ALPHA OMEGA: Effect of low doses of n-3 fatty acids on cardiovascular diseases in post-MI patients 

---

	Presenter
	Presenter | see Discussant report | read the press release
Kromhout, Daan (Netherlands)

List of Authors:
Daan Kromhout, Erik J. Giltay, Johanna M. Geleijnse, on behalf of the Alpha Omega Trial study group

Abstract:

Background:
A recent meta-analysis of randomized controlled trials of fish oil supplementation with at least 1 gram of the n-3 fatty acids eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) per day, carried out in cardiac patients, showed a 20% reduction in cardiac mortality. Animal experimental research showed that low-doses of these fatty acids produce large changes in heart membrane composition and prevent cardiac arrhythmia, in particular ventricular fibrillation. For alpha-linolenic acid (ALA), the n-3 fatty acid from vegetable origin, less evidence is available for a protective effect on cardiovascular diseases. We designed the Alpha Omega Trial to examine the effects of low-doses of n-3 fatty acids (EPA+DHA and/or ALA) in margarine on cardiovascular diseases (CVD) in stable Dutch post-MI patients. The Alpha Omega Trial is registered with ClinicalTrials.gov, Identifier NCT00127452.
Hypotheses tested:
1. A low-dose of EPA+DHA or ALA is inversely associated with major cardiovascular events
2. A low-dose of EPA+DHA or ALA is inversely related to fatal coronary heart disease
3. A low-dose of EPA+DHA or ALA is protective against clinical sequelae of ventricular fibrillation
Methods:
The Alpha Omega Trial is a multicenter, randomized, double-blind, placebo-controlled trial with a 2x2 factorial design resulting in 4 treatment groups (see Figure).

Subjects were enrolled in the trial from April 2002 to December 2006 and were randomly assigned to different margarine spreads that provided low doses of n-3 fatty acids or placebo and followed for 40 months. Dosages were approximately 400 mg/day for EPA+DHA and approximately 2 g/day for ALA. A total of 4,837 post-MI patients were recruited in collaboration with cardiologists of 32 hospitals in the Netherlands. We included men and women aged 60-80 years with a documented history of MI up to 10 years before randomization. Margarine spreads were targeted at an average intake of 20 g/day. In active treatment arms, oleic acid was exchanged for either 400 mg of EPA+DHA, 2 g of ALA or the combination of 400 mg of EPA+DHA and 2 g ALA. The ratio of EPA to DHA in the margarine that contained fish oil was 3:2. The four margarines were similar in taste, texture and color.
At baseline data were collected on demographic factors, lifestyle, medical history, and current health status. Subjects were physically examined including anthropometry, blood pressure, heart rate, and blood sampling. Compliance was assessed at baseline, and after 20 and 40 months of follow-up, by measuring fatty acids in plasma cholesteryl esters in three different samples of 810 patients each. Self-reported non-fatal cardiovascular events, including MI, cardiac arrest, and stroke and cardiac interventions (PCI, CABG, ICD) were verified against hospital records. Vital status was monitored and there was no loss to follow-up. Information on the causes of death was obtained from the death certificate and the patient’s GP, and fatal events were coded by an independent Endpoint Adjudication Committee. Statistical analysis concerning n-3 fatty acids and CVD were carried out by an independent statistician, according to a predefined Statistical Analysis Plan (see www.alphaomegatrial.com) .
Results:
The cohort included 3,783 (78%) men and 1,054 (22%) women. The patients were on average 69 years old. Approximately 17% of the patients were smokers, and 80% of the men and 54% of the women used alcohol. The prevalence of overweight was 53% and of obesity 24%. Almost all patients (98%) used antithromotic agents, 90% used antihypertensive drugs and 86% lipid modifying drugs, mainly statin. Anti-arrhythmic drugs were used by 3%. Diabetes was present in 1,014 (21%) patients.
During 40 months of follow-up, 671 patients developed a major cardiovascular event. The risk of major cardiovascular events during follow-up was not reduced in patients who received EPA+DHA or ALA, compared to placebo. In women who received ALA, however, there was a borderline significant 27% (HR=0.73; 95% CI: 0.51 to 1.03) reduction in this primary endpoint. We also carried out an exploratory analysis in patients with diabetes who showed a 49% (HR=0.51; 95% CI: 0.27 to 0.97) reduction in coronary heart disease mortality in patients who received EPA+DHA. In these patients EPA+DHA reduced also the number of ventricular arrhythmia-related events by 49% (HR=0.51; 95% CI: 0.24 to 1.11) and ALA by 61% (HR=0.39; 95% CI: 0.17 to 0.88). Adverse events did not differ among the study groups.
Conclusions:
From this trial we conclude that low doses of n-3 fatty acids do not reduce major cardiovascular events, except a possible beneficial effect of ALA in women, which needs confirmation in other trials. The results obtained in an exploratory analysis in diabetic patients at high risk for recurrent cardiovascular events suggest that n-3 fatty acids may prevent ventricular arrhythmia-related events. Also this protective effect of n-3 fatty acids in diabetic patients needs further confirmation.

---

	Discussant
	Discussant | see Presenter abstract
Tavazzi, Luigi (Italy)

Report

Background:
Several studies have investigated the role of EPA and DHA, especially with respect to CV disease prevention. The DART trial found a 29% reduction in mortality among 1015 men randomly advised to eat at least two portions of oily fish per week, compared with a comparable number of men not so advised. No reduction in nonfatal myocardial infarction (MI) was observed.

Subsequently, the GISSI-Prevenzione trial tested a n-3 PUFA supplement (850 mg/day of EPA+ DHA as purified ethyl esters) among 11,323 patients with recent MI. At 3.5 years, n-3 PUFA treatment significantly reduced the primary endpoint (combined death, nonfatal MI, and stroke) compared with the control group. This benefit resulted largely from a 45% reduction in sudden cardiac death (SCD). No reduction in non fatal events was observed.

The Japan EPA Lipid Intervention Study (JELIS) examined the effects of EPA ethyl esters (1.8 g/day) or usual care and showed a significant reduction of the primary endpoint of the study (i.e., major cardiovascular events, defined as the development of unstable angina, fatal or nonfatal MI, SCD, and/or revascularization) in over 18.000 hypercholesterolemic Japanese men and women, all on low-dose statins, who were followed for 4.6 years. No reduction of SCD was observed in JELIS, but the rate of SCD in the trial (and in Japan) was so low that the study had no statistical power to show it. GISSI-HF was a randomized, large scale, double-blind, placebo-controlled trial showing in the adjusted analysis that n-3 PUFA (850–882 mg/d) significantly reduced mortality by 9% (P=0.041) and by 8% all-cause mortality or hospitalization for a cardiovascular cause (P=0.009) in 7046 patients in class II–IV chronic heart failure (HF).

For alpha-LNA there is some evidence for a beneficial effect in the cardiovascular system although a recent review concluded that LNA supplementation did not reduce cardiac-related mortalities. A meta-analysis of 5 prospective cohort studies showed an inverse relationship between alpha-LNA intake and fatal CHD, indicating an overall risk reduction around 20% that was borderline statistically significant. Data on the effect of ALA intake on stroke are scanty.

Comment:
Why n-3 PUFA treatment gave neutral results in a trial [The Alpha-Omega trial (AOT)] conducted in a country with low-intake of n3-PUFA and therefore expected to have a marked benefit on cardiovascular disease?

A number of explanations can be considered.
Firstly, were the tested hypotheses sound enough? The answer may be yes for the EPA+DHA hypothesis, while the contrary can be said for the alpha-LNA hypothesis. However, the AOT results seem to go in the opposite direction, with more “evidence” produced for alpha-LNA as compared to EPA+DHA.
Secondly, the dose of EPA+DHA used in this trial was less than half the amount tested in previous studies. Was it too low? The level of EPA and DHA in AOT in plasma were approximately changed from 2% to 3 mol % in the AOT patients. The same figure in GISSI-Prevenzione and GISSI-HF were 4-5 mol% at baseline and 7-8 mol% during the course of the study. Baseline levels of EPA and DHA were definitely higher in the JELIS trial (approximately 7 mol%).
Thirdly, did flaws in the assumptions on which the trial design was based affect the trial conduct? The AOT investigators decided to change their primary endpoint as well as the study duration during the course of the study. The original primary endpoint was incidence of CHD mortality and it was expected to be 4% per year in 4,000 post-MI patients aged 60 to 80. The study was supposed to have 80% power for a 25% relative risk reduction of the primary endpoint during 3 years of treatment for EPA-DHA and 67% power for a 20% reduction for ALA. Because the rate of fatal CHD events was lower than expected, the AOT investigators decided an increase in sample size to a total of 4,837 subjects and an extended duration of intervention from 36 to 40 months. For the same reason, in 2009 (the study started in 2002), the original primary endpoint of “fatal CHD” was replaced by “major cardiovascular events” including all non fatal CV events, and revascularization procedures. The insufficiency of sample size is not surprising. For instance, the rate of CHD mortality observed in the post-AMI patients of GISSI-Prevenzione (which was started in 1993 and published in 1999, i.e., 3 years before the kick-off of the AOT trial) was almost half that expected in the chronic post-MI patients enrolled (3,7 years after AMI, median) in AOT.

Moreover, the extended primary end-point, including many different components driven by the softer components (non fatal events and particularly coronary revascularization) exposed the trial to the major risk of dilution. Though the authors do not report the absolute figures of the events in each group, the question is whether the components of the revised composite primary endpoint are specific enough for the specific mechanisms of action of n-3 PUFA. Since most of the benefit of n-3 PUFA seems to be due to their antiarrhythmic effects, the choice of a composite primary endpoint is particularly challenging, because the strength of the benefit (if any) on each of the components could be markedly different.

In conclusion, the AOT investigators should be praised for conducting such an independent, large-scale clinical trial with limited governmental funds. In particular the food-based nature of the trial should be commended as an attempt to base on science the widespreading commercial advising. However the relatively small sample size, the low-dose of EPA-DHA and the probably unavoidable difficulties related to the (too?) pragmatic approach of the trial do not allow to draw conclusions about the effect of n-3 PUFA in the studied population.

The content of this article reflects the personal opinion of the author/s and is not necessarily the official position of the European Society of Cardiology.