Good Calories, Bad Calories (13 page)

Once politics, the public, and the press had decided on the benefits of low-fat diets, science was left to catch up. In the early 1970s, when NIH

administrators opted to forgo a $1 bil ion National Diet-Heart Study that might possibly be definitive and to concentrate instead on a half-dozen studies, at a third of the cost, they believed the results of these smal er studies would be sufficiently persuasive to conclude publicly that low-fat diets would prolong lives. The results of these studies were published between 1980 and 1984.

Four of these studies tried to establish relationships between dietary fat and health within populations—in Honolulu, Puerto Rico, Chicago (Stamler and Shekel e’s second Western Electric study), and Framingham, Massachusetts. None of them succeeded. In Honolulu, the researchers fol owed seventy-three hundred men of Japanese descent and concluded that the men who developed heart disease seemed to eat slightly more fat and saturated fat than those who didn’t, but the men who died seemed to eat slightly less fat and slightly less saturated fat than those who didn’t. This observation was made in Framingham and Puerto Rico as wel . In 1981, investigators from the three studies published an article in the journal Circulation discussing the problem.

They said it posed a dilemma for dietary advice, but not an insurmountable one. The fact that the men in Puerto Rico and Honolulu who remained free of heart disease seemed to eat more starches suggested that it might be a good idea to recommend that we al eat more starch, as McGovern’s Dietary Goals actual y had. And because the advice should never be to eat more calories, we would have to eat less fat to avoid gaining weight.

When one is reading this report, it’s hard to avoid the suspicion that once the government began advocating fat reduction in the American diet it changed the way many investigators in this science perceived their obligations. Those who believed that dietary fat caused heart disease had always preferential y interpreted their data in the light of that hypothesis. Now they no longer felt obliged to test any hypothesis, let alone Keys’s. Rather, they seemed to consider their obligation to be that of “reconciling [their] study findings with current programs of prevention,” which meant the now official government recommendations. Moreover, these studies were expensive, and one way to justify the expense was to generate evidence that supported the official advice to avoid fat. If the evidence didn’t support the recommendations, then the task was to interpret it so that it did.*16

The other disconcerting aspect of these studies is that they suggested (with the notable exception of three Chicago studies reported by Jeremiah Stamler and col eagues) low cholesterol levels were associated with a higher risk of cancer. This link had original y been seen in Seymour Dayton’s VA Hospital trial in Los Angeles, and Dayton and others had suggested that polyunsaturated fats used to lower cholesterol might be the culprits. This was confirmed in 1972 by Swiss Red Cross researchers. In 1974, the principal investigators of six ongoing population studies—including Keys, Stamler, Wil iam Kannel of Framingham, and the British epidemiologist Geoffrey Rose—reported in The Lancet that the men who had developed colon cancer in their populations had “surprisingly” low levels of cholesterol, rather than the higher levels that they had initial y expected. In 1978, a team of British, Hungarian, and Czech researchers reported similar findings from a sixteen-thousand-man clinical trial of a cholesterol-lowering drug. By 1980, this link between cancer and low cholesterol was appearing in study after study. The most consistent association was between colon cancer and low cholesterol in men. In the Framingham Study those men whose total cholesterol levels were below 190 mg/dl were more than three times as likely to get colon cancer as those men with cholesterol greater than 220; they were almost twice as likely to contract any kind of cancer than those with cholesterol over 280 mg/dl.

This finding was met with “surprise and chagrin,” Manning Feinleib, a National Heart, Lung, and Blood Institute (NHLBI) epidemiologist, told Science.

This association was considered sufficiently troublesome that the NHLBI hosted three workshops between 1980 and 1982 to discuss it. In this case, however, the relevant administrators and investigators did not consider it sufficient to pay attention only to the positive evidence (that low cholesterol was associated with an increased risk of cancer even in clinical trials) and reject the negative evidence as irrelevant or erroneous, as they had when implicating high cholesteral as a cause of heart disease. Instead, they searched the literature and found a few studies—including a Norwegian study published a decade earlier in a Scandinavian journal supplement—that reported no link between low cholesterol and cancer. As a result, the NHLBI concluded that the evidence was inconsistent, only “suggestive” that “low cholesterol may be in some way associated with cancer risk,” said Robert Levy after the first workshop. After the second workshop, by which time the Framingham, Honolulu, and Puerto Rico studies had reported the same association, the NHLBI administrators stil considered the results inconclusive: “The findings do not represent a public health chal enge; however, they do present a scientific chal enge,” they wrote. Levy did tel the journal Science that this low-cholesterol/cancer link might make those investigators who were arguing that everyone’s cholesterol should be as low as possible “a little more cautious.”

After the third workshop, Levy and his NHLBI col eagues concluded that the evidence stil didn’t imply cause and effect. They believed that high cholesterol caused heart disease and that low cholesterol was only a sign of people who might be cancer-prone, perhaps because of a genetic predisposition. This seemed like an arbitrary distinction, and it was certainly based on assumptions more than facts. The NHLBI administrators acknowledged that further research would be necessary to clarify “the perplexing inconsistencies.” Stil , the evidence did “not preclude, countermand, or contradict the current public health message which recommends that those with elevated cholesterol levels seek to lower them through diets.”

In the early 1970s, the National Heart, Lung, and Blood Institute had bet its heart-disease prevention budget on two enormous trials that held out hope of resolving the controversy.

The first was the Multiple Risk Factor Intervention Trial, known as MRFIT and led by Jeremiah Stamler. The goal of MRFIT was to “throw the kitchen sink” at heart disease: to convince the subjects to quit smoking, lower their cholesterol, and lower their blood pressure—the multiple-risk-factor interventions. The MRFIT investigators tested the cholesterol of 362,000 middle-aged American men and found twelve thousand (the top 3 percent) whose cholesterol was so high, more than 290 mg/ml, that they could be considered at imminent risk of having a heart attack. The MRFIT investigators believed that these men were so likely to succumb to heart disease that preventive measures would be even more likely to demonstrate a benefit. (If men with lower cholesterol were included, or if women were included, the study would require a considerably greater number of subjects or a longer fol ow-up to demonstrate any significant benefit.) These twelve thousand men were randomly divided between a control group—told to live, eat, and address their health problems however they desired—and a treatment group—advised to quit smoking, take medication to lower their blood pressure if necessary, and eat a low-fat, low-cholesterol diet, which meant drinking skim milk, using margarine instead of butter, eating only one or two eggs a week, and avoiding red meat, cakes, puddings, and pastries. Al twelve thousand were then fol owed for seven years, at a cost of $115 mil ion.

The results were announced in October 1982, and a Wall Street Journal headline captured the situation succinctly: “Heart Attacks: A Test Col apses.”

There had been slightly more deaths among those men who had been counseled to quit smoking, eat a cholesterol-lowering diet, and treat their high blood pressure than among those who had been left to their own devices.*17

The second trial was the $150 mil ion Lipid Research Clinics (LRC) Coronary Primary Prevention Trial. The trial was led by Basil Rifkind of the NHLBI and Daniel Steinberg, a specialist on cholesterol disorders at the University of California, San Diego. The LRC investigators had screened nearly half a mil ion middle-aged men and found thirty-eight hundred who had no overt signs of heart disease but cholesterol levels sufficiently high—more than 265

mg/dl—that they could be considered imminently likely to suffer a heart attack. Half of these men (the control group) were told to eat fewer eggs and less fatty meats and drink less milk, and were given a placebo pil to take daily. The other half (the treatment group) were counseled to eat the same cholesterol-lowering diet, but they were also given a cholesterol-lowering drug cal ed cholestyramine. Both groups had been told to diet, because the LRC

investigators considered it unethical to withhold al treatment from the control group, given their high cholesterol levels and high risk of heart disease. It was an odd decision for two reasons. First, the LRC trial had been approved in the early 1970s in lieu of the National Diet-Heart Study that was necessary to establish the safety and effectiveness of a cholesterol-lowering diet; the LRC investigators had no proof that such a diet would benefit their subjects, rather than harm them. Second, both groups were told to diet, the trial could determine only the effectiveness of the drug—the single variable that differed between them.

In January 1984, the results of the trial were published in The Journal of the American Medical Association. Cholesterol levels dropped by an average of 4 percent in the control group—those men taking a placebo. The levels dropped by 13 percent in the men taking cholestyramine. In the control group, 158 men suffered nonfatal heart attacks during the study and 38 men died from heart attacks. In the treatment group, 130 men suffered nonfatal heart attacks and only 30 died from them. Al in al , 71 men had died in the control group and 68 in the treatment group. In other words, cholestyramine had improved by less than .2 percent the chance that any one of the men who took it would live through the next decade. To cal these results “conclusive,” as the University of Chicago biostatistician Paul Meier remarked, would constitute “a substantial misuse of the term.” Nonetheless, these results were taken as sufficient by Rifkind, Steinberg, and their col eagues so they could state unconditional y that Keys had been right and that lowering cholesterol would

save lives.

Rifkind and his col aborators also concluded that the cholesterol-lowering benefits of a drug applied to diet as wel . Although the trial included only middle-aged men with cholesterol levels higher than those of 95 percent of the population, Rifkind and his col eagues concluded that those benefits “could and should be extended to other age groups and women and…other more modest elevations of cholesterol levels.” As Rifkind told Time magazine, “It is now indisputable that lowering cholesterol with diet and drugs can actual y cut the risk of developing heart disease and having a heart attack.”

Pete Ahrens cal ed this extrapolation from a drug study to a diet “unwarranted, unscientific and wishful thinking.” Thomas Chalmers, an expert on clinical trials who would later become president of the Mt. Sinai School of Medicine in New York, described it to Science as an “unconscionable exaggeration of al the data.” In fact, the LRC investigators acknowledged in their JAMA article that their attempt to ascertain a benefit from diet alone had failed.

Rifkind later explained the exaggerated claims. For twenty years, he said, those who believed in Keys’s hypothesis had argued that lowering cholesterol would prevent heart attacks. They had spent hundreds of mil ions of dol ars trying to prove it, in the face of extreme skepticism. Now they had demonstrated that lowering cholesterol had reduced heart-disease risk and maybe even saved lives. They could never prove that cholesterol-lowering diets would do the same—that would be too expensive, and MRFIT, which might have implied such a conclusion, had failed—but now they had established a fundamental link in the causal chain from lower cholesterol to cardiovascular health. With that, they could take the leap of faith from cholesterol-lowering drugs to cholesterol-lowering diets. “It’s an imperfect world,” Rifkind said. “The data that would be definitive is ungettable, so you do your best with what is available.”

With publication of the LRC results, the National Heart, Lung, and Blood Institute launched what Robert Levy cal ed “a massive health campaign” to convince the public of the benefits of lowering cholesterol, whether by diet or drug, and the media went along. Time reported the LRC findings in a story headlined “Sorry, It’s True. Cholesterol Real y Is a Kil er.” The article about a drug trial began, “No whole milk. No butter. No fatty meats. Fewer eggs…” In March, Time ran a fol ow-up cover story quoting Rifkind as saying that the LRC results “strongly indicate that the more you lower cholesterol and fat in your diet, the more you reduce your risk of heart disease.” Anthony Gotto, president of the American Heart Association, told Time that if everyone went along with a cholesterol-lowering program, “we wil have [atherosclerosis] conquered” by the year 2000.