Fish Oil and Heart Disease

Rizos EC, Ntzani EE, Bika E, Kostapanos MS, Elisaf MS. Association between omega-3 fatty acid supplementation and risk of major cardiovascular disease events. A systematic review and meta-analysis. JAMA. 2012;308:1024-1032.

 

The authors conducted a systematic review and meta-analysis to assess the role of omega-3 fatty acids on major cardiovascular outcomes. Data were combined from 20 randomized clinical trials that included a total of 68,680 individuals. All clinical trials evaluated the effect of various sources of omega-3 fatty acids, including dietary, dietary supplement, and prescription omega-3 fatty acids, on cardiovascular outcomes.

 

With the exception of 2 trials in which omega-3 administration was based on dietary counseling, the mean omega-3 dose was 1.51 g/d (0.77 g/d EPA and 0.60 g/d DHA). Half of the studies used a dose of 1.0 g/day or greater. The median treatment duration was 2 years, and the maximum duration was 6.2 years. The majority of studies included patients with existing cardiovascular disease. The authors set a statistical significance threshold of 0.0063 instead of the standard 0.05 to account for the substantial variation in study design, patient population, dose, and form of omega-3 intervention among clinical trials.

 

Major cardiovascular outcomes assessed included all-cause mortality, cardiac death, sudden death, myocardial infarction, and stroke. Initially, 3,635 studies were screened and 3,385 were eliminated because they did not include major cardiovascular endpoints, were not randomized, were review articles, or were irrelevant for other reasons. The remaining 250 clinical trials were retrieved and full-text articles examined; 230 were eliminated because they did not meet 1 or more of the eligibility requirements. The final analysis included 20 studies. For the outcome of all-cause mortality, 17 studies with 63,279 patients and 6,295 deaths were analyzed; for cardiac death, 13 studies with 56,407 patients and 3,480 events were analyzed; for sudden death, 7 studies including 41,751 patients and 1,030 deaths were analyzed; for myocardial infarction, 13 studies with 53,875 patients and 1,755 events were analyzed; and for stroke, 9 studies with 52, 589 patients and 1,490 events were analyzed. For each clinical outcome, the relative risk and the absolute risk reduction were calculated using pre-determined statistical models.

 

 

No statistically significant (P<0.0063) association was observed between omega-3 fatty acid intake and all-cause mortality, cardiac death, sudden death, myocardial infarction, or stroke. The authors conclude that their findings do not justify the routine use of omega-3 supplements in clinical practice or guidelines supporting the use of omega-3 fats.

 

The results showed a 4% relative risk (RR) reduction for all-cause mortality (P=0.17), 9% reduction for cardiac death (P=0.01), 13% reduction for sudden death (P=0.06), and 11% reduction for myocardial infarction (P=0.14), indicating that supplementation with omega-3 fatty acids was associated with benefit. With regard to stroke, there was no trend towards benefit. The authors did not employ standard statistical methods in an effort to account for the heterogeneous nature of the studies included in the analysis.

 

Systematic reviews and meta-analyses, like this one by Rizos et al, are an attempt to combine data from multiple studies to generate an integrated result. However, several conditions are critical to conducting a proper meta-analysis, and small violations of those conditions can lead to misleading results. One major condition is that the studies need to be similar in sample size, patient population (healthy or diseased), and intervention (dose and form). As the degree of differences between studies increases, the potential for meta-analysis results to be misleading is amplified.

 

This meta-analysis has no bearing on the importance of omega-3 fatty acids as essential nutrients. Authoritative sources such as the World Health Organization, American Heart Association, and the US Institute of Medicine’s Food Nutrition Board. (See table.) In addition, the 2010 Dietary Guidelines for Americans acknowledge that “moderate evidence shows that 8 oz per week of a variety of seafood, which provide an average consumption of 250 mg per day of EPA and DHA, is associated with reduced cardiac deaths among individuals with and without pre-existing cardiovascular disease,” and offers these recommendations:

 

Experts in the field of omega-3 fatty acids and cardiovascular disease have raised legitimate concern about the subjective nature of the author’s conclusions and call into question the use of a statistical significance threshold of 0.0063. If the authors applied the standard statistical significance threshold of 0.05, the results of the analysis would have shown fish oil supplementation significantly reduced the risk for cardiac death by 9% (RR: 0.91; 95% CI: 0.85–0.98). Similarly, there would have been statistically significant risk reduction for the other outcomes if the statistical threshold of 0.05 had been applied. It is noteworthy that a systematic review and meta-analysis that demonstrated that omega-3 fatty acids benefit cardiovascular outcomes would not be unique or new,^1,2 and therefore would not be a candidate for publication in a top-tier medical journal. 

 

In addition to the fishy statistics used in this analysis, clinicians should consider that the trend toward benefit demonstrated by omega-3 fatty acids was in addition to participants’ already being treated with drugs, lifestyle interventions, and cardiac rehabilitation programs. In other words, any effect detected from omega-3s is above and beyond what is obtainable with state-of-the-art interventions. In the absence of other treatments, it is possible that the effect of fish oil would be more pronounced. However, prohibiting patients’ access to the standard of medical care would be considered unethical and not likely part of a study design to be approved in humans. A conundrum for omega-3 researchers is that as cardiovascular care gets more advanced, the threshold of proof is raised for any treatment to show additional benefit.

 

Another important concept omitted in the interpretation of these data is that omega-3 fatty acids are nutrients, not drugs. There are distinct differences between the evidence that can be obtained from testing drugs using randomized, controlled trails and the evidence for the testing of supplemental nutrients, such as omega-3 fatty acids. For example, the analysis did not assess the baseline omega-3 status or biomarkers of omega-3 intake throughout the study. Dietary intake of fish, flax, and other sources of omega-3 fatty acids at baseline and throughout a study will impact treatment and placebo groups as well as study conclusions. Studies that included patients already receiving therapy for cardiovascular disease may mask the modest and more long-term benefits of omega-3 fatty acids.

 

In addition, the review did not explore the importance of balancing the omega-3 to omega-6 fatty acid ratio. What if the health benefit of omega-3 fats results from the influence of optimally balanced omega-3 to omega-6 on systemic inflammation over a span of decades? The authors make no mention of these important factors in their discussions, but instead make the headline-grabbing conclusion that their findings “do not justify the use of omega-3 as a structured intervention in everyday clinical practice or guidelines supporting dietary omega-3 PUFA administration.” I believe that it is borderline irresponsible to oppose guidelines that encourage omega-3 administration in a population that does not consume the recommended amounts of food containing EPA and DHA.³ Nutrition-oriented clinicians are acutely aware that cold-water fatty fish such as anchovies, sardines, and herring are not favorite foods of most Americans or are not readily available due to financial or other constraints. It is for these reasons that omega-3 supplementation is warranted to achieve the benefit of these essential nutrients.

 

It would have been much more scientifically responsible and accurate for the authors to provide a conclusion that recognized the limitations of the analysis and the incrementally growing nature of science. For example, even a conservative researcher might conclude from this analysis that for mature patients (average age of 63) with existing cardiovascular disease, administration of 1–2 g of omega-3 fatty acids, when added to standard of care for heart disease, moderately influenced risk for major cardiovascular outcomes. However, these data do not address the effects of higher doses administered for a longer period of time, treating patients earlier in the disease process, or in patients who forgo or are non-compliant with prescriptions.

 

The recent systematic review and meta-analysis by Rizos et al should not change clinicians’ efforts to get patients to include a variety of sources of omega-3 fatty acids, including dietary supplements to fill dietary gaps, and to integrate omega-3 supplements into cardiovascular care for optimal outcome.