Use of varenicline versus bupropion and risk of psychiatric adverse events

Aim To investigate whether varenicline use was associated with increased risk of psychiatric adverse events, compared with bupropion, another drug used for smoking cessation. Designsetting and participants We conducted a registry-based cohort study in Denmark, 2007-10, comparing new users of varenicline and bupropion in unmatched and 1:1 propensity score-matched analyses. Measurements Using Cox regression, we estimated the hazard ratio (HR) of any psychiatric adverse event (emergency department visit or in-patient admission with a psychiatric diagnosis) within 30 days following treatment initiation. The unmatched and matched analyses correspond to conventional crude and fully adjusted analyses, respectively. Findings In unmatched analyses, there were 106 (0.18%) psychiatric adverse events among 59790 varenicline users (rate 22 events per 1000 person-years), compared with 46 (0.26%) events among 17936 bupropion users (rate 31 per 1000); the HR was 0.69 [95% confidence interval (CI): 0.49-0.98]. In propensity score-matched analyses, 39 (0.22%) events occurred among 17935 varenicline users (rate 27 per 1000), compared with 46 (0.26%) events among 17935 bupropion users (rate 31 per 1000); varenicline was not associated with increased risk of psychiatric adverse events (HR 0.85, 95% CI: 0.55-1.30). The overall rate of psychiatric adverse events was substantially higher among participants with a history of psychiatric disorder than in patients without such history; the risk associated with varenicline did not differ significantly by history of psychiatric disorder. Conclusions In Denmark, the risk of psychiatric adverse events diagnosed during an emergency department visit or in-patient admission was not significantly higher with varenicline use compared with bupropion.