How To Treat the Patient With Antiphospholipid Antibody Syndrome Who Needs Anticoagulation?

When asked to reflect on the publications of the past year that I will remember most, I have to admit that I first thought of Dr. Keith Stewart’s 2019 editorial in ASH Clinical News titled “The Biggest Losers.”¹  While not about clotting (unless one considers the venous thromboembolism risks of long-distance air travel), this was a hilarious and thought-provoking read, and one that I recommend colleagues check out.

Dr. Stewart’s publication has, of course, no bearings on my clinical patient management. What have, however, impacted my clinical practice are three publications on antiphospholipid syndrome (APS) from the past 15 months,^2,5,6  and these are the articles I will discuss.

The TRAPS study²  (Table; online only) was reviewed in detail in the March/April 2018 and November/December 2018 issues of The Hematologist. This randomized trial compared rivaroxaban with the vitamin K antagonist (VKA) warfarin in patients with APS who were “triple positive” for antiphospholipid antibodies (APLAs) and had had a thrombotic event (venous, arterial, or microthrombosis). After enrollment of 120 patients, the study was prematurely terminated because of an excess of events — a composite of thromboembolism, major bleeding, and vascular death — in the rivaroxaban arm. Eleven events (19%) were reported in the rivaroxaban arm, and two events (3%) in the warfarin arm (HR, 6.7; 95% CI, 1.5-30.5; p=0.01). Most importantly, thromboembolic events occurred in seven patients (12%) randomized to rivaroxaban, with no event in the warfarin group; all thrombotic events were arterial. Major bleeding occurred in four patients (7%) in the rivaroxaban group and in two patients (3%) in the warfarin group. The authors of the study concluded that in high-risk patients with APS, the use of rivaroxaban was associated with an increased risk of events (thrombotic or bleeding) and thus, showed no benefit over VKA, but instead demonstrated excess risks.

Of note, this study may have had significant bias toward VKA performing better than rivaroxaban as it may have preferentially enrolled patients who were expected to do well on VKA. Unfortunately, the authors did not report in their baseline characteristics table what anticoagulant the patients had been on before study enrollment and how long they had been on prior anticoagulation. VKAs are known to have a substantial failure rate, with recurrent thrombosis occurring in up to 11.1 percent of patients.^3,4  In clinical practice, such patients are often switched to another anticoagulant and would likely not have been approached for enrollment onto the TRAPS study given their history of VKA failure. Therefore, this study likely enrolled patients who, by their history of having tolerated a VKA well, were expected to continue to do well on a VKA. On the other hand, numerous patients randomized to the rivaroxaban arm were likely rivaroxaban-naïve and thus never had a chance to show rivaroxaban failure. This may have led to a bias of the observed lower thrombosis rate in VKA compared to the rivaroxaban group. The same limitation applies to the study by Dr. Josep Ordi-Ros and colleagues.⁵ 

Therefore, the only scientifically accurate clinical conclusion that can be drawn from the TRAPS study as well as the study by Dr. Ordi-Ros and colleagues is that patients with APS and a history of thrombosis who have done well on a VKA are best continued on a VKA as they have a low VKA failure rate, but may have a higher thrombosis rate if switched to rivaroxaban. However, in patients with newly diagnosed APLA positivity or newly diagnosed bona fide APS (i.e., repeatedly positive APLA tests at least 3 months apart), it is currently unknown whether a VKA or a direct oral anticoagulant (DOAC) is a better choice; either one can lead to anticoagulant failure.

The randomized rivaroxaban versus VKA study⁵  (Table) was similar in design to the TRAPS study, but it included patients with any type of APS (i.e., single-, double-, or triple-positive APLA tests; at least 60.5% of the enrolled patients were “triple positive”). The primary efficacy endpoint was the proportion of patients with new thrombotic events: 11.6 percent in the rivaroxaban-treated group (10 arterial and 2 venous events) versus in 6.3 percent in the VKA group (3 arterial and 3 venous events). The primary safety outcome was major bleeding, which occurred in 6.3 percent in the rivaroxaban group versus 7.4 percent in the VKA group. The authors concluded that rivaroxaban showed a non–statistically significant near doubling of the risk for recurrent thrombosis. Noteworthy again is the fact that the thrombotic events in the rivaroxaban group were predominately arterial, mostly strokes.

The caveats regarding this study are the same as those discussed for the TRAPS trial; due to the study design, patients were enrolled who had tolerated a VKA well prior to randomization, whereas patients randomized to rivaroxaban were rivaroxaban-naïve. Thus, this study may also have been unintentionally biased toward enrolling patients preselected for a low risk of VKA failure. The results, therefore, are potentially biased toward suggesting that patients with APS with VKA do better on a VKA than on rivaroxaban. My conclusion from this study is the same as the one from the TRAPS study: Patients with APS and a history of thrombosis who have done well on a VKA are best continued on a VKA because they have a low VKA failure rate, but a higher rate if switched to rivaroxaban. It is important to highlight that the strong statement in the “Summary for Patients,”⁵  that “rivaroxaban should not be used to prevent blood clots in patients with APS,” is scientifically not accurate or justifiable, given the study limitations discussed here. I am concerned about this strong statement, from a patient management as well as medical-legal point of view.

The retrospective Japanese study⁶  compared the outcomes of anti-Xa DOACs (edoxaban, n=12; rivaroxaban, n = 5; apixaban, n=1) and warfarin in a “real world utilization” setting (Table). Of the 206 identified patients with APS, 18 had been treated with an anti-Xa DOAC; 36 controls were treated with a VKA were matched by age, gender, coexistence of systemic lupus erythematosus, and the presence of antiplatelet therapy from the pool of the available 206 patients with APS. The primary endpoint was event-free survival (EFS) for five years; events were defined as recurrence of arterial or venous thrombosis and severe bleeding. “Triple positivity” was present in 33.3 percent of the DOAC-treated patients, and 38.9 percent of the warfarin-treated patients. Two analyses were done; as 14 of 18 DOAC-treated patients had been switched from warfarin to a DOAC (because they “wished to avoid restrictions around food containing a lot of vitamin K”), an analysis was possible that compared EFS while on warfarin with EFS while on a DOAC in the same patient. EFS was significantly shorter during DOAC treatment than during warfarin therapy (HR, 12.1; 95% CI, 1.7-248.0). Additionally, a comparison between patients treated with Xa-inhibitors and those on warfarin also showed a significantly shorter EFS in the DOAC-treated group (HR, 11.9; 95% CI, 2.9-56.0). The authors concluded that “factor Xa inhibitors may not be recommended for APS.”

Caveat: This was a retrospective and small study (only 18 DOAC patients enrolled) with the usual limitations of a retrospective and small study. The results have wide confidence intervals. A fair conclusion from this study is that anticoagulation failure may occur on anti-Xa DOACs as well as on warfarin.

In terms of noteworthy ongoing research, the ASTRO-APS study, which randomized APS patients with thrombosis to apixaban or warfarin, is ongoing but no longer recruiting due to a loss of funding.^7,8  The follow up of the 49 enrolled patients ends in April 2020, and results will hopefully be presented at the International Society on Thrombosis and Haemostasis meeting in Milan, Italy, in July 2020. I expect additional useful information from this study that will help us discuss the most suitable anticoagulant for patients with APS. Of note, the study design needed modifications while enrollment was ongoing because of a possibly higher than expected rate of stroke among patients with a history of stroke who were randomized to apixaban, raising concern that a higher rate of arterial thrombosis may be seen with apixaban than with warfarin. However, results need to be awaited before any solid conclusion can be drawn. Unfortunately, this study is small, which will limit the strength of its findings.

My Clinical Management Conclusions

In conclusion, given the heterogeneity of APS patients, in respect to their thrombotic events (arterial or venous) as well as the number of positive test assays and type and titer of their APLA tests, one has to be careful to not take a “one-management-fits-all” approach. Individual antithrombotic decision-making in APS patients is needed. The sidebar below summarizes my current approach.