Direct oral anticoagulant use and subsequent start of proton pump inhibitors as proxy for gastric complaints

Purpose Dabigatran use has been linked to gastrointestinal complaints, but it is unknown if this leads to more use of proton pump inhibitors (PPI). Furthermore, it is unknown whether gastrointestinal complaints occur more frequently in dabigatran users compared with other direct oral anticoagulant (DOACs) users. We investigated the association between DOAC use (dabigatran, rivaroxaban, or apixaban) and subsequent PPI initiation as a proxy for gastrointestinal complaints. Methods In this population‐based observational study with an active‐comparator new user study design, anonymised dispensing data from Community Pharmacies in the Netherlands from 2012 to 2016 were used. Patients initiating DOAC for the treatment of atrial fibrillation without any PPI use before or at time of DOAC initiation were included. The outcome measure, subsequent PPI initiation, was determined in 28553 DOAC users. Results The patients initiating dabigatran (10 942), apixaban (4897), or rivaroxaban (12714) were comparable for age (mean 69 years), sex (62% men), socioeconomic class, and concomitant medication use. The risk of PPI initiation in apixaban versus rivaroxaban users was similar (adjusted hazard ratio 1.06; 95% confidence interval 0.96‐1.31) The adjusted hazard ratio of initiating PPI for dabigatran users was 1.21 (95% confidence interval 1.14‐1.29) compared with rivaroxaban/apixaban users. The cumulative incidence of PPI initiation at 6 months of follow‐up for patients using dabigatran was 13.0%, and 10.0% for those using rivaroxaban/apixaban, yielding a number needing treatment of 33. Conclusions Proton pump inhibitor initiation occurred frequently in incident DOAC users but more often in patients treated with dabigatran than in those treated with rivaroxaban or apixaban.