"The Big Fat Surprise": How Science Is Corrupted By Government Money And Government Process
/I haven't made a practice of reviewing books here, but it's time for an exception. A few weeks ago a reader named Nina Teicholz, who is also an accomplished writer in the health arena, wrote to offer me a copy of her recent book, titled "The Big Fat Surprise: Why Butter, Meat & Cheese Belong in a Healthy Diet." I took her up on the offer, and shortly after that the book arrived in my mailbox. And then I actually read it, which is somewhat remarkable given that I'm about 20 books behind at any given time.
This is a book that you really need to read. And not just because you will finally learn some real information about what constitutes a healthy diet. This is also a story of the corruption of science by government money and government process that has application far beyond just the field of nutrition.
Today, all Americans deeply know, seemingly as part of their cultural heritage, that a healthy diet is low in fat. Indeed, not just low in overall fat, but particularly low in saturated fat, the kind that is found in large amounts in butter, milk and cheese. Go to the market, and seemingly every product has a "low fat" option, generally accompanied by the phrase "heart healthy!" Eat out with your friends, and observe them demonstrating their virtue over you by eating nothing but a salad, perhaps accompanied by a smidgeon of low-fat chicken or fish. (Maybe this phenomenon is less prevalent outside of Manhattan. But a great thing about living in Manhattan is how easy it is to horrify your friends by eating an intentionally high-fat meal in their presence.)
I've been on to the scam of the low fat diet for some time (see, for example, here from July 2013). But it's actually harder than you might think to get your hands on real information on this subject. Hundreds upon hundreds of articles just take it as a complete given that low fat equals heart healthy, without ever questioning where this idea ever came from. Besides, the American Heart Association has been on the low fat bandwagon for decades, and same for the U.S. government dietary guidelines. Hey, everybody knows it!
And now into the groupthink leaps Ms. Teicholz, with a thoroughly-researched tome on the development of the so-called diet-heart hypothesis. She tells the story of how over the course of a few decades from about the 1950s to the 80s that hypothesis rapidly swept away all dissenters and rose to uncontested dominance in the field. Oh, even as the evidence relating to the hypothesis ranged from ambiguous at best to completely adverse in many instances. But the hypothesis was backed by a small group of intensely committed scientists and journalists who managed to get control over the principal scientific organizations (notably the American Heart Association in this instance) and the principal government grant-givers (here NIH and NHLBI), and used that control to marginalize and ultimately de-fund their opponents.
I won't attempt to go too deeply into the details, but a few examples will give some of the flavor. The diet-heart hypothesis owes its triumph more than anything else to the work from the 1940s to the 1980s of one Ancel Keys, a highly energetic and ambitious professor from the University of Minnesota. In the mid-50s Keys proposed testing the hypothesis by doing a first-ever international multi-country epidemiological study. In 1956 he won from the U.S. Public Health Service to fund the proposed study a $200,000 annual grant -- at the time an enormous sum. The countries he chose to study were Italy, Greece, Yugoslavia, Finland, the Netherlands, Japan and the United States. And the results:
What Keys found, as he had hoped, was a strong correlation between the consumption of saturated fat and deaths from heart disease. (page 38)
But here's the thing: from previous work by Yerushalmy and Hellebore (1950) Keys already knew rough numbers for fat consumption and heart disease rates for 22 countries, which included five of the seven that Keys selected to study (the exceptions were Greece and Yugoslavia). And the five countries for which these data were known all fit neatly along an upward-sloping line from low fat/low heart disease (Japan) on one end, to high fat/high heart disease (United States) on the other end. Meanwhile, there were plenty of countries that were known not to fit into this neat pattern. And these were not small or insignificant countries. Examples were France, Switzerland and Mexico. Well, they were just omitted from the study. Teicholz:
[Keys] might have selected a European country to challenge his fat hypothesis, like Switzerland or France (or Germany or Norway or Sweden). Instead, he chose only those nations (based on national statistics) that seemed likely to confirm it. . . . Attempting to explain why Keys did not seek out countries that would offer more challenges to his ideas, [Keys colleague] Blackburn said, "Keys just had a personal aversion to being in France and Switzerland."
Sure. To this day, in books I have read on the supposed support for the diet-heart hypothesis, the Keys seven-country study is cited as among the most important evidence.
By the early 70s the need for some more definitive proof or refutation of the diet-heart hypothesis had become clear, but the main government funders (here mainly NIH) were daunted by the prospective cost. Were they really going to spend literally billions to provide controlled diets for years to tens of thousands of people in two groups? Instead, they opted to do two not-inexpensive studies, but kind of on the cheap. $250 million was allocated to the two studies. First up was the Multiple Risk Factor Intervention Trial, or MRFIT (pronounced "Mister Fit"), a massive test that ran from 1973 to 1982. Here was the plan: they screened 361,000 middle-aged American males to find some 12,000 with particularly high blood cholesterol, greater than 290 mg/dL. They divided these into two groups, an "intervention" group and a control group. The intervention group got multiple such interventions, including counseling to quit smoking, medication to lower high blood pressure, and advice on a low fat, low cholesterol diet. They drank skim milk, used margarine instead of butter, limited eggs to two of fewer per week, and avoided meat and desserts. The control group's members were left to do as they pleased. The men were followed for seven years. Results?
The results, announced in September 1982, were a disaster for the diet-heart hypothesis. Although men in the intervention group had been spectacularly successful in changing their diets, quitting smoking, and reducing their blood pressure, they died at slightly higher rates than the controls.
Oops. Teicholz details how the results of MRFIT were then systematically suppressed and forgotten. Meanwhile, the remainder of the $250 million went to a study called Lipid Research Clinic Coronary Primary Prevention Trial, abbreviated as LRC. This study divided 3800 men who had high blood cholesterol again into two groups, both of which were counseled to reduce fat in the diet, but one of which also received an early cholesterol-lowering drug called cholestyramine. In other words, diet was not part of what was tested; the test was of the cholesterol-lowering drug. The results were reported in 1984. The treatment group had "slightly fewer heart attacks" (page 128), but total mortality for the two groups was almost identical. The treatment group suffered additional deaths from such things as suicide, accidents, and cancer. And commenting on other similar studies, Teicholz says:
Other cholesterol-lowering studies where diet had been the only intervention consistently found higher rates of cancer and gallstones in the experimental group . . . . In addition, populations found to have very low cholesterol, such as the Japanese, suffer from higher rates of strokes and cerebral hemorrhage compared to groups whose average cholesterol is higher. (page 129)
But somehow the extremely ambiguous results of the LRC study -- a study that did not even test diet at all -- got spun into being the definitive proof of the diet-heart hypothesis. In 1984 NHLBI organized a massive conference specifically intended to generate a scientific "consensus" on the causes of heart disease. The conference was organized by Basil Rifkind and chaired by Daniel Steinberg, two of the leaders of the LRC study. The conference ended with a so-called consensus statement that admitted of no ambiguities:
The conference "consensus" statement, which Steinberg read out on the last morning of the event, was not a measured assessment of the complicated role that diet might play in a little-understood disease. Instead, there was "no doubt," he stated, that reducing cholesterol through a low-fat, low-saturated-fat diet would "afford significant protection against coronary heart disease" for every American over the age of two. (page 132)
And thus the low fat diet got an imprimatur of "consensus" from a small amount of highly ambiguous data. Nothing since has been able to dislodge it. Example: a long-running study of essentially everybody in the town of Framingham, Massachusetts reported follow-up results in the 90s. Here are some:
In 1961, after six years of study, the Framingham investigators announced their first big discovery: that high total cholesterol was a reliable predictor for heart disease. . . . However, thirty years later [1991], in the Framingham follow-up study . . . it turned out that the predictive power of total cholesterol was not nearly as strong as study leaders had originally thought. For men and women with cholesterol between 205 and 264 milligrams per deciliter (mg/dL), no relationship between these numbers and heart disease risk could be found. . . . [F}or men aged forty-eight to fifty-seven, those with cholesterol in the midrange (183-222 mg/dL) had greater risk of heart attack death than those with higher cholesterol (222-261 mg/dL). (pages 64-65)
Teicholz describes several more studies with results that would either refute the diet-heart hypothesis or are highly troubling in other respects, notably showing higher total mortality for people on low cholesterol diets from causes including cancer, stroke and violence. But so what? Since when is it the actual evidence that counts when "science" turns on control of government funds and government-sponsored committees that control grants and publication rights. The small clique of intense promoters of the diet-heart hypothesis -- notably Keys, Rifkind and Steinberg -- gradually got control of the key associations (e.g., American Heart Association) and government funders (e.g., NIH and NHLBI) and saw to it that their opponents lost their funding, were no longer invited to conferences, and could no longer get their papers published. And thus, the day has long since been carried. It's just a part of our culture. Everybody knows that a low fat diet is "heart healthy"!
Teicholz overall is a lot nicer about the state of corruption than I would be. But low key can be very effective. I'm just hoping that someone other than me pays attention.
Of course, readers will quickly recognize that this story bears remarkable resemblance to the story of the anthropogenic global warming hypothesis. The AGW story is even far worse because all the proposed solutions are catnip to leftists looking for reasons to promote government takeover of everything. Is there any hope of ever again looking at the evidence to see if the hypothesis has been refuted? Perhaps if that occurs in the arena of nutrition, the good sense will spread.