Moderate dose oral anticoagulant therapy in patients with the antiphospholipid syndrome? No

Authors


Frederick R. Rickles, Professor of Medicine, The George Washington University, Washington, DC, USA.
Tel.: +1 301 634 7090; fax: +1 301 634 7049; e-mail: frickles@faseb.org

Patients with antiphospholipid antibodies have a high risk of recurrent thrombosis. Optimal intensity of oral anticoagulant therapy to prevent recurrence (as measured by the range of the International Normalized Ratio [INR]) has not been determined for these patients. A retrospective study by Rosove and Brewer [1] suggests that a high INR range should be targeted for these patients. In this setting, Crowther et al. [2] and now Finazzi et al. in this issue of the journal [3] provide data from randomized-controlled trials of warfarin anticoagulation, comparing high-intensity (target INR, 3.1–4.0) with moderate-intensity (target INR, 2.0–3.0) in patients with thrombosis and the antiphospholipid syndrome. In patients with a first episode of venous or arterial thrombosis, they found a nonsignificant difference in thrombotic recurrence, in favor of the moderate-intensity regimen (hazard ratio 3.1; 95% confidence interval 0.6–15.0; P = 0.15 in the Crowther study and 1.97; 0.39–6.83; P = 0.5043 in the Finazzi study).

However, over the study duration in the Crowther study (mean follow-up of 2.7 years), efficacy endpoints were attained in only eight of the 114 patients enrolled (six in the high-intensity group and two in the moderate-intensity group). Similarly, the safety endpoint of major bleeding was documented by Crowther et al. in only four patients assigned the moderate-intensity regimen and in three patients assigned the high-intensity regimen (hazard ratio 1.0; 95% confidence interval 0.2–4.8; P = 0.96). In the study published in this issue of the journal by Finazzi et al., major hemorrhage was observed in two patients in the high-intensity group and three patients receiving conventional anticoagulation therapy (hazard ratio 0.66; confidence interval 0.11–3.96; P = 0.6518). Combining the data of the two studies in a ‘meta-analysis’ led Finazzi et al. to say that a ‘significant excess of minor bleeding (hazard ratio of 2.30, 95% confidence interval 1.16–4.58, P = 0.02) as well as a borderline significant estimate of an excess thrombotic risk with high-intensity anticoagulation (hazard ratio 2.49, 95% confidence interval 0.93–6.67, P = 0.07) is observed’.

Both groups of investigators conclude that the high-intensity regimen was neither more effective in preventing recurrent thrombosis nor more likely to cause major bleeding, compared with the moderate-intensity regimen. The Crowther study provided the ‘basis for the recommendation of an INR of 2.0–3.0 as the preferred intensity of anticoagulant treatment with VKA (vitamin K antagonists)’, ‘…for the prevention of recurrent thromboembolism in patients with persistently positive antiphospholipid antibodies…’ in the most recent edition of the ACCP Conference on Antithrombotic and Thrombolytic Therapy: Evidence-Based Guidelines [4]. These guidelines are utilized widely to define the standard of care in many academic centers, at least in the USA. Therefore, the importance of carefully weighing the value of the primary evidence for the recommendations cannot be underestimated.

Both of the studies were well designed and might have provided important information to guide the practitioner had they been extended over a longer duration of follow-up or if a larger number of patients had been enrolled. The small number of patients experiencing the endpoints assessed, however, limits the applicability of the conclusions from both studies; even when the endpoints are combined in the Finazzi study or when all of the endpoints are combined from the two studies. Both studies suffer from beta or type II error, where insufficient thrombotic events were observed to exclude the possibility that the higher dose of warfarin (high-intensity regimen) might be more effective. Although counterintuitive to suggest that the higher intensity regimen might produce excess thrombosis, when combined, the studies do suggest, however, that the higher dose of warfarin is associated with more bleeding events, a conclusion underscored by several previous randomized-controlled trials [5].

Several additional points regarding the interpretation of these studies should be taken into consideration, some of which were included in the correspondence that the Crowther study provoked [6]:

  • Too few patients with arterial thrombosis were studied.
    •   –Thus, the data may not be generalizable for this subgroup.
  • Patients with previous thrombotic recurrence while receiving warfarin were excluded.
    •   –This group may be more likely to be protected from a second recurrence by high-intensity warfarin therapy.
  • Patients with stroke were excluded.
    •   –This important and frequent subgroup also may require the higher anticoagulant dose.
  • Patients in the high-intensity group in the Crowther study received ‘subtherapeutic’ levels of anticoagulation 43% of the time, compared with 19% in the moderate-intensity group. Similar data can be found in the Finazzi study, where the intensity of anticoagulation overlaps between the two study groups.
    •   –This suggests that the high dose anticoagulation regimen was not adequately attained in this study group.
  • six out of eight recurrences in the Crowther study occurred in patients whose INR was 2–3.
    •   –Does this indicate that a higher dose, regardless of treatment ‘group’, prevents recurrence?

In their reply to several Letters to the Editor [6], Crowther et al. admitted that ‘additional studies will be required to prove that moderate-intensity warfarin is not inferior to high-intensity warfarin’. Finazzi et al. state toward the end of their discussion that ’as the rate of thrombosis in the high-intensity anticoagulation arm was higher than the control arm in both Crowther's and our trial, the prolongation of recruitment or follow-up was unlikely to reverse such [an] unfavorable trend’. Crowther et al. go on to state that ‘…such studies would be prohibitively large and probably not feasible…’. We suggest that the lesson of cancer trials cut short before long-term reversal of survival curves should not be lost on investigators designing studies of patients with the antiphospholipid syndrome. If future studies are designed to include higher-risk patients (such as some of those excluded from both of the current studies) and a longer period of follow-up, the number of events will be higher, thus permitting the investigators to reach a successful conclusion with a feasible study. For now, we urge caution in applying the published conclusions to all patients with antiphospholipid syndrome and thrombosis. We continue to utilize high-intensity warfarin for high risk patients (e.g. those with arterial thrombosis, recurrent thrombosis, post-thrombotic syndrome, less than complete recanalization of the affected vessel, and persistently elevated fibrin d-dimer levels).

Acknowledgements

The authors thank Dr Samuel Simmens, Interim Director, Biostatistical Center Medical Center Unit, The George Washington University, Washington, D.C. for his critical appraisal of this commentary.

Ancillary