PDF: ASTRO-APS
(2022)Objective
Apixaban versus warfarin for prevention of recurrent thrombosis in patients with thrombotic antiphospholipid syndrome (TAPS).
Study Summary
Intervention
Apixaban 2.5 mg BID (escalated to 5 mg BID after the 25th patient per DSMB) vs. warfarin (INR 2–3) for 12 months in TAPS patients. Open-label, blinded-endpoint (PROBE) design.
Study Design
Arms: Apixaban vs Warfarin
Outcome
• Primary safety (major bleeding + CRNMB): 0 events with apixaban vs. 1 major bleed (vaginal hemorrhage, INR 2.9) with warfarin
• After excluding patients with prior arterial thrombosis: 2 strokes on apixaban (n=17) vs. 0 on warfarin (n=16)
• Trial terminated early due to inadequate accrual and loss of funding, not safety
• Patient satisfaction (ACTS) was higher with apixaban at all intervals (Month 1 p<0.05; Months 3–12 p<0.01)
Bottom Line
In this small pilot RCT terminated early due to inadequate accrual and loss of funding (not for safety), all 6 thrombotic events (ischemic strokes) occurred in the apixaban arm and none in the warfarin arm, reinforcing concerns that DOACs may not be suitable alternatives to warfarin in TAPS. Conclusions are limited by early termination, small sample size, and multiple protocol modifications.
Major Points
- Multicenter PROBE-design pilot trial comparing apixaban and warfarin in 48 patients with thrombotic APS (TAPS); originally aimed to enroll 200.
- Trial was terminated prior to target enrollment due to inadequate accrual and the resultant loss of funding — not stopped by the DSMB for safety (the DSMB explicitly recommended continued enrollment after the strokes were observed).
- Multiple DSMB-driven protocol modifications: (1) apixaban dose escalated from 2.5 mg to 5 mg BID after the 25th patient; (2) after the 30th patient, previously enrolled patients with a history of arterial thrombosis were returned to warfarin and future enrollees with prior arterial thrombosis were excluded; (3) brain MRI (stroke detection protocol) was required, with exclusion of patients showing radiographic evidence of prior stroke or white matter changes disproportionate for age.
- No major or clinically relevant non-major bleeds occurred in the apixaban arm; 1 major bleed (vaginal hemorrhage, INR 2.9) in the warfarin arm.
- Patients on apixaban reported significantly higher satisfaction with anticoagulation (ACTS) at every time point (Month 1 p<0.05; Months 3, 6, 9, 12 p<0.01).
Study Design
- Study Type
- Randomized, open-label, blinded-endpoint (PROBE)
- Randomization
- Yes
- Blinding
- Blinded outcome adjudication
- Sample Size
- 48
- Follow-up
- 12 months
- Centers
- 2
- Countries
- United States
Primary Outcome
Definition: Composite of clinically overt thrombosis (arterial and venous) and vascular death
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 0 events (0 per 1000 person-years) | 6 ischemic strokes (318 per 1000 person-years) | - | Cox proportional-hazards models did not converge; comparative statistics not reported |
Limitations & Criticisms
- Early termination prior to target enrollment (48 of planned 200) due to inadequate accrual and loss of funding — not stopped for safety, though the excess strokes were the most clinically concerning finding
- Small sample size (n=48) limited statistical power; Cox proportional-hazards models did not converge, precluding formal between-arm comparisons
- Multiple mid-trial protocol modifications: apixaban dose escalation (2.5 → 5 mg BID), return of already-enrolled arterial-thrombosis patients to warfarin, exclusion of future patients with prior arterial thrombosis, and MRI-based exclusion of patients with radiographic evidence of prior stroke or disproportionate white matter changes
- Open-label design may introduce bias despite blinded adjudication
- Heterogeneous cohort (definite, likely, and historical APS) and near-single-center enrollment limit generalizability
Citation
Woller SC, Stevens SM, Kaplan D, et al. Blood Adv. 2022 Mar 22;6(6):1661–1670.