As a proponent of consuming fatty fish (sardines, salmon, etc.), I was interested to read the new fish oil study; as an opponent of meta-analyses, however, not so much. A meta-analysis is a type of study whereby the researcher thinks of something they want to prove, then cherry picks studies that best support their point. Or perhaps I’m just biased. Nonetheless,
Association between fish consumption, long chain omega 3 fatty acids, and risk of cerebrovascular disease: systematic review and meta-analysis (Chowdhury et al., 2012)
In brief, regarding whole fish consumption, 3 servings per week reduced stroke risk by 6% and 5 servings by 12%. Surprisingly, there was no effect of fish oil pills that contained ~1.8 grams of long chain omega 3 fatty acids. What this study lacks is any information about the dose of EPA and DHA (the major bioactive fatty acids in fatty fish); and with 38 studies analyzed, I’m not about to try to figure it out (sorry team)… a serving of fish can have anywhere from 0 to 1 gram of EPA and DHA; 1.8 grams of long chain omega 3 fatty acids can have anywhere from 0 to 1.8 grams of EPA and DHA. Therefore, I’ll resort to reviewing two of my favorite fish studies of all time: DART and GISSI. For a more detailed review of fish oils and these studies, check out The poor, misunderstood calorie (chapter 9).
DART was a 2 year long randomized intervention trial in men who just had a heart attack. A “primary prevention trial” includes healthy subjects with the goal of preventing any incidents; a “secondary prevention trial” includes patients with a disease and the goal is to prevent worsening of their disease. DART was the latter, and tested 3 dietary interventions: 1) eat less saturated fat; 2) eat more fatty fish; and 3) eat more cereal fibre.
The interventions were simple and effective; here’s what the patients reported eating:
Findings: 1) reducing saturated fat intake had no effect on mortality; 2) eating more fatty fish significantly decreased mortality; and 3) eating more fibre increased mortality.
Translation: eat more salmon and sardines; and eat fewer grains. This isn’t heresy, just get your fiber elsewhere, like from fruits and veggies.
GISSI was another secondary prevention trial, although they opted for fish oil pills instead of the dietary route and included more patients. Here’s who died:
Those who received fish oil pills experienced a 14% reduction in all-cause mortality.
What can DART & GISSI tell us about the new Chowdhury study? Mea culpa! I may have been too harsh on the concept of a meta-analysis.
Disclaimer: there were a lot of differences between DART and GISSI. I’m going to ignore most of them.
DART tested 3 servings of fatty fish per week, which provides a weekly dose of approximately 3,000 mg EPA & DHA. GISSI tested 1 gram of fish oil daily, which provides a weekly dose of approximately 6,160 mg EPA & DHA. Thus, the dose of EPA & DHA from fish oil pills in GISSI was twice as much as that from fatty fish in DART; but the reduction in mortality was two times greater in DART. Half the dose, twice the benefits… Fatty fish FTW!
So some of Chowdhury’s findings can be at least partly justified (fatty fish are good for you). But why no effect of fish oil pills? Chowdhury included both primary and secondary prevention trials; perhaps fish oil pills are more beneficial to unhealthier patients, ie, those who have suffered a heart attack. Thus, including primary prevention trials would dilute/attenuate any perceived benefits of fish oil pills. Indeed, this is what is seen for statins: no effect on mortality in primary prevention trials, but reduced mortality in [some] secondary prevention trials.
Conclusions: 1) eat more salmon or sardines. If your concerned about heavy metals, farmed vs. wild, or interested in the EPA and DHA content of salmon, see Fish blog, take I. And 2) rely on prescription fish oil pills, krill oil, or high quality OTC fish oil supplements only if you absolutely cannot tolerate eating fish. Especially if you’re unhealthy.