Evidence Based Review Article

The Journal of Informed Pharmacotherapy 2003;14:202.

Does Long-Term, Low-Intensity Warfarin Therapy Reduce the Risk of Recurrent Venous Thromboembolism?

Reviewer: Jeff Nagge, Cynthia Jackevicius
Reviewer’s e-mail address: jeff.nagge@utoronto.ca
,  cynthia.jackevicius@uhn.on.ca
Reviewer's profession/specialty: Pharmacy/Cardiology/Anticoagulation

Original Citation

Ridker PM, Goldhaber SZ, Danielson E, Rosenberg Y, Eby CS, Deitcher SR, Cushman M, Moll S, Kessler CM, Elliott CG, Paulson R, Wong T, Bauer KA, Schwartz BA, Miletich JP, Bounameaux H, Glynn RJ; PREVENT Investigators.  Long-term, low-intensity warfarin therapy for the prevention of recurrent venous thromboembolism.   N Engl J Med 2003 Apr 10;348(15):1425-34.  PubMed Citation

Overall Study Question

This multicentre, randomized, double blind trial was conducted to determine if long-term, low intensity warfarin therapy (target INR 1.5 to 2.0) would safely reduce the risk of recurrent venous thromboembolism (RVTE) in patients who had a previous idiopathic venous thrombosis. A secondary purpose of the trial was to establish whether patients with identifiable thrombophilic mutations such as factor V Leiden or G20210A prothrombin mutation would differentially benefit from warfarin prophylaxis.

Patients who were at least 30 years old with objectively confirmed idiopathic venous thromboembolism (VTE) were eligible for enrollment into the trial. Idiopathic VTE was defined as VTE that did not occur within 90 days after surgery or trauma. After completing three months of uninterrupted oral anticoagulation and a 28-day open-label run-in phase, patients were randomly allocated to receive either warfarin prophylaxis with a target INR between 1.5 and 2.0, or matching placebo.

The primary outcome for the trial was the time to the first confirmed RVTE event following randomization. Secondary outcome measures included the occurrence of major and minor bleeding episodes, death, and stroke. The trial had a planned follow-up of 4 years, but was stopped by the data and safety monitoring board after a mean of 2.1 years of follow-up due to the strong evidence of efficacy.

What was the declared relationship between authors and sponsors for this study?

The study was sponsored by grants from the National Heart, Lung and Blood Institute. Bristol-Myers Squibb provided study drug and placebo free of charge. Two of the authors (Drs. Ridker and Glynn) declared that they had received grants from Bristol-Myers Squibb.

Are the Results of the Study Valid?

1. Was assignment of patients randomized?

Yes. After three months of uninterrupted oral anticoagulation with full dose warfarin, patients participated in a 28-day open-label run-in phase to ensure that they could have their warfarin dose titrated to achieve an INR between 1.5 and 2.0 without exceeding a dose of 10 mg oral per day. The run-in phase was also conducted to exclude patients who were considered to be less than 85% compliant.

The randomization of patients completing the run-in successfully was stratified according to clinical site, time since the index event (< 6 months or > 6 months) and whether or not the index event was the patient’s first VTE.

2. Were all patients who entered the trial properly accounted for and attributed at its conclusion?

Cannot determine. Over the course of the trial, 578 patients had entered the 28-day run-in phase. At the early termination of the trial, 13 patients were still in the run-in phase, 57 patients had not completed the run-in process successfully and were not randomized, 253 patients had been assigned to receive placebo and 255 had been assigned to receive low-intensity warfarin.  It appears that follow-up data for the primary outcome of RVTE was available for all 508 randomized patients.  However, from the results presented by the investigators it is not possible to determine if follow-up assessment for the secondary end points was completed for all 508 patients.

The primary analysis was an intention-to-treat comparison of the two treatment groups.

3. Were patients, their clinicians, and study personnel 'blind' to treatment?

Yes. The investigators employed a matching placebo, and dose adjustments in the placebo group were made based on sham INR values generated by the data coordinating centre.

A committee of clinicians who were unaware of treatment group assignment reviewed all end points. Endpoints were only considered confirmed when objective criteria had been met. While the endpoints of recurrent VTE and major hemorrhage were well defined, the authors did not define minor hemorrhage.


4. Were the groups similar at the start of the trial?

Unknown. The summary information provided suggests that clinical characteristics and the frequency of several known VTE risk factors were similar in the two groups at baseline. Unspecified characteristics of interest include the incidence of deep vein thrombosis only at presentation, the number of patients at baseline with suspected or confirmed pulmonary embolism, the number of patients receiving unfractionated heparin versus low-molecular weight heparin acutely, and the number of patients in each group receiving daily aspirin therapy.

5. Aside from the experimental intervention, were the groups treated equally?

Yes. All patients returned for office visits once every two months. At each office visit, dose adjustments were performed if necessary for both groups, and any clinical events that had occurred since the previous visit were evaluated. No surveillance for asymptomatic VTE was undertaken.

What were the Results?

1. How large was the treatment effect? 

The trial was terminated early after a planned interim analysis by the independent data and safety monitoring committee as there was strong evidence of differences in efficacy between groups and the monitoring boundary specified in advance had been crossed. The intention-to-treat analysis revealed that 37 of the 253 patients assigned to placebo (14.6%) (7.2 per 100 person-years) and 14 of the 255 patients (5.5%) (2.6 per 100 patient-years) assigned to low-intensity warfarin developed a confirmed RVTE. The calculated hazard ratio was 0.36 (95% CI 0.19-0.67, p<0.001), with a resultant risk reduction of 64%, an absolute risk reduction of 9.1%, and a number needed to treat of 11 patients.

For the 77 patients with factor V Leiden or prothrombin mutation, there was a 75% risk reduction, which was not significantly different than that of the whole population.  Among those who were receiving the assigned study drug at the time of the recurrent event (on-treatment analysis), the risk reduction with low-intensity warfarin was 76% (HR 0.24, 95% CI 0.10-0.54). Although the authors stated that their primary outcome was time to first RVTE, a time-related analysis was not presented in the results.

Major hemorrhage was defined as any bleeding episode that led to hospitalization, transfusion of packed red cells or hemorrhagic stroke. Minor bleeding was not defined. Major bleeding occurred in 2 patients in the placebo group (0.8%) and 5 patients in the warfarin group (2.0%) for a nonsignificant HR of 2.53 (95% CI 0.49-13.03, p=0.25), an absolute risk increase of 1.2% and a number needed to harm of 83 patients.  Minor bleeding occurred in 34 patients in the placebo group (13.4%) and 60 patients in the warfarin group (23.5%) for an almost two-fold significantly higher rate of minor bleeding for those taking warfarin (HR 1.92, 95% CI 1.26-2.93, p=0.002, absolute risk increase of 10.1%, number needed to harm of 10 patients). The authors calculated a composite endpoint of RVTE, major bleeding episodes or death and found that there was a statistically significant difference in favour of the warfarin group (HR 0.52, 95% CI 0.31-0.87, p=0.01).

2. How precise was the estimate of the treatment effect?

The confidence intervals for the endpoint of RVTE, minor bleeding and the composite endpoint are reasonably narrow indicating precision of the treatment effect.  However, the small number of major bleeding endpoints results in wide confidence intervals and more uncertainty of the true estimate of major bleeding.

Will the Results Help Me in Caring for My Patients?

1. Can the results be applied to my patient care?

Yes. The results of this study can be generalized to the majority of patients with an idiopathic venous thrombosis. The inclusion and exclusion criteria were not exceedingly restrictive and were reasonable. Of note, patients with lupus anticoagulant or antiphopholipid antibodies were excluded.  We do not know the number of patients who were screened and did not meet the entry criteria. However, only about 10% of patients (n=57) did not complete the open-label run-in phase, meaning that few patients had issues with not having a stable INR or being non-compliant. The run-in criteria used do mean that the patients enrolled in the trial may be somewhat easier to manage than those seen in routine practice.
As discussed previously, some baseline characteristics were not available in order to describe the population fully. Of note, the average age of the patients (53 years) what somewhat younger than the average age of other recent trials of anticoagulation in idiopathic VTE. (1)

2. Were all clinically important outcomes considered?

Yes. Recurrent VTE, major bleeding, minor bleeding and death were assessed. Also, a composite endpoint was derived encompassing these endpoints and trying to balance the risks and benefits of therapy with low-intensity warfarin. This study importantly only assessed clinically significant RVTE events, as there was no surveillance for asymptomatic thrombosis. Screening for asymptomatic thrombosis has been a major criticism of many of the VTE primary prevention studies. However, the authors do not indicate whether the RVTE were proximal or distal events. Proximal events are known to be of greater significance with higher potential for adverse patient outcomes. (2)  The authors also did not investigate quality of life or economic endpoints which are important in this situation.

3. Are the likely treatment benefits worth the potential harms and costs?

The number needed to treat to prevent one recurrent VTE is 11 patients, which does not seem like a large number of patients to achieve a benefit. This must be weighed against the potential risk of bleeding. If 1,000 patients were treated with low-intensity warfarin for 2 years, there would be 91 less episodes of recurrent VTE, however, offset by 112 episodes of bleeding (12 major bleeds and 102 minor bleeds). While rates of major bleeding are lower than in regular-intensity warfarin trials, minor bleeding episodes may still detract from patients taking anticoagulation therapy.1 This trial likely underestimates the true bleeding risk due to the run-in period where patients with noncompliance or difficulty with management were excluded. This needs to be taken into consideration when weighing the risks and benefits of this therapy. In addition, the lack of definition of minor bleeding makes the interpretation more difficult.

The overall composite endpoint used by the authors, which combines both efficacy and toxicity does indicates a highly significant benefit of low-intensity warfarin over placebo. Monitoring was done rather infrequently which makes this regimen more favourable in terms of convenience. Despite this, over 20% of patients in both arms discontinued therapy during the study (mean follow-up 2.1 years). This indicates that adherence rates to therapy in the long-term may be an issue as it is with other chronic therapies. Sustained benefit will be an issue with any chronic therapy with poor adherence, and rates in practice would likely be worse than in a clinical trial. A potential reason for discontinuation may be the high rate of minor bleeding in the study, with about one-quarter of patients experiencing this outcome. In addition, even with low-intensity warfarin, drug interactions and other issues still need to be assessed and complicate the convenience of this regimen. Cost of warfarin is relatively inexpensive and in comparison to the costs of a RVTE and potential complications, rather minor. An economic analysis of this type of regimen would be warranted to determine the cost-effectiveness.



Venous thromboembolism that occurs in the absence of a transient risk factor, otherwise known as idiopathic VTE, is associated with a risk of recurrence of at least 10% per year. (3)  The severity of a recurrent event may range from an asymptomatic calf vein thrombosis to a fatal pulmonary embolism.

To prevent recurrences of VTE, the 6th American College of Chest Physicians (ACCP) consensus conference on anticoagulation recommends that patients with a first episode of idiopathic VTE receive oral anticoagulation with a target INR of between 2.0 and 3.0 for at least 6 months4. Since the publication of the ACCP guidelines, Agnelli and colleagues reported that a longer duration of therapy with warfarin (target INR between 2.0 to 3.0) reduced the incidence of recurrent events at the expense of a greater incidence of bleeding (1). In order to find the optimal balance between antithrombotic efficacy and the risk of bleeding, investigators are now focusing on the intensity of oral anticoagulation for preventing recurrences of VTE.

The current trial demonstrates that extending the duration of treatment with low-intensity warfarin reduces recurrences of VTE without clearly increasing the incidence of major bleeding requiring hospitalization. This was achieved despite monitoring INR only once every two months. It should be noted that the incidence of major bleeding may have been different had the trial not ended almost two years earlier than planned and had not excluded difficult to manage patients.

An important methodological issue that warrants mention is that patients received only three months of full intensity warfarin therapy despite the ACCP guidelines that recommend at least 6 months of therapy after a first episode of idiopathic VTE. This could place the placebo group at a higher risk of recurrent VTE, a possibility supported by the fact that the majority of recurrent episodes in the placebo group appeared to occur within the first few months after discontinuation of warfarin therapy.

Several outstanding questions remain in this field. First, how does extended duration low intensity warfarin therapy compare to both extended duration full intensity warfarin and placebo in terms of balancing the prevention of recurrent VTE with the increased risk of bleeding?  To date, there have been no trials that compare all three options. Second, what role will the new fixed-dose oral antithrombotic agents, such as ximelagatran, play in the prevention of recurrent VTE?  These questions are currently being addressed in clinical trials. Until the results of these trials become available, it appears that low-intensity warfarin therapy (target INR between 1.5 and 2.0) should be offered to patients who have completed at least 3-6 months of full intensity warfarin therapy for an episode of idiopathic VTE.  However, the potential bleeding risks must be assessed for individual patients as this risk is lessened with low-intensity warfarin, but not eliminated.

Reviewer Competing Interests

None declared.


  1. Agnelli G, Prandoni P, Santamaria MG et al. Three months versus one year of oral anticoagulant therapy for idiopathic deep vein thrombosis. New Engl J Med 2001; 345(3):165-9.
  2. Moser KM. Venous thromboembolism. Am Rev Resp Dis 1990;141:235-49.
  3. Kearon C. Natural History of Venous Thromboembolism. Circulation 2003;107:I-22 - I-30.
  4. Hyers TM, Agnelli G, Hull RD, et al. Antithrombotic therapy for venous thromboembolic disease. Chest 2001;119:Suppl:176S-193S.

Copyright © 2003 by the Journal of Informed Pharmacotherapy. All rights reserved.