The Cochrane Database of Systematic Reviews has just updated its own review: Omega-3 fatty acids for the primary and secondary prevention of cardiovascular disease (Abdelhamid, Brown, Brainard, et al., 2018). Here’s our take on it:
Michael Pollan, the brilliant food writer, reckoned you could sum up what to do about nutrition and diets in 7 words; “Eat food, not too much, mostly plants.” That sums up both what’s best for humans and what’s best for the planet.
We reckon you can sum up what’s wrong with evidence-based medicine (EBM) in 10 words; “Evidence is a waste of data; systematic reviews are palimpsests.” You can use that as a knife to quickly dissect this study.
There are many things wrong with this review. Somebody’s PR department has spun the review’s “no clear evidence of benefit” into “evidence of no benefit” – absence of evidence becoming evidence of absence. And clearly the media were entirely happy to take that one and run with it.
Systematic Reviews are Palimpsests
What’s a palimpsest? Back when things got written on vellum, an animal skin, not on paper, you didn’t throw it away; you recycled it and wrote over the original. It was called a palimpsest. A systematic review gives an opportunity to write over the conclusions of a whole list of papers with your new version of the truth. You do that by the way that
you select and exclude them. For instance there was a meta-analysis (that’s a systematic review with more numbers) in 2005 that concluded that vitamin E supplements significantly increased the risk of death (Miller, Pastor-Barriuso, Dalal, et al., 2005). The way they did that was to rule out any study with less than 10 deaths – when fewer deaths was exactly the outcome they were supposed to be looking for. The reason they gave for doing that was “because we anticipated that many small trials did not collect mortality data.” We’re not buying it; they used it as a trick to enable them to get the negative result they wanted – to over-write the findings of a long list of original studies.
And here we have authors doing the very same thing in this omega-3 study – and upping the ante slightly. Now the threshold is 50 deaths. Fewer than that and your study is ruled out of the final, supposedly least biased, analysis…on the grounds that it’s more biased. We don’t know how they could keep a straight face while saying (our interpretation); “The studies with fewer deaths showed more benefit from omega-3s, so we excluded them.”
At least that’s what happened back in 2004 when the first version of this came out (Hooper, Thompson, Harrison, et al.,2004). But this is the 8th update (we think) and they no longer bother to tell you about what they included or excluded in detail, so we can only assume that if they had changed that exclusion they would have told us.
The weird thing is that they are allowed to do it. Nutrition researcher Dr. Steve Hickey has shown that in systematic reviews there is generally control for bias in the included studies, but none for bias in the actual review and its authors. (Hickey & Noriega, 2009; Hickey, Hickey & Noriega, 2013) They found not one example of adequate blinding among 100 Cochrane reviews (like this one); they could all be palimpsests. Do we know that they are fake? No, but it doesn’t matter: what we do know is that we can’t trust them. Nor can we trust this Cochrane review. Things haven’t changed since 2004.
Evidence is a Waste of Data
Evidence is what lawyers and courts use to find someone Guilty or Not Guilty, and we all know how that can go wrong. It’s a binary system: you’re either one or the other. But at least if you’re on trial all the evidence should be about you and whether you did the crime. In EBM the evidence is all about populations, not about individuals. When a doctor tells you “There’s a 1 in 3 chance this treatment will work” he is required to base that on big studies, or even systematic reviews. You don’t, and you can’t, know what that means for you because very likely you don’t fit the population profile. As Steve Hickey (again) said, the statistical fallacy underlying all this states that you have one testicle and one ovary – because that’s the population average!
The authors of this study update started off with about 2100 papers that looked relevant. They then excluded 90 per cent of them for various reasons – some of them good reasons, some not. A smarter way to work would be to data-mine them and look for useful information about sub-groups and sub-effects in all the papers. Is there a particular reason omega-3s might work for you and not for others? Perhaps you can’t stand fish, or are allergic to them, and so are deficient in omega-3s. But the review system doesn’t allow it, it insists on overall conclusions (about populations), and that’s a colossal waste of data.
It also confounds the overall finding of the review – it biases it in fact. Here’s an example: while most subgroups that made it to the final analysis showed a small reduction in risk from taking omega-3s in one form or another (pills, food, whatever), those who got it from supplemented foods, which we understand means stuff like margarine with added omega-3, showed a 4.3-fold death risk increase! The problem here is that the effects of omega-3 fatty acids cannot be studied alone as if they were a drug. What counts are all the other components of the diet that affect a person’s health. Processed foods and drinks that contain many unhealthy ingredients can’t be made healthy by adding small doses of vitamins, minerals, and omega-3 fatty acids. In fact, many processed foods that contain small doses of vitamins and other essential nutrients are unhealthy because they contain large doses of sugar, salt, and harmful ingredients such as preservatives, dyes, and other non-food items.
Why Lipids are so Important
Part of the problem is that lipids are truly complicated, and not many people, patients, doctors or even scientists, understand them well. You need a good understanding of lipid metabolism to appreciate the difference in metabolism and impact between alpha-linolenic acid (ALA, in food such as oily fish) and extracted oils such as EPA and DHA that are only found at high levels in omega-3 supplements. At these levels they are effectively new to nature; nobody, indeed no mammal, was exposed to really high doses of DHA until we invented fish oil supplements (Cortie & Else, 2012). Miss that fact and you miss the difference between having people eat fresh oily fish or just using omega-3 margarine!
We know from a variety of studies that a diet containing generous portions of green leafy and colorful vegetables and fruits, moderate portions of eggs, fish, and meat, and supplements of adequate doses of essential nutrients (vitamins and minerals) is effective at lowering the risk for cardiovascular disease. Adequate doses of both omega-3 (in flax oil, walnuts, fish) and omega-6 (in seed oils such as canola, soybean, peanut) fatty acids are essential for health. Although essential, omega-6 fatty acids are thought to contribute to inflammation throughout the body whereas omega-3 fatty acids are anti-inflammatory. Omega-3 fatty acids are essential for most body organs including the brain but are found in lower levels than omega-6 fatty acids in most vegetables. Risk for cardiovascular disease can be lowered by adequate doses of vitamins C (3,000-10,000mg/d), D (2,000-10,000 IU/d), E (400-1,200 IU/d), and magnesium (300-600 mg/d) in addition to an excellent diet that includes an adequate dose of omega-3 fatty acids (Case, 2017).
Originally published in the Orthomolecular Medicine News Service on August 6, 2018.
OMNS Archive – http://orthomolecular.org/resources/omns/index.shtml
Abdelhamid, A, Brown TJ, Brainard JS, et al., (2018) Omega 3 fatty acids for the primary and secondary prevention of cardiovascular disease. Cochrane Database of Syst Rev. 7:CD003177
Case HS (2017) Orthomolecular Nutrition for Everyone. Turner Publication Co., Nashville, TN. ISBN-13: 978- 1681626574
Cortie CH, Else, PL, (2012) Dietary docosahexaenoic acid (22:6) incorporates into cardiolipin at the expense of linoleic acid (18:2): Analysis and potential implications. International Journal of Molecular Sciences, 13(11): 15447-15463.
Hickey S, Noriega LA. Implications and insights for human adaptive mechatronics from developments in algebraic probability theory, IEEE, UK Workshop on Human Adaptive Mechatronics (HAM), Staffs, 15-16 Jan 2009.
Hickey S, Hickey A, Noriega LA, (2013) The failure of evidence-based medicine? Eur J Pers Centered Healthcare 1: 69-79.
Hooper L, Thompson RL, Harrison RA, et al.. (2004) Omega 3 fatty acids for prevention and treatment of cardiovascular disease. Cochrane Database Syst Rev. (4):CD003177.
Miller ER, Pastor-Barriuso R, Dalal D, et al., (2005) Review Meta-Analysis?: High-Dosage Vitamin E Supplementation May Increase. Annals of Internal Medicine, 142(1), pp.37-46.