Non-steroidal anti-inflammatory drugs versus acetaminophen and risk of venous thromboembolism: An active comparator new user cohort study.

Abstract

Previous observational studies have reported an increased risk of venous thromboembolism (VTE) among non-steroidal anti-inflammatory drug (NSAID) users compared with non-users. However, these studies may have been subject to bias due to unmeasured confounders related to patient conditions requiring NSAIDs. We employed an active comparator new user cohort design, using acetaminophen as the active comparator. New users of either NSAIDs or acetaminophen aged 18-65 years were identified from a Japanese health insurance claims database. The adjusted hazard ratio (aHR) of VTE among new NSAID and acetaminophen users within 60 days of prescription was calculated using propensity score overlap weighting and Cox regression. To examine how a naïve comparison between users and non-users of NSAIDs might affect the results, we repeated the analysis using non-users of NSAIDs as the comparator group instead of acetaminophen users. Among 4,282,421 new NSAID users and 2,728,202 new acetaminophen users, 1504 (0.022%) developed VTE during follow-up. New NSAID users had a significantly lower incidence of VTE compared with new acetaminophen users (aHR, 0.70; 95% confidence interval [CI], 0.62-0.80). When compared with non-users, NSAID users had a significantly higher incidence of VTE (aHR, 3.18; 95% CI, 2.85-3.55). Although new NSAID users had a lower incidence of VTE than new acetaminophen users, they had a higher incidence compared with non-users. Assuming that acetaminophen does not increase VTE risk, these findings suggest that NSAIDs themselves may not increase VTE risk.