Dietary Fats and Heart Health: Big Numbers, but Questions Linger

Cardiovascular Disease

ABSTRACT & COMMENTARY

Dietary Fats and Heart Health: Big Numbers, but Questions Linger

By Howell Sasser, PhD

Associate, Performance Measurement, American College of Physicians, Philadelphia, PA

Dr. Sasser reports no financial relationships relevant to this field of study.

Synopsis: A meta-analysis and a clinical trial, both published recently, found that the evidence of cardiovascular benefit with increased consumption of omega-3 fatty acids is weak at best. Large study samples make these results compelling, but caution is warranted in interpreting and applying them in practice.

Sources: Chowdhury R, et al. Association of dietary, circulating, and supplement fatty acids with coronary risk: A systematic review and meta-analysis. Ann Intern Med 2014;160:398-406.

AREDS2 Writing Group. Effect of long-chain omega-3 fatty acids and lutein + zeaxanthin supplements on cardiovascular outcomes: Results of the Age-Related Eye Disease Study 2 (AREDS2) randomized clinical trial. JAMA Intern Med 2014;174:763-771.

Chowdhury and colleagues conducted a systematic review of evidence published before July 2013. They identified 32 observational and 27 experimental studies that reported the risk of coronary disease associated with measured levels of circulating fatty acids or with fatty acid supplementation.

The observational studies included dietary surveys and assessment of fatty acid levels in blood (measured in whole blood, plasma, serum, or red blood cells, depending on the study) or adipose tissue. Follow-up was between 5 and 23 years. All were judged to be of high or medium quality, and all adjusted for a number of potential confounders, including age, sex, tobacco use, history of diabetes, and blood pressure. The experimental studies included dietary and supplement-based interventions. All included random assignment in their designs. The outcome of interest, risk of coronary disease, was constructed as a composite endpoint composed of fatal or non-fatal myocardial infarction (MI), coronary heart disease, coronary insufficiency, coronary death, angina, and angiographic coronary stenosis.

Meta-analytic regression models were used to produce risk estimates aggregated over all studies. Separate sets of models were produced for observational studies of circulating fatty acids (OBS-C), observational studies of dietary fatty acid intake, and interventional studies. The "population" of each model varied by the number of studies that reported a specific fatty acid, ranging between 140,000 and 420,000 person-years in the observational studies, and between 7100 and 23,000 participants in the intervention studies.

The models were assessed for potential bias by stratification on sex, year of study entry, the dietary assessment tool used, duration of follow-up, how the outcome was defined (i.e., which elements of the composite outcome each individual study included), and statistical factors involved in the way the models were constructed. No significant differences in the estimates were noted.

Relevant results of the various models are shown in Table 1. Reductions in cardiovascular risk with rising omega-3 consumption were generally modest. The main exceptions to the lack of statistical significance were the results for eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA) when measured through dietary intake. Interestingly, the risks with rising consumption of saturated fats and trans-fats, though elevated, were also small and statistically non-significant — 1.06 (95% CI 0.86-1.30) and 1.05 (95% CI 0.76-1.44), respectively.

A second paper reports the results of a large randomized trial. The Age-Related Eye Disease Study 2 (AREDS2) was a clinical trial of dietary supplementation for the slowing of progressive blindness in older adults. Participants were enrolled on the basis of having moderate or advanced macular degeneration in at least one eye rather than on the basis of their cardiovascular health. However, potential study enrollees with existing cardiovascular disease (CVD) were eligible if their condition was judged to be "stable," and if their initial cardiac event was more than 12 months prior to enrollment.

The combination of EPA and DHA was one of the trial’s experimental interventions, and the investigators collected data on cardiovascular events as well as eye data among study participants. The cardiac endpoints recorded included MI, hospitalized acute coronary syndrome, coronary artery bypass surgery, hospitalized congestive heart failure, unexpected death, resuscitated cardiac arrest, cardiac angioplasty or stent placement, cardioverter-defibrillator placement, transient ischemic attack, ischemic or hemorrhagic stroke, or carotid artery stenting, angioplasty, or endarterectomy. Several composite endpoints were designated before the trial began, as shown in Table 2.

The study enrolled 4203 participants who were randomly assigned in a factorial fashion to EPA+DHA (350 mg and 650 mg per day, respectively) plus placebo, another active intervention (lutein + zeaxanthin) plus placebo, both active agents, or double placebo. The median age at enrollment was 74 (range 50-85). The median duration of treatment — and of follow-up — was 4.8 years.

Table 2 shows the risk of cardiovascular events among those receiving EPA+DHA as compared with all others in the study (those receiving lutein + zeaxanthin plus placebo, and those receiving double placebo). Results for the main outcome (CVD morbidity and mortality) were similar when the analysis was restricted to a comparison of those receiving both interventions with those receiving neither (relative hazard = 0.89; 95% CI, 0.67-1.19).

Commentary

These findings are compelling, if for no other reason, because of the large sample sizes and rigorous methods of the studies that produced them. However, there are some issues that one should bear in mind before deciding how to act on them.

The first relates to the timing and duration of "treatment" and the likely trajectory of its effects. Most, but not all, scientific studies are designed to show the time-limited (< 10 years) effects of time-limited therapies. This leads to a bias toward interventions that produce changes large enough to be measurable in a short time. In the case of dietary fats and heart disease, this seems problematic. There is evidence that fatty streaks in the arteries develop in adolescence and plaques begin to form in young adulthood.^1,2 A dietary change introduced in mid-life may not have the same effect as one begun at a young age and maintained over time. More to the point, its impact may not be measured accurately in a study which lasts only a few years, regardless of where in the life course the study takes place. Interestingly, the findings of prominent long-term longitudinal studies that have included dietary measurements have been divided on the value of omega-3 fats — the Chicago Western Electric and Nurses’ Health Studies showed protective effects, but the Health Professionals’ Follow-Up and Physicians’ Health Studies did not.^3-6 Diet is a lifelong "process," as are its effects. Even studies that are careful to recruit only those who are free of (clinical) CVD at baseline have serious issues with bias if the average age of those recruited is in the typical clinical CVD window.

A second issue relates to the place of omega-3 fats in the totality of the diet. It is arguably the case that a key assumption of diet research involves the role of omega-3s as markers for broader eating patterns. This can take two forms. One is the assumption that those who consume higher amounts of these fats also consume higher amounts of other desirable foods (whole grains, fruits and vegetables, low-fat protein) and lower amounts of undesirable foods (refined grains, high-fat meat and dairy, refined sugar). The second form assumes that omega-3-rich foods crowd out other less desirable protein sources (in other words, tuna steak replaces cheeseburgers in an otherwise unchanged volume of protein consumption).

These premises may be valid in some cases, but there is evidence that they are not always so. A recent study found that statin users consume more calories and fat on average than non-users, and that statin users’ consumption of calories and fat has increased over the past 15 years (i.e., within-group change) while non-users consumption has not changed markedly.⁷ It seems reasonable to ask if a similar pattern has taken place with respect to omega-3 consumption as it has become popularized and commercialized. If getting a daily dose of omega-3s, perhaps in the form of fish oil capsules, is seen as inoculation against the effects of less healthy foods, an omega-3-rich diet may not equate to a generally healthy diet.

The total diet is also important in connection with omega-3s’ hypothesized mechanism of action. It is known that omega-3 and omega-6 fats "compete" to form eicosanoids, short-lived compounds with both beneficial and harmful (specifically inflammatory) effects.⁸ There is evidence that eicosanoids formed from metabolism of omega-3s are anti-inflammatory (or at least less inflammatory), so the ratio of omega-3 to omega-6 fats may have consequences for the condition of tissues such as vascular endothelium which in turn play a role in cardiovascular disease. This means that intake of EPA and DHA cannot be viewed in isolation, but rather must be considered in proportion to omega-6 intake. Risk data for varying omega-3/omega-6 ratios were not reported by either paper.

Additional cautions apply specifically to the AREDS2 Study. We should note that it was designed and powered to test the effect of an intervention on a different disease process. Even though a cardiovascular endpoint was defined before the study began, there may have been too few events to produce a statistically convincing finding. The authors do not report any post-hoc power calculations, so we do not know if or to what extent the study was underpowered for its cardiovascular outcomes. The enrollment criteria related to CVD also appear to have been quite lenient — "stable" disease covers a wide range of types and severities of pathology. For these reasons, the AREDS2 results are best viewed as suggestive but not conclusive.

Less as a limitation than a source of speculation is the difference in findings by Chowdhury’s group between total omega-3 and EPA and DHA. EPA and DHA were borderline-significant in the dietary analyses and showed greater risk reductions than total omega-3. This suggests that EPA and DHA have different — and perhaps more desirable — effects than other fatty acids in the same class. Had the AREDS2 study shown a similar magnitude of reduction, or even simply a statistically significant reduction, this would seem more convincing. As it is, we may conclude that the final verdict on these specific fatty acid subclasses has not yet been rendered.

Despite these limitations and conflicting findings, it is not possible to dismiss these studies’ results lightly. What, then, are clinicians and patients to make of them? It is clear that manipulation of dietary fats is not a "silver bullet" for reducing cardiovascular risk. Even if the risk reductions reported in these two studies were statistically significant, they would not be of the same magnitude reported for drug therapies or exercise.⁹ It also seems unlikely that mid- to late-life changes in diet can undo the effects of eating habits sustained over many years. Although some observational data included in the meta-analysis covered as much as 23 years of follow-up, the lack of strong and consistent findings only further strengthens the case against a meaningful protective effect.

Yet it is surely unwise to abandon all recommendations about dietary fats. As in many areas of health and wellness, perhaps the best course is to encourage patients to form a balanced "portfolio" of risk reduction strategies — diet, exercise, drug therapy, and stress control, among others — at as young an age as possible. While we await the final word on dietary fats, a combined approach may provide a greater net benefit than the sum of its parts, and avoidance of overemphasis on any one element will help to prevent neglect of the others.

References

1. McGill HC, et al. Origin of atherosclerosis in childhood and adolescence. Am J Clin Nutr 2000;72:1307s-1315s.
2. National Heart, Lung, and Blood Institute. Expert Panel on Integrated Guidelines for Cardiovascular Health and Risk Reduction in Children and Adolescents. Full Report. 2012. http://www.nhlbi.nih.gov/guidelines/cvd_ped/peds_guidelines_full.pdf. Accessed May 30, 2014.
3. Daviglus ML, et al. Fish consumption and 30-year risk of fatal myocardial infarction. N Engl J Med 1997;365:1046-1053.
4. Hu FB, et al. Fish and omega-3 fatty acid intake and risk of coronary heart disease in women. JAMA 2002;287:1815-1821.
5. Ascherio A, et al. Dietary intake of marine n-3 fatty acids, fish intake, and the risk of coronary disease among men. N Engl J Med 1995;332:977-982.
6. Albert CM, et al. Fish consumption and risk of sudden cardiac death. JAMA 1998;279:23-28.
7. Sugiyama T, et al. Different time trends of caloric and fat intake between statin users and nonusers among US adults: Gluttony in the time of statins? JAMA Intern Med Published online April 24, 2014. doi:10.1001/jamainternmed.2014.1927 [Epub ahead of print].
8. Wall R, et al. Fatty acids from fish: The anti-inflammatory potential of long-chain omega-3 fatty acids. Nutr Rev 2010;68:280-289.
9. Green DJ, et al. Exercise and cardiovascular risk reduction: Time to update the rationale for exercise? J Appl Physiol 2008;105:766-768.