Idraparinux versus Standard Therapy for Venous Thromboembolic Disease

Study Organization

We conducted two separate, randomized, open-label trials: one compared the efficacy and safety of idraparinux with those of standard therapy in patients with deep venous thrombosis (the DVT Study), and the other made a similar comparison of therapies in patients with pulmonary embolism (the PE Study). The two trials were sponsored by Sanofi-Aventis.

The steering committee (including two representatives of the sponsor) had final responsibility for the study designs, protocols, statistical analysis, study oversight, verification of the data, and data analysis. The protocols were approved by the institutional review board at each center. The data were gathered and maintained by the sponsor. All suspected outcome events were classified by a central adjudication committee, whose members were unaware of treatment assignments. An independent data and safety monitoring board periodically reviewed the studies' outcomes and advised the steering committee. The writing committee wrote the first and subsequent drafts of the manuscript and vouches for the accuracy and completeness of the reported data.

Patients

Consecutive patients over 18 years of age who presented with acute symptomatic deep venous thrombosis or pulmonary embolism were eligible. Patients presenting with lower-extremity symptoms were considered to be candidates for participation in the DVT Study, and those presenting with chest symptoms were considered to be candidates for participation in the PE Study.

The criteria for deep venous thrombosis were a calf trifurcation or more proximal vein that was not compressible on ultrasonography or an intraluminal filling defect on venography.^5,6 Criteria for pulmonary embolism were an intraluminal filling defect in subsegmental or more proximal pulmonary arteries on spiral computed tomography (CT) or pulmonary angiography, a high-probability finding on a ventilation–perfusion lung scan, or a nondiagnostic finding with documented deep venous thrombosis.^5,6 Patients without chest symptoms in whom deep venous thrombosis was diagnosed were not routinely tested for pulmonary embolism.

Patients were ineligible if they met one or more of the following criteria: receipt of a therapeutic dose of low-molecular-weight heparin or unfractionated heparin administered for more than 36 hours before randomization; treatment with thrombolysis, embolectomy, or a vena cava filter required for the current episode; another indication for a vitamin K antagonist; pregnancy or breast-feeding; a creatinine clearance of less than 10 ml per minute; uncontrolled hypertension (systolic blood pressure >180 mm Hg or diastolic blood pressure >110 mm Hg); or a life expectancy of less than 3 months.

After giving written informed consent, patients were randomly assigned to receive either idraparinux or standard therapy with the use of a computerized voice-response system. Randomization was stratified according to center and intended treatment duration (3 or 6 months on the basis of the perceived risk of recurrence, as assessed by the treating physician).

Treatment Regimens

Patients who were assigned to the idraparinux group received a once-weekly subcutaneous dose of 2.5 mg. For patients with a creatinine clearance of less than 30 ml per minute (as calculated with the Cockcroft–Gault formula), the second and subsequent doses were 1.5 mg.

Patients who were assigned to receive standard therapy received tinzaparin, enoxaparin, or intravenous heparin adjusted for the activated partial-thromboplastin time (ratio, 1.5 to 2.5), followed by warfarin or acenocoumarol (international normalized ratio [INR], 2.0 to 3.0), which was started within 24 hours after randomization. During initial treatment, INRs were determined frequently. Heparin was discontinued when the INR was 2.0 or more for 2 consecutive days and the patient had received at least 5 days of initial treatment. Thereafter, the INR was determined at least once per month.

Surveillance and Follow-up

Patients were contacted weekly up to week 4 and at weeks 7 and 13 (as well as at week 26 in the 6-month stratum). For all patients, a single additional visit was scheduled 3 months after the end of the study period (regardless of whether the discontinuation was planned or premature). At each contact, a checklist was used to elicit information on symptoms and signs of recurrent venous thromboembolism and bleeding. Patients were instructed to report to the study center immediately if any of these symptoms occurred. In case of suspected recurrent pulmonary embolism or deep venous thrombosis, the protocol required objective testing.

Outcome Assessment

The primary efficacy outcome was symptomatic recurrent venous thromboembolism, defined as objectively documented recurrent pulmonary embolism, deep venous thrombosis, or death attributed to pulmonary embolism. The criteria for diagnosis of recurrent pulmonary embolism were one or more of the following findings: a new intraluminal filling defect on spiral CT or pulmonary angiography, a cutoff of a vessel of more than 2.5 mm in diameter on pulmonary angiography, a new perfusion defect of at least 75% of a segment with corresponding normal ventilation (high probability), a new non–high-probability perfusion defect associated with deep venous thrombosis as documented by ultrasonography or venography, or a new pulmonary embolism confirmed at autopsy. The criteria for the diagnosis of recurrent deep venous thrombosis were one or more of the following findings: a new noncompressible venous segment or a substantial increase (4 mm or more) in the diameter of the thrombus during full compression in a previously abnormal segment on ultrasonography or a new intraluminal filling defect on venography.

Table 1. Table 1. Definition of Major and Clinically Relevant Bleeding.

The main safety outcomes were clinically relevant bleeding (major or clinically relevant nonmajor hemorrhage) and death from all causes. Bleeding was defined as major or as clinically relevant with the use of criteria described previously (Table 1).^3,5,6 Death was classified as due to pulmonary embolism, bleeding, cancer, or other established diagnoses. Pulmonary embolism was considered the cause of death if there was objective documentation or if the cause of death was unexplained and pulmonary embolism could not be confidently ruled out.

Statistical Analysis

For each study, we assumed a 4% incidence of the primary efficacy outcome in the standard-therapy group at 3 months.⁷ We hypothesized that idraparinux would be at least as effective as the standard treatment. In previous studies involving patients with venous thromboembolism who received inadequate treatment, the reported incidence of recurrence was approximately 20%.⁸ With the use of noninferiority analysis, idraparinux was to be considered at least as effective as standard therapy if the upper limit of the 95% confidence interval for the odds ratio was less than 2. This corresponded to a preservation of at least 54% of the minimal effect, as derived from a literature review.^7,8 A sample of 1100 patients per group would provide a power of 90% to show noninferiority with a one-sided type I error of 0.025.⁹

Near the end of the planned recruitment periods, lower-than-expected overall rates of recurrent venous thromboembolism were observed in each study. According to the protocols, without unblinding of the steering committee, we considered increasing the sample to 1450 per group to maintain the power of each study at 80%. This increase was implemented for the DVT Study but not for the PE Study on the basis of a recommendation from the data and safety monitoring board.

The primary analysis of efficacy for each study was based on the incidence of recurrent venous thromboembolism at 3 months; odds ratios and 95% confidence intervals were calculated with the use of the normal approximation of the log odds ratio. For each study, the null hypothesis that we tested was that the upper limit of the 95% confidence interval for the odds ratio was greater than 2.0. Planned secondary analyses of efficacy included the calculation of the cumulative incidence of recurrent venous thromboembolism with the use of nonparametric Kaplan–Meier estimates and a comparison of these incidences at 6 months in the 6-month treatment stratum with the use of a Cox proportional-hazards model.

If the prespecified criteria for noninferiority were met, we planned to test for a reduced incidence of clinically relevant bleeding at 3 months and 6 months with the use of the chi-square test and to compare the incidences of major bleeding at 3 months and 6 months. All analyses included all randomized patients. All reported P values are two-sided, except for those in the primary efficacy analyses.

Patients

Between May 2003 and November 2004, a total of 6054 patients were screened for the DVT Study, of whom 2904 were randomly assigned to study groups (1452 to the idraparinux group and 1452 to the standard-therapy group). For the PE Study, 4628 patients were screened and 2215 were randomly assigned (1095 to the idraparinux group and 1120 to the standard-therapy group). A diagram of the enrollment of patients and randomization is shown in the Supplementary Appendix, which is available with the full text of this article at www.nejm.org.

Table 2. Table 2. Demographic and Clinical Characteristics of the Patients.

The baseline characteristics of the patients are presented in Table 2. The mean age of the patients was 58 years in the DVT Study and 62 years in the PE Study. Men accounted for 54% of patients in the DVT Study and 48% of patients in the PE Study. The mean interval between the onset of symptoms and randomization was 8 days in both trials. On the basis of the clinician's judgment, a treatment duration of 3 months was planned for 22% of patients in the DVT Study and for 9% of patients in the PE Study; 6 months of treatment was planned for the remaining patients (see the Supplementary Appendix).

Treatment and Follow-up

Table 3. Table 3. Characteristics of Treatment in Each Study.

Data on treatment with idraparinux and initial and subsequent treatment in the standard-therapy group are shown in Table 3. In the standard-therapy groups of both studies, more than 80% of patients had an INR of 2.0 or more at the end of initial treatment (82.8% in the DVT Study and 84.6% in the PE Study). Reasons for premature discontinuation of treatment are listed in Table 3. Follow-up with respect to the primary efficacy outcome was similar in all four groups in the two studies and was complete in 99.3% of patients in the DVT Study and 99.1% of patients in the PE Study.

Recurrent Venous Thromboembolism

Table 4. Table 4. Clinical Outcomes.

Figure 1. Figure 1. Cumulative Incidence of Venous Thromboembolic Events.

The graphs show comparisons between the idraparinux group and the standard-therapy group for patients with deep-vein thrombosis (the DVT Study, Panel A) and those with pulmonary embolism (the PE Study, Panel B).

In the DVT Study, 191 patients in the idraparinux group had at least one episode of suspected recurrent venous thromboembolism before day 92, of whom 42 had a confirmed recurrence. In the standard-therapy group, the corresponding numbers were 145 and 43, respectively. Thus, the incidence of recurrent venous thromboembolism at day 92 was 2.9% in the idraparinux group and 3.0% in the standard-therapy group (Table 4), for an odds ratio of 0.98 (95% confidence interval [CI], 0.63 to 1.50) in the idraparinux group. Hence, the prespecified noninferiority criterion was met (P<0.001). Between day 92 and 183, an additional four events occurred in the 6-month stratum of each treatment group. For this stratum, the hazard ratio during the 6-month study period was 1.01 (95% CI, 0.66 to 1.55), which also met the noninferiority criterion. The occurrence of events over time is shown in Figure 1. The types of events are listed in Table 4.

In the PE Study, 171 patients in the idraparinux group had at least one episode of suspected recurrent venous thromboembolism before day 92, of whom 37 had a confirmed recurrence. In the standard-therapy group, the corresponding numbers were 108 and 18, respectively. Thus, the incidence of recurrent venous thromboembolism at day 92 was 3.4% in the idraparinux group and 1.6% in the standard-therapy group (Table 4), for an odds ratio of 2.14 (95% CI, 1.21 to 3.78) in the idraparinux group. Hence, the prespecified noninferiority criterion was not met (P=0.59). Furthermore, because the lower limit of the 95% confidence interval is above 1.0, the idraparinux regimen was inferior to standard treatment. Between day 92 and 183, an additional five events occurred in the 6-month stratum of the idraparinux group versus four in the standard-therapy group. Hence, the hazard ratio during the 6-month period was 2.09 (95% CI, 1.22 to 3.57), which did not meet the noninferiority criterion. The occurrence of events over time is shown in Figure 1. The difference in incidence of recurrence between the two groups originated mainly during the first 2 weeks of treatment. The types of events are detailed in Table 4.

A comparison of the relative treatment effect for the primary efficacy outcome between the two studies indicated that the different efficacies of idraparinux as compared with standard therapy were unlikely to be due to chance (P=0.03).

Bleeding Complications

In the DVT Study, the incidence of clinically relevant bleeding at day 92 was 4.5% in the idraparinux group and 7.0% in the standard-therapy group (P=0.004) (Table 4). At day 183, the incidence of clinically relevant bleeding in the 6-month stratum was 8.3% in the idraparinux group and 8.1% in the standard-therapy group (P=0.85). The corresponding rates of major bleeding were 0.8% and 1.2%, respectively, at day 92 (P=0.35) and 1.9% and 1.5% at day 183 (P=0.50).

In the PE Study, the incidence of clinically relevant bleeding at day 92 was 5.8% in the idraparinux group and 8.2% in the standard-therapy group (Table 4). At day 183, these rates in the 6-month stratum were 7.7% and 9.7%, respectively. The corresponding rates of major bleeding were 1.1% and 2.1% at day 92 and 1.4% and 2.8% at day 183.

Adverse Events

Table 5. Table 5. Adverse Events.

Adverse events that were reported in the two trials, including those leading to permanent discontinuation of the study drug, are listed in Table 5. In the DVT Study, cancer was reported more often as a cause of discontinuation in the idraparinux group than in the standard-therapy group, although no consistent tumor site was noted; this difference was not noted in the PE Study. Rash, dermatitis, and urticaria were reported more often in the idraparinux groups of both trials. In the PE Study, cardiac failure was reported more often in the idraparinux group than in the standard-therapy group (in 28 vs. 16 patients). In the DVT Study, cardiac failure was more frequent in the standard-therapy group than in the idraparinux group (in 6 vs. 18 patients).

Rate of Death

In the DVT Study, the rate of death at day 92 was 2.3% in the idraparinux group and 2.0% in the standard-therapy group (P=0.61) (Table 4). At day 183, the death rate in the 6-month stratum was 4.9% in the idraparinux group and 3.9% in the standard-therapy group (P=0.25).

In the PE Study, the death rate at day 92 was 5.1% in the idraparinux group and 2.9% in the standard-therapy group (P=0.006) (Table 4). At day 183, the death rate in the 6-month stratum was 6.4% in the idraparinux group and 4.4% in the standard-therapy group (P=0.04). The causes of death are detailed in Table 4.

We conducted two parallel trials comparing idraparinux with standard anticoagulant therapy for the treatment of deep venous thrombosis and pulmonary embolism. The findings differed between the two trials. In patients with deep venous thrombosis, the efficacy of idraparinux in preventing subsequent thromboembolic events was similar to that of standard therapy. In contrast, in patients with pulmonary embolism, the efficacy of idraparinux was inferior to that of standard therapy. This difference in efficacy was due to an excess of early fatal and nonfatal recurrences of pulmonary embolism and was associated with an increase in total mortality. Bleeding rates in the idraparinux groups were similar to or lower than those in the standard-therapy groups.

If the apparent difference in efficacy that we observed in our trials is real, this observation challenges the concept that the same anticoagulant regimen is adequate for both deep venous thrombosis and pulmonary embolism.¹ This concept is based on a large body of clinical experience in which the same fixed doses of subcutaneous low-molecular-weight heparin or fondaparinux had similar efficacy for deep venous thrombosis and pulmonary embolism.^2,5-7,10 By contrast, earlier observations suggested a difference in pharmacokinetic and pharmacodynamic responses to unfractionated heparin in patients with pulmonary embolism, who required larger doses than did patients with deep venous thrombosis.^11,12 Preliminary analyses of results of the first week of therapy in our studies did not suggest a substantial difference in pharmacokinetics between the two idraparinux groups (data not shown). Furthermore, a plausible explanation for the observed differences is lacking, especially since many patients with deep venous thrombosis probably had concurrent asymptomatic pulmonary embolism, which was not systematically ruled out by objective testing at study entry.

It should be noted that in the PE Study, the rate of recurrence of thromboembolism in the standard-therapy group was lower than anticipated. In most of the contemporary studies, the recurrence rate has been 4 to 5% in the first 3 months, with recurrent pulmonary embolism accounting for about two thirds of events.^2,6,10 In the PE Study, the recurrence rate in the standard-therapy group was only 1.6% at 3 months, and recurrences were pulmonary embolism in less than half of the patients (44%), whereas the observed rates in the idraparinux group were consistent with those in previous studies.^6,10 One possible explanation for the low event rate in the standard-therapy group is selection of a low-risk population. However, the demographic profile of patients in the PE Study was similar to the profiles both in the DVT Study and in previous studies. Although the trials had an open-label design, outcome adjudication was blinded; we therefore do not believe that the event rates were influenced by observer bias.

One other observation requires comment. Although in the DVT Study, the rate of death at day 183 was higher in the idraparinux group than in the standard-therapy group, the difference was not significant and was mainly explained by an excess of deaths from cancer. It should be noted that at study entry there was no stratification for cancer and there were slightly more patients with cancer in the idraparinux group of this study. For the PE Study, the excess mortality both at 3 and 6 months is explained by an excess of deaths from pulmonary embolism as well as other deaths (Table 3). The excess of deaths from pulmonary embolism reinforces concern about the decreased efficacy of idraparinux in patients with pulmonary embolism in this setting.

Another factor that was not specifically addressed in these two trials, but which should be considered in evaluating the risks and benefits of idraparinux, is the absence of a specific antidote for this anticoagulant that could be administered during bleeding. Although this is also true of some other anticoagulants used in the treatment of venous thromboembolic disease, it is a particular liability for an agent with such a long duration of action. In some cases, patients with major bleeding were treated with recombinant activated factor VII or with activated prothrombin complex concentrates.

In conclusion, in two clinical trials, idraparinux was compared with standard therapy for the treatment of deep venous thrombosis and pulmonary embolism. Although the efficacy of idraparinux was not inferior to that of standard therapy in the DVT Study, it was inferior in the PE Study.

The writing committee's affiliations are as follows: the Department of Vascular Medicine, Academic Medical Center, University of Amsterdam, Amsterdam (H.R.B.); King's College Hospital, London (A.T.C.); the Division of Pulmonary and Critical Care Medicine, University of Washington School of Medicine, Seattle (B.D.); INSERM CIE3, Université Saint-Etienne, Centre Hospitalier Universitaire de Saint-Etienne, Hôpital Bellevue, Service de Médecine Interne et Thérapeutique, Saint-Etienne, France (H.D.); Flinders Medical Centre, Bedford Park, SA, Australia (A.S.G.); Faculty of Health Sciences, McMaster University, Hamilton, ON, Canada (M.G.); Sanofi-Aventis, Antony, France (G.P.); U.O. Angiologia I, Malattie Thromboemboliche, Fondazione Istituto di Ricovero e Cura a Carattere Scientifico Policlinico San Matteo, Pavia, Italy (F.P.); Department of Epidemiology, Care and Public Health Research Institute, University of Maastricht, and Department of Clinical Epidemiology and Medical Technology Assessment, Academic Hospital, Maastricht, the Netherlands (M.H.P.); and College of Public Health, University of Oklahoma Health Sciences Center, Oklahoma City (G.E.R.).