Eleven RCTs (n=15,806) were included.
No difference in nonfatal MI was shown between n-3 polyunsaturated fatty acid enriched diets and control diets or placebo (RR 0.8, 95% CI: 0.5, 1.2), based on 9 RCTs between which there was statistically significant heterogeneity. A pooled analysis of 8 RCTs showed a significant reduction in fatal MI (RR 0.7, 95% CI: 0.6, 0.8) with no statistically significant heterogeneity.
Data on sudden death available from 5 trials showed a benefit (RR 0.7, 95% CI: 0.6, 0.9) with no statistically significant heterogeneity. Overall mortality was also significantly reduced (RR 0.8, 95% CI: 0.7, 0.9), based on 9 RCTs with no statistically significant heterogeneity.
The summary end points were similar for dietary (2 RCTs) and supplement (9 RCTs) interventions.
All but one trial scored 3 or more for quality out of a possible 5. In sensitivity analyses blinding of patients, providers or outcome assessment did not explain differences in treatment effect between studies. There appeared to be evidence of publication bias and, hence, possible overestimation of the treatment effects.