Omega-3s meta-analysis didn't look at best studies, experts say

On Monday, the Journal of the American Medical Association released a study that concluded that omega-3s were not helpful to patients who already have cardiovascular disease. But industry sources are quick to point out the meta-analysis flaws.   

A meta-analysis released Monday by the Journal of the American Medical Association concluded that omega-3s intake didn’t help patients who already have cardiovascular disease (CVD). Industry sources took issue with the structure of meta-analysis and disagreed with its conclusions, citing a number of supportive studies that were excluded.

The meta-analysis, conducted by four Korean researchers, looked at 1,007 articles from which they gleaned 14 randomized, double-blind, placebo-controlled studies covering a total of 20,485 patients with a history of CVD. They concluded that there was “insufficient evidence of a secondary preventive effect” for these patients.

All of these patients were already sick and were using various pharmaceutical treatments. It’s a huge confounding factor, one that the meta-analysis did not control for. And even 20,000-plus patients, given the differences between the chosen studies, may be too small a sample size under those conditions to parse out whatever effects omega-3s might be exerting.

"That’s one of the issues with most of the secondary prevention studies for omega-3s," said Adam Ismail, executive director of the Global Organization for EPA and DHA Omega-3.

“That doesn’t mean omega-3s have no benefit in that population. It means you have to have a really big population size in order to see an effect because they are already giving things that you know have an effect. So any additional effect is likely to be smaller than just giving them omega-3s and nothing else."

Duffy McKay, vice president of scientific & regulatory affairs for the Council for Responsible Nutrition (CRN), also took issue with the study’s methodology.

“The dose range [of the studies chosen] was significant,” he said. “You go all the way from 400 mg to just shy of 6 grams. When you start to get to 1,000mg a day you’re adding a nice load of omega-3s and usually you see some effects. When you get up toward three grams you start to affect inflammatory processes.

“Are those appropriately combinable in a meta-analysis? That leaves me with some questions.”

Meta-analysis excluded most important omega-3 studies

Both Ismail and McKay noted that the structure of the meta-analysis, focused on placebo-controlled studies, prevented the inclusion of two of the most powerful pieces of information supporting the connection of omega-3s and cardiovascular protection—the GISSI and JELIS studies. Both of these studies were intervention, or “open label” studies, meaning there was no attempt to control for placebo effect.

The GISSI study, conducted in Italy in the mid 1990s, took a three-and-a-half year look at patients with CVD. One group took their prescribed treatments and nothing else, one added vitamin E, while a third group added 1g daily of fish oil. The study found no effect for the vitamin E, but the fish oil group showed a strong, positive response. The researchers concluded, “dietary supplementation with n-3 PUFA (polyunsaturated fatty acids) led to a clinically important and statistically significant benefit.”

The JELIS study was another positive piece of the literature excluded by the structure of the meta-analysis. This study, conducted in the late 1990s in Japan, looked at more than 18,000 patients with CVD, half of whom took statin drugs only with the other half taking statins and 1,800mg of EPA daily. The conclusion: “EPA is a promising treatment for prevention of major coronary events, and especially non-fatal coronary events.”

Focus on 'totality of evidence,' not pieces

These studies were excluded on the basis of their design; the meta-analysis focused only on placebo-controlled studies. But McKay said relying only on this type of study in the field of human nutrition can lead up a blind alley.

“In these kind of studies you don’t have a true placebo group,” he said. “Unless it’s controlled for, the placebo group [could be] exposed to omega-3s to some level.”

McKay said it is important to focus on the strong totality of evidence backing the effects of omega-3s.

“We have strong mechanisms [of action], we have all this evidence that [omega-3s are] under consumed. We have a lot of human trials that are positive both for cardiovascular and other endpoints,” he said.

McKay said the design of the meta-analysis shows, “the weaknesses of trying to take a magnifying glass and look at one small set of studies and try to draw conclusions.”

Ismail agreed, saying, “What I really fundamentally disagree with is the structure of the study.

“If you really want to assess omega-3s in secondary heart disease prevention you need to get rid of the confounding factors. You need to look at the appropriate time and the appropriate dosage. And that didn’t happen.”

The CRN and the Natural Products Association (NPA) issued responses to the meta-analysis. Read CRN's response here and NPA's here.

Hide comments


  • Allowed HTML tags: <em> <strong> <blockquote> <br> <p>

Plain text

  • No HTML tags allowed.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Lines and paragraphs break automatically.