The efficacy of fish oil in decreasing restenosis following percutaneous transluminal coronary angioplasty (PTCA) remains controversial despite seven published reports of randomized trials involving 951 patients. We performed a meta-analysis to determine whether these trials, viewed in aggregate, demonstrate a significant benefit. We evaluated rates of restenosis two to 12 months after PTCA and calculated an estimate of the overall effect and 95% confidence interval (CI). The typical odds ratio (treatment versus control) was 0.71 (95% CI 0.54, 0.94), P = 0.016 (two-tailed). The data show a strong and highly significant (P less than .0001) relationship between daily fish oil dose and gastrointestinal side effects. While compatible with a small to moderate benefit of fish oil on rates of restenosis, these results require confirmation in a randomized clinical trial large enough to distinguish reliably between a clinically meaningful benefit and a null result.