Efficacy and safety of direct oral factor Xa inhibitors compared with warfarin in patients with morbid obesity: a single-centre, retrospective analysis of chart data

Abstract

Background: Because studies of direct oral anticoagulants in patients with venous thromboembolism and non-valvular atrial fibrillation have had minimal representation of morbidly obese patients (ie, body-mass index [BMI] ≥40 kg/m²), their efficacy and safety in this population are unclear. We investigated whether apixaban and rivaroxaban are as effective and safe as warfarin in morbidly obese patients.

Methods: We did a single-centre, retrospective analysis of chart data for all adult patients aged at least 18 years at Montefiore Medical Center (Bronx, NY, USA) with a BMI of at least 40 kg/m² who were prescribed apixaban, rivaroxaban, or warfarin for either venous thromboembolism or atrial fibrillation between March 1, 2013, and March 1, 2017. Patients who had both venous thromboembolism and atrial fibrillation were excluded, as were patients with indications other than atrial fibrillation and venous thromboembolism. Outcomes of recurrent venous thromboembolism, stroke, and bleeding were measured from the first prescription date to the earliest of a thrombotic event, medication discontinuation, death, or end of study on June 30, 2017. Analyses were stratified by anticoagulation indication and adjusted for comorbidities, CHA₂DS₂-VASc score, and age where appropriate. Outcome rates were compared using Pearson's χ² or Fisher's exact test. Time-to-event analyses accounting for length of follow-up were used to compare risks of outcomes.

Findings: We obtained data for 795 patients: 150 prescribed apixaban, 326 rivaroxaban, and 319 warfarin. In 366 patients prescribed an anticoagulant for venous thromboembolism, the incidence of recurrent venous thromboembolism was similar between the apixaban, rivaroxaban, and warfarin cohorts (1/47 [2·1%, 95% CI 0·0-6·3], 3/152 [2·0%, 0·0-4·2], and 2/167 [1·2%, 0·0-2·9], respectively; p=0·74). Incidence of major bleeding in this patient group was also similar between the treatment cohorts (1/47 patients on apixaban [2·1%, 95% CI 0·0-6·3], 2/152 on rivaroxaban [1·3%, 0·0-3·1], and 4/167 on warfarin [2·4%, 0·1-4·7]; p=0·77). In 429 patients prescribed an anticoagulant for atrial fibrillation, incidence of stroke was similar between the treatment cohorts (1/103 patients on apixaban [1·0%, 95% CI 0·0-2·9], 4/174 on rivaroxaban [2·3%, 0·1-4·5], and 2/152 on warfarin [1·3%, 0·0-3·1], p=0·71). In this patient group, major bleeding occurred in 3/103 patients on apixaban (2·9%, 95% CI 0·0-6·2), 5/174 on rivaroxaban (2·9%, 0·4-5·4), and 12/152 on warfarin (7·9%, 3·6-12·2); p=0·063. Time-to-event analyses showed that risk of all outcomes in patients with venous thromboembolism, and stroke and composite bleeding in patients with atrial fibrillation, were similar between the anticoagulant cohorts.

Interpretation: Our retrospective study provides further evidence of similar efficacy and safety between the direct oral anticoagulants apixaban and rivaroxaban, and warfarin in morbidly obese patients with atrial fibrillation and venous thromboembolism. These data, if confirmed in prospective studies, might enable patients with a BMI of at least 40 kg/m² to benefit from more convenient, and possibly safer, anticoagulants.

Funding: None.