFH for the preventionof VTE in patients with malignancy who areundergoing major surgery. The ENOXACANtrial evaluated 631 patients undergoing plannedelective curative abdominal or pelvic surgery forcancer who had been randomly assigned toreceive either UFH 5000 U subcutaneously 3times/day or enoxaparin 40 mg subcutaneouslyonce/day. Prophylaxis was started 2 hours beforesurgery and continued for 10 days in bothgroups. Bilateral venography was performed inall patients at 10 days or sooner if a patientdeveloped clinical symptoms of VTE. Allpatients were followed for 3 months.

At 10 days, no significant difference was notedin the combined rate of DVT, pulmonaryembolism, or death in patients exposed to LMWHor UFH (14.7% vs 18.2%, 95% CI -9.2–2.3,p=NS). The rate of major bleeding was alsosimilar (4.1% vs 2.9%, p=NS). At 3 months, nosignificant difference in mortality was observedbetween the two groups. In the subgroup ofpatients undergoing colorectal surgery andconsidered to be at higher risk for VTE than thegeneral study population, the rate of VTE was18.8% in patients treated with UFH and 16.6% inpatients treated with enoxaparin. In addition, inthe subgroup of patients whose operations lastedlonger than 4 hours, the VTE rate was 13.9% inpatients treated with UFH and 7.8% in patientstreated with enoxaparin. This study was used tosupport the role of an LMWH in prevention of

443


cancer-associated thrombosis in patientsundergoing surgery.

Leonardi MJ, McGory ML, Ko CY. A systematicreview of deep venous thrombosis prophylaxis incancer patients: implications for improvingquality. Ann Surg Oncol 2007;14:929–36.

This systematic review was an evaluation ofrandomized controlled trials that investigated therole of VTE prophylaxis in patients with cancerwho were undergoing surgery. Its aim was todefine the overall rate of DVT, the efficacy ofhigh- versus low-dose heparin, and the rate ofbleeding complications associated withpharmacologic prophylaxis in this population.Twenty-six trials were included, all of whichinvolved some aspect of pharmacologic VTEprophylaxis, started 2 hours before surgery, in7639 patients. The overall DVT rate in thecontrol or placebo groups of evaluable studieswas 35.2% and was reduced to 12.7% when aheparin product was used for VTE prophylaxisand to 5% when heparin was combined with amechanical method of VTE prophylaxis. Asubgroup analysis found that high-dose LMWHgreater than 3400 U/day was associated with alower rate of DVT than lower dose LMWH (7.9%vs 14.5%, p<0.0001), as was higher dose UFH5000 U 3 times/day compared with lower doseUFH 5000 U twice/day (8% vs 13.4%, p=0.0132).No significant differences were detected betweenLMWH and UFH with respect to effectiveness inreducing the rate of DVT compared with placeboor control, or in minor bleeding (10% overall),major bleeding (1% overall), or bleedingrequiring discontinuation of pharmacologicprophylaxis (3% overall). This review lendssupport to the general need for VTE prophylaxisin patients undergoing surgery and remindsreaders of regulatory requirements that will drivethe need to ensure adequate prophylaxis and toevaluate postoperative VTE rates.

Akl EA, Terrenato I, Barba M, et al. Low-molecular-weight heparin vs unfractionatedheparin for perioperative thromboprophylaxis inpatients with cancer. Arch Intern Med 2008;168:1261–9.

Using a more sophisticated meta-analyticmethodology than the systematic reviewdescribed above, the authors of this studyevaluated the combined results of 14 randomizedcontrolled trials that specifically compared anLMWH with UFH in the prevention of VTE in5822 patients undergoing cancer surgery. The

primary objective was to evaluate mortality,which was not significantly different in patientstreated with an LMWH compared with thosetreated with UFH (RR 0.89, 95% CI 0.61–1.28);in addition, no significant differences in mortalityrate were observed when LMWH was comparedwith UFH given twice/day or 3 times/day. Anumber of subanalyses were conducted, the mostsignificant of which was a comparison of LMWHwith different dosing frequencies of UFH.Regardless of the diagnostic strategy used todetect DVT, LMWH was superior to UFH (RR0.72, 95% CI 0.55–0.84). This benefit wassignificant in the subgroup receiving UFHtwice/day (RR 0.66, 95% CI 0.44–0.99) but notin the subgroup receiving UFH 3 times/day (RR0.78, 95% CI 0.53–1.15). Of importance,however, no trials directly compared twice-dailywith thrice-daily UFH regimens. No significantdifferences in minor bleeding, major bleeding,hematoma formation, bleeding requiring surgicalintervention, or other measures of bleeding werefound between groups. This meta-analysissuggests that LMWH is equivalent to UFH given3 times/day for VTE prophylaxis in patientsundergoing cancer surgery, and that any detecteddifference is driven by comparisons of LMWHwith UFH given twice/day, in which LMWH is infact superior.

Bergqvist D, Agnelli G, Cohen AT, et al.Duration of prophylaxis against venousthromboembolism with enoxaparin after surgeryfor cancer. N Engl J Med 2002;346:975–80.

The ENOXACAN study published in 1997compared the efficacy and safety of enoxaparinwith that of UFH for VTE prophylaxis in patientsundergoing elective cancer surgery and found asignificant benefit to using LMWH. Recognizingthat the risk of VTE extends beyond the period ofhospitalization after major surgery, theENOXACAN II trial, undertaken by the samestudy group, aimed to determine the value ofextending the duration of VTE prophylaxis. Atotal of 332 patients requiring open, elective,curative surgery for abdominal or pelvic cancerreceived enoxaparin 40 mg subcutaneouslyonce/day for 6–10 days as routine VTEprophylaxis. The patients were subsequentlyrandomly assigned to continue either enoxaparinor placebo for an additional 21 days. With use ofan intent-to-treat analysis, extended-durationprophylaxis reduced the rate of venographicallydetected VTE between days 25 and 31 from 12%in the placebo group to 4.8% in the enoxaparin

444


group (p=0.02). The overall RR was 60%. At 3months, the difference persisted, with 13.8% inthe placebo group and 5.5% in the enoxaparingroup (p=0.01). The rate of major bleeding wasvery low and not significantly different betweenthe two groups.

Pregnancy

Bates SM, Greer IA, Pabinger I, Sofaer S, HirshJ. Venous thromboembolism, thrombophilia,antithrombotic therapy, and pregnancy: AmericanCollege of Chest Physicians evidence-basedclinical practice guidelines, 8th ed. Chest2008;133:S844–86.

This article is part of the ACCP Evidence-BasedClinical Practice Guidelines, which were updatedin 2008. The article provides informativerecommendations with supporting evidenceregarding the use of antithrombotic agents, aswell as the management of VTE and thrombo-philia, during pregnancy. It includes keyrecommendations for VTE prophylaxis andtreatment in pregnancy, mechanical heart valvesand thrombophilia in pregnancy, effects ofvarious antithrombotic exposure in utero, fetalcomplications of antithrombotic therapy, use ofantithrombotics in nursing women, andimplications of women’s preferences and valuesduring pregnancy. These updated guidelinesinclude strong evidence for substituting warfarinwith UFH or LMWH in pregnant women andscreening for antiphospholipid antibodies, as wellas contemporary recommendations for pregnantwomen with thrombophilia, including thepractice of VTE prophylaxis in the ante- andpostpartum periods. New sections include directthrombin inhibitor and pentasaccharide use inutero, thrombolysis use during pregnancy, andVTE after cesarean section.

Carp H, Dolitzky M, Inbal A. Thromboprophylaxisimproves the live birth rate in women withconsecutive recurrent miscarriages and hereditarythrombophilia. J Thromb Haemost 2003;1:433–8.

Hereditary thrombophilias have beenassociated with fetal loss and are often the causeof recurrent miscarriages. This case-controlstudy of 85 patients examines whether the use ofenoxaparin 40 mg subcutaneously once/day iseffective at increasing live birth rates amongwomen with a history of recurrent miscarriagesand a hereditary thrombophilia. Women wereincluded in the study if they had three or moreprevious and consecutive pregnancy losses in the

first or second trimester. Results showed thatpatients treated with enoxaparin throughoutpregnancy had a significantly higher live birthrate compared with controls (70.2% vs 43.8%,p<0.02). This beneficial effect was apparentprimarily among those women with a history ofno previous live births (p=0.008), in whomenoxaparin resulted in a 10-fold increase in theOR for a live birth.

This study also analyzed the proportion of livebirths for different types of documentedhereditary thrombophilia. Enoxaparin was foundto increase live birth rates in each of thethrombophilias tested, except for those womenwith the prothrombin gene mutation. Thegreatest beneficial effect was seen in thosewomen with factor V Leiden. Although largerstudies are required before thromboprophylaxiswith enoxaparin becomes the standard ofpractice in this patient population, this studyprovides data that enoxaparin may have a role inpreventing further pregnancy losses in womenwith a history of recurrent miscarriages andhereditary thrombophilia.

Bauersachs RM, Dudenhausen J, Faridi A, et al.Risk stratification and heparin prophylaxis toprevent venous thromboembolism in pregnantwomen. Thromb Haemost 2007;98:1237–45.

Women are in a hypercoagulable state and atincreased risk for thrombosis during pregnancy.When additional risk factors (such as previousidiopathic VTE, thrombophilia, or both) arepresent, the thrombosis risk is further increased,most significantly within the first trimester. Themost optimal VTE prevention strategy in thispatient population is unknown. This large,prospective study evaluated a risk assessmentstrategy and prophylaxis regimen for thosepregnant women at increased VTE risk becauseof a history of VTE and thrombophilia. Thestudy assigned 810 pregnant women to one ofthree risk groups according to individual risk atthe time of presentation. The most common riskfactors were thrombophilias (~75%), history ofprevious VTE (~60%), and family history of VTE(~31%). All patients received daily doses ofsubcutaneous dalteparin, although the frequency,dose, and length of prophylaxis varied amongrisk groups.

The results demonstrated that this particularrisk stratification and prophylaxis strategy inpregnant women at increased risk for VTEresulted in low levels of symptomatic VTE events.Objectively confirmed VTE by ultrasonography,

445


CT, or ventilation-perfusion scan occurred in0.6% of women, and only in those patientsassigned to be at high or very high risk. Thepercentage of women experiencing a clinicallyrelevant bleed was low (~5%), and minimalcomplications associated with dalteparin usewere reported. This study provides evidence thatusing the defined risk stratification tool withdalteparin as antepartum VTE prophylaxis inpregnant women with additional risk factors is aneffective approach to decreasing VTE events.

James AH. Prevention and management ofvenous thromboembolism in pregnancy. Am JMed 2007;120:S26–34.

During pregnancy, the risk of VTE is increased,yielding an estimated 2 events/1000 births. Thisreview article discusses the literature supportingrecommendations for the prevention ormanagement of VTE during pregnancy and in thepostpartum period. The physiologic changes thatoccur during pregnancy resulting in ahypercoagulable state are described, as well asadditional risk factors for peripartum VTE. Thetop four risk factors are thrombophilia, history ofthrombosis, antiphospholipid syndrome, andlupus. Despite the increased risk for thrombosisduring pregnancy, most women do not requireroutine anticoagulation because the risksoutweigh the benefits. Recommendations forprophylaxis depending on risk level are discussedin this article, as are various dosing regimens forUFH and LMWH according to the patient’sindividual risk. Given various changes thatoccur during pregnancy, such as increased bloodvolume, increased volume of distribution,increased glomerular filtration, and increasedprotein binding, vigilant agent selection andmonitoring are warranted. The literatureconcerning complications with various heparintherapies are discussed, as are data supportingfondaparinux use. The article reviews how tomanage anticoagulants during the time ofdelivery and in the postpartum period, as manywomen will still require anticoagulation for thefirst 6 weeks of the postpartum period.

Obesity and Bariatric Surgery

Kalfarentzos F, Fotini S, Yarmenitis S, et al.Prophylaxis of venous thromboembolism usingtwo different doses of low-molecular-weightheparin (nadroparin) in bariatric surgery: aprospective randomized trial. Obes Surg2001;11:670–6.

Obese patients undergoing bariatric surgerywarrant VTE prophylaxis because they are athigh risk for developing a fatal pulmonaryembolism or postthrombotic syndrome.However, a consensus does not exist on the mosteffective prophylactic approach, including agentselection, dosage, and schedule. It is difficult toextrapolate data from other comparative trialsgiven that the inclusion of this patient populationis limited. This prospective, randomized, pilottrial evaluated the safety and efficacy of a lowerfixed dose compared with a higher fixed dose ofnadroparin for the prophylaxis of postoperativethromboembolism in patients undergoing gastricbypass surgery for morbid obesity (body massindex [BMI] > 36 kg/m2). According to a parallelstudy design, 30 patients received subcutaneousnadroparin 0.6 ml/day (5700 antifactor Xa U),and another 30 patients received subcutaneousnadroparin 1.0 ml/day (9500 antifactor Xa U).Both doses were started 1 day before surgery.The primary end point of rate of VTE wasevaluated by using compression ultrasonography,and results showed no thrombotic events in anypatients after surgery, at discharge, or at the 3-and 6-month follow-up periods in either group.No bleeding events were reported in the lowerdose group, but in the higher dose group twopatients developed major hemorrhages. Thestudy provides evidence that the once-dailyadministration of a lower fixed dose ofnadroparin was as effective as the higher dose inpreventing VTE and was safer, because of fewerbleeding events, in this morbidly obesepopulation. This pilot, single-center studyprovides the rationale for developing a larger,randomized study within this patient population,to indicate that a single daily dose of nadroparin,unadjusted for BMI, may be sufficient for VTEprophylaxis in those patients undergoingbariatric surgery.

Scholton DJ, Hoedema RM, Scholten SE. Acomparison of two different prophylactic doseregimens of low-molecular-weight heparin inbariatric surgery. Obes Surg 2002;12:19–24.

Patients undergoing bariatric surgery are athigh risk for VTE. Morbid obesity is anindependent risk factor for sudden death fromacute pulmonary embolism. The LMWHs havebeen increasingly used for VTE prophylaxis, buttheir efficacy in morbidly obese patients isunknown. This study explored LMWH use inmorbidly obese patients undergoing eitherprimary or revisional bariatric surgery at an

446


active bariatric surgical practice. The study useda multimodal approach for VTE prophylaxis in481 patients based on the recommendation forhighest-risk patients as assessed by a thrombosisrisk assessment tool. Patients were to receive acombination of early ambulation, GCS, IPC, andtwice-daily administration of enoxaparinbeginning 2 hours before surgery and continueduntil the patients were fully ambulatory or untildischarge, whichever came first. Patientsreceived enoxaparin either 30 or 40 mg every 12hours. Primary bariatric surgery was performedin most of the patients (91.5%), whose averageBMI was approximately 51 kg/m2. Patients whoreceived the higher dose of enoxaparin haddecreased VTE events compared with thosereceiving the lower enoxaparin dose (two DVTsvs one DVT and four nonfatal pulmonaryembolisms, p<0.01). There was one bleedingcomplication in each group, neither of which waslife-threatening. This study provides data that inmorbidly obese patients undergoing bariatricsurgery, enoxaparin 40 mg twice/day may bemore effective in preventing VTE than the lowerdose regimen.

Frederiksen SG, Hedenbro JL, Norgren L.Enoxaparin effect depends on body-weight andcurrent doses may be inadequate in obesepatients. Br J Surg 2003;90:547–8.

This study sought to correlate antifactor Xaactivity with body weight after an injection ofenoxaparin. An LMWH is a first-line choice forVTE prophylaxis for various surgical procedures,but the efficacy of LMWH in preventing clinicallysignificant VTE events has not been studiedadequately in obese patients. Such a study isunlikely to be performed given the large numberof patients that would be necessary to detect adifference due to the low frequency of events.However, the effect of a given enoxaparin dosecan be monitored through plasma antifactor Xalevels. This study recruited 17 patients under-going surgical procedures, 10 of whom wereundergoing bariatric surgery. Antifactor Xa levelswere drawn hourly for 6 hours, then every 2hours for two more levels, after an enoxaparin40-mg subcutaneous dose was given 20 hoursbefore surgery. The study found no resultantVTE events and no excessive bleeding in any ofthe patients. Peak levels of antifactor Xa werefound to occur 3–5 hours after enoxaparin wasadministered, and a strong negative correlationwas noted between body weight and antifactorXa activity after enoxaparin was administered.

Further studies will be needed to address thesafety of adjusting LMWH dosages to achievesimilar antifactor Xa activity among patients ofvarious weights.

Quebbemann B, Akhondzadeh M, Dallal R.Continuous intravenous heparin infusionprevents peri-operative thromboembolic eventsin bariatric surgery patients. Obes Surg2005;15:1221–4.

Morbidly obese patients undergoing bariatricsurgery are at high risk for VTE. However,numerous factors complicate the selection ofVTE prophylaxis in this population, includingdecreased venous return from pneumo-peritoneum, poor effectiveness of GCS, andunclear pharmacodynamics of heparinadministration. Researchers in this study used aprotocol of continuous, non–weight-based,intravenous UFH infusion, starting 1 hour beforesurgery, for VTE prophylaxis. The UFH infusionswere started at 400 U/hour and continued untilthe day of discharge. No UFH bolus was given,nor were dosage adjustments made based onweight, activated partial thromboplastin time, orantifactor Xa levels. In addition, IPCs were used,and ambulation was strongly encouraged. Noroutine surveillance studies were performed;however, all clinically evident VTE events wererecorded. Results showed that only 1 (0.12%) of822 patients receiving the UFH infusiondeveloped a pulmonary embolism after thegastric bypass procedure, and UFH therapy wastemporarily held or stopped in 5% of patientsbecause of postoperative bleeding. The authorsconcluded that a continuous low-dose intra-venous UFH infusion is associated with a lowrate of VTE events and hemorrhage and,therefore, is a potential option for VTE prophylaxisin this patient population.

Hamad GG, Choban PS. Enoxaparin forthromboprophylaxis in morbidly obese patientsundergoing bariatric surgery: findings of theprophylaxis against VTE outcomes in bariatricsurgery patients receiving enoxaparin (PROBE)study. Obes Surg 2005;15:1368–74.

The goal of the Prophylaxis Against VTEOutcomes in Bariatric Surgery Patients ReceivingEnoxaparin (PROBE) study was to retrospec-tively gather information on symptomatic VTEand bleeding events in patients who underwentbariatric surgery and were given various regimensof enoxaparin for VTE prophylaxis in five studycenters throughout the United States. The study

447


included 668 patients with an averagepreoperative BMI of 49.6 kg/m2, most of whom(84.9%) underwent open gastric bypass surgerywith mean postoperative follow-up of 10.5 days.Each center used a different enoxaparin dosingregimen for prophylaxis, including enoxaparin30 mg started preoperatively, 30 mg/day startedafter hospital discharge, 40 mg/day startedpostoperatively, and 40 mg twice/day startedpostoperatively. The duration of prophylaxis alsovaried, ranging from 12 hours to 10 days.

Cumulatively, 1.0% of patients had VTE eventsafter surgery (0.1% DVT and 0.9% pulmonaryembolism), all of which occurred after thecessation of prophylaxis. The highest rates ofVTE were found in those patients who receivedenoxaparin 30 mg/day beginning after discharge.Among those patients who received enoxaparinperioperatively, no DVTs were reported and 0.7%had a pulmonary embolism. All patients whoexperienced a VTE had one or more additionalrisk factors for VTE aside from severe obesity andabdominal surgery, the most frequent beingadvanced age and smoking. Few bleedingcomplications occurred. The six reportednonfatal bleeding events were observed in thosecenters that administered 40 mg/day or 40 mgtwice/day.

This study confirms that morbidly obesepatients undergoing bariatric surgery are atsubstantial risk for VTE complications, despitereceiving prophylaxis. Although minimal datafrom randomized controlled trials exist regardingthe optimal enoxaparin regimen for prophylaxisin this population, the data from this studyindicate the necessity for additional risk profilingfor such patients, describe regimens that may besuperior to others, and suggest that extendedprophylaxis may be warranted.

Kucher N, Leizorovicz A, Vaitkus PT, et al.Efficacy and safety of fixed low-dose dalteparinin preventing venous thromboembolism amongobese or elderly hospitalized patients: a subgroupof the PREVENT trial. Arch Intern Med 2005;165:341–5.

The original PREVENT trial found that a fixeddose of dalteparin was effective in reducing VTEevents by 45% (RR 0.55, 95% CI 0.38–0.80) inpatients aged 40 years or older who requirehospitalization for an acute medical illness. Thissubgroup analysis of the PREVENT trial wasconducted to determine whether this fixed dosewas as effective in obese patients as in nonobesepatients. The authors retrospectively analyzed

data from 3706 hospitalized patients included inthe PREVENT trial who received dalteparin 5000IU/day or placebo. Obesity was defined as a BMIgreater than 30 kg/m2 for men and 28.6 kg/m2 forwomen, and 30.4% of the overall population metthese criteria.

In obese patients, the primary end point—acomposite of symptomatic VTE, fatal pulmonaryembolism, sudden death, or asymptomaticproximal DVT by day 21— occurred in 2.8% ofthe dalteparin group compared with 4.3% in theplacebo group (RR 0.64, 95% CI 0.32–1.28). Innonobese patients, the primary end pointoccurred in 2.8% of the dalteparin group and5.2% of the placebo group (RR 0.53, 95% CI0.34–0.82). Dalteparin was therefore effective inobese patients, with an RRR of 36% across BMIsubgroups, except for those patients with a BMIgreater than 40 kg/m2. This indicates thatperhaps a fixed low dose of dalteparin may not beappropriate in the morbidly obese population,particularly in those with additional risk factors.Dalteparin was deemed safe in the obesepopulation, in that no increase in mortality (4.6%vs 2.7%, p=0.14) or increase in major hemorrhage(0% vs 0.7%, p>0.99) versus placebo, wasobserved by day 21. This study’s findings imply alow fixed dose of dalteparin is safe and effectivein preventing VTE among hospitalized patientswith acute medical illness and is not significantlyaffected by BMI, unless patients have a BMIgreater than 40 kg/m2.

Kothari SN, Lambert PJ, Mathiason MA. Acomparison of thromboembolic and bleedingevents following laparoscopic gastric bypass inpatients treated with prophylactic regimens ofunfractionated heparin or enoxaparin. Am J Surg2007;194:709–11.

The risk of postoperative bleeding must beweighed against the risk of postoperative VTEevents in those patients undergoing laparoscopicgastric bypass surgery. This study prospectivelyevaluated 30-day VTE and bleeding events in 476patients, half of whom received enoxaparin andthe other half received UFH for VTE prophylaxis.The enoxaparin regimen consisted of 40 mgsubcutaneously before surgery, then 40 mg on theevening of postoperative day 0 and twice/dayuntil discharge. The UFH regimen consisted of5000 U subcutaneously before surgery, nothingon the evening of postoperative day 0, and 5000U 3 times/day until discharge. All patients werealso treated with IPC and early ambulation.

No DVT events occurred in either cohort

448


(p=0.999), and one pulmonary embolismoccurred in the UFH group (p=0.999).Postoperative transfusions were required in 5.9%of the enoxaparin group, compared with 1.3% inthe UFH group (p=0.011). In addition, 1.7% ofpatients in the enoxaparin group requiredreexploration for bleeding, compared with nonein the UFH group (p=0.123). Both regimenswere effective in preventing VTE; however, thisstudy suggests UFH as the preferred option forVTE prophylaxis in this patient populationbecause of decreased bleeding complications.

Renal Insufficiency and Dialysis

Sanderink GJ, Guimart CG, Ozoux ML, JariwalaNU, Shukla UA, Boutouyrie BX. Pharmaco-kinetics and pharmacodynamics of theprophylactic dose of enoxaparin once daily over4 days in patients with renal impairment.Thromb Res 2002;105:225–31.

In this open-label, multicenter, parallel-groupstudy, the impact of renal function on thepharmacokinetic and pharmacodynamic profile ofenoxaparin 40 mg once/day was evaluated incohorts of 12 patients each who had normal, mild,moderate, or severe renal impairment. Bothantifactor Xa and antifactor IIa activity weremeasured over 24–36 hours on days 1 and 4 byusing chromogenic assays. A small and gradualincrease in the measured 3-hour postdose peakantifactor Xa activity after the fourth dose wasobserved as renal function declined. At 24 hours,no significant difference in antifactor Xa activitywas apparent among the degrees of renalinsufficiency, with the exception of those withsevere renal impairment (creatinine clearance < 30ml/min), who had a 39% lower clearance byantifactor Xa compared with the healthy controls(p=0.0001). No clear association was notedbetween the antifactor Xa activity level andobserved bleeding with the prophylactic dose.Although no conclusion was made about mild-to-moderate renal insufficiency, the results supporteda potential need to reduce the dosage of enoxa-parin in patients with severe renal impairment.The impact of severe renal impairment requiringhemodialysis on enoxaparin was not addressed.

Mahé I, Aghassarian M, Drouet L, et al.Tinzaparin and enoxaparin given at prophylacticdose for eight days in medical elderly patientswith impaired renal function: a comparativepharmacokinetic study. Thromb Haemost2007;97:581–6.

The pharmacokinetics of enoxaparin 40 mg(28 patients) and that of tinzaparin 4500antifactor Xa U (27 patients) were explored inthis prospective, open-label, randomized, two-hospital, parallel-design analysis of a consecutiveseries of patients older than 75 years withconcurrent renal impairment, defined as aestimated creatinine clearance of 20–50ml/minute. Despite a higher antifactor Xaactivity dose of tinzaparin, it still yielded lowermean antifactor Xa activity at days 1 and 8 of0.44 and 0.46 U/ml (p=0.296) compared with0.55 and 0.67 U/ml for enoxaparin (p<0.001).Independently, each agent did not show asignificant correlation between estimatedcreatinine clearance and measured antifactor Xaactivity. However, it should be noted that thesample was small. This suggests that thepresence of renal insufficiency may prolong theclearance of enoxaparin to a greater extent thantinzaparin, and that each agent should beconsidered independently when determiningwhether to reduce the dosage as renal functiondeclines.

Lim W, Dentali F, Elkeboom JW, Crowther MA.Meta-analysis: low-molecular-weight heparin andbleeding in patients with severe renalinsufficiency. Ann Int Med 2006;144:673–84.

In this meta-analysis, antifactor Xa levels andthe frequency of bleeding in patients with acreatinine clearance of either 30 ml/minute orless, or greater than 30 ml/minute were comparedfor treatment or prophylactic dosing of tinzaparin,dalteparin, or enoxaparin. Direct observations orsubgroups from randomized trials that includedpatients with renal insufficiency (excluding thoserequiring hemodialysis) who had received anLMWH were included. Either the Modificationof Diet in Renal Disease or Cockcroft-Gaultequation was used to quantify renal function,creating a potential bias in the meta-analysis.Eighteen studies met the inclusion criteria. Themeasured peak antifactor Xa level at 4 hours wassignificantly higher if the creatinine clearancewas below 30 ml/minute, a finding mostinfluenced by enoxaparin compared withtinzaparin or dalteparin. The use of LMWH wasassociated with an increase in the risk for majorbleeding in patients with creatinine clearance lessthan 30 ml/minute compared to those with acreatinine clearance greater than 30 ml/minute(5.0% vs 2.4%, OR 2.25, 95% CI 1.19–4.27,p=0.013). Major bleeding was also increasedwhen a standard therapeutic dose of enoxaparin

449


was used (8.3% vs 2.4%, OR 3.88, 95% CI1.78–8.45) but was not increased when anempirically adjusted dosage of enoxaparin wasused (0.9% vs 1.9%, OR 0.58, 95% CI 0.09–3.78,p=0.23 for heterogeneity). Not enough data wereavailable to determine whether adjustments inthe dosages of dalteparin or tinzaparin mightreduce the risk of bleeding in patients with renalimpairment.

Lim W, Cook DJ, Crowther MA. Safety andefficacy of low-molecular-weight heparins forhemodialysis in patients with end-stage renalfailure: a meta-analysis of randomized trials. JAm Soc Nephrol 2004;15:3192–206.

Clinical trials frequently exclude patients withrenal failure who require hemodialysis, yet thesepatients have an increased risk for bleeding whenexposed to heparins. These patients may requirelocal anticoagulation to maintain the patency ofthe dialysis circuit or systemic anticoagulationfor treatment or prevention of venous or arterialthrombosis. In this meta-analysis, the charac-teristics of LMWH administered for prevention ofextracorporeal thrombosis in patients with end-stage renal disease requiring long-termhemodialysis were explored. Randomized trialspublished between 1966 and 2004 that used theLMWHs dalteparin (eight trials), nadroparin(five), tinzaparin (three), and enoxaparin (two)during hemodialysis were identified andcompared with respect to bleeding rates and thedevelopment of dialysis-circuit thrombosis.Dosing in antifactor Xa U varied, but was mostlyin the established range for prophylaxis. Notethat ideal targets for antifactor Xa activity in thispopulation have not been established.

Most bleeding was considered minor, primarilyoccurring at vascular access sites. In the 11 trialsthat compared an LMWH with UFH, nosignificant difference in the combined major andminor bleeding rate was observed, with an RRRof 0.96 favoring UFH (95% CI 0.27–3.43,p=0.95). Prevention of thrombosis of theextracorporeal circuit favored UFH, but it wasnot significant, with an RR of 1.15 (95% CI0.7–1.91). However, the UFH and LMWH dosesadministered were not consistent among thetrials. The characteristics of LMWH in thesetting of acute renal failure and associatedapproaches to renal replacement were notexplored. There remains very little literatureabout the use of LMWHs in patients who requiredialysis; therefore, this meta-analysis is animportant consideration for clinicians.

Long-Distance Travel

Scurr JH, Machin SJ, Bailey-King S, et al.Frequency and prevention of symptomless deepvein thrombosis in long-haul flights: arandomised trial. Lancet 2001;357:1485–9.

The true risk of VTE in patients traveling onflights longer than 6 hours in duration isunknown. These investigators attempted todefine the risk of VTE in low-risk patients andestablish the effectiveness of GCS for VTEprevention. Travelers at low risk for VTE andseated in economy class for a flight lasting morethan 8 hours were randomly assigned to nomechanical prophylaxis or below-knee GCS(20–30 mm Hg). Travelers were advised to placethe stockings on before boarding and to removethem immediately on arrival. Low-riskvolunteers were enrolled by excluding those withprevious VTE, cardiorespiratory disease, andmalignancy. The presence of VTE by duplexultrasonography was the primary end point.Secondary end points were described assuperficial thrombophlebitis and D-dimerelevation.

Twelve (10%) of 116 passengers not wearingGCS developed an asymptomatic DVT in the calf.None of the 115 passengers wearing GCSdeveloped a VTE. Thrombophlebitis wasdetected in none of the control patients and in3% of people wearing GCS. No relationship in D-dimer was observed, possibly due to the shorthalf-life of D-dimer and the lag time from arrivalto drawing blood. These data established thedegree of risk associated with VTE in low-riskpatients (~1 in 10). In addition, GCS wereassociated with a reduction in asymptomatic VTEin flights lasting longer than 8 hours.

Cesarone MR, Belcaro G, Nicolaides AN, et al.Venous thrombosis from air travel: the LONFLIT-3 study—prevention with aspirin vs low-molecular-weight heparin (LMWH) in high-risksubjects: a randomized trial. Angiology2002;53:1–6.

The Prevention of Venous Thrombosis inLong-Haul Flights (LONFLIT) trials have studiedvarious questions regarding VTE prevention inthe setting of extended travel times. TheLONFLIT-3 trial evaluated the effectiveness oftwo pharmacologic methods for VTE preventioncompared with a control group. Travelers weredeemed high risk by including those with aprevious VTE, coagulation disorders, severeobesity, malignancy, and large varicose veins.

450


Participants were randomly assigned to receiveplacebo, aspirin 400 mg/day for 3 days starting12 hours before departure, or one dose ofsubcutaneous enoxaparin 1 mg/kg given 2–4hours before departure. All passengers wereeducated on additional preventive measures suchas hydration, in-flight exercises, and avoidance ofrestrictive baggage and clothes. The primaryoutcome was the occurrence of DVT andsuperficial thrombosis confirmed by duplexultrasonography. Thromboses were detected in4.8%, 3.6%, and 0.6% (p<0.002) of passengersassigned to placebo, aspirin, and enoxaparin,respectively. In addition, 85% of thromboticevents occurred in passengers ticketed innonaisle seats. No adverse effects were notedwith enoxaparin; however, 13% of the patientsexperienced gastrointestinal symptoms. Thistrial underscores the need for prophylaxis inhigh-risk patients traveling on long flights, therelative ineffectiveness of aspirin for VTEprevention, and potential need for preferredseating for those at risk for thrombosis. Thesedata also demonstrate the benefits ofpharmacologic prophylaxis when used incombination with other strategies such ashydration and in-flight exercises.

Belcaro G, Cesarone MR, Nicolaides AN, et al.Prevention of venous thrombosis with elasticstockings during long-haul flights: the LONFLIT-5 JAP study. Clin Appl Thromb Hemost2003;9:197–201.

The LONFLIT-5 study evaluated theeffectiveness of GCS in high-risk patientstraveling on flights lasting from 11–13 hours.Travelers were deemed high risk if they had aprevious VTE, coagulation disorders, severeobesity, malignancy, and large varicose veins.Participants were randomly assigned to controlor below-knee GCS (14–17 mm Hg). Subjectswere advised to wear the GCS 3–4 hours beforetheir flight and to remove them after deplaning.All travelers were educated on a mild exerciseregimen, avoidance of restrictive seating, andhydration. The primary outcome was defined asVTE by ultrasonographic scanning. D-dimer andfibrinogen serum levels were measured beforedeparture (within 12 hrs) and after arrival(within 4 hrs). Among the patients assigned toGCS, one developed a distal DVT (0.97%),whereas six patients (5.8%, p<0.0025) in thecontrol group developed a DVT. Among womenwho experienced an event, all had been takinglow-dose oral contraceptive therapy for at least

12 months. D-dimer and fibrinogen levels weresignificantly different between subjects withthrombosis and those without thrombosis. Thesedata demonstrate the effectiveness of GCS inhigh-risk travelers for the prevention ofasymptomatic DVT. In addition, the high rate ofDVT in the control group illustrates the need forincreased awareness of VTE on very long flights.

Clarke M, Hopewell S, Juszczak E, Eisinga A,Kjeldstrøm M. Compression stockings forpreventing deep vein thrombosis in airlinepassengers. Cochrane Database Syst Rev2006;(2):CD004002.

This Cochrane review assesses the effectivenessof GCS in passengers on flights lasting at least 4hours. Ten randomized trials, representing atotal of 2856 people, were included in thisanalysis. Nine of the trials examined theeffectiveness of bilateral GCS, and one trialinvestigated unilateral GCS. Eight of the 10 trialsincluded experiences from the LONFLIT studies.The primary outcome was the diagnosis ofsymptomatic or asymptomatic DVT byultrasonography or venography. Secondaryanalyses included the diagnosis of pulmonaryembolism, death, superficial vein thrombosis,and adverse effects. A total of 50 peopledeveloped asymptomatic DVT (three in the GCSgroups and 47 in the placebo group). None ofthe participants experienced symptomatic DVT,pulmonary embolism, or adverse effects. A totalof 16 people developed superficial veinthrombosis (four in the GCS groups and 12 inthe placebo group). This review confirms thatuse of GCS reduces asymptomatic DVT fromapproximately 10/1000 to 2 or 3/1000. Athorough review of potential pathophysiologicmechanisms responsible for the development ofthromboses during air travel is included.

Philbrick JT, Shumate R, Siadaty MS, BeckerDM. Air travel and venous thromboembolism: asystematic review. J Gen Intern Med 2007;22:107–14.

Although many tertiary resources are availablefor VTE prevention during long flights, most ofthese analyses focus on a single prophylacticstrategy. These investigators included 25 trialsthat reported primary outcomes for variouspreventive VTE measures during long-distancetravel. Four variables were significantlypredictive of VTE: method of screening(ultrasonography, OR 390%, p<0.0001), type ofprimary outcome (DVT only, OR 23%; all VTE,

451


OR 21%, p<0.0001), clinical risk of thrombosis(high risk, OR 3.6%, p<0.0001), and duration oflong distance travel (> 8 hrs, OR 2.3%,p<0.0001). Overall, asymptomatic VTE wasmore common than symptomatic DVT andpulmonary embolism. Graduated compressionstockings and LMWH prevented VTE associatedwith long-distance travel, whereas aspirinprovided no clinical benefit. The authorsrecommend hydration and frequent in-flightexercises for all patients. Travelers with lowthrombotic risk on a flight lasting less than 6hours require no prophylaxis. Those passengerswith at least one risk factor for thrombosis orflying longer than 6 hours should consider GCSor LMWH. This article serves as an excellentresource with a very thorough summary (detailedtable spanning three pages) of all the major trialsinvestigating long-flight VTE prophylaxis.

Mechanical Prophylaxis

Cornwell EE, Chang D, Velmahos G, et al.Compliance with sequential compression deviceprophylaxis in at-risk trauma patients: aprospective analysis. Am Surg 2002;68:470–3.

Sequential compression devices (SCDs) havebecome an alternative in the prevention of VTE.Their popularity has grown in patients withcontraindications to UFH; however, anecdotalevidence has suggested that compliance withthese devices is poor. The investigators evaluatedcompliance with SCDs in nonambulatory traumapatients. Six observations were made during a24-hour period to assess the presence andfunctioning of SCDs in 227 patients. Fullcompliance was defined as SCDs that were onand functioning at all six observations. Only 42patients (19%) were classified as fully compliant.Of the 1343 available observations, SCDs were onand functioning in only 712 observations (53%).The most common periods of noncompliancewere the mid-morning and early afternoon. Inthose patients reported as noncompliant, onethrombotic risk factor was present in 83% andmultiple risk factors in 41%. Some cliniciansview SCDs as a prophylactic option relativelyeasy to use and order. This trial demonstrates theneed for continued education and may highlightkey hospital shifts to initially target. Theeffectiveness of SCDs has been reported;however, these findings have occurred in thewell-controlled environment of a clinical trial.

Clarke-Pearson DL, Dodge RK, Synan I,McClelland RC, Maxwell GL. Venous thrombo-

embolism prophylaxis: patients at high risk to failintermittent pneumatic compression. ObstetGynecol 2003;101:157–63.

Current guidelines suggest IPC as an option forVTE prophylaxis in patients undergoinggynecologic surgery. The investigators conducteda retrospective review to identify patients whofailed IPC. In those patients who experienced anevent, an association between VTE risk factorsand the occurrence of VTE was measured. Theoverall frequency of VTE in 1862 patientsundergoing gynecologic surgery was 1.3%. Theseevents consisted of 9 DVTs and 15 pulmonaryembolisms. After multivariable regressionanalysis, three independent prognostic factorswere identified: cancer (OR 3.8, p=0.001),history of DVT (OR 5.3, p=0.006), and age olderthan 60 years (OR 2.6, p=0.04). Patients withtwo or three of these significant risk factors had a3.2% rate of VTE compared with 0.6% in patientswith zero or one risk factor. These data affordclinicians the ability to screen or stratify patientsundergoing gynecologic surgery by VTE risk.Clinicians treating women with more than onerisk factor should consider pharmacologic ratherthan mechanical strategies to prevent VTE.

Prandoni P, Lensing AW, Prins MH, et al. Below-knee elastic compression stockings to prevent thepost-thrombotic syndrome: a randomized,controlled trial. Ann Intern Med 2004:141;249–56.

The management of DVT often includes thegoal of pulmonary embolism prevention.However, postthrombotic syndrome occurs innearly 30% of patients with a proximal DVT andis often underrecognized. These investigatorsevaluated the efficacy of GCS for the preventionof postthrombotic syndrome over 2 years.Patients who had developed VTE were randomlyassigned to no stockings or GCS with 30–40 mmHg pressure. Patients also received standardpharmacologic therapy for 3–6 months. Astandardized scale was used to evaluate thepresence and severity of postthromboticsyndrome. Postthrombotic syndrome waspresent in 44 (49%) of 90 patients assigned to nostockings compared with 23 (26%) of 90 patientsusing GCS. Of those control patients whodeveloped postthrombotic syndrome, 10 caseswere considered severe compared with threecases in the patients using GCS. Most diagnosesof postthrombotic syndrome (40% in the controlgroup and 21% in the GCS group) were made inthe first 6 months, with most patients receiving

452


oral anticoagulation. Two independent riskfactors for the development of postthromboticsyndrome were identified: ipsilateral DVT (HR3.32, p=0.04) and 10-year increments of age (HR1.36, p=0.003). This landmark trial confirms thehigh rate of postthrombotic syndrome (50%),demonstrates the effectiveness of GCS incombination with anticoagulation (approxi-mately 50% reduction), and identifies factorsassociated with development of postthromboticsyndrome.

The PREPIC Study Group. Eight-year follow-upof patients with permanent vena cava filters inthe prevention of pulmonary embolism: thePREPIC (prevention du risque d’emboliepulmonaire par interruption cave) randomizedstudy. Circulation 2005;112:416–22.

Long-term safety and efficacy data on venacava filters have been scarce and unreliable. ThePrevention du Risque d’Embolie Pulmonaire parInterruption Cave (PREPIC) trial, the largestvena cava filter trial to our knowledge, reported areduction in pulmonary embolism at the 2-yearfollow-up. However, this benefit was over-shadowed by an increase in DVT and nomortality benefit. The PREPIC 8-year follow-upwas performed to assess the long-term safety andefficacy of permanent vena cava filter insertion.Patients with a diagnosis of acute proximal DVTand considered high risk for the development ofpulmonary embolism were enrolled. Pulmonaryembolism occurred in 9 patients with a filter(6.2%) and 24 patients without a filter (15.1%,p=0.008). Deep vein thrombosis was discoveredin 57 patients with a filter (35.7%) and 41patients without a filter (25.7%, p=0.042). Nosignificant difference in rates of postthromboticsyndrome and mortality were observed. Ofinterest, 24 patients without a filter (46%)developed a pulmonary embolism despitereceiving vitamin K antagonist therapy. Severalsubgroups of patients were identified as high riskfor the development of pulmonary embolism,including patients with a pulmonary embolism atthe time of diagnosis, patients with cancer, andnonsurgical patients who developed VTE. Thedelicate balance of safety and efficacy is evidentin the PREPIC trial. Filter insertion resulted in anumber needed to treat of 12 for pulmonaryembolism, but this was offset by a numberneeded to harm of 10 for DVT. The PREPIC trialjustifies the need for extended prophylaxis, aswell as combination strategies, in high-riskpatients.

Amaragiri SV, Lees TA. Elastic compressionstockings for prevention of deep vein thrombosis.Cochrane Database Syst Rev 2000;(1):CD001484.

Many of the trials evaluating GCS for theprevention of DVT were small and underpowered.This Cochrane review evaluated the effectivenessof GCS in 16 randomized trials involving morethan 2000 patients. This review was designed tomeasure the effectiveness of GCS in hospitalizedpatients; however, most patients were surgicalrather than medically ill patients. Trials wereorganized into two groups based on whether GCSwere used alone or in combination withanticoagulation. In patients randomized to GCSalone, 15% of patients in the GCS group and 29%in the control group developed a VTE(p<0.00001). In trials using dual prophylacticstrategies, VTE was experienced in 3% of thepatients wearing GCS plus another methodcompared with 14% of the control patients. Ineither scenario, GCS appear to be an effectivestrategy to reduce the rate of VTE. Several keypatient demographics and GCS variables wereunderrepresented in this review, and care shouldbe taken when extrapolating the results. Inaddition to the fact that no low-risk patients wereincluded in this review, the effect of GCSduration, length, and pressure were not assessed.Also, the GCS were compared with control ratherthan anticoagulation, limiting the results of theanalysis.

Urbankova J, Quiroz R, Kucher N, et al.Intermittent pneumatic compression and deepvein thrombosis prevention: a meta-analysis inpostoperative patients. Thromb Haemost2005;94:1181–5.

Although pharmacologic VTE prophylaxis isthe preferred strategy in moderate- or high-riskpatients undergoing surgery, many of thesepatients have contraindications to anticoagu-lation. Investigators performed a meta-analysisto evaluate the effectiveness of IPC in patientswho had surgery. A total of 2270 patients from15 trials (surgeries included general, orthopedic,neurosurgery, oncologic, and urologic) wereeligible for inclusion. Initiation of IPC varied intrials and was classified as preoperative (beforeanesthesia), perioperative (immediately afteranesthesia), and postoperative (< 24 hrs aftersurgery). The RRR of DVT was 60% (RR 0.40,95% CI 0.29–0.56, p<0.001) in patients wearingIPC compared with no IPC. The use of IPC didnot reduce pulmonary embolism compared withthe control group (p=0.7). In addition, the

453


occurrence of DVT did not differ significantlybetween patients for whom IPC was startedbefore surgery and those for whom IPC wasstarted after surgery. These data ensure cliniciansthat patients undergoing surgery do benefit frominitiation of IPC for VTE prophylaxis. Thisreview also demonstrates that the timing of IPCinitiation around surgery is not clinicallyimportant.

Mazzone C, Chiodo Grandi F, Sandercock P, etal. Physical methods for preventing deep veinthrombosis in stroke. Cochrane Database SystRev 2004;(4):CD001922.

Although uncommon, DVT and pulmonaryembolism are clinically important complicationsof stroke. Anticoagulants reduce the risk of VTE;however, this benefit is offset by the risk ofhemorrhage. This Cochrane review was conductedto assess the effectiveness of mechanicalprophylaxis for VTE prevention in patients withstroke. Despite the large number of studiesconsidered for this analysis, the stringentinclusion criteria allowed only two studies (123patients) to be included. Overall, mechanicalprophylaxis methods were not associated with adecreased rate of DVT in survivors (OR 0.54,95% CI 0.18–1.57) or in those who died (OR1.54, 95% CI 0.5–4.77). This review demonstratesthe disparity of data in this high-risk group ofpatients. Although no firm conclusions can bedrawn from this analysis regarding the efficacy ofmechanical prophylaxis, the strategy appears safein patients with stroke.

Extended Prophylaxis

Eikelboom JW, Quinlan DJ, Douketis JD.Extended-duration prophylaxis against venousthromboembolism after total hip or kneereplacement: a meta-analysis of the randomisedtrials. Lancet 2001;358:9–15.

This trial is discussed in the OrthopedicSurgery section; however, a discussion in thissection is also warranted. This meta-analysis ofnine studies involving 3999 patients helpedestablish the current ACCP recommendations toconsider extending prophylaxis beyond 10 daysand up to 35 days after THR and TKR surgeries.Although all individual studies included in theanalysis had shown a reduction in symptomaticVTE after 4–6 weeks of extended prophylaxiscompared with in-hospital prophylaxis, statisticalsignificance was reached in only two of ninestudies. The combined results showed a

significant decrease in symptomatic VTE (3.3%vs 1.3%, OR 0.38). Similar RR was demonstratedfor proximal DVT (9.1% vs 2.9%, OR 0.33) andsymptomatic pulmonary embolism (0.6% vs0.2%, OR 0.43), but the pulmonary embolismend point did not reach significance secondary tothe relatively low event rate. Major bleeding wasrare (0.2%) overall in the out-of-hospital periodand did not differ significantly between groups,but there was a small but significant increase inminor bleeding (3.7% vs 2.5%) in the extendedprophylaxis group. Results did not varyaccording to the duration of in-hospitalprophylaxis (4–15 days) or whether or notmandatory venography was required in theindividual trial before hospital discharge.

An important prespecified subgroup analysisshowed that the reduction in symptomatic eventswas greater for extended prophylaxis versus in-hospital prophylaxis in patients undergoing THR(1.3% vs 4.3%, OR 0.33) compared with thoseundergoing TKR (1.0% vs 1.4%, OR 0.74). Itshould be noted that only two studies examinedboth TKR and THR, whereas seven examinedTHR exclusively. However, the findings areconsistent with those of previous studies, whichdemonstrated that a greater percentage of VTEafter TKR are asymptomatic and that the timecourse of symptomatic events with TKR occursearlier than with THR. Hence, the ACCPrecommendations for extended prophylaxis withTHR (grade 1A) are stronger than those for TKR(grade 2B).

Hull RD, Pineo GF, Stein PD, et al. Extendedout-of-hospital low-molecular-weight heparinprophylaxis against deep venous thrombosis inpatients after elective hip arthroplasty: asystematic review. Ann Intern Med 2001;135:858–69.

This review included six of the nine studiesincluded in the meta-analysis above but provideda more focused examination, since it was limitedto elective THR. The LMWHs were the soleextended prophylaxis option, and all trials usedplacebo as the comparator. Bilateral ascendingvenography was the only diagnostic test allowedfor identifying DVT in these studies. The studieswere all double-blind and involved more than1950 patients receiving once-daily enoxaparin(three studies) or once-daily dalteparin (threestudies) for 18–29 days after hospitalization.Combined results showed significant reductionsin favor of extended LMWH prophylaxis for allDVT (22.5% vs 7.9%, RR 0.41), proximal DVT

454


(11.2% vs 3.0%, RR 0.31), and symptomatic DVT(4.2% vs 1.4%, RR 0.36). Unlike the previousmeta-analysis, these benefits were realizedwithout a significant difference in minor bleeding(2.66% with LMWH vs 2.44% with placebo).Major bleeding again was rare, with report ofonly one case in the placebo group and none inthe LMWH group during the out-of-hospitalperiod. The review provides valuable infor-mation from the individual trials such as thenumber of patients with previous VTE or cancerand the proportion undergoing successfulvenography. The length of inpatient prophylaxisdid not impact the subsequent results.

The relatively low 15 patients needed to treatto prevent one proximal DVT or 45 to preventone symptomatic VTE with minimal bleedingrisk in the posthospital period in this reviewprovide the basis for the consideration ofextended prophylaxis for THR, and serves as abenchmark of comparison for extendedprophylaxis in other clinical scenarios. Mostcost-effectiveness evaluations have concludedthat extended prophylaxis for THR is cost savingor a worthwhile expense given the eventsprevented. Oral vitamin K antagonists have sincedemonstrated efficacy with extended prophylaxisin THR, but their benefit in clinical trials hasbeen offset by a higher risk of bleeding than withLMWH.

Eriksson BI, Lassen MR, for the Pentasaccharidein Hip Fracture Surgery Plus Investigators.Duration of prophylaxis against venousthromboembolism with fondaparinux after hipfracture surgery: a multicenter, randomized,placebo-controlled, double-blind study. ArchIntern Med 2003;163:1337–42.

The Pentasaccharide in Hip-Fracture Surgery(PENTHIFIRA) Plus trial was a multicenter trialthat examined 656 patients receiving 6–8 days offondaparinux 2.5 mg/day after undergoing hipfracture surgery and followed up by random-ization in a double-blind fashion to receiveplacebo or continued fondaparinux for 19–23days. Bilateral venography was performed at25–32 days. This population was selected afterpast studies had shown a higher rate of fatalpulmonary embolism in patients undergoing hipfracture surgery than in those undergoing THRand a majority of symptomatic events occurringbeyond day 11 after surgery. Significantreductions were seen with extendedfondaparinux for all VTE (35.0% vs 1.4%) andproximal DVT (15.8 vs 0.9%). For symptomatic

VTE, the reduction was significant (2.7% vs0.3%) but with much fewer events. Symptomaticpulmonary embolism was observed in threepatients, all receiving placebo, with one beingfatal. Major bleeding (2.4% vs 0.6%, p=0.06),minor bleeding (1.5% vs 0.6%), and need fortransfusions (8.9% vs 6.1%) were allinsignificantly higher with fondaparinux thanplacebo. Probably owing to the advanced ageand underlying frailty of the patient population,the mean length of hospitalization during thisstudy was 13.5 days. A lower percentage ofpatients than seen in THR studies, 65%, hadadequate venography to evaluate DVT. Thisstudy demonstrates a very impressive RRR ofapproximately 90% for fondaparinux in hipfracture surgery for all efficacy end points. Theabsolute RR of 2.4% for symptomatic VTE issimilar to that seen for LMWH extendedprophylaxis in THR. As a result of this trial, theACCP recommends that VTE prophylaxis beextended beyond 10 days and up to 35 days forhip fracture surgery (grade 1A), andfondaparinux carries the only grade 1A forextended prophylaxis in patients undergoing hipfracture surgery.

Bergqvist D, Agnelli G, Cohen AT, et al.Duration of prophylaxis against venousthromboembolism with enoxaparin after surgeryfor cancer. N Engl J Med 2002;346:975–80.

This trial is discussed in the Cancer section,but a discussion here is also warranted. In one ofthe first major trials to examine extendedprophylaxis outside of orthopedic surgery,ENOXACAN II, 332 patients having undergoneplanned curative surgery for abdominal or pelviccancer and subsequent prophylaxis with 6–10days of open-label enoxaparin 40 mg/day wererandomly assigned to either continuation ofenoxaparin 40 mg or placebo daily for 21 moredays. Over 80% of the surgeries were performedon the gastrointestinal tract, whereas approxi-mately 8% involved female reproductive organs.Bilateral venography was then performed at25–31 days. Patients were followed 3 months, inpart to see if extended prophylaxis merelydelayed, rather than prevented, the onset of VTE.A significant reduction in the primary end pointof all VTE was seen with enoxaparin (12.0% vs4.8%, RR 60%, p=0.02) compared with placebo.However, of the 28 events reported during the21-day double-blind period, 27 of them wereasymptomatic and 24 were distal vein DVTs.Only one pulmonary embolism was detected

455


during that time frame. Therefore, with theserelatively low event rates, the absolute RRs forproximal DVT (1.8% vs 0.6%, absolute RR 1.2%,number needed to treat 83) and symptomaticVTE (0.6% vs 0%, number needed to treat 167)were not significant. Bleeding events during the21-day double-blind period were not significantlydifferent between enoxaparin and placebo (majorbleeding 0.4% vs 0% and minor bleeding 4.7% vs3.6%). During the period between venographyand the end of the 3-month follow-up, three ofthe four symptomatic VTE events that occurredwere from the group that had received placebo.This demonstrated that extended prophylaxishelped to prevent VTE formation and did notmerely delay its onset.

Hull RD, Schellong SM, Tapson VF, et al.Extended-duration thromboprophylaxis inacutely ill medical patients with recent reducedmobility: methodology for the EXCLAIM study. JThromb Thrombolysis 2006;22:31–8.

This article details the methodology of theExtended Clinical Prophylaxis in Acutely IllMedical Patients (EXCLAIM) trial, which wasdesigned after results of previous studies inmedically ill patients demonstrated that anincreased risk of VTE may continue for anextended duration and suggested a potentialmortality benefit if an adequately powered trialwere performed. The EXCLAIM trial enrolledmore than 5100 hospitalized patients in 264centers in 18 countries who had recent reducedmobility and at least one other risk factor forVTE. All patients received open-label enoxaparin40 mg/day for a mean ± SD of 10 ± 4 days beforebeing randomly assigned to double-blind therapywith either continued enoxaparin 40 mg orplacebo daily for an additional 28 ± 4 days. Atthe end of the double-blind treatment period, allpatients underwent bilateral compressionultrasonographic testing, which was differentfrom most trials in which bilateral venographywas used. Patients were then followed for anadditional mean ± SD of 142 ± 10 days forsymptomatic events.

Some of the results of the trial have beenannounced but have not been formally publishedas of the time of this writing. A planned,blinded, interim analysis allowed the steeringcommittee to redefine some inclusion criteria toallow enrollment of patients with a higher riskfor VTE (recent immobility plus age > 75 yrs,history of VTE, or previous diagnosis of cancer).How this affected who constituted the final study

population awaits further examination. This isespecially important in light of the announcedresults (presented at the International Society ofHematology meeting in 2007) that demonstratedan impressive, significant 44% RRR in all VTE(4.9% vs 2.8%) and 73% in symptomatic VTE(1.1% vs 0.3%) in favor of prolonged LMWHprophylaxis. It may be argued that ultrasono-graphic examination is more reflective of currentpractice, but that greater absolute reductions mayhave been realized by using a more sensitive testsuch as venography. However, unlike most othertrials of extended prophylaxis, significantly moremajor (0.15% vs 0.6%) and minor (3.7% vs 5.2%)bleeding events were observed with extendedprophylaxis, and the all-cause mortality at 6months did not differ significantly betweengroups. The final results of the trial are awaited,including subgroup analyses to determinespecific populations in whom extendedprophylaxis is most beneficial.

Quality Improvement Initiatives

Goldhaber SZ, Tapson VF, for the DVT FREESteering Committee. A prospective registry of5451 patients with ultrasound-confirmed deepvein thrombosis. Am J Cardiol 2004;93:259–62.

Insights on the epidemiology of DVT wereexplored in this prospective analysis from 183U.S. hospitals that used a registry of patients whohad a DVT confirmed with ultrasonographicexamination. There were no exclusion criteria.A total of 5451 surgical and medical patientswere enrolled and were reclassified as having theDVT in either the in-hospital or outpatientsetting. Diagnosis of symptomatic DVT wassooner if the patient was in the hospital (after 1day of symptoms) compared with the outpatientsetting (after 3 days of symptoms). Use ofLMWH as a bridge to warfarin was the mostcommon approach used in either setting fortreatment. One of the most notable findings wasthat 71% of patients with a DVT had not receivedprophylaxis within the previous 30 days.However, it should be taken into account thatpresence of bleeding concerns and hesitation touse prophylaxis were not considered. Regardless,this large multicenter study clearly points out theneed to improve VTE prophylaxis.

Tapson VF, Decousus H, Pini M, et al, for theIMPROVE Investigators. Venous thrombo-embolism prophylaxis in acutely ill hospitalizedmedical patients: findings from the international

456


medical prevention registry on venous thrombo-embolism. Chest 2007;132:936–45.

This international observational study used aregistry to assess how VTE prophylaxis was usedin acutely ill medical patients. Data wereabstracted after hospital discharge of 15,156patients in 52 hospitals from 12 countries. Ofnote is that 55% of the 24,585 initially assessedpatients were excluded from the analysis.Prophylaxis was administered in 61% of patients,with 52% and 43% of patients from the UnitedStates and other countries, respectively, receivingprophylaxis according to the ACCP guidelines inplace at that time. In the United States, UFH wasthe most common form of prophylaxis comparedwith LMWH in other countries. Overall, thisanalysis points out the need to improveprophylaxis on a global basis.

Agency for Healthcare Research and Quality,U.S. Department of Health and Human Services.Surgical care improvement project (SCIP)guidelines. Available from www.qualitymeasures.ahrq.gov. Accessed September 12, 2008.

Surgical procedures significantly increase therisk of developing VTE. Despite the over-whelming evidence, including numeroussupporting trials, the use of VTE prophylaxiscontinues to be inadequate. The Surgical CareImprovement Project (SCIP) focuses ondecreasing postoperative complications andimproving postoperative outcomes. To minimizethese risks, including postoperative VTE, theCenters for Medicare and Medicaid Servicesalong with the Joint Commission have developedmeasures to ensure VTE prophylaxis was ordered(SCIP 1) and provided within 24 hours beforeand after surgery (SCIP 2). Additional measuresare in development to assess the total number ofhospitalized adult patients who had a subsequentadmission within 30 days for conditions relatedto VTE. Health systems may need to expandmanagement plans beyond the time the patient iswithin their sphere of influence, as the riskperiod for VTE may continue after discharge, toavoid both being penalized with reductions inpay-for-performance as well as having tointernally cover the cost of care for the subsequentVTE. The Centers for Medicare and MedicaidServices have recently announced that it will nolonger pay for hospital-acquired “never events,”including VTE after orthopedic surgery.

The Joint Commission. National consensusstandards for prevention and care of venous

thromboembolism (VTE). Available fromwww.jointcommission.org/PerformanceMeasurement/PerformanceMeasurement/VTE.htm.Accessed September 12, 2008.

In 2005, the Joint Commission and the NationalQuality Forum collaborated on developingconsensus standards for the prevention or care ofVTE. The measures encompassed prevention,treatment, and outcomes. In the summer of2008, the National Quality Form endorsed six ofthe original 19 initial measures. Implementationof these measures is scheduled to occur in the fallof 2009. Among various requirements, thesemeasures will encourage the provision of VTEprophylaxis during admission and at transfer intoor out of the ICU. They will also require appro-priate dosing and monitoring (including plateletcounts) during UFH therapy. In addition,discharge instructions for individuals whodeveloped VTE during hospitalization, andassessment of the rate of potentially preventableVTE will be required of all health care institutions.

The Joint Commission. 2008 national patientsafety goals hospital program. Available fromht tp : / /wwwjo intcommiss ion .org /Pa t i entSafety/NationalPatientSafetyGoals/08_hap_npsgs.htm. Accessed September 12, 2008.

The Joint Commission has recognized theimportance of providing safety and quality ofhealth care through various RR strategies.Anticoagulation therapy has been recognized as asource of frequent drug-related adverse events,many of which may cause harm to the patient.Starting in January 2009, National Patient SafeGoals 3E requires that all health care settingsimplement programs to ensure the safe use ofanticoagulants. The nine goals encompassvarious aspects of the use and delivery ofanticoagulants, as well as the education ofproviders and patients regarding antithrombotictherapy. Although only three classes ofanticoagulants are currently targeted (UFH,LMWH, and warfarin), a comprehensive programwill include defined programs for the use ofdirect thrombin inhibitors as well. This initiativehas positioned the use of anticoagulation therapyas a critical issue for health systems and isresulting in the development of improvedmanagement programs across the United States.

Kucher N, Koo S, Quiroz R, et al. Electronicalerts to prevent venous thromboembolismamong hospitalized patients. N Engl J Med2005;352:969–77.

457


The benefits of an electronic alerting systemwere evaluated in this observational analysis ofpatients randomly assigned to no intervention(1251 patients) and those whose managingphysician received an electronic alert (1255patients) regarding individualized risk of VTEand links to VTE prophylaxis guidelines andorder sets. The use of mechanical (10.0% vs1.5%, p<0.001) or pharmacologic (23.6% vs13.0%, p<0.001) prophylaxis was higher whenthe physician was notified than when nonotification was received. The primary end pointwas development of VTE at 90 days, whichoccurred in 4.9% of the intervention groupversus 8.2% of the control group. Because thestudy was nonblinded, potential diagnostic biascannot be excluded. Nevertheless, this study isthe most robust of many intervention studies andsuggests that alerting systems can improve therate of VTE prophylaxis and lower the risk ofclinical VTE.

Tooher R, Middleton P, Pham C, et al. Asystematic review of strategies to improveprophylaxis for venous thromboembolism inhospitals. Ann Surg 2005;241:397–415.

This review explored strategies from 30 studiespublished between 1996 and 2003 that were usedto improve VTE prophylaxis. Studies notshowing a positive impact were excluded. Onlyone randomized controlled trial was located, withthe remainder including audits related toimplementation of a practice guideline or protocol.Adherence to guidelines in general was found tobe consistently poor, with no more than 50%receiving prophylaxis. Higher VTE prophylaxisrates were observed when implementing anactive or multiple strategy approach. Presence ofan audit process and providing feedback orhaving an active reminder in place appeared tohave the highest success rates. Because of thenature of how the assessments were conducted,notable bias could be present. The samples ofthe studies were not large enough to provide hardoutcomes such as reduction in the rate of VTE.Nevertheless, the findings are significant, and thearticle provides an excellent review of themultiple studies that have evaluated variousstrategies to improve VTE prophylaxis rates.

Yu HT, Dylan ML, Lin J, Dubois RW. Hospitals’compliance with prophylaxis guidelines forvenous thromboembolism. Am J Health SystPharm 2007;64:69–76.

The ACCP guidelines are considered to be akey document that guides approaches to VTEprophylaxis. In this retrospective report of anational data warehouse comprising data on 1.3million hospitalized patients, by using selectedInternational Classification of Diseases, NineRevision codes, the compliance rate to theguidelines was assessed in individuals older than40 years. Of the 123,304 admissions that met theinclusion and exclusion criteria, only a 13.3%compliance rate was noted. Compliance washigher in selected populations such as orthopedicsurgery, but 16% or less in the remainingpatients. Only 23.4% received some form ofprophylaxis. These observations support theneed for improved use of appropriate VTEprophylaxis, which is a focus of pendinginitiatives from the Joint Commission. Further,the SCIP guidelines, which have so far used therecommendations of the ACCP conference todefine prophylaxis, will also create incentives forfacilities to be compliant.

References

1. Geerts WH, Bergqvist D, Pineo GF, et al. Prevention of venousthromboembolism: American College of Chest Physiciansevidence-based clinical practice guidelines, 8th ed. Chest2008;133(suppl 6):S381–453.

2. Zhan C, Miller MR. Excess length of stay, charges, and mortalityattributable to medical injuries during hospitalization. JAMA2003;290:1868–74.

3. Dobesh PP, Brouse SD, Johnson DC, et al. Key articles andguidelines relative to treatment of patients with acute coronarysyndromes. Pharmacotherapy 2004;24:105–44.

4. Cheng JWM, Frank L, Garrett SD, Lu Y, Sanoski CA, WhiteCM. Key articles and guidelines in pharmacotherapeuticmanagement of arrhythmias. Pharmacotherapy 2004;24:248–79.

5. de Denus S, Hardy AM, Olsen KL, Robinette B. Key articles andguidelines in the management of hypertension. Pharmaco-therapy 2004;24:1385–99.

6. Ng TMH, Carter O, Guillory G, et al. High-impact articlesrelated to the pharmacotherapeutic management of systolic heartfailure. Pharmacotherapy 2004;24:1594–633.

7. Ito MK, Cheung RJ, Gupta EK, et al. Key articles, guidelines,and consensus papers relative to the treatment ofdyslipidemias–2005. Pharmacotherapy 2006;26:939–1010.

8. Dobesh PP, Brouse SD, Dager WE, Spinler SA, Stacy Z,Wiggins BS. Key articles and guidelines in the management ofacute coronary syndromes and in percutaneous coronaryintervention: 2007 update. Pharmacotherapy 2007;27:1722–58.

458

https://www.researchgate.net/publication/283600488_Prevention_of_venous_thromboembolism_American_College_of_Chest_Physicians_Evidence-Based_Clinical_Practice_Guidelines?el=1_x_8&enrichId=rgreq-c212c3af8f49ba065adb23a8d9bff502-XXX&enrichSource=Y292ZXJQYWdlOzI0MjM2NDIwO0FTOjEzOTA5NDA2MzU4NzMyOEAxNDEwMTc0MDE3MTMy




https://www.researchgate.net/publication/9061205_Excess_Length_of_Stay_Charges_and_Mortality_Attributable_to_Medical_Injuries_During_Hospitalization?el=1_x_8&enrichId=rgreq-c212c3af8f49ba065adb23a8d9bff502-XXX&enrichSource=Y292ZXJQYWdlOzI0MjM2NDIwO0FTOjEzOTA5NDA2MzU4NzMyOEAxNDEwMTc0MDE3MTMy



https://www.researchgate.net/publication/6495695_Key_Articles_and_Guidelines_in_Pharmacotherapeutic_Management_of_Arrhythmias?el=1_x_8&enrichId=rgreq-c212c3af8f49ba065adb23a8d9bff502-XXX&enrichSource=Y292ZXJQYWdlOzI0MjM2NDIwO0FTOjEzOTA5NDA2MzU4NzMyOEAxNDEwMTc0MDE3MTMy



https://www.researchgate.net/publication/5805743_Key_Articles_and_Guidelines_in_the_Management_of_Acute_Coronary_Syndromes_and_in_Percutaneous_Coronary_Intervention_2007_Update?el=1_x_8&enrichId=rgreq-c212c3af8f49ba065adb23a8d9bff502-XXX&enrichSource=Y292ZXJQYWdlOzI0MjM2NDIwO0FTOjEzOTA5NDA2MzU4NzMyOEAxNDEwMTc0MDE3MTMy




key articles and guidelines for the prevention of venous thromboembolism

Documents