Aim To evaluate the long-term efficacy and potential risk of psychiatric side effects of varenicline for smoking cessation compared with placebo or nicotine replacement therapy. Subject and methods Systematic search of electronic databases (MEDLINE, EMBASE, Cochrane Central Register, SCI) up to March 2011. Two reviewers independently determined the eligibility of randomized controlled trials comparing varenicline with placebo, or nicotine replacement therapy with follow-up of at least 12 months. Information was independently extracted by 2 reviewers.Results Ten trials involving 6,375 smokers were included in the meta-analysis. The pooled risk ratios (RR) for continuous abstinence was 2.83 (95% CI: 2.20-3.63) at 52 weeks for varenicline (1 mg, twice per day) versus placebo. Varenicline seemed to be more effective in smokers with chronic obstructive pulmonary disease (RR, 3.33) than smokers with cardiovascular diseases (RR, 2.64) or health smokers (RR, 2.52), non-Asian smokers than Asian smokers (2.98 vs. 1.94), elder smokers than younger smokers (2.87 vs. 2.52), female smokers than male smokers (2.98 vs. 1.94). The five predominant reported adverse events for varenicline compared to placebo were vomiting, nausea, abnormal dreams, constipation, and dysgeusia. There was no sufficient evidence that varenicline was associated with an increased risk of psychiatric side effects (RR, 1.45, 95% CI: 0.90-2.32). Conclusion Varenicline therapy compared with placebo is associated with a favorable effect on smoking cessation at the end of 52 weeks. However, the psychiatric adverse events related with varenicline should be further studied with larger qualified study. People with preexisting mental illnesses should be prudently treated with varenicline.