The use of omega-3 fatty acids and other supplements as opposed to a healthy diet to prevent cardiovascular disease is controversial and the effects beyond current drugs are disputed. There has always been a tendency in medicine to take a lifestyle-related risk protective factor and try to bottle it for convenience and commercial advantage. The most famous instance is the vast use of anti-oxidant vitamins as opposed to eating fruit and vegetables which are known to be associated with reduced rates of cardiovascular disease.^[1,2]

In contrast to a proper diet, anti-oxidant vitamin supplements show no benefit in large-scale randomised placebo-controlled trials.^[3] This, however, has not stopped the classical techniques of public relations and news management being applied to protect a multibillion dollar market.

• Deny the story.
• Allege bias – ‘the trial was not representative’.
• Quibble on details – ‘inadequate methods, duff statistics, inclusion of flawed trials etc’.
• Shift the goalposts – ‘the trial would work in a subpopulation’.
• Use anecdotal reports – ‘use basic science mechanistic studies; post hoc analyses’.
• If all else fails admit the truth but only when convenient.

Gradually both doctors and patients are slowly beginning to believe that anti-oxidant vitamins are of no use. So they switch to a new panacea as they still wish to believe. This new cure-all for cardiovascular disease is omega-3 fatty acids and particularly those derived from fish; a belief based on the Eskimo (Inuit) diet.^[4] So what is the evidence?

First of all, omega-3 (n-3) fatty acids are a heterogeneous group of molecules which are poorly synthesised in man.^[5] The common forms include docosahexaenoic acid (DHA) (22:6; n-3) and eicosapentaenoic acid (EPA) (20:6; n-3) which are often contrasted to omega-6 (n-6) fatty acids which include γ-linolenic (18:3; n-6) and the polysubstrate for inflammatory mediators-arachidonic acid (20:4; n-6).^[6,7] Omega-3 fatty acids are found in plants and in fish but as many plants also contain high levels of omega-6 fatty acids fish consumption is recommended as a primary source. Omega-3 fatty acids are supposed to exert their beneficial effects through reducing sympathetic overactivity, enhancing nitric oxide-mediated vasodilation, reducing monocyte adhesion and reducing levels of arachidonic acid-derived mediators including thromboxane A2 as well as reducing thrombomodulin and von Willebrand factor and to also reduce insulin resistance probably through actions at the peroxisomal proliferator-activating receptor (PPAR)-gamma.^[5,8,9] The PPAR effects may also be responsible for their action in reducing triglycerides but they are multiple gene activators affecting PPAR-α, PPAR-γ and PPAR-δ but also farnesoid-X receptors, Liver-X receptor and hepatic nuclear factor (HNF)-4α.^[5,10] The prime action seems to be PPAR-α with secondary actions on PPAR-γ after conversion through 15-lipoxygenase. They reduce hepatic lipogenesis by downregulating sterol receptor element-binding protein 1c, upregulation of hepatic and muscle fatty acid oxidation via PPARs and an increase in glycogen synthesis by reducing HNF-4α. There is increasing evidence that DHA and EPA may have differential effects and thus considering all omega-3 fatty acids to be similar may be an over-simplification.^[11,12]

Numerous observational studies have documented an inverse relationship between questionnaire-assessed omega-3 fatty acid concentrations allied on a few occasions with measurement of membrane and tissue concentrations with reduced rates of cardiovascular disease.^[13–16] The common recommended intake is 25–60 g of fish high in omega-3 fatty acids or one portion of such fish weekly or monthly. Yet other sources of omega-3 are almost equally beneficial – nuts, soybean oil being associated with 30–60% reductions in cardiovascular disease.^[15] Overall fatal coronary heart disease (CHD) is reduced by 38% (18–54%) stroke by 31% (12–46%) with high as opposed to low fish intake being associated with a 17% (10–24%) reduction in fatal myocardial infarction (MI) and 14% (8–19%) reduction in any CHD. One meta-analysis shows that omega-3 fatty acid consumption is associated with a 23% reduction in mortality and 32% in cardiovascular mortality – similar to statins.^[17–19] However, when reviewed in detail the studies are less clear. Prospective studies show unclear results with some positive [e.g. the Diet And Reinfarction Trial (DART)^[20]] and some negative [Dietary Angina Reduction Trial (DART-2)^[21,22]]. Another meta-analysis of dietary fish intake in 15,806 patients in 11 trials suggested a 17% (8–25%) reduction in CHD mortality and 25% (22–29%) reduction in non-fatal MI with some as opposed to no fish intake in people without pre-existing CHD.^[17] A further meta-analysis of 11 studies and 13 cohorts comprising 222,364 patients over 11.2 years found a 11% (+1% to 21%) reduction in CHD mortality with increased benefit at higher intakes (> 5/week) with a 38% (18–54%) reduction with fish consumption five times per week.^[14] Only five studies could be assessed for non-fatal MI and these showed no consistent benefit with fish intake. The relationship between fish consumption and risk reduction was complex in another analysis with the suggestion of a partial plateau in effect with 1–2 portions of oily fish per week.^[16] In primary prevention each extra serving per week reduced risk by 4%.^[16] Yet, this same analysis was limited in patients with pre-existing disease as all the trials used commercial supplements at large doses. One potential problem with a high fish intake is a parallel increase in mercury exposure which may itself predispose to CHD but no consistent effect has been seen in epidemiological studies examining the role of mercury exposure.^[16]

The data for stroke is based on a meta-analysis of nine cohorts which shows a 9% (-6% to 21%) reduction in stroke for any as opposed to no fish consumption with a weak dose proportional effect increasing to 31% (12–46%) reduction in those consuming fish > 5/week.^[23] Yet other studies only suggest a 2% increment with extra portions.^[24] Yet as these analyses mention higher fish intake is associated with higher social class, more exercise, less smoking and less obesity and so the dietary studies may not be unconfounded.

The data for α-linolenic acid (ALA), the precursor to DHA or EPA are weaker.^[25] A number of flax oil trials were confounded by changes in protocols, short durations, small size. The best known is the Lyon Heart Health study where a 5% (1.8 g) ALA-enriched diet given in an open randomised fashion to 605 patients post-MI resulted in a 73% (31–88%) reduction at 2 years in cardiac death and non-fatal MI which was maintained to 4 years.^[26] Yet even this trial was confounded by parallel changes in saturated fat, fish and anti-oxidant intakes.

The argument for supplementation is driven by the results of the Grupo Italiano per lo Studio della Sopravvivenza nell'Infarcto miocardio – prevenzione (GISSI-P) study^[27] which is the largest study in the field. In this study, 11,324 patients with recent myocardial infarcts (< p =" 0.048)" p =" 0.053)" p =" 0.02;" nnt =" 157/year)" p =" 0.008;" nnt =" 159/year)">

More recently, the Japan EPA Lipid Intervention Study (JELIS)^[28] investigated the effect of 1.8 g EPA added to new background 10–20 mg pravastatin or 5–10 mg simvastatin in 18,645 patients at least 6 months post-MI with total cholesterol > 6.5 (LDL-C > 4.4 mmol/l). The primary end-point was cardiovascular death, fatal and non-fatal MI and stroke, and percutaneous coronary intervention. The design assumed event rates of 2.13% per year in secondary prevention (n = 3664) and 0.58% per year in primary prevention (n = 14,981). Unusually for international trials but similar to MEGA^[29] the study recruited 69% women. Statin therapy was effective in reducing LDL-C by 25% in both groups to 3.25 mmol/l. The overall results showed a 19% (5–31%) reduction in events (3.5% → 2.8%; p = 0.01; NNT = 658/year). At 4.6 years there was a non-significant 18% (+6% to 37%) reduction in the primary prevention group (1.7% → 1.4%; p = 0.11; NNT = 1538/year) while events were reduced by 19% (0–34%) in the secondary prevention group (10.7% → 8.7%; p = 0.048; NNT = 375/year). There was no inter-group heterogeneity, but it is noticeable that as in MEGA the greatest event rate and degree of reduction was seen in men were events were reduced by 24% (6–38%) while women showed a non-significant 13% (+13% to 32%) reduction in events. The event reduction was driven by a 24% reduction in unstable angina in the whole cohort comprising reduced angina in the secondary prevention group and in fatal and non-fatal MI and angina in the primary prevention group.

There is little data on the role of omega-3 fatty acids in peripheral arterial disease.^[30] One meta-analysis exist of six studies with 313 patients comparing omega-3 fatty acid supplementation with placebo lasting up to 2 years. No significant differences were seen in ankle brachial pressure index, pain-free or maximal walking distance. Blood viscosity levels decreased. As in other studies gastrointestinal side effects were observed and LDL-C levels were increased by 0.80 mmol/l (0.34–1.26 mmol/l).

Omega-3 fatty acids have anti-arrhythmic effects in animal models and cell culture.^[31] They have negative chronotropic and inotropic effects on cardiomyocytes by affecting sodium- and calcium-channel function and secondary effects on cytosolic-free calcium levels.^[32] Recently a number of trials have examined their effects in cardiac arrhythmias. In small-scale studies of patients undergoing bypass grafting omega-3 fatty acids were associated with a 68% (2–90%) reduction in new atrial fibrillation in an uncontrolled study of 160 patients.^[33] In contrast in a study of 200 patients with implanted cardiac ventricular defibrillators (ICDs) 2-year therapy showed an increase in ventricular tachycardia (VT) at 2 years but patients were not taking class I or class III anti-arrhythmics and had a high rate of activation.^[34] More recently, the Study on Omega-3 Fatty acid and ventricular Arrhythmia^[35] randomised 546 patients with low fish intakes and with an ICD and > 1 episode of ventricular arrhythmia within the last year to 2 g of fish oil (900 mg EPA) as opposed to corn oil for 12 months. There was no difference in the primary end-point of death, VT or ventricular fibrillation (VF) although a trend (p = 0.09) was seen to reduced rates in the small subgroup of patients with prior myocardial infarcts. In the Fatty Acid Arrhythmia Trial^[36] 402 patients with an ICD were randomised to 2.6 g of omega-3 fatty acids or olive oil. At 12 months there was borderline 28% reduction in new VT or VF (p = 0.057). However, 35% of patients discontinued the treatment and a per protocol analysis suggested a 38% (p = 0.03) reduction in new arrhythmias. However, it can be seen that these studies are unlikely to be adequately powered to draw unequivocal conclusions about the role of omega-3 fatty acids in VF.

On the basis of GISSI-P omega-3 fatty acid supplements are recommended in many guidelines^[37,38] and in the UK NICE guidelines if <>[39] However, the main benefit with omega-3 fatty acids occurred early prior to major statin initiation.^[27] It is notable that in GISSI-P there seemed to be a reduction in acute coronary syndromes and arrhythmias in the first 6 months.^[40] In JELIS, the bulk of the benefits seem to occur after 2.5 years but although statin therapy was universal, it was systematically under-dosed by modern guidelines, and unfortunately there was no acute initiation of omega-3 supplements to allow a comparison with the results achieved earlier in GISSI-P. The question that remains is what would be the effect of omega-3 fatty acids added to optimal statin therapy in a high-risk group? Basic calculation suggest that if statins are used as in the Treatment to new Targets study^[41] or other dose comparison studies then the NNT will rise to 480/year (i.e. 97 for a typical 5-year period) which is of borderline utility for health economic purposes.

There remains a need for further studies of adequate size, rigorous design and utilising optimal background therapy. The OMEGA trial because of report in 2008 has randomised 3800 patients within 2–5 days of MI to 1 g DHA–EPA or placebo with a primary end-point of sudden cardiac death at 1 year.^[42] The GISSI – heart failure study is examining the role of statin and or 1 g DHA–EPA in 7000 patients with heart failure in a 2 × 2 design using a primary end-point of total mortality and hospital admission.^[43] In patients with diabetes the ASCEND trial,^[44] 10,000 patients are being randomised to 1 g DHA–EPA or aspirin in a 2 × 2 design to investigate the effects of omega-3 fatty acids on major adverse cardiac events. The results are expected in 2012.

So what can one conclude? There seems to be a clear benefit for a diet or the lifestyle associated with eating fish. The data on supplements is confusing and does not necessarily show clear benefits when added to optimal-lipid management with statins which definitely reduce cardiovascular events^[45] and may themselves have some anti-arrhythmic effects.^[46] So the recipe for a healthy life is fish but no chips. And may be some (red) wine.