Low-Dose Aspirin for Preventing Recurrent Venous Thromboembolism

Study Design

The Aspirin to Prevent Recurrent Venous Thromboembolism (ASPIRE) study was a double-blind, randomized, placebo-controlled study of the use of low-dose aspirin in patients who had had a first-ever unprovoked venous thromboembolism and who had completed initial anticoagulation therapy. We randomly assigned patients who had completed anticoagulation therapy to aspirin, at a dose of 100 mg daily, or placebo; randomization was performed through a central Web-based randomization system, with stratification according to center and duration of initial oral anticoagulation therapy (≤26 weeks or >26 weeks). Enteric-coated aspirin, in 100-mg tablets, and matching placebo were provided without charge by Bayer HealthCare Pharmaceuticals; the company played no other role in the study and was not involved in the collection or analysis of the data or in the preparation of the manuscript. Patients were asked to take one tablet daily for a minimum of 2 years. The maximum duration of treatment was subsequently capped at 4 years.

Patients

Male and female patients were eligible for inclusion if they were at least 18 years of age and had had a first unprovoked episode of objectively diagnosed symptomatic deep-vein thrombosis involving the popliteal vein or more proximal leg veins or an acute pulmonary embolism. Venous thromboembolism was considered to be unprovoked if it occurred in the absence of the following transient risk factors during the preceding 2 months: confinement to bed for more than 1 week, major surgery, trauma requiring a cast, pregnancy or the puerperium, and the use of the oral contraceptive pill or hormone-replacement therapy. All patients were required to have completed initial anticoagulation therapy with heparin followed by warfarin (or an effective alternative anticoagulant). The duration of the initial anticoagulation therapy had to be between 6 weeks and 24 months; however, it was recommended that a target international normalized ratio of 2 to 3 be maintained with warfarin therapy for 6 to 12 months.

Patients were not eligible for inclusion if the first unprovoked episode of venous thromboembolism had occurred more than 2 years before enrollment; if they had an indication or contraindication for the use of aspirin, other antiplatelet therapy, or a nonsteroidal antiinflammatory drug; if they had an indication for continuing oral anticoagulation therapy; or if they had other medical problems that would interfere with participation in the trial or limit life expectancy. A detailed description of the eligibility criteria is provided in the study protocol, available with the full text of this article at NEJM.org.

Follow-up

Patients attended follow-up visits at 1 month and 6 months after randomization and every 6 months thereafter and were contacted by telephone or e-mail at the 3-month mark between visits. All patients who were enrolled after a first episode of unprovoked deep-vein thrombosis underwent venous ultrasound examination within 1 month after randomization to determine whether there was residual thrombus, in order to distinguish between residual thrombosis and a recurrence of thrombosis in subsequent assessments. Patients were instructed to report to their study center immediately if symptoms suggestive of a recurrence of venous thromboembolism, bleeding, adverse effects of a study drug, or other clinically significant changes developed. Information for each patient was included up to 4 years after enrollment or up to the scheduled final visit, whichever came first. The final visit was scheduled for each site during the first quarter of 2012, after the decision was made to close the trial.

Outcome Measures

The primary outcome of the study was a recurrence of venous thromboembolism, defined as a composite of symptomatic, objectively confirmed deep-vein thrombosis, nonfatal pulmonary embolism, or fatal pulmonary embolism. Prespecified secondary outcomes were major vascular events (a composite of venous thromboembolism, myocardial infarction, stroke, or cardiovascular death) and a measure of the net clinical benefit (a reduction in the rate of the composite of venous thromboembolism, myocardial infarction, stroke, major bleeding, or death from any cause). The risks of arterial thrombosis and cardiovascular death are increased in patients with unprovoked venous thromboembolism^15,16; these events were included as secondary outcomes because they are clinically important and are likely to be influenced by aspirin therapy. The diagnosis of a recurrence of venous thromboembolism required the presence of new symptoms and objective evidence on appropriate imaging of new thrombosis that was not identified on previous imaging (as detailed in the study protocol). Pulmonary embolism was considered to be the cause of death if it was confirmed at autopsy or if death was preceded by a recurrence of symptomatic deep-vein thrombosis or pulmonary embolism and the cause of death could not be attributed to an alternative diagnosis. All episodes of venous thromboembolism, myocardial infarction, and stroke and the causes of death were adjudicated by an independent outcome assessment committee whose members were unaware of the group assignments.

The primary safety outcome was bleeding, either major or clinically relevant nonmajor bleeding. Major bleeding was defined as overt bleeding that was associated with a decrease in hemoglobin of at least 2 g per deciliter or that necessitated transfusion of 2 or more units of blood, involved a critical site (e.g., retroperitoneal or intracranial bleeding), was disabling, required surgical intervention, or contributed to death. Bleeding episodes that did not meet the definition of major bleeding were considered to be clinically relevant only if they led to discontinuation of the study drug for more than 14 days.

With plans developed a priori to subsequently pool the results of the ASPIRE trial and the Warfarin and Aspirin (WARFASA) trial,¹⁴ the protocols of the two trials were prospectively harmonized to ensure that the randomized treatments were identical and that eligibility criteria and outcome definitions were similar.

Study Oversight

The study protocol was written by members of the executive committee of the trial (see the Supplementary Appendix, available at NEJM.org) and was approved by the University of Sydney human research ethics committee, as well as by the ethics committees at each participating site. Written informed consent was obtained from all patients before they underwent randomization. Clinical data were collected from electronic case-report forms with the use of an InForm clinical trial database (Oracle). The National Health and Medical Research Council (NHMRC) Clinical Trials Centre was responsible for the collection, maintenance, integrity, and confidentiality of all data. The executive committee was responsible for all aspects of the conduct of the study. An independent safety and data monitoring committee reviewed the progress of all aspects of the study, including data on safety, and met annually. All primary and secondary events were adjudicated by an independent event adjudication committee whose members were unaware of the group assignments. The final analysis plan and adjudication of all primary and secondary events were completed before unblinding of results according to group assignments. The analysis was performed at the NHMRC Clinical Trials Centre. The first draft of the manuscript was written by the first two authors and the last author, and all the authors contributed to the final manuscript and attest to the accuracy of the data and to the fidelity of the study to the protocol. No one who is not an author contributed to the manuscript.

Statistical Analysis

The ASPIRE study was originally designed to recruit 3000 patients, which would have given the study 90% power to detect a relative-risk reduction of 30% in the incidence of venous thromboembolism with aspirin as compared with placebo, assuming an event rate of 7% per year in the placebo group, at a two-sided alpha level of 0.05. The study commenced in 2003, but because of slow recruitment, the target sample size was reduced in 2005 to 1500 patients with an expected duration of follow-up of 4 years, and plans were made (with interim trial results concealed) to combine the final results with those of the WARFASA trial¹⁴ in a prospectively planned meta-analysis (Australian New Zealand Clinical Trials Registry number, ACTRN12611000684921). The ASPIRE study with the reduced sample size, when combined with the WARFASA study, had 80% power to detect a 30% reduction in the primary outcome of recurrence of venous thromboembolism (and the ASPIRE study alone had 80% power to detect a 32% reduction in the secondary outcome of all major vascular events). Recruitment closed in August 2011 because of declining recruitment rates and limited resources, with an additional 12 months of follow-up planned for the last patient enrolled. The study follow-up was subsequently closed on March 31, 2012, after publication of the results of the WARFASA study, since it was believed that continuation of the trial would provide limited additional information if patients decided to switch to open-label aspirin on the basis of the results of the WARFASA trial.

In the primary analysis, we compared the two study groups with respect to the first occurrence of symptomatic and objectively confirmed deep-vein thrombosis, nonfatal pulmonary embolism, or fatal pulmonary embolism using an intention-to-treat approach and including events up to the final scheduled visit or up to a maximum of 4 years from the time of randomization. Data from patients who withdrew consent or who were lost to follow-up were censored at the time of the last follow-up assessment. All patients who stopped using the study drug continued to be followed and were included in the intention-to-treat analysis. Survival curves were estimated with the use of the Kaplan–Meier procedure and were compared with the use of log-rank tests.

In the primary analysis, treatment effects (with their 95% confidence intervals) were estimated by means of Cox regression analysis unadjusted for patient risk factors. In a similar analysis, we evaluated the effect of aspirin on prespecified secondary outcomes of major vascular events (a composite of venous thromboembolism, myocardial infarction, stroke, or cardiovascular death) and net clinical benefit (reduction in the rate of the composite of venous thromboembolism, myocardial infarction, stroke, major bleeding, or all-cause death). Additional adjusted analyses of primary and secondary outcomes incorporated adjustment for prespecified characteristics (age, sex, smoking history, body-mass index [BMI], type of first unprovoked event, and duration of initial anticoagulation therapy). Interactions in Cox models were used to assess differences in the effect of aspirin across prespecified subgroups defined according to age, sex, duration of initial anticoagulation therapy, BMI, and type of first unprovoked event. In a time-to-event analysis that included only data from patients while they were receiving the study drug, data were censored at the time of the first discontinuation of the study drug for 90 or more days without recommencement and excluded outcome events that occurred after discontinuation. In addition, we estimated the efficacy of aspirin in a fully adherent group by adjusting the treatment effect in the intention-to-treat analysis for the nonadherence rates averaged over the study period; the nonadherence rates were defined as the proportion of patients assigned to aspirin who discontinued it and the proportion of patients assigned to placebo who initiated antiplatelet or anticoagulation treatment.¹⁷ We combined the results from the ASPIRE and WARFASA studies by performing a meta-analysis of the log hazard ratios for the treatment effect on vascular events from each study weighted by their inverse variances.¹⁸ All analyses were performed with the use of SAS software, version 9.3 (SAS Institute).

Patients

Table 1. Table 1. Baseline Characteristics of the Patients.

From May 2003 through August 2011, a total of 822 patients underwent randomization at 56 sites in five countries (Fig. S1 in the Supplementary Appendix). Twelve patients (6 in the placebo group and 6 in the aspirin group) who were enrolled after a diagnosis of first unprovoked proximal deep-vein thrombosis and who were included in the analysis were subsequently found to be ineligible after a review of the records: 8 (4 in each group) were found not to have had proximal deep-vein thrombosis, 2 (1 in each group) had a provoked venous thromboembolism, 1 in the placebo group had an indication for long-term anticoagulation, and 1 in the aspirin group had had a prior unprovoked venous thromboembolism. The baseline characteristics of the patients did not differ significantly between the two groups (Table 1). A total of 447 patients enrolled in the ASPIRE study (54%) were men, the median age was 54 years, 36% had a BMI (the weight in kilograms divided by the square of the height in meters) of 30 or higher, 15% reported that a first-degree relative had had a venous thromboembolism, 5% had a prior provoked venous thromboembolism, and 2% had active cancer (with non-melanoma skin cancer not included) at the time of randomization. The index event was proximal deep-vein thrombosis alone in 57% of the patients, pulmonary embolism alone in 28%, and co-occurring deep-vein thrombosis and pulmonary embolism in 14%. A total of 73% of the patients had received anticoagulation therapy for at least 6 months before randomization. The median time between the cessation of anticoagulation therapy and random assignment in the ASPIRE study was 7 days in both groups. The median duration of follow-up was 37.2 months.

Recurrent Venous Thromboembolism

Figure 1. Figure 1. Risks of First Recurrent Venous Thromboembolism and Major Vascular Events.

Panel A shows the cumulative risk of a first recurrence of venous thromboembolism, defined as symptomatic deep-vein thrombosis or pulmonary embolism. Panel B shows the cumulative risk of major vascular events (a composite of recurrent venous thromboembolism, myocardial infarction, stroke, or cardiovascular death). The inset shows the same data on an enlarged y axis.

Table 2. Table 2. Outcome Events, According to Study Group.

During the follow-up period, the primary outcome of recurrent venous thromboembolism occurred in 73 of 411 patients (18%) assigned to placebo and 57 of 411 (14%) assigned to aspirin (a rate of 6.5% per year vs. 4.8% per year; hazard ratio with aspirin, 0.74; 95% confidence interval [CI], 0.52 to 1.05; P=0.09) (Figure 1A and Table 2). After adjustment for baseline characteristics, the hazard ratio was 0.72 (95% CI, 0.51 to 1.01; P=0.06) (Fig. S2 in the Supplementary Appendix). There were 132 episodes of nonfatal venous thromboembolism and 2 cases of fatal venous thromboembolism in 130 patients. One fatal pulmonary embolism occurred in each group. In 103 of the 134 cases (77%), the recurrent venous thromboembolism event was unprovoked. Most of the recurrent deep-vein thrombosis events without a concurrent pulmonary embolism occurred in patients who had been enrolled with a diagnosis of first unprovoked deep-vein thrombosis alone (70%, 57 of 82 events). Similarly, 73% of the recurrent pulmonary embolism events (35 of 48 events) — with or without concurrent deep-vein thrombosis — occurred in patients who were enrolled with a diagnosis of unprovoked pulmonary embolism (with or without deep-vein thrombosis). In 100 patients (58 in the placebo group and 42 in the aspirin group), venous thromboembolism recurred for the first time while they were receiving the study drug or within 7 days after discontinuation of the study drug, whereas in 30 patients recurrences happened after discontinuation of the study drug. The analysis of data from patients while they were receiving the study drug showed a significant benefit with aspirin (event rate of 7.6% per year with placebo vs. 4.8% per year with aspirin; hazard ratio, 0.65; 95% CI, 0.44 to 0.96; P=0.03). The risk of a recurrence of venous thromboembolism was higher during the first year of follow-up (10.6% with placebo and 4.9% with aspirin) than in subsequent years (Figure 1A). The treatment effects were similar within prespecified subgroups (Fig. S3 in the Supplementary Appendix).

Secondary Outcomes and Bleeding Events

The secondary outcome of major vascular events (a composite of venous thromboembolism, myocardial infarction, stroke, or cardiovascular death) occurred in 88 patients assigned to placebo and 62 assigned to aspirin (a rate of 8.0% per year vs. 5.2% per year; hazard ratio with aspirin, 0.66; 95% CI, 0.48 to 0.92; P=0.01) (Figure 1B). Clinically relevant bleeding occurred in 8 patients assigned to placebo (6 of whom had an episode of major bleeding) and 14 assigned to aspirin (8 of whom had an episode of major bleeding). In 2 patients, both in the placebo group, the major bleeding was fatal. The rates of all bleeding episodes did not differ significantly between the study groups. The analysis of net clinical benefit, defined as a reduction in the rate of the composite of venous thromboembolism, myocardial infarction, stroke, major bleeding, or death from any cause, showed that aspirin was associated with a reduction of 33% in that composite outcome, with an event rate of 9.0% per year in the placebo group as compared with 6.0% per year in the aspirin group (hazard ratio with aspirin, 0.67; 95% CI, 0.49 to 0.91; P=0.01).

Adverse Events and Discontinuation of Study Drug

Adverse events leading to hospitalization occurred in 117 patients (28%) assigned to placebo and in 102 patients (25%) assigned to aspirin (Table S1 in the Supplementary Appendix). During the follow-up period, 132 patients assigned to placebo and 117 patients assigned to aspirin discontinued the study drug (a rate of 15.1% per year vs. 11.9% per year; hazard ratio with aspirin, 0.79; 95% CI, 0.62 to 1.01; P=0.06) (data not shown). More patients in the placebo group than in the aspirin group (32 vs. 21) discontinued the study drug because of an indication for thromboprophylaxis, whereas more patients in the aspirin group than in the placebo group (14 vs. 2) discontinued the study drug owing to gastrointestinal adverse effects or bleeding (Table S2 in the Supplementary Appendix). The median length of time that patients received a study drug was 27.2 months, and the median total follow-up time was 37.2 months. A total of 68 patients assigned to placebo and 54 assigned to aspirin initiated open-label antiplatelet therapy or anticoagulation therapy before a defined vascular event occurred. The times to permanent discontinuation of the study drug and commencement of antiplatelet or anticoagulant medications are shown in Fig. S4 in the Supplementary Appendix. The combined rate of nonadherence to the study drug in the placebo and aspirin groups, averaged over the study period, was 22%: 15% in the aspirin group discontinued the study drug and 7% in the placebo group initiated antiplatelet or anticoagulation treatment (data not shown). The risk reduction with aspirin after adjustment of the intention-to-treat estimate for this nonadherence rate was 33%, as compared with the unadjusted intention-to-treat estimate of risk reduction of 26%.

Although the results of the ASPIRE trial did not show a significant reduction in the primary outcome of recurrent venous thromboembolism with aspirin as compared with placebo in patients who had had a first unprovoked venous thromboembolism, it did show that aspirin reduced the secondary composite outcome of major vascular events by 34% without increasing bleeding and resulted in a significant net clinical benefit. Furthermore, the estimated reduction of 26% (95% CI, −5 to 48) in the rate of recurrence of venous thromboembolism with aspirin is consistent with results from the recently reported WARFASA study,¹⁴ which showed a reduction of 42% (95% CI, 7 to 64) in the rate of recurrence of venous thromboembolism.

Figure 2. Figure 2. Hazard Ratios for Venous Thromboembolism, Major Vascular Events, and Clinically Relevant Bleeding.

The number of events of venous thromboembolism, major vascular events, and clinically relevant bleeding and the associated hazard ratios in the Aspirin to Prevent Recurrent Venous Thromboembolism (ASPIRE) study and the Warfarin and Aspirin (WARFASA) study are shown with data from each study separately and with pooled data. Venous thromboembolism refers to the first recurrence of symptomatic deep-vein thrombosis or pulmonary embolism. The outcome of major vascular events is a composite of venous thromboembolism, myocardial infarction, stroke, or cardiovascular death. Clinically relevant bleeding includes major or clinically relevant nonmajor bleeding.

With fewer patients recruited than originally planned, the ASPIRE trial by itself was not powered to show a significant reduction in the primary outcome, but when combined, as prospectively planned, with the WARFASA study (in which the patients had baseline characteristics that were similar to those of the ASPIRE patients),¹⁴ a clear effect is evident. The combined results of the WARFASA and ASPIRE trials show a highly significant reduction of 32% in the rate of recurrence of venous thromboembolism (P=0.007) and a reduction of 34% in the rate of major vascular events (P=0.002), with no excess of bleeding (Figure 2).

Owing to the relatively high rate of discontinuation of the study drug in the ASPIRE study, the estimated treatment effect is likely to have underestimated the potential benefit of aspirin therapy. The estimated effect in the ASPIRE study was 35% for patients while they were receiving aspirin, an estimate that is consistent with the intention-to-treat estimate in the ASPIRE study after adjustment for nonadherence and is also consistent with the intention-to-treat estimate from the WARFASA trial.

For patients who discontinue anticoagulation, the risk of a late recurrence of venous thromboembolism after a first unprovoked event remains high: approximately 10% in the first year and 30% after 10 years. ^1-4 Recurrent venous thromboembolism is associated with a case fatality rate of 5 to 10%^4,19,20 and a risk of post-thrombotic syndrome that is increased by a factor of 6.^1,21,22 Treatment with vitamin K antagonists prevents the recurrence of venous thromboembolism, but many patients are unwilling to accept extended therapy because of the risk of bleeding and the inconvenience.

In this context, aspirin, although substantially less effective than warfarin,⁵ provides an attractive alternative because it is simple and inexpensive and its safety profile is well documented. Patients who have had a first unprovoked event of venous thromboembolism appear to be at greater risk for arterial thrombosis and cardiovascular death,^15,16 and an added appeal of aspirin is that it has been associated with an overall reduction in the risk of major thrombotic events (arterial and venous) and cardiovascular death. The ASPIRE study suggests that for every 1000 patients treated for 1 year, aspirin can be expected to be associated with 17 fewer episodes of recurrent venous thromboembolism and 28 fewer major thrombotic events, at the cost of 5 nonfatal bleeding episodes.

In conclusion, the findings of the ASPIRE study, especially when considered together with data from the WARFASA study, provide consistent evidence that low-dose aspirin is beneficial in preventing recurrent venous thromboembolism and major vascular events in patients who have had a first episode of unprovoked venous thromboembolism. Thus, aspirin is an attractive option for such patients once they have completed an initial course of anticoagulation therapy.