This subgroup analysis of ARISTOPHANES study used multiple datasources to compare S/SE, MB, and their respective components among NVAF patients (pts) prescribed NOACs stratified by index dosage.

A retrospective observational study of NVAF pts initiating api, dabi, riva or warfarin from 01/01/2013–09/30/2015 was conducted using CMS Medicare data and four other US commercial claims databases, covering>180 million beneficiaries. After propensity score matching (PSM) in each database between standard-dose NOACs (5mg BID api–150mg BID dabi, 5mg BID api–20mg QD riva, and 150mg BID dabi–20mg QD riva) and lower-dose NOACs (2.5mg BID api-75mg BID dabi, 2.5mg BID api–10 or 15mg QD riva, and 75mg BID dabi–10 or 15mg QD riva), the resulting patient records were pooled. Cox models were used to estimate hazard ratios of S/SE and MB. S/SE included ischemic stroke, hemorrhagic stroke, and SE; MB included gastrointestinal (GI) bleeding, intracranial hemorrhage (ICH), and other MB.

Standard-dose api pts had a similar rate of S/SE but a lower rate of MB vs. standard-dose dabi; lower-dose api pts had a lower rate of S/SE and MB versus lower-dose dabi. In both standard and lower dose analyses, api was associated with a lower rate of S/SE and MB compared to riva. Compared to standard-dose riva, standard-dose dabi was associated with a similar rate of S/SE and lower rate of MB. Lower-dose dabi was associated with a higher rate of S/SE and similar rate of MB compared to lower-dose riva.

Among NVAF pts, both standard- and lower-dose api pts had lower rates of MB compared to corresponding doses of dabi and riva, respectively; both doses of api demonstrated lower rates of S/SE vs. corresponding doses of riva. The comparisons between dabi and riva showed varying results for S/SE and MB across dosage levels. Dose selection criteria cannot be ascertained from the current data sources. Future studies of pts who were appropriately dosed are warranted.