Apixaban dosing in hemodialysis - can drug level monitoring mitigate controversies?

Inconsistent study results and contradictory recommendations from health authorities regarding the use of apixaban in patients on hemodialysis have generated considerable uncertainty among clinicians, making investigations of appropriate dosing an unmet need.

We analyzed pre-dialysis apixaban drug levels from a tertiary care dialysis unit, comparing 2.5 mg once versus twice daily dosing. We applied mixed-effects models including dialysis modality, adjusted standard Kt/V, ultrafiltration, and dialyzer characteristics. We included an exploratory analysis of bleeding events and compared the drug levels of our dialysis patients to those from non-CKD reference populations taking the standard dose of 5 mg twice daily.

We analyzed 143 drug levels from 24 patients. Mean (SD) age at first drug level measurement was 64.7 (15.9) years (50 % female), median (IQR) follow-up was 12.5 (5.5 - 21) months. For the apixaban 2.5 mg once and twice daily groups, median (IQR) drug levels were 54.4 (< 40 - 72.1) and 71.3 (48.8 - 104.1) ng/mL respectively (P < 0.001). Levels were below the detection limit in 30 % (with 2.5 mg once daily) and 14 % (with 2.5 mg twice daily) respectively. Only dosing group (twice versus once daily) was independently associated with higher drug levels (P = 0.002). Follow-up did not suggest accumulation. The 95^th percentile of drug levels did not exceed those of non-CKD populations taking 5 mg twice daily. Median (IQR) drug levels before a bleeding (8 episodes) were higher than those without a subsequent bleeding: 111.6 (83.1 - 129.3) versus 54.8 (< 40 - 77.1) ng/mL (P < 0.001). Concomitant antiplatelet therapy was used in 86% of those with bleeding events versus 6% without bleeding events (P < 0.001).

Drug monitoring may be a contributory tool to increase patient safety. Despite non-existing target ranges, drug levels on both edges of the spectrum (e.g. below detectability or beyond the 95^th percentiles of reference populations) may improve decision-making in highly individualized risk-benefit analyses.