OBJECTIVE To examine the efficacy and tolerability of ethyl-eicosapentaenoate (EPA-E) monotherapy for major depressive disorder (MDD). METHODS Fifty-seven adults with DSM-IV MDD were randomly assigned from January 2003 until June 2006 to receive 1 g/d of eicosapentaenoic acid (EPA) or placebo for 8 weeks in a double-blind, randomized, controlled pilot study. Response criteria were on the basis of the 17-item Hamilton Depression Rating Scale (HDRS-17). Subjects' plasma lipid profiles were examined by gas chromatography. RESULTS Thirty-five subjects (63% female; mean +/- SD age = 45 +/- 13 years) were eligible for the intent-to-treat (ITT) analysis. In the ITT sample, mean +/- SD HDRS-17 scores decreased from 21.6 +/- 2.7 to 13.9 +/- 8.9 for the EPA group (n = 16) and from 20.5 +/- 3.6 to 17.5 +/- 7.5 for the placebo group (n = 19) (P = .123); the effect size for EPA was 0.55. ITT response rates were 38% (6/16) for EPA, and 21% (4/19) for placebo (P = .45). Among the 24 study completers, mean +/- SD HDRS-17 scores decreased from 21.3 +/- 3.0 to 11.1 +/- 8.1 for the EPA group and from 20.5 +/- 3.8 to 16.3 +/- 6.9 for the placebo group (P = .087); the effect size for EPA was 0.73. Completer response rates were 45% (5/11) for EPA, and 23% (3/13) for placebo (P = .39). Among EPA subjects, baseline n-6/n-3 ratio was associated with decrease in HDRS-17 score (r = -0.686, P = .030) and with treatment response (P = .032); change in n-6/n-3 ratio was associated with change in HDRS-17 score (r = .784, P = .032). Side effects, reported in 2 EPA subjects and 5 placebo subjects, were exclusively gastrointestinal, mild, and not associated with discontinuation. CONCLUSIONS EPA demonstrated an advantage over placebo that did not reach statistical significance, possibly due to the small sample and low completer rates, which were the major study limitations. BACKGROUND clinicaltrials.gov Identifier: NCT00096798.