Poor concordance among drug compendia for proposed interactions between enzyme-inducing antiepileptic drugs and direct oral anticoagulants.

Abstract

To assess concordance regarding proposed interactions between enzyme-inducing antiepileptic drugs (EI-AEDs) and direct oral anticoagulants (DOACs) in leading, international drug compendia. We measured consistency of interaction reporting for each DOAC with a group of potent EI-AEDs among eight provider and consumer-focused drug compendia. Discrepant severity ranking systems were consolidated on a 0 to 4 scale. Percent agreement (PA) and Scott/Fleiss' Kappa (к) were used to quantify inter-compendia agreement on interaction listings, with linear weighting when consolidated severity rankings were taken into account (wPA and wк, respectively). For dabigatran, rivaroxaban, apixaban, and edoxaban, poor inter-compendia concordance was observed for interaction listings with EI-AEDs, with and without accounting for severity rankings (wPA: 0.54-0.72/wк: -0.08-0.03, and PA 0.47-0.79/к: -0.09-0.09, respectively). Conversely, betrixaban was consistently listed as not interacting with EI-AEDs in almost all assessed compendia, despite overlap in P-glycoprotein-based transport with other DOACs. Only 6/20 (30%) EI-AED/DOAC interactions were listed in all eight databases, and even in these six cases, severity rankings were universally discordant. Extreme inconsistencies in interaction reporting (some compendia assigning the highest possible severity ranking, while others reported no interactions) were observed in half of the individually examined interactions (10/20). Drug compendia are highly inconsistent in the inclusion and reported severity of interactions between EI-AEDs and DOACs. Generation of high-quality, real-world evidence from large-scale outcome studies is imperative to resolve discordance and provide clarity for clinical guidelines.