Author's Response to Letters to the Editor Regarding the Article Entitled "Anticoagulation Therapy for Venous Thromboembolism"

David K. Cundiff


January 03, 2005

Dr. Hoyt refers to the 6 placebo-controlled, randomized trials of low-molecular-weight heparins (LMWHs) cited in the Seventh American College of Chest Physicians (ACCP) Conference on Antithrombotic and Thrombolytic Therapy regarding the prevention of venous thromboembolism (VTE).[1] My VTE review dealt with the treatment of deep vein thrombosis (DVT) and pulmonary emboli (PE) for which no placebo-controlled trials of LMWHs or thrombolytics have ever been published. Likewise, the .5% major bleeding rate in the PREVENT Thromboprophylaxis trial is not in the 1% to 33% major bleeding-rate range that I quoted for VTE treatment trials.

Since Dr. Hoyt brought up the prophylaxis in MEDical patients with ENOXaparin (MEDENOX) and PREVENT Thromboprophylaxis trials and US Food and Drug Administration (FDA) regulators 4 years ago and argued with me that randomized, controlled trials (RCTs) showing the safety and efficacy of anticoagulants in VTE prevention justify anticoagulant treatment for VTE despite flawed or inconclusive treatment RCTs, let me briefly address these 2 trials and the topic of prevention of VTE in medical and surgical inpatients. In MEDENOX trial,[2] neither the incidence of fatal PE by day 14 (0 of 362 in the placebo group vs 0 of 711 in the enoxaparin groups [combining the 20-mg/day and 40-mg/day groups]), the overall incidence of death by day 14 (16 of 362 [4.4%] in the placebo group vs 27 of 761 [3.8%] in the enoxaparin group, P = .48), nor the overall death rate by 80-110 days (50 of 362 [13.9%] in the placebo group vs 92 of 711 [12.9%] in the enoxaparin groups, P = .69) are significantly different. Likewise, In the PREVENT Thromboprophylaxis trial,[3] neither the incidence of fatal PE by day 21 (2 of 1829 in the placebo group vs 0 of 1807 in the dalteparin group, P = .16), the overall incidence of sudden death by day 21 (3 of 1829 in the placebo group vs 5 of 1807 in the dalteparin group, P = .47), nor the overall death rate by 90 days (103 of 1747 in the placebo group vs 107 of 1715 in the dalteparin group, P = .67) are significantly different.

Although it is also not statistically significant in these 2 trials, 0 of 362 placebo-treated patients died of bleeding by day 80-110 vs 3 of 711 (.4%) enoxaparin-treated patients,[2] and 1 of 1829 placebo-treated patient vs 2 of 1807 (.1%) dalteparin-treated patients died of bleeding by day 21.[3] This is consistent with meta-analyses of thromboprophylaxis trials showing significantly increased rates of major bleeding in patients receiving vitamin K inhibitors, unfractionated heparin and LMWHs in orthopaedic surgery patients,[4] higher doses of LMWHs in general surgery patients,[5] unfractionated heparin in general medical patients,[6] unfractionated heparin in acute myocardial infarction patients,[7] and unfractionated heparin and LMWHs in stroke patients.[8] Overall, in surgical patients, heparin and LMWHs significantly increase the rate of major bleeding compared with placebo or no prophylaxis (relative risk [RR] = 1.75, 95% confidence interval [CI] = 1.53-2.01) and compared with aspirin (RR = 1.24, 95% CI = 1.12-1.37).[9] Published meta-analyses of surgical and medical patients receiving prophylactic anticoagulation vs placebo have not included the RR of fatal bleeding.

About 12 million Americans receive heparin or LMWHs for prophylaxis each year.[10] Based on LMWHs increasing the risk of fatal bleeding by 3 of 711 cases as in MEDENOX or 1 of 1807 patients as in the PREVENT Thromboprophylaxis trial and 12 million hospitalized Americans receiving prophylactic LMWHs, over 50,000 and over 6600 excess deaths due to bleeding would occur, respectively. Published safety data on heparin and LMWH uses for prophylaxis of hospitalized patients do not rule out that this many or more Americans may die each year of bleeding due to current ACCP guidelines for prophylactic anticoagulants.

According to the CIs around the risk ratios of total mortality at 80-110 days in MEDENOX (50 of 362 placebo-treated patients vs 92 of 711 enoxaparin-treated patients, RR = .9274, 95% CI = .6404-1.3430) and total mortality at 90 days in the PREVENT Thromboprophylaxis trial (103 of 1747 vs 107 of 1715 dalteparin-treated patients, RR = 1.062, 95% CI = .8034-1.4041), these trials only assure us that at the 95% level of statistical confidence, out of every 12 million people receiving enoxaparin and dalteparin prophylaxes, no more than 560,000 and 280,000 excess deaths will occur because of anticoagulation prophylaxis, respectively.

Dr. Hoyt is correct that fatal PE can occur in young people with spinal cord injuries or patients with acquired or inherited thrombophilia. Breckenridge and Ratnoff[11] documented a death rate from PE of 2.7 of 1,000,000 per year in normal, healthy, males and nonpregnant females between 15 and 45 years old. However, 95% to 96.5% of autopsy-proven PE deaths are in patients with previous terminal illnesses.[12,13] In an autopsy series from exclusively postoperative surgical patients, only 31% would have been expected to survive 5 years had the in fatal pulmonary emboli (FPE) not occurred.[14]

The HealthGrades Study of medical errors assumes that all cases of postoperative VTE in Medicare patients result from medical errors, presumably including the failure to anticoagulate prophylactically.[15] The discussion above suggests that this is an unfair assumption.

Published guidelines recommend[16] and do not advise[9] the routine use of heparin or LMWH for acute coronary syndromes. The question is not whether heparin prevents deaths vs placebo -- it does. The issue is whether the additional bleeding risk of heparin is warranted when aspirin significantly reduces deaths with much less risk of major bleeding. The main meta-analysis on the topic concluded, "The clinical evidence from randomised trials does not justify the routine addition of either intravenous or subcutaneous heparin to aspirin in the treatment of acute myocardial infarction.[7]" Routine heparin use in the first 2 weeks after a presumed ischemic stroke has not been associated with any net reduction in the proportion of patients who are dead or dependent at 6 months.[9]

Given the documented increased rate of major bleeding in surgical patients given prophylactic anticoagulants, the issue of efficacy of prophylactic heparin and LMWHs in reducing overall mortality is of great importance. In the recently released ACCP thromboprophylaxis guidelines, the claim that low-dose heparin reduces postoperative mortality in general surgery patients is based on an overview of trials published in 1988, showing a reduction in FPE (14 of 6355 vs 39 of 5847, P = .0002) and overall mortality (4.2% vs 3.2%; odds ratio [OR] = .8; number needed to treat [NNT] = 97).[1,17] For a number of reasons, this evidence does not apply to current general surgery patients:

  • In these older trials (1970s and early 1980s), prompt mobilization of patients postoperatively and the use of mechanical methods of prophylaxis, such as graded compression hoses, were much less commonly used than currently.

  • These trials excluded patients who were at high risk for either VTE or adverse outcomes.

  • The incidence of FPE in the controls (.7%)[17] was higher than the average incidence of FPE in unanticoagulated patients when more recent data are included (.5%).[9]

  • As documented in my VTE review, the clinical-pathological correlation of FPE is extremely poor. Consequently, the reported rates of FPE without having autopsy results in all fatal cases are meaningless.[18]

  • In these elective surgery trials, the incidence of VTE development in anticoagulant-treated vs nontreated patients was typically compared for the 7-10 days that the patients recovered in the hospital. However, much of the VTE that develops postoperatively does so after anticoagulants are stopped and patients discharged.[1]

  • The effects of heparin-induced thrombocytopenia with thrombosis (HITT) or other adverse reactions to anticoagulants after hospital discharge may have negated any predischarge benefit.

  • Data collected from 20 RCTs show that low-dose heparin significantly increases the rate of wound hematomas (RR = 1.56, 95% CI: 1.27-1.97).[19]

  • This is an "overview" of RCTs, not a meta-analysis because mortality data on 10% of the randomized patients are not available.

  • The more recent (2001) meta-analysis, including all available RCTs in general surgery, comparing LMWH with placebo or no treatment (n = 5456) showed no significant mortality reduction with LMWH.[5]

In a meta-analysis of RCTs involving VTE prophylaxis of total hip replacement patients, the crude risks of clinically important bleeding (usually wound hematoma) were 0% for compression stockings, .3% for controls, .4% for aspirin, 1.8% for LMWH, and 2.6% for unfractionated heparin.[20] Because of the seriousness of wound hematomas and the frequently associated infections in these patients, the Scottish Intercollegiate Guidelines Network (SIGN) changed the hip fracture and hip and knee replacement surgical VTE prophylaxis guidelines from recommending anticoagulants to advising aspirin in 2002.[9] They based this change largely on an RCT of aspirin vs placebo in hip fracture and hip arthroplasty patients, which showed a reduction in FPE with aspirin comparable with that of anticoagulants,[21] and a cohort study showing a lower than previously demonstrated rate of VTE in joint replacement patients treated with aspirin, mechanical leg compression, and leg exercises.[22]

The onset of heparin-induced thrombocytopenia (HIT) usually occurs 5-10 days following initiation of heparin, and the threshold of thrombocytopenia typically happens 7-14 days after heparin is started. People at the highest risk of thrombosis due to HIT (1% to 5%) include postoperative orthopaedic, cardiac, and vascular surgery patients. Considerable variability exists, so about 10% of HIT patients do not have thrombocytopenia but have strong serologic evidence of antiplatelet antibodies associated with thrombosis, and in 3% to 5% of HIT patients, the thrombocytopenia begins after the heparin has been stopped. If HIT develops, the absolute risk of thrombosis is 30% to 75% and about 5% have fatal thrombosis.[23] The thrombosis can be venous or arterial.

Of 2094 surgical patients (0-90 days postoperative) who developed DVT in the "DVT-FREE" prospective registry from 186 medical centers, 718 of 2094 (34.3%) developed the DVT as outpatients (median time to DVT = 21 days). And of postoperative patients with surgery > 14 days before DVT in this registry, 376 of 787 (48%) had received prophylaxis.[24] This high rate of anticoagulant prophylaxis in surgical patients who later develop DVT suggests but does not prove that the prophylaxis may not reduce DVT incidence or may increase the risk of DVT.

A randomized trial that illustrates the potential extent of delayed VTE in surgical patients receiving anticoagulant prophylaxis involves fondaparinux, a factor Xa inhibitor that was recently approved by the FDA on the basis of comparisons with heparin and LMWHs in noninferiority trials.[25,26,27] Eriksson and Lassen[28] reported that 9 of 326 postoperative hip fracture patients developed symptomatic VTE with fondaparinux given for 6-8 days post-hip-fracture surgery compared with 1 of 330 who received a 1-month course of postoperative fondaparinux prophylaxis. Patients were not followed after 1 month to ascertain the post-anticoagulant VTE incidence.[28]

Of people from the DVT-FREE prospective registry developing DVTs as medical or surgical inpatients, 1147 of 2726 (42%) received prophylaxis within 30 days before diagnosis.[24] The DVT-FREE registry did not track the proportion of these patients who would have been considered at high risk for VTE or the probability of a high-risk patient receiving anticoagulant prophylaxis. Consequently, these registry data do not rule out that adverse reactions to anticoagulants may have caused DVTs in some of the 42% of inpatients who had previously received prophylaxis.

Consequently, the question arises as to whether HIT with thrombosis,[23] post-warfarin hypercoagulability,[29,30] and other adverse effects of prophylactic anticoagulants given postoperatively cause a significant proportion of VTE cases. Two studies described below relate to this question.

To my knowledge, only 1 study monitoring the development of VTE in hospitalized patients was large enough (about 80,000 hospital discharges over the period of a 2-year registry) to produce useful data concerning the development of symptomatic VTE, particularly FPE. Goldhaber and colleagues[31] reported on the effectiveness of anticoagulant prophylaxis for the usual spectrum of inpatient indications (except neurosurgery) in preventing symptomatic VTE and death by FPE. The study period included the index hospitalization and 30 days after discharge, allowing for the inclusion of cases of late-occurring HITT and rebound hypercoagulability from vitamin K inhibitor withdrawal causing VTE and FPE, which the randomized trials did not do. Of the 384 VTE cases found, 201 had received prophylaxis and 183 had not. Twelve of the 13 FPE cases occurred in patients receiving anticoagulant prophylaxis.

Goldhaber and colleagues[31] did not track the percentage of patients who had risk factors justifying anticoagulant prophylaxis or the percentage of patients with VTE risk factors who received it, so estimates from the literature are in order. About 12 million/37 million hospitalized patients (32%) received heparin or LMWH for VTE prophylaxis in the United States in 2000.[10,32] In Goldhaber's hospital (Brigham and Women's Hospital, Boston, Massachusetts) in the early 1990s, only 29% of medical intensive care unit (MICU) patients with VTE risk factors received anticoagulant prophylaxis.[33] The proportion of surgical and medical inpatients at high risk of VTE who receive anticoagulant prophylaxis significantly increased over the 1990s with worldwide heparin sales growing at about 15% per year.[34] However, current published estimated proportions of high-VTE-risk inpatients receiving anticoagulant prophylaxis are not available but probably approach 50%.

Factors complicating the estimation of the proportion of hospitalized patients at risk for VTE and the proportion of those at risk who receive anticoagulant prophylaxis are the following:

  • In hospitalized patients, there is no consensus on the criteria for determining the risk level required for anticoagulant prophylaxis;

  • A prospective study of DVT development in 100 consecutive patients from the Brigham and Women's Hospital MICU found that traditionally recognized risk factors failed to identify patients who developed DVT[33];

  • Prospective studies of non-MICU medical patients and surgical patients showing that VTE risk-factor analysis significantly helps outcomes in anticoagulant prophylaxis trials are not available;

  • An unknown but potentially significant percentage of VTE cases may arise in hospitalized patients who are not at high VTE risk; and

  • When reporting the number of patients receiving VTE prophylaxis, many studies do not distinguish among anticoagulant prophylaxis, mechanical prophylaxis, and inferior vena caval filters.

Given these limitations with an estimated proportion of hospitalized patients at high risk for VTE and the frequency of high-risk patients receiving anticoagulant prophylaxis, let's refer back to Goldhaber's 2-year chart review of Brigham and Women's Medical Center patients.[31] If two thirds of the 80,000 hospital patients had VTE prophylaxis indications and half of the high-risk patients received anticoagulants, the ORs of anticoagulant prophylaxis/no prophylaxis for developing VTE and FPE would markedly favor no anticoagulation: 2.21 (95% CI, 1.80-2.70) and 24.0 (95% CI, 3.12-185), respectively. Changing the proportion of patients with prophylaxis indications or the percentage of high-risk patients receiving anticoagulants changes these ORs negligibly.

At my request, Dr. Goldhaber told me the anticoagulation prophylaxis status of the patients who died with PE contributing to death[10] and those dying of causes unrelated to PE.[8] However, I am not at liberty to publish the number. Consequently, we cannot assess the effect of anticoagulation prophylaxis on the all-important total death rate in Goldhaber's VTE patients. In accordance with these data, Lindblad and colleagues,[14] from Malmo, Sweden, found that the incidence of previous prophylaxis with anticoagulants in postoperative patients with autopsy-proven FPE increased from 50% in 1971-1975 to 76% in 1976-1980, 85% in 1979-1983, and 94% in 1984-1988. The proportion of surgical patients in their hospital receiving anticoagulant prophylaxis was not reported. The best study that I can find for comparison is a medical record review from the late 1980s from 16 central Massachusetts hospitals, which disclosed that 25,410 of 151,349 (17%) of hospital patients were at high risk for VTE (based on age, length of hospitalization, and an additional risk factor), and of those at high risk, 32% received prophylaxis (range, 9% to 56%).[35] The data of Goldhaber and associates[31] and Lindblad and colleagues[14] support the hypothesis that anticoagulant prophylaxis in hospitalized patients with VTE risk factors increases the rates of FPE due to adverse effects of anticoagulants.

Several points in the most recent ACCP guidelines for treatment of VTE disease by Buller and colleagues[36] (referenced by Dr. Hoyt) should be discussed. Their basis of declaring that quality RCT evidence ("grade 1A") supports their recommendation for treatment of DVT with anticoagulants is the placebo-controlled, randomized trial of patients with clinically diagnosed PE by Barritt and Jordan[37] in 1960. As noted in my review, the Barritt and Jordan[37] trial did not report the number of their 35 patients who had DVT. Buller did not address the numerous flaws in this trial, mentioned in my review, that render it worthless as the evidence-basis for anticoagulant therapy of DVT or PE. They neglected to mention the 2 small, randomized, placebo-controlled trials of anticoagulant treatment of DVT patients that were inconclusive.[38,39] They also erroneously described the randomized trial of phenylbutazone vs heparin and acenocoumarol in DVT patients by Nielsen and colleagues,[40,41] saying that both groups initially received heparin. In this trial, 2 of 48 patients receiving standard anticoagulants died vs 0 of 42 receiving only phenylbutazone.

Concerning the "Deep vein thrombosis prevention and care project" of the Joint Commission on Healthcare Organization Accreditation and the National Quality Forum (referenced by Dr. Hoyt), I found from Jerod Loeb, PhD, Joint Commission Vice President for Research, and Kenneth Kizer, MD, MPH, NQF CEO, that they would inform the project-steering committee of the existence of my review but not require that they address my review in their evaluation process or issue a written point-by-point rebuttal or analysis in their report or elsewhere.

Regarding the point from Dr. Markowicz about "adverse liability decision if a maloccurence takes place, regardless of this data, or even causality," the loss of my medical license over such a maloccurence demonstrates that it can happen. Currently, because of the US FDA approval of anticoagulants as the standard VTE treatment, failure to prescribe them in a patient with VTE opens any physician up to a malpractice suit if the patient dies of FPE, whereas a fatal intracranial bleed or FPE despite anticoagulant treatment would not as likely result in a suit. I sent the link to the review to Drs. Robert Temple and Lilia Talarico, FDA regulators responsible for the approval of heparins and vitamin K inhibitors for the treatment of VTE. FDA Ombudsman Warren Rumble replied for them that they read the review and would not be contesting any points in it or its conclusion that anticoagulant therapy for VTE should be reconsidered.

I emailed the 5 National Institutes of Health (NIH) scientists responsible for NIH-funded anticoagulation trials (Yves Rosenberg, MD; Michael Domanski, MD; Denise Simons-Morton, MD; Peter Savage, MD; and Barbara Alving, MD), asking them to submit written comments to Medscape about my review. Through the NIH ombudsman Howard Gadlin, I found that they read my review and would not comment.

Dr. Sarsfield raises the issue of the possible effectiveness of urokinase or other thrombolytic drug treatment of VTE to prevent postphlebitic syndrome. Some studies on this topic support exploring this approach further,[42,43] but others indicate that convincing evidence supporting this strategy does not exist.[44,45] The FDA has not granted an indication to any thrombolytic drug for this purpose.

Because no anticoagulation experts from academia, government, or the pharmaceutical industry ventured to comment on this review, challenging the safety and efficacy of anticoagulants in VTE treatment, allow me to respond to 6 rebuttal arguments favoring anticoagulant treatment of VTE. Seven Cochrane Collaboration VTE peer reviewers raised these issues while critiquing a systematic review of anticoagulant therapy for VTE, which I am in the process of coauthoring.[46] After we analyzed the approximately 50 non-RCT articles that they thought supported anticoagulant therapy in VTE, none of the peer reviewers disputed any of our critiques of the literature that they asked us to include. At that point, the editor decided to exclude all non-RCT data from the review and encouraged us to argue these issues in other journals. Consequently, this material will not be published elsewhere.

1. "Continuous IV heparin is better than intermittent heparin."

In the 9 randomized trials addressing this issue, neither the risk of bleeding nor the rates of recurrent VTE are significantly different.[47,48,49,50,51,52,53,54,55]

2. "The more rapid the therapeutic PTT the better the long term outcome."

Although 5 randomized trials have all found a strong relationship between attaining a partial thromboplastin time (PTT) in the therapeutic range in the first 24 hours,[52,53,56,57,58] these trials did not randomize the patients into those whose PTT goal was intentionally subtherapeutic vs those dosed with the aim of achieving a therapeutic PTT within 24 hours. The protocols for initial heparin dose and for dose adjustments were the same for those rapidly achieving therapeutic PTTs and those taking more than 24 hours for the PTT to become therapeutic. Consequently, delay in attaining a therapeutic PTT may simply be a marker for larger bulk of clot, more thrombin circulating, and higher risk of VTE recurrence. The association of delayed therapeutic PTT with recurrent VTE does not, therefore, support the efficacy of anticoagulants.

3. "Randomized trials of anticoagulant vs none after initial anticoagulant treatment show efficacy of anticoagulants."

Comparing the VTE recurrence rates over time, RCTs show a pattern that suggests a clinical rebound hypercoagulability phenomenon after stopping vitamin K antagonists that parallels the recognized rebound hypercoagulability seen in vitro.[29] In the first 2 months after discontinuing vitamin K antagonists, the VTE recurrence rate is much higher than the subsequent 2-monthly recurrence rate -- from 7.0% dropping to 1.0%[59] and 8.6% dropping to .73%.[60] In an RCT comparing long-term, low-dose warfarin vs placebo in previously anticoagulated DVT patients who had been off of warfarin for > 1 month (ie, VTE recurrence rate prestudy = 0), the VTE recurrence rate in the first 2 months after discontinuing vitamin K antagonists was 4.3% and subsequently dropped to .47% per 2 months.[61] This pattern in these 3 trials signifies that rebound hypercoagulability causes most of the early VTE recurrences in the groups withdrawn from warfarin.[62] In these 3 trials, patients randomized to continue warfarin were never observed after warfarin was discontinued to rule out rebound hypercoagulation.

In RCTs studying different durations of vitamin K antagonists that follow both the shorter and longer duration groups of patients after the anticoagulation is stopped, the VTE recurrence rates do not differ significantly. In a comparison of warfarin given for 3 months vs 1 year, Agnelli and colleagues[63] followed patients for 37.5 months on average. They observed no difference in the VTE recurrence rate between those who continued for 1 year vs those who discontinued at 3 months (21 of 133 vs 21 of 134, respectively, RR = .99, 95% CI .57- 1.73). In patients treated with heparin followed by 4 weeks of warfarin for DVT, Levine and colleagues[64] compared 8 more weeks of warfarin with discontinuing anticoagulants after just 4 weeks. During the 8 weeks following randomization, 9 of 105 (8.6%) of the placebo patients developed recurrent VTE compared with 1 of 109 (.9%) of the warfarin patients, P = .009. Over the entire 11 months of follow-up, 12 of 105 placebo patients developed recurrence compared with 7 of 109 warfarin patients, P = .3.

4. "Clinical trials show the efficacy of heparins and vitamin K antagonists in treating symptomatic calf vein thrombosis."

In a randomized trial of patients with calf vein DVT, recurrences were seen in 0 of 23 patients treated with 3 months of warfarin vs 8 of 29 not anticoagulated. (Six of 29 were symptomatic.) One out of 23 patients given warfarin had a symptomatic DVT 15 days after stopping warfarin, suggesting rebound hypercoagulation. At 90 days, abnormal perfusion on lung scan occurred in 1 of 23 in the warfarin group vs 3 of 29 of untreated patients. (One of 3 of these PE episodes was symptomatic but not fatal.) No deaths occurred in either group.[65] A study by Moser and LeMoine[66] found that 0 of 21 patients with calf vein DVT and no anticoagulant treatment had symptomatic PE or silent PE on lung scans, and none developed physical findings or symptoms of postphlebitic syndrome. Because of this much less than conclusive evidence, the ACCP guidelines for calf DVT make anticoagulation optional.[67]

5. "LMWHs have been shown to be efficacious and safe."

No randomized trials of LMWHs with unanticoagulated controls were found in an extensive literature search or from contacting pharmaceutical companies and anticoagulation experts. Although 2 randomized comparisons of LMWHs vs unfractionated heparin suggest fewer recurrent VTE events and less bleeding with LMWH,[68,69] most studies have not supported these findings in the treatment of VTE.[70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90] LMWHs are therapeutically equivalent to heparins, but that does not prove that they reduce the rates of symptomatic recurrence or death in VTE patients or that their benefits exceed their risks.

6. "Acenocoumarol alone is inferior to acenocoumarol and heparin in proximal DVT."

In a small (n = 120) RCT, symptomatic recurrences or extensions of VTE were borderline statistically significantly more in patients not treated with heparin (12 of 60 vs 4 of 60, P = .058).[91] However, because acenocoumarol and other vitamin K inhibitors used alone have not themselves been shown to be more effective than placebo in VTE, this trial does not prove the efficacy of anticoagulants. In this trial, the superiority of combined heparin and acenocoumarol depends on a difference in end points (ie, extension of thrombosis, recurrence of thrombosis, and PE), which occurred from week 4 to week 22 of the trial (ie, 3 of 60 events in the combined therapy group and 9 of 60 events in the acenocoumarol-alone group). There was no statistically significant difference in end points in the first 7 days of the study during the time patients received the heparin (ie, 1 of 60 event in the combined therapy group and 3 of 60 events in the acenocoumarol-alone group, P = .31). Even if this trial was interpreted to demonstrate that heparin reduces VTE recurrences in proximal DVT patients, it cannot be concluded that it prevents deaths from PE, because no patient died in the group without heparin. Indeed, deaths and major complications may be increased because heparins carry risks of major and fatal bleeding.

I wrote to the CEOs of some of the pharmaceutical companies that make anticoagulants for VTE (Bristol-Myers Squibb -- warfarin [Coumadin], LEO Pharma -- tinzaparin [Innohep], Aventis Pharma -- enoxaparin [Lovenox], and fondaparinux [Aristra]), asking that they send a convincing rebuttal to my review or withdraw their medications from the indication of treatment of VTE. Hopefully, responses from the anticoagulation experts will be forthcoming.

I agree with Dr. Hoyt that, "Until better studies and therapies are available most of us will choose the current recommendations put forth by the Seventh ACCP Conference." However, this review is a challenge to the ACCP Conference participants and other anticoagulation experts in academia and government to either convincingly rebut the points that I have made or reconsider the role of anticoagulation in the treatment of VTE and, at Dr. Hoyt's prompting, the prophylaxis of VTE.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: