September 11th, 2012

More Evidence That Omega-3 Supplements Don’t Work

Once again researchers have failed to find any clinical benefit for omega-3 supplements. In a new meta-analysis and systematic review published in JAMA, Evangelos Rizos and colleagues analyzed 20 randomized controlled trials including  68,680 patients and found no significant effect on any of the endpoints:

  • all-cause mortality: relative risk (RR) 0.96, CI 0.91 – 1.02
  • cardiac death: RR 0.91, CI 0.85 – 0.98 (not significant after correction for multiple comparisons)
  • sudden death: RR 0.87, CI 0.75 – 1.01
  • MI: RR 0.89, CI 0.76 – 1.04
  • stroke: RR 1.05, CI 0.93 – 1.18

The authors report that they found no evidence supporting a beneficial effect related to either lowering triglycerides or reducing sudden death. Regarding triglycerides, they write, “the proposed protective role of omega-3 PUFAs by lowering triglyceride levels is not supported by our study, because our findings do not support an advantage of higher (triglyceride-lowering) doses compared with lower doses of omega-3.” In addition, no benefit was found in preventing sudden death, “thus rejecting a distinct antiarrhythmic mediated omega-3 PUFA effect,” although the authors acknowledge that the evidence was “underpowered to detect a small underlying effect.”

Although early studies suggested a benefit for omega-3 supplements, the accumulation of evidence has resulted in a consistent failure to confirm this benefit, according to the authors. Current evidence, they conclude, does “not justify the use of omega-3 as a structured intervention in everyday clinical practice or guidelines supporting dietary omega-3 PUFA administration.”

2 Responses to “More Evidence That Omega-3 Supplements Don’t Work”

  1. Judith Andersen, AB, MD says:

    Eugene Stead, MD, a Georgia native, trained in Boston, and for many years the Chair of Medicine at Duke University, once told a Duke medical student, presenting data from a research project in endocrinology on Research Day, that, although the statistical analysis suggested no difference between the outcomes in “treated’ vs “untreated” patients, “when one line on a graph of means is always below the other line , even if the confidence intervals overlap, maybe there is something to be learned.” He suggested that either the study was underpowered to show the difference or that the wrong question was being asked– and invited the brave but intimidated student to elaborate. Apart from the stroke issue, although the significance of the difference is not convincing, the “line” for the cardiovascular outcomes mean for omega-3 use is consistently below that of the control groups. I don’t think this is accidental.

    We already know, from many studies of various interventions, that prevention of cardiac and cerebrovascular events do not consistently track together: agents effective in preventing repeat events in CAD, particularly post-intervention, do not necessarily exhibit the same benefits in stroke prevention, that there are gender differences in response– and that there are adverse consequences to “overtreatment ” in pursuit of prevention. And we know that the vascular biologies of the heart and the brain, and their reactions to insult, are not identical or even consistently close. And that study populations in Japan, Scandinavia, central Europe, the many components of Asia, South America, and the northeastern and southern United States are not remotely comparable. Multigenic inheritance, other personal factors including physical activity, dietary preferences and food availability, food vitamin supplementation, environmental exposures and behavioral choices are critical epidemiologic factors. Single agent clinical interventions, even when the agent has been shown to have favorable in vitro and experimental animal effects, have often proven unimpressive. Absence of a critical cofactor ( omega-3 fatty acids are not the only possible beneficial component of an Alaskan diet, just the most potentially profitable),

    What led us to omega-3 popularity was the identification of an “orphan “population consuming foods rich in omega-3s who exhibited a mild bleeding disorder and very little cardiovascular disease, and the subsequent research findings of alteration in vascular and plasma prostaglandin distribution, with a predominance of omega-3 derivatives, more “prostacyclin-like” antithrombotic physiology with less prothrombotic prostaglandin-2 series effects in those whose diets were rich in omega 3 fatty acids.

    We’re often not good at the details. Native populations with anomalous dietary intake and favorable physiology are not eating omega-3 fatty acids. They are eating wild-caught foods: fish with variable diets, including other fish and plant life, aquatic mammals whose diets are primarily small organisms, etc., and their diets are not the only source of their cardiovascular health. The fact that they catch the fish, land the whale, collect and eat the wild plants is the point. We go to the grocery store, buy our food and our supplements and drive home.

    This meta-analysis is very important, in that it demonstrates that a dietary supplement is not a substitute for a life-style. And that by relying on medication and dietary supplements, we are missing the biologic point. Our current medical system relies on medication interventions to neutralize or reverse the adverse effects of our bad biologic behavior. An obese patient with hypertension and metabolic syndrome should use antihypertensives, metformin, etc, but should also be doing the other difficult things to improve his/her health, dietary change and exercise paramount among them. We write prescriptions easily, perform corrective interventions expensively, and rarely really spend the time necessary to convince a moderately overweight hypertensive patient to begin a serious program of exercise and dietary discretion. And we fight an upstream battle with our own well-being at stake: time and reimbursement. Prescriptions are easy; a patient’s failure to improve increases income; interventions are expensive and useful to a medical center’s bottom line; overall human health and well-being is not much enhanced by many of the things we do. Dissing omega-3s may be useful (but they are not harmful), but it would be good if we had a realistic plan to provide an, admittedly difficult but important, alternative pathway to health to our patients that emphasized a return to the things that are missing in our current care plans and require constant reinforcement.

  2. I find it interesting that despite the relatively low dose of omega-3 used in most of the studies and the multiple confounding facts in this meta-analysis, there is a trend for benefit in every category except stroke. Despite this encouraging finding, we are throwing out fish oil derived omega-3 fatty acids as having no value.

    I feel that this again represents a misuse of a meta-analysis. We are over interpreting this result and the conclusion reached is likely in error.

    When will the NIH have courage enough to fund a fish oil study using at least 2 gms of omega-3 fatty acid in a setting without high dose statin? I suspect that such a study would find a significantly different answer than what is seen in this meta-analysis.