Good Calories, Bad Calories (9 page)

They could not agree on the dietary-fat hypothesis; they could barely discuss it, as Henry Blackburn had noted, because they were seeing two dramatical y different bodies of evidence.

Another revealing example of selection bias was the reanalysis of a study begun in 1957 on fifty-four hundred male employees of the Western Electric Company. The original investigators, led by the Chicago cardiologist Oglesby Paul, had given them extensive physical exams and come to what they cal ed a “reasonable approximation of the truth” of what and how much each of these men ate. After four years, eighty-eight of the men had developed symptoms of coronary heart disease. Paul and his col eagues then compared heart disease rates among the 15 percent of the men who seemingly ate the most fatty food with the 15 percent who seemingly ate the least. “Worthy of comment,” they reported, “is the fact that of the 88 coronary cases, 14 have appeared in the high-fat intake group and 16 in the low-fat group.”

Two decades later, Jeremiah Stamler and his col eague Richard Shekel e from Rush–Presbyterian–St. Luke’s Medical Center in Chicago revisited Western Electric to see how these men had fared. They assessed the health of the employees, or the cause of death of those who had died, and then considered the diets each subject had reportedly consumed in the late 1950s. Those who had reportedly eaten large amounts of polyunsaturated fats, according to this new analysis, had slightly lower rates of coronary heart disease, but “the amount of saturated fatty acids in the diet was not significantly associated with the risk of death from [coronary heart disease],” they reported. This alone could be considered a refutation of Keys’s hypothesis.

But Stamler and Shekel e knew what result they should have obtained, or so they believed, and they interpreted the data in that light. Their logic is worth fol owing. “Although most attempts to document the relation of dietary cholesterol, saturated fatty acids, and polyunsaturated fatty acids to serum cholesterol concentration in persons who are eating freely have been unsuccessful,” they explained, “positive results have been obtained in investigations besides the Western Electric Study.” They then listed four such studies: a new version of Keys’s study on Japanese men in Japan, Hawaii, and California; a study of men living for a year at a research station in Antarctica; a study of Tarahumara Indians in the Mexican highlands; and one of infants with a history of breast-feeding. To Stamler and Shekel e, these four studies provided sufficiently compel ing support for Keys’s hypothesis that they could interpret their own ambiguous results in a similar vein. “If viewed in isolation,” they explained, “the conclusions that can be drawn from a single epidemiologic study are limited. Within the context of the total literature, however, the present observations support the conclusion that the [fat] composition of the diet affects the level of serum cholesterol and the long-term risk of death from [coronary heart disease, CHD] in middle-aged American men.”

The New England Journal of Medicine published Stamler’s analysis of the Western Electric findings in January 1981, and the press reported the results uncritical y. “The new report,” stated the Washington Post, “strongly reinforces the view that a high-fat, high-cholesterol diet can clog arteries and cause heart disease.” Jane Brody of the New York Times quoted Shekel e saying, “The message of these findings is that it is prudent to decrease the amount of saturated fats and cholesterol in your diet.” The Western Electric reanalysis was then cited in a 1990 joint report by the American Heart Association and the National Heart, Lung, and Blood Institute, entitled “The Cholesterol Facts,” as one of seven “epidemiologic studies showing the link between diet and CHD [that] have produced particularly impressive results” and “showing a correlation between saturated fatty acids and CHD,” which is precisely what it did not do.*7

In preventive medicine, benefits without risks are nonexistent. Any diet or lifestyle intervention can have harmful effects. Changing the composition of the fats we eat could have profound physiological effects throughout the body. Our brains, for instance, are 70 percent fat, mostly in the form of a substance known as myelin that insulates nerve cel s and, for that matter, al nerve endings in the body. Fat is the primary component of al cel membranes.

Changing the proportion of saturated to unsaturated fats in the diet, as proponents of Keys’s hypothesis recommended, might wel change the composition of the fats in the cel membranes. This could alter the permeability of cel membranes, which determines how easily they transport, among other things, blood sugar, proteins, hormones, bacteria, viruses, and tumor-causing agents into and out of the cel . The relative saturation of these membrane fats could affect the aging of cel s and the likelihood that blood cel s wil clot in vessels and cause heart attacks.

When we consider treating a disease with a new therapy, we always have to consider potential side effects such as these. If a drug prevents heart disease but can cause cancer, the benefits may not be worth the risk. If the drug prevents heart disease but can cause cancer in only a tiny percentage of individuals, and only causes rashes in a greater number, then the tradeoff might be worth it. No drug can be approved for treatment without such consideration. Why should diet be treated differently?

The Seven Countries Study, which is considered Ancel Keys’s masterpiece, is a pedagogical example of this risk-benefit problem. The study is often referred to as “landmark” or “legendary” because of its pivotal role in the diet-heart controversy. Keys launched it in 1956, with $200,000 yearly support from the Public Health Service, an enormous sum of money then for a single biomedical research project. Keys and his col aborators cobbled together incipient research programs from around the world and expanded them to include some thirteen thousand middle-aged men in sixteen mostly rural populations in Italy, Yugoslavia, Greece, Finland, the Netherlands, Japan, and the United States. Keys wanted populations that would differ dramatical y in diet and heart-disease risk, which would al ow him to find meaningful associations between these differences. The study was prospective, like Framingham, which means the men were given physical examinations when they signed on, and the state of their health was assessed periodical y thereafter.

Results were first published in 1970, and then at five-year intervals, as the subjects in the study aged and succumbed to death and disease. The mortality rates for heart disease were particularly revealing. Expressed in deaths per decade, there were 9 heart-disease deaths for every ten thousand men in Crete, compared with 992 for the lumberjacks and farmers of North Karelia, Finland. In between these two extremes were Japanese vil agers at 66 per ten thousand, Belgrade faculty members and Rome railroad workers at 290, and U.S. railroad workers with 570 deaths per ten thousand.

According to Keys, the Seven Countries Study taught us three lessons about diet and heart disease: first, that cholesterol levels predicted heart-disease risk; second, that the amount of saturated fat in the diet predicted cholesterol levels and heart disease (contradicting Keys’s earlier insistence that total fat consumption predicted cholesterol levels and heart disease with remarkable accuracy); and, third, a new idea, that monounsaturated fats protected against heart disease. To Keys, this last lesson explained why Finnish lumberjacks and Cretan vil agers could both eat diets that were 40

percent fat but have such dramatical y different rates of heart disease. Twenty-two percent of the calories in the Finnish diet came from saturated fats, and only 14 percent from monounsaturated fats, whereas the vil agers of Crete obtained only 8 percent from saturated fat and 29 percent from monounsaturated fats. This could also explain why heart-disease rates in Crete were even lower than in Japan, even though the Japanese ate very little fat of any kind, and so very little of the healthy monosaturated fats, as wel . This hypothesis could not explain many of the other relationships in the study

—why eastern Finns, for instance, had three times the heart disease of western Finns, while having almost identical lifestyles and eating, as far as fat was concerned, identical diets—but this was not considered sufficient reason to doubt it. Keys’s Seven Countries Study was the genesis of the Mediterranean-diet concept that is currently in vogue, and it prompted Keys to publish a new edition of his 1959 best-sel er, Eat Well and Stay Well, now entitled How to Eat Well and Stay Well the Mediterranean Way.

Despite the legendary status of the Seven Countries Study, it was fatal y flawed, like its predecessor, the six-country analysis Keys published in 1953

using only national diet and death statistics to support his points. For one thing, Keys chose seven countries he knew in advance would support his hypothesis. Had Keys chosen at random, or, say, chosen France and Switzerland rather than Japan and Finland, he would likely have seen no effect from saturated fat, and there might be no such thing today as the French paradox—a nation that consumes copious saturated fat but has comparatively little heart disease.

In 1984, when Keys and his col eagues published their report on the data after fifteen years of observation, they explained that “little attention was given to longevity or total mortality” in their initial results, even though what we real y want to know is whether or not we wil live longer if we change our diets.

“The ultimate interest being prevention,” they wrote, “it seemed reasonable to suppose that measures control ing coronary risk factors would improve the outlook for longevity as wel as for heart attacks, at least in the population of middle-aged men in the United States for whom [coronary heart disease] is the outstanding cause of premature death.” Now, however, with “the large number of deaths accumulated over the years,” they realized that coronary heart disease accounted for less than one-third of al deaths, and so this “forced attention to total mortality.”

Now the story changed: High cholesterol did not predict increased mortality, despite its association with a greater rate of heart disease. Saturated fat in the diet ceased to be a factor as wel . The U.S. railroad workers, for instance, had a death rate from al causes lower—and so a life expectancy longer

—than the Finns, the Italians, the Yugoslavs, the Dutch, and particularly the Japanese, who ate copious carbohydrates, fruits, vegetables, and fish. Only the vil agers of Crete and Corfu could stil expect to live significantly longer than the U.S. railroad workers. Though this could be explained by other factors, it stil implied that tel ing Americans to eat like the Japanese might not be the best advice. This was why Keys had begun advocating Mediterranean diets, though evidence that the Mediterranean diet was beneficial was derived only from the vil agers of Crete and Corfu in Keys’s study, and not from those who lived on the Mediterranean coast of Yugoslavia or in the cities of Italy.

In discussions of dietary fat and heart disease, it is often forgotten that the epidemiologic tools used to link heart disease to diet were relatively new and had never been successful y put to use previously in this kind of chal enge. The science of epidemiology evolved to make sense of infectious diseases, not common chronic diseases like heart disease. Though the tools of epidemiology—comparisons of populations with and without the disease—had proved effective in establishing that a disease such as cholera is caused by the presence of micro-organisms in contaminated water, as the British physician John Snow demonstrated in 1854, it is a much more complicated endeavor to employ those same tools to elucidate the subtler causes of chronic disease. They can certainly contribute to the case against the most conspicuous determinants of noninfectious diseases—that cigarettes cause lung cancer, for example. But lung cancer was an extremely rare disease before cigarettes became widespread, and smokers are thirty times as likely to get it as nonsmokers. When it comes to establishing that someone who eats copious fat might be twice as likely to be afflicted with heart disease—a very common disorder—as someone who eats little dietary fat, the tools were of untested value.

The investigators attempting these studies were constructing the relevant scientific methodology as they went along. Most were physicians untrained to pursue scientific research. Nonetheless, they decided they could reliably establish the cause of chronic disease by accumulating diet and disease data in entire populations, and then using statistical analyses to determine cause and effect. Such an approach “seems to furnish information about causes,”

wrote the Johns Hopkins University biologist Raymond Pearl in his introductory statistics textbook in 1940, but it fails, he said, to do so.

“A common feature of epidemiological data is that they are almost certain to be biased, of doubtful quality, or incomplete (and sometimes al three),”

explained the epidemiologist John Bailar in The New England Journal of Medicine in 1980. “Problems do not disappear even if one has flawless data, since the statistical associations in almost any nontrivial set of observations are subject to many interpretations. This ambiguity exists because of the difficulty of sorting out causes, effects, concomitant variables, and random fluctuations when the causes are multiple or diffuse, the exposure levels low, irregular, or hard to measure, and the relevant biologic mechanisms poorly understood. Even when the data are general y accepted as accurate, there is much room for individual judgment, and the considered conclusions of the investigators on these matters determine what they wil label ‘cause’…”

The only way to establish cause and effect with any reliability is to do “control ed” experiments, or control ed trials, as they’re cal ed in medicine. Such trials attempt to avoid al the chaotic complexities of comparing populations, towns, and ethnic groups. Instead, they try to create two identical situations

—two groups of subjects, in this case—and then change only one variable to see what happens. They “control” for al the other possible variables that might affect the outcome being studied. Ideal y, such trials wil randomly assign subjects into an experimental group, which receives the treatment being tested—a drug, for instance, or a special diet—and a group, which receives a placebo or eats their usual meals or some standard fare.