Paroxetine Versus Placebo and Other Agents for Depressive Disorders: A Systematic Review and Meta-Analysis

Martin A. Katzman, M.D.; Andrea C. Tricco, M.Sc.; Diane McIntosh, M.D.; Marie J. Filteau, M.D.; Pierre Bleau, M.D.; Pratap R. Chokka, M.D.; Kevin D. Kjernisted, M.D.; Hiram Mok, M.D.; and Ba' Pham, M.Sc.

Objective: To compare paroxetine with placebo and other antidepressants across multiple efficacy and tolerability outcomes.

Data sources: Searches were conducted in MEDLINE (1966-2004), EMBASE (1980-2004), CINAHL (1982-2004), all Evidence-Based Medicine Reviews (1991-2004), HealthSTAR (1975-2004), BIOSIS (1980-2004), and PsycINFO (1840-2004). Medical Subject Headings (MeSH) included "paroxetine" OR "Paxil" exploded. The searches were not restricted by language, publication type, or study design.

Study selection: A study report was included if it described a randomized trial of paroxetine versus placebo or other antidepressants for patients with depressive disorders. Records were screened independently by 2 reviewers under the supervision of another reviewer.

Data extraction: Three investigators abstracted data, including study design, trial characteristics, and psychiatric assessment tools, using a prespecified form. Two investigators assessed quality of reporting using Jadad's scale.

Data synthesis: We included 62 unique randomized controlled trials. Paroxetine yielded consistently and significantly better remission (rate difference [RD]: 10% [95% CI = 6 to 14]), clinical response (RD: 17% [95% CI = 7 to 27]), and symptom reduction (effect size: 0.2 [95% CI = 0.1 to 0.3]) than placebo. Such consistency in the evidence base was not observed between paroxetine and other antidepressants. Pairwise comparisons of paroxetine and venlafaxine, mirtazapine, mianserin, or fluoxetine yielded inconsistent results across efficacy outcomes. Controlled-release paroxetine was the only antidepressant with significantly fewer dropouts due to adverse events than immediate-release paroxetine (RD: 5% [95% CI = 0.1 to 11]).

Conclusions: There were no significant and valid differences between paroxetine and other antidepressants to suggest that multiple modes of action improve clinical outcomes.

(J Clin Psychiatry 2007;68:1845-1859)

Received Dec. 19, 2006; accepted March 19, 2007. From the Faculty of Medicine, Department of Psychiatry, University of Toronto, Ontario; the Department of Clinical Sciences, Northern Ontario School of Medicine, Sudbury and Thunder Bay; the Departments of Psychology and Community Health, Lakehead University, Thunder Bay, Ontario (Dr. Katzman); the Chalmers Research Group, Children's Hospital of Eastern Ontario Research Institute, Ottawa (Ms. Tricco and Mr. Pham); the Institute of Population Health, University of Ottawa, Ontario (Ms. Tricco); the Department of Psychiatry, University of British Columbia, Vancouver (Drs. McIntosh, Kjernisted, and Mok); the Department of Psychiatry, Université Laval, Quebec City, Quebec (Dr. Filteau); the Department of Psychiatry, McGill University, Montreal, Quebec (Dr. Bleau); the Department of Psychiatry, University of Alberta, Edmonton (Dr. Chokka); and the Biostatistics and Epidemiology Group, GlaxoSmithKline Canada, Mississauga, Ontario (Mr. Pham), Canada.

Funding for this research was obtained as part of an unrestricted educational grant from GlaxoSmithKline Canada.

This research was presented as a poster presentation at the 158th annual meeting of the American Psychiatric Association, May 21-26, 2005, Atlanta, Ga., and at the 21st International Conference on Pharmacoepidemiology and Therapeutic Risk Management, August 21-24, 2005, Nashville, Tenn.

Acknowledgments and financial disclosure are listed at the end of the article.

Corresponding author and reprints: Martin A. Katzman, M.D., S.T.A.R.T. Clinic for Mood and Anxiety Disorders, 790 Bay St., Suite 900, Toronto, Ontario M5G 1N8, Canada (e-mail: mkatzman@startclinic.ca).