Omega-3 fatty acids for the primary and secondary prevention of cardiovascular disease

Background Researchers have suggested that omega-3 polyunsaturated fatty acids from oily fish (long-chain omega-3 (LCn3), including eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA)), as well as from plants (alpha-linolenic acid (ALA)) benefit cardiovascular health. Guidelines recommend increasing omega-3-rich foods, and sometimes supplementation, but recent trials have not confirmed this. Objectives To assess effects of increased intake of fish-and plant-based omega-3 for all-cause mortality, cardiovascular (CVD) events, adiposity and lipids. Search methods We searched CENTRAL, MEDLINE and Embase to April 2017, plus ClinicalTrials. govandWorldHealth Organization International Clinical Trials Registry to September 2016, with no language restrictions. We handsearched systematic review references and bibliographies and contacted authors. Selection criteria We included randomised controlled trials (RCTs) that lasted at least 12 months and compared supplementation and/or advice to increase LCn3 or ALA intake versus usual or lower intake. Data collection and analysis Two review authors independently assessed studies for inclusion, extracted data and assessed validity. We performed separate randomeffects meta-analysis for ALA and LCn3 interventions, and assessed dose-response relationships through meta-regression. Main results We included 79 RCTs (112,059 participants) in this review update and found that 25 were at low summary risk of bias. Trials were of 12 to 72 months' duration and included adults at varying cardiovascular risk, mainly in high-income countries. Most studies assessed LCn3 supplementation with capsules, but some used LCn3-or ALA-rich or enriched foods or dietary advice compared to placebo or usual diet. LCn3 doses ranged from 0.5g/d LCn3 to > 5 g/d (16 RCTs gave at least 3g/d LCn3). Meta-analysis and sensitivity analyses suggested little or no effect of increasing LCn3 on all-cause mortality (RR 0.98, 95% CI 0.90 to 1.03, 92,653 participants; 8189 deaths in 39 trials, high-quality evidence), cardiovascular mortality (RR 0.95, 95% CI 0.87 to 1.03, 67,772 participants; 4544 CVDdeaths in 25 RCTs), cardiovascular events (RR 0.99, 95% CI 0.94 to 1.04, 90,378 participants; 14,737 people experienced events in 38 trials, high-quality evidence), coronary heart disease (CHD) mortality (RR 0.93, 95% CI 0.79 to 1.09, 73,491 participants; 1596 CHD deaths in 21 RCTs), stroke (RR 1.06, 95% CI 0.96 to 1.16, 89,358 participants; 1822 strokes in 28 trials) or arrhythmia (RR 0.97, 95% CI 0.90 to 1.05, 53,796 participants; 3788 people experienced arrhythmia in 28 RCTs). There was a suggestion that LCn3 reduced CHD events (RR 0.93, 95% CI 0.88 to 0.97, 84,301 participants; 5469 people experienced CHD events in 28 RCTs); however, this was not maintained in sensitivity analyses -LCn3 probably makes little or no difference to CHD event risk. All evidence was of moderate GRADE quality, except as noted. Increasing ALA intake probably makes little or no difference to all-cause mortality (RR 1.01, 95% CI 0.84 to 1.20, 19,327 participants; 459 deaths, 5 RCTs), cardiovascular mortality (RR 0.96, 95% CI 0.74 to 1.25, 18,619 participants; 219 cardiovascular deaths, 4 RCTs), and CHD mortality (1.1% to 1.0%, RR 0.95, 95% CI 0.72 to 1.26, 18,353 participants; 193 CHD deaths, 3 RCTs) and ALA may make little or no difference to CHD events (RR 1.00, 95% CI 0.80 to 1.22, 19,061 participants, 397 CHD events, 4 RCTs, low-quality evidence). However, increased ALA may slightly reduce risk of cardiovascular events (from 4.8% to 4.7%, RR 0.95, 95% CI 0.83 to 1.07, 19,327 participants; 884 CVD events, 5 RCTs, low-quality evidence with greater effects in trials at low summary risk of bias), and probably reduces risk of arrhythmia (3.3% to 2.6%, RR 0.79, 95% CI 0.57 to 1.10, 4,837 participants; 141 events, 1 RCT). Effects on stroke are unclear. Sensitivity analysis retaining only trials at low summary risk of bias moved effect sizes towards the null (RR 1.0) for all LCn3 primary outcomes except arrhythmias, but for most ALA outcomes, effect sizes moved to suggest protection. LCn3 funnel plots suggested that adding in missing studies/results would move effect sizes towards null for most primary outcomes. There were no dose or duration effects in subgrouping or meta-regression. There was no evidence that increasing LCn3 or ALA altered serious adverse events, adiposity or lipids, except LCn3 reduced triglycerides by similar to 15% in a dose-dependant way (high-quality evidence).