(B11) Comparisons of non-vitamin K antagonist oral anticoagulants and warfarin for incident depressive disorders in patients with atrial fibrillation: A nationwide cohort study using target trial emulation framework

Publication Number: 623

Background: Vitamin K has been shown to reduce ceramides and pro-inflammatory processes in the hippocampus, which can have positive effects on alleviating anxious and depressive symptoms. Research suggests that non-vitamin K antagonist oral anticoagulants (NOACs) are associated with lower levels of these symptoms compared to warfarin. As these mental health conditions are risk factors for suicide, it is important to understand the differences in the risk profiles of suicidal attempts and suicide among AF patients receiving oral anticoagulants.

Objectives: To evaluate non-vitamin K antagonist oral anticoagulants and warfarin for the risk of suicidal attempts and suicide among patients with atrial fibrillation.

Methods: We conducted a target trial emulation study using Taiwan's NHIRD database to investigate the effects of oral anticoagulants on patients with AF. We included patients aged over 20 years and newly received oral anticoagulant treatment between 2012 and 2020. To ensure balanced characteristics and emulate randomized allocation in the target trial, we implemented propensity score approaches with fine stratification weighting. The primary outcome was composite suicidal outcomes, including suicidal attempts and suicide, during the study period. Both as-started analysis and on-treatment analysis were performed for causal contrasts of interest. To estimate the PS-weighted hazard ratio (HR) with 95% confidence intervals (CIs) for each outcome, we took competing events into account and applied both cause-specific hazards models and subdistribution hazards models.

Results: We included a total of 103,768 (71.74%) patients newly receiving NOACs and 40,877 (28.26%) patients newly receiving warfarin with a mean age of 74.43 years (SD, 10.97) and 70.58 years (SD, 12.66), respectively. After propensity score with fine stratification weighting within the two treatment groups, the weighted population comprised 103,768 patients in NOACs groups and 40,877 patients in warfarin groups. The overall mean follow‐up duration in the NOAC and warfarin groups were 2.75 and 4.31 years, respectively. Among the weighted population with targeting the average treatment effect (ATE), the cumulative incidence rate of composite suicidal outcomes was 4.16 per 1000 person-years in NOACs groups and 4.89 per 1000 person-years in warfarin groups, respectively. NOACs users had a significantly lower hazard of composite suicidal outcomes (HR, 0.85; 95%CI, 0.75 to 0.97 for ATE) compared to warfarin users. Our sensitivity analysis, which excluded patients with chronic kidney diseases at baseline, also supported this finding. In on-treatment analysis, the effect estimates remained larger, with lower trends for the hazard of composite suicidal outcomes for NOACs users compared to warfarin users (HR, 0.79; 95%CI, 0.60 to 1.05 for ATE).

Conclusions: In conclusion, our findings suggested that, compared to warfarin users, NOACs users had a significantly lower hazard of composite suicidal outcomes among patients with AF. Considering the negative impact of suicide on the healthcare of patients with AF, NOACs may be a preferable treatment option over warfarin in reducing the risk of composite suicidal outcomes for patients who require oral anticoagulant therapy. This clinical information can aid in optimizing patient care and enhancing oral anticoagulant treatment outcomes.