During the 1970s, two Danish researchers, Hans Olaf Bang and Jorn Dyerberg, published a series of studies (starting with this one) in which they connected the low rate of heart disease among the Inuit of Greenland to their diet, which contained large amounts of fish and other marine animals.

Thus began the nutritional recommendation that the rest of us should eat plenty of oily (fatty) fish to protect ourselves from heart disease, the leading cause of death in the U.S. The American Heart Association, for example, urges us to consume fatty fish at least twice a week (more if you have diagnosed heart disease) or, if that’s not possible, to take fish oil supplements with omega-3 fatty acids.

Recently, several large and well-designed clinical trials have reported ambiguous or negative results regarding the heart-protective effects of both fish oil and fish-oil supplements. But that hasn’t changed the recommendations — or slowed down the sales of fish-oil supplements. More than $1.2 billion of the supplements were reportedly sold last year in the United States alone.

Now, though, it also turns out that those landmark studies by Bang and Dyerberg — which continue to be cited each year in the medical literature as proof of the heart-healthy properties of omega-3 fatty acids — are based on, well, fishy data. For, as a fascinating article published earlier this summer in the Canadian Journal of Cardiology points out, Bang and Dyerberg never actually investigated the cardiovascular health of the Inuits. Their findings relied on speculation, not data.

A comprehensive review of the research

The lead author of the new article is Dr. George Fodor of the University of Ottawa Heart Institute. He and his colleagues reviewed all studies published in the medical literature before January 2014 that had reported (in either English or Danish) on the incidence of heart disease among Greenland’s Inuit (Eskimo) population.

They found 48 studies. A review of the evidence in those studies, Fodor and his co-authors write, “leads us to the conclusion that Eskimos have a prevalence of [heart disease] similar to non-Eskimo population, they have excessive mortality due to cerebrovascular strokes, their overall mortality is twice as high as that of non-Eskimo populations, and their life expectancy is approximately 10 years shorter than the Danish population.”

How, then, did Bang and Dyerberg come to such a different conclusion? And why is their version still the one that predominates in the medical literature?

Invalid and inaccurate reports

First, say Fodor and his co-authors, Bang and Dyberg conducted their studies only around the small town of Uummannaq (also spelled Umanak), which is located 310 miles north of the Arctic Circle. With 1,300 inhabitants, the town represents about 2.3 percent of Greenland’s population, and some of the settlements around Uummannaq are more than 100 miles away from the nearest medical facility.

Second, Bang and Dyberg relied on heart-disease statistics provided by Greenland’s chief medical officer for the years 1963-1967 and 1973-1976 — statistics that fell far short of telling the entire heart-disease story of the Inuit, as Fodor and his co-authors explain:

These reports were based on death certificates and hospital admissions. Concerns over the validity and accuracy of death certificate and mortality statistics in Greenland have been raised in a number of reports. According to the Deputy [Chief Medical Officer] in Greenland [in the early 1970s], Flemming Mikkelsen, 30% of the total population lived in outposts and small settlements where no medical officer was stationed. If a person died in one of these areas, the certificate would be completed by the nearest medical officer, based on information provided by a medical auxiliary or some other “competent” person.

Thus, 20% of death certificates were completed without a doctor having examined the patient or the body. [Two epidemiologists] also pointed out [in 1980] that there was a specific concern with mortality data and hospital admission statistics in Greenland, because doctors had limited diagnostic facilities and the study population was widely scattered with few possibilities of communication during certain seasons. Therefore, the reported data are likely an underestimation of the true magnitude of the disease in this area.

Indeed, a 1986 study reported that only one in seven deaths in Greenland occurred in a hospital equipped with the facilities to accurately determine whether a death was the result of heart disease.

That study alone should have led the medical community to question the validity of Bang and Dyberg’s findings, say Fodor and his colleagues.

The perpetuation of a myth

As for the other issue — why researchers and others in the medical profession continue to uncritically cite the Bang and Dyerberg studies as proof that the Inuit have a lower risk of heart disease — Fodor and his co-authors suggest (rather kindly) that it may reflect a “misinterpretation of the original findings or an example of confirmation bias.”

Or perhaps, they add, such willful ignoring of the facts represents “a trend of applying less rigorous standards of scientific evidence when reporting about non-pharmacological (ie, lifestyle) interventions.”

Either way, the “Eskimo diet” myth has been with us for four decades. And given the huge amount of money riding on its perpetuation, I suspect the myth will remain with us for some time to come.

At the end of the paper, Fodor and his colleagues offer a very apt quote from the 17th century English philosopher and essayist Francis Bacon: “Man prefers to believe what he prefers to be true.”

Unfortunately, Fodor’s paper is behind a paywall. But you’ll find an abstract on the Canadian Journal of Cardiology’s website.