BACKGROUND:Neuropsychiatric safety and relative efficacy of varenicline, bupropion, and transdermal nicotine patch (NRT) in those with psychiatric disorders are of interest. METHODS:We performed secondary analyses of safety and efficacy outcomes by psychiatric diagnosis in EAGLES (Evaluating Adverse Events in a Global Smoking Cessation Study), a 12-week, randomized, double-blind, triple-dummy, placebo- and active (NRT)-controlled trial of varenicline and bupropion with 12-week follow-up, in a subset population, n = 4092, with a primary psychotic (n = 390), anxiety (n = 792), or mood (n = 2910) disorder. Primary end-point parameters were incidence of prespecified moderate and severe neuropsychiatric adverse events (NPSAEs) and weeks 9 to 12 continuous abstinence rates (9-12CAR). RESULTS:The observed NPSAE incidence across treatments was 5.1% to 6.3% in those with a psychotic disorder, 4.6% to 8.0% in those with an anxiety disorder, and 4.6% to 6.8% in those with a mood disorder. Neither varenicline nor bupropion was associated with significantly increased NPSAEs relative to NRT or placebo in the psychiatric cohort or any psychiatric diagnostic subcohort. There was a significant effect of treatment on 9-12CAR (P < 0.0001) and no significant treatment-by-diagnostic subcohort interaction (P = 0.24). Abstinence rates with varenicline were superior to bupropion, NRT, and placebo, and abstinence with bupropion and NRT was superior to placebo. Within-diagnostic subcohort comparisons of treatment efficacy yielded estimated odds ratios for 9-12CAR versus placebo of greater than 3.00 for varenicline, greater than 1.90 for bupropion, and greater than 1.80 for NRT for all diagnostic groups. CONCLUSIONS:Varenicline, bupropion, and nicotine patch are well tolerated and effective in adults with psychotic, anxiety, and mood disorders. The relative effectiveness of varenicline, bupropion, and NRT versus placebo did not vary across psychiatric diagnoses.