by Mary Enig, PhD, and Sally Fallon
First published in Nexus Magazine,
Dec ’98-Jan ’99 and Feb ’99-Mar ’99
Enig speaks out
When Mary Enig, a graduate student at the University of Maryland, read the McGovern committee report, she was puzzled. Enig was familiar with Kummerow’s research and she knew that the consumption of animal fats in America was not on the increase—quite the contrary, use of animal fats had been declining steadily since the turn of the century. A report in the Journal of American Oil Chemists—which the McGovern Committee did not use—showed that animal fat consumption had declined from 104 grams per person per day in 1909 to 97 grams per day in 1972, while vegetable fat intake had increased from a mere 21 grams to almost 60.14 Total per capita fat consumption had increased over the period, but this increase was mostly due to an increase in unsaturated fats from vegetable oils—with 50 percent of the increase coming from liquid vegetable oils and about 41 percent from margarines made from vegetable oils. She noted a number of studies that directly contradicted the McGovern Committee’s conclusions that “there is . . . a strong correlation between dietary fat intake and the incidence of breast cancer and colon cancer,” two of the most common cancers in America. Greece, for example, had less than one-fourth the rate of breast cancer compared to Israel but the same dietary fat intake. Spain had only one-third the breast cancer mortality of France and Italy but the total dietary fat intake was slightly greater. Puerto Rico, with a high animal fat intake, had a very low rate of breast and colon cancer. The Netherlands and Finland both used approximately 100 grams of animal fat per capita per day but breast and colon cancer rates were almost twice in the Netherlands what they are in Finland. The Netherlands consumed 53 grams of vegetable fat per person compared to 13 in Finland. A study from Cali, Columbia found a fourfold excess risk for colon cancer in the higher economic classes, which used less animal fat than the lower economic classes. A study on Seventh-Day Adventist physicians, who avoid meat, especially red meat, found they had a significantly higher rate of colon cancer than non-Seventh Day Adventist physicians. Enig analyzed the USDA data that the McGovern Committee had used and concluded that it showed a strong positive correlation with total fat and vegetable fat and an essentially strong negative correlation or no correlation with animal fat to total cancer deaths, breast and colon cancer mortality and breast and colon cancer incidence—in other words, use of vegetable oils seemed to predispose to cancer and animal fats seemed to protect against cancer. She noted that the analysts for the committee had manipulated the data in inappropriate ways in order to obtain mendacious results.
Enig submitted her findings to the Journal of the Federation of American Societies for Experimental Biology (FASEB), in May, 1978, and her article was published in the FASEB’s Federation Proceedings15 in July of the same year—an unusually quick turnaround. The assistant editor, responsible for accepting the article, died of a heart attack shortly thereafter. Enig’s paper noted that the correlations pointed a finger at the trans fatty acids and called for further investigation. Only two years earlier, the Life Sciences Research office, which is the arm of FASEB that does scientific investigations, had published the whitewash that had ushered partially hydrogenated soybean oil onto the GRAS list and removed any lingering constraints against the number one ingredient in factory-produced food.
The food giants fight back
Enig’s paper sent alarm bells through the industry. In early 1979, she received a visit from S. F. Reipma of the National Association of Margarine Manufacturers. Reipma was visibly annoyed. He explained that both his association and the Institute for Shortening and Edible Oils (ISEO) kept careful watch to prevent articles like Enig’s from appearing in the literature. Enig’s paper should never have been published, he said. He thought that ISEO was “watching out.”
“We left the barn door open,” he said, “and the horse got out.”
Reipma also challenged Enig’s use of the USDA data, claiming that it was in error. He knew it was in error, he said, “because we give it to them.”
A few weeks later, Reipma paid a second visit, this time in the company of Thomas Applewhite, an advisor to the ISEO and representative of Kraft Foods, Ronald Simpson with Central Soya and an unnamed representative from Lever Brothers. They carried with them—in fact, waved them in the air in indignation—a two-inch stack of newspaper articles, including one that appeared in the National Enquirer, reporting on Enig’s Federation Proceedings article. Applewhite’s face flushed red with anger when Enig repeated Reipma’s statement that “they had left the barn door open and a horse got out,” and his admission that Department of Agriculture food data had been sabotaged by the margarine lobby.
The other thing Reipma told Enig during his unguarded visit was that he had called in on the FASEB offices in an attempt to coerce them into publishing letters to refute her paper, without allowing Enig to submit any counter refutation as was normally customary in scientific journals. He told Enig that he was “thrown out of the office”—an admission later confirmed by one of the FASEB editors. Nevertheless, a series of letters did follow the July 1978 article.16 On behalf of the ISEO, Applewhite and Walter Meyer of Procter and Gamble criticized Enig’s use of the data; Applewhite accused Enig of extrapolating from two data points, when in fact she had used seven. In the same issue, John Bailar, Editor-in-Chief of the Journal of the National Cancer Institute, pointed out that the correlations between vegetable oil consumption and cancer were not the same as evidence of causation and warned against changing current dietary components in the hopes of preventing cancer in the future—which is of course exactly what the McGovern Committee did.
In reply, Enig and her colleagues noted that although the NCI had provided them with faulty cancer data, this had no bearing on the statistics relating to trans consumption, and did not affect the gist of their argument—that the correlation between vegetable fat consumption, especially trans fat consumption, was sufficient to warrant a more thorough investigation. The problem was that very little investigation was being done.
University of Maryland researchers recognized the need for more research in two areas. One concerned the effects of trans fats on cellular processes once they are built into the cell membrane. Studies with rats, including one conducted by Fred Mattson in 1960, indicated that the trans fatty acids were built into the cell membrane in proportion to their presence in the diet, and that the turnover of trans in the cells was similar to that of other fatty acids. These studies, according to J. Edward Hunter of the ISEO, were proof that “trans fatty acids do not pose any hazard to man in a normal diet.” Enig and her associates were not so sure. Kummerow’s research indicated that the trans fats contributed to heart disease, and Kritchevsky—whose early experiments with vegetarian rabbits were now seen to be totally irrelevant to the human model—had found that trans fatty acids raise cholesterol in humans.17 Enig’s own research, published in her 1984 doctoral dissertation, indicated that trans fats interfered with enzyme systems that neutralized carcinogens and increased enzymes that potentiated carcinogens.18
How much trans fat is “normal”?
The other area needing further investigation concerned just how much trans fat there was in a “normal diet” of the typical American. What had hampered any thorough research into the correlation of trans fatty acid consumption and disease was the fact that these altered fats were not considered as a separate category in any of the data bases then available to researchers. A 1970 FDA internal memo stated that a market basket survey was needed to determine trans levels in commonly used foods. The memo remained buried in the FDA files. The massive Health and Human Services NHANES II (National Health and Nutrition Examination Survey) survey, conducted during the years 1976 to 1980, noted the increasing US consumption of margarine, french fried potatoes, cookies and snack chips—all made with vegetable shortenings—without listing the proportion of trans.
Enig first looked at the NHANES II data base in 1987 and when she did, she had a sinking feeling. Not only were trans fats conspicuously absent from the fatty acid analyses, data on other lipids made no sense at all. Even foods containing no trans fats were listed with faulty fatty acid profiles. For example, safflower oil was listed as containing 14% linoleic acid (a double bond fatty acid of the omega-6 family) when in fact it contained 80%; a sample of butter crackers was listed as containing 34% saturated fat when in fact it contained 78%. In general, the NHANES II data base tended to minimize the amount of saturated fats in common foods.
Over the years, Joseph Sampagna and Mark Keeney, both highly qualified lipid biochemists at the University of Maryland, applied to the National Science Foundation, the National Institutes of Health, the US Department of Agriculture, the National Dairy Council and the National Livestock and Meat Board for funds to look into the trans content of common American foods. Only the National Livestock and Meat Board came through with a small grant for equipment; the others turned them down. The pink slip from National Institutes of Health criticized items that weren’t even relevant to the proposal. The turndown by the National Dairy Council was not a surprise. Enig had earlier learned that Phil Lofgren, then head of research at the Dairy Council, had philosophical ties to the lipid hypothesis. Enig tried to alert Senator Mettzanbaum from Ohio, who was involved in the dietary recommendations debate, but got nowhere.
A USDA official confided to the Maryland research group that they “would never get money as long as they pursued the trans work.” Nevertheless they did pursue it. Sampagna, Keeney and a few graduate students, funded jointly by the USDA and the university, spend thousands of hours in the laboratory analyzing the trans fat content of hundreds of commercially available foods. Enig worked as a graduate student, at times with a small stipend, at times without pay, to help direct the process of tedious analysis. The long arm of the food industry did its best to put a stop to the group’s work by pressuring the USDA to pull its financial support of the graduates students doing the lipid analyses, which the University of Maryland received due to its status as a land grant college.
In December of 1982, Food Processing carried a brief preview of the University of Maryland research19 and five months later the same journal printed a blistering letter from Edward Hunter on behalf of the Institute of Shortening and Edible Oils.20 The University of Maryland studies on trans fat content in common foods had obviously struck a nerve. Hunter stated that the Bailar, Applewhite and Meyer letters that had appeared in Federation Proceedings five years earlier, “severely criticized and discredited” the conclusions reached by Enig and her colleagues. Hunter was concerned that Enig’s group would exaggerate the amount of trans found in common foods. He cited ISEO data indicating that most margarines and shortenings contain no more than 35% and 25% trans respectively, and that most contain considerably less.
What Enig and her colleagues actually found was that many margarines indeed contained about 31% trans fat—later surveys by others revealed that Parkay margarine contained up to 45% trans—while many shortenings found ubiquitously in cookies, chips and baked goods contained more than 35%. She also discovered that many baked goods and processed foods contained considerably more fat from partially hydrogenated vegetable oils than was listed on the label. The finding of higher levels of fat in products made with partially hydrogenated oils was confirmed by Canadian government researchers many years later, in 1993.21
Final results of Enig’s ground-breaking compilation were published in the October 1983 edition of the Journal of the American Oil Chemists Society.22 Her analyses of more than 220 food items, coupled with food disappearance data, allowed University of Maryland researchers to confirm earlier estimates that the average American consumed at least 12 grams of trans fat per day, directly contradicting ISEO assertions that most Americans consumed no more that six to eight grams of trans fat per day. Those who consciously avoided animal fats typically consumed far more than 12 grams of trans fat per day.
Cat and mouse games in the journals
The ensuing debate between Enig and her colleagues at the University of Maryland, and Hunter and Applewhite of the ISEO, took the form of a cat and mouse game running through several scientific journals. Food Processing declined to publish Enig’s reply to Hunter’s attack. Science Magazine published another critical letter by Hunter in 1984,23 in which he misquoted Enig, but refused to print her rebuttal. Hunter continued to object to assertions that average consumption of trans fat in partially hydrogenated margarines and shortenings could exceed six to eight grams per day, a concern that Enig found puzzling when coupled with the official ISEO position that trans fatty acids were innocuous and posed no threat to public health.
The ISEO did not want the American public to hear about the debate on hydrogenated vegetable oils—for Enig this translated into the sound of doors closing. A poster presentation she organized for a campus health fair caught the eye of the dietetics department chairman who suggested she submit an abstract to the Society for Nutrition Education, many of whose members are registered dietitians. Her abstract concluded that “. . . meal plans and recipes developed for nutritionists and dieticians to use when designing diets to meet the Dietary Guidelines, the dietary recommendation of the American Heart Association or the Prudent Diet have been examined for trans fatty acid content. Some diet plans are found to contain approximately 7% or more of calories as trans fatty acids.” The Abstract Review Committee rejected the submission, calling it “of limited interest.”
Early in 1985 the Federation of American Societies for Experimental Biology (FASEB) heard more testimony on the trans fat issue. Enig alone represented the alarmist point of view, while Hunter and Applewhite of the ISEO, and Ronald Simpson, then with the National Association of Margarine Manufacturers, assured the panel that trans fats in the food supply posed no danger. Enig reported on University of Maryland research that delineated the differences in small amounts of naturally occurring trans fats in butter, which do not inhibit enzyme function at the cellular level, and man-made trans fats in margarines and vegetable shortenings which do. She also noted a 1981 feeding trial in which swine fed trans fatty acid developed higher parameters for heart disease than those fed saturated fats, especially when trans fatty acids were combined with added polyunsaturates.24 Her testimony was omitted from the final report, although her name in the bibliography created the impression that her research supported the FASEB whitewash.25
In the following year, 1986, Hunter and Applewhite published an article exonerating trans fats as a cause of atherosclerosis in the prestigious American Journal of Clinical Nutrition26, whose sponsors, by the way, include companies like Procter and Gamble, General Foods, General Mills, Nabisco and Quaker Oats. The authors once again stressed that the average per capita consumption of trans fatty acids did not exceed six to eight grams. Many subsequent government and quasi government reports minimizing the dangers of trans fats used the 1986 Hunter and Applewhite article as a reference.
Enig testified again in 1988 before the Expert Panel on the National Nutrition Monitoring System (NNMS). In fact she was the only witness before a panel, which began its meeting by confirming that the cause of America’s health problems was the overconsumption of “fat, saturated fatty acids, cholesterol and sodium.” Her testimony pointed out that the 1985 FASEB report exonerating trans fatty acids as safe was based on flawed data.
Behind the scenes, in a private letter to Dr. Kenneth Fischer, Director of the Life Sciences Research Office (LSRO), Hunter and Applewhite charged that “the University of Maryland group continues to raise unwarranted and unsubstantiated concerns about the intake of and imagined physiological effects of trans fatty acids and . . . they continue to overestimate greatly the intake of trans acids by typical Americans.” “No one other than Enig,” they said, “has raised questions about the validity of the food fatty acid composition data used in NHANES II and. . . she has not presented sufficiently compelling arguments to justify a major reevaluating.”
The letter contained numerous innuendos that Enig had mischaracterized the work of other researchers and had been less than scientific in her research. It was widely circulated among National Nutrition Monitoring System agencies. John Weihrauch, a USDA scientist, not an industry representative, slipped it surreptitiously to Dr. Enig. She and her colleagues replied by asking, “If the trade association truly believes ‘that trans fatty acids do not pose any harm to humans and animals’. . . why are they so concerned about any levels of consumption and why do they so vehemently and so frequently attack researchers whose finding suggest that the consumption of trans fatty acids is greater than the values the industry reports?”
Maryland researchers argued that trans fats should be included in food nutrition labels; the Hunter and Applewhite letter asserted that “there is no documented justification for including trans acids . . . as part of nutrition labeling.”
During her testimony Enig also brought up her concerns about other national food databases, citing their lack of information on trans. The Food Consumption Survey contained glaring errors—reporting, for example, consumption of butter in amounts nearly twice as great as what exists in the US food supply, and of margarine in quantities nearly half those known to exist in the food supply. “The fact that the data base is in error should compel the Congress to require correction of the data base and reevaluation of policy flowing from erroneous data,” she argued, “especially since the congressional charter for NHANES was to compare dietary intake and health status and since this data base is widely used to do just that.” Rather than “correction of the data base,” [The] National Nutritional Monitoring System officials responded to Enig’s criticism by dropping the whole section pertaining to butter and margarine from the 1980 tables.
Enig’s testimony was not totally left out of the National Nutritional Monitoring System final report, as it had been from the FASEB report three years earlier. A summary of the proceedings and listing of panelists released in July of 1989 by Director Kenneth Fischer announced that a transcript of Enig’s testimony could be obtained from Ace Federal Reporter in Washington DC.27 Unfortunately, his report wrongly listed the date of her testimony as January 20, 1988, rather than January 21, making her comments more difficult to retrieve.
The Enig-ISEO debate was covered by the prestigious Food Chemical News and Nutrition Week 28—both widely read by Congress and the food industry, but virtually unknown to the general public. National media coverage of dietary fat issues focused on the proceedings of the National Heart, Lung and Blood Institute as this enormous bureaucracy plowed relentlessly forward with the lipid hypothesis. In June of 1984, for example, the press diligently reported on the proceedings of the NHLBI’s Lipid Research Clinics Conference, which was organized to wrap up almost 40 years of research on lipids, cholesterol and heart disease.
The problem with the 40 years of NHLBI-sponsored research on lipids, cholesterol and heart disease was that it had not produced many answers—at least not many answers that the NHLBI was pleased with. The ongoing Framingham Study found that there was virtually no difference in coronary heart disease “events” for individuals with cholesterol levels between 205 mg/dL and 294 mg/dL—the vast majority of the US population. Even for those with extremely high cholesterol levels—up to almost 1200 mg/dL, the difference in CHD events compared to those in the normal range was trivial.29 This did not prevent Dr. William Kannel, then Framingham Study Director, from making claims about the Framingham results. “Total plasma cholesterol” he said, “is a powerful predictor of death related to CHD.” It wasn’t until more than a decade later that the real findings at Framingham were published—without fanfare—in the Archives of Internal Medicine, an obscure journal. “In Framingham, Massachusetts,” admitted Dr. William Castelli, Kannel’s successor “the more saturated fat one ate, the more cholesterol one ate, the more calories one ate, the lower people’s serum cholesterol. . . we found that the people who ate the most cholesterol, ate the most saturated fat, ate the most calories weighed the least and were the most physically active.”30
NHLBI’s Multiple Risk Factor Intervention Trial (MRFIT) studied the relationship between heart disease and serum cholesterol levels in 362,000 men and found that annual deaths from CHD varied from slightly less than one per thousand at serum cholesterol levels below 140 mg/dL, to about two per thousand for serum cholesterol levels above 300 mg/dL, once again a trivial difference. Dr. John LaRosa of the American Heart Association claimed that the curve for CHD deaths began to “inflect” after 200 mg/dL, when in fact the “curve” was a very gradually sloping straight line that could not be used to predict whether serum cholesterol above certain levels posed a significantly greater risk for heart disease. One unexpected MRFIT finding the media did not report was that deaths from all causes—cancer, heart disease, accidents, infectious disease, kidney failure, etc.—were substantially greater for those men with cholesterol levels below 160 mg/dL.31
Lipid Research Clinics Trial
What was needed to resolve the validity of the lipid hypothesis once and for all was a well-designed, long-term diet study that compared coronary heart disease events in those on traditional foods with those whose diets contained high levels of vegetable oils—but the proposed Diet-Heart study designed to test just that had been cancelled without fanfare years earlier. In view of the fact that orthodox medical agencies were united in their promotion of margarine and vegetable oils over animal foods containing cholesterol and animal fats, it is surprising that the official literature can cite only a handful of experiments indicating that dietary cholesterol has “a major role in determining blood cholesterol levels.” One of these was a study involving 70 male prisoners directed by Fred Mattson32—the same Fred Mattson who had pressured the American Heart Association into removing any reference to hydrogenated fats from their diet-heart statement a decade earlier. Funded in part by Procter and Gamble, the research contained a number of serious flaws: selection of subjects for the four groups studied was not randomized; the experiment inexcusably eliminated “an equal number of subjects with the highest and lowest cholesterol values;” twelve additional subjects dropped out, leaving some of the groups too small to provide valid conclusions; and statistical manipulation of the results was shoddy. But the biggest flaw was that the subjects receiving cholesterol did so in the form of reconstituted powder—a totally artificial diet. Mattson’s discussion did not even address the possibility that the liquid formula diet he used might affect blood cholesterol differently than would a whole foods diet when, in fact, many other studies indicated that this is the case. The culprit, in fact, in liquid protein diets appears to be oxidized cholesterol, formed during the high-temperature drying process, which seems to initiate the buildup of plaque in the arteries.33 Powdered milk containing oxidized cholesterol is added to reduced fat milk—to give it body—which the American public has accepted as a healthier choice than whole milk. It was purified, oxidized cholesterol that Kritchevsky and others used in their experiments on vegetarian rabbits.
The NHLBI argued that a diet study using whole foods and involving the whole population would be too difficult to design and too expensive to carry out. But the NHLBI did have funds available to sponsor the massive Lipid Research Clinics Coronary Primary Prevention Trial in which all subjects were placed on a diet low in cholesterol and saturated fat. Subjects were divided into two groups, one of which took a cholesterol-lowering drug and the other a placebo. Working behind the scenes, but playing a key role in both the design and implementation of the trials, was Dr. Fred Mattson, formerly of Procter and Gamble.
An interesting feature of the study was the fact that a good part of the trial’s one-hundred-and-fifty-million-dollar budget was devoted to group sessions in which trained dieticians taught both groups of study participants how to choose “heart-friendly” foods—margarine, egg replacements, processed cheese, baked goods made with vegetable shortenings, in short the vast array of manufactured foods awaiting consumer acceptance. As both groups received dietary indoctrination, study results could support no claims about the relation of diet to heart disease. Nevertheless, when the results were released, both the popular press and medical journals portrayed the Lipid Research Clinics trials as the long-sought proof that animal fats were the cause of heart disease. Rarely mentioned in the press was the ominous fact that the group taking the cholesterol-lowering drugs had an increase in deaths from cancer, stroke, violence and suicide.34
LRC researchers claimed that the group taking the cholesterol-lowering drug had a 17% reduction in the rate of CHD, with an average cholesterol reduction of 8.5%. This allowed LRC trials Director Basil Rifkind to claim that “for each 1% reduction in cholesterol, we can expect a 2% reduction in CHD events.” The statement was widely circulated even though it represented a completely invalid representation of the data, especially in light of the fact that when the lipid group at the University of Maryland analyzed the LRC data, they found no difference in CHD events between the group taking the drug and those on the placebo.
A number of clinicians and statisticians participating in a 1984 Lipid Research Clinics Conference workshop, including Michael Oliver and Richard Krommel, were highly critical of the manner in which the LRC results had been tabulated and manipulated. The conference, in fact, went very badly for the NHLBI, with critics of the lipid hypothesis almost outnumbering supporters. One participant, Dr. Beverly Teter of the University of Maryland’s lipid group, was delighted with the state of affairs. “It’s wonderful’” she remarked to Basil Rifkind, study coordinator, “to finally hear both sides of the debate. We need more meetings like this” His reply was terse and sour: “No we don’t.”
National Cholesterol Consensus Conference
Dissenters were again invited to speak briefly at the NHLBI-sponsored National Cholesterol Consensus Conference held later that year, but their views were not included in the panel’s report, for the simple reason that the report was generated by NHLBI staff before the conference convened. Dr. Teter discovered this when she picked up some papers by mistake just before the conference began, and found they contained the consensus report, already written, with just a few numbers left blank. Kritchevsky represented the lipid hypothesis camp with a humorous five-minute presentation, full of ditties. Edward Ahrens, a respected researcher, raised strenuous objections about the “consensus,” only to be told that he had misinterpreted his own data, and that if he wanted a conference to come up with different conclusions, he should pay for it himself.
The 1984 Cholesterol Consensus Conference final report was a whitewash, containing no mention of the large body of evidence that conflicted with the lipid hypothesis. One of the blanks was filled with the number 200. The document defined all those with cholesterol levels above 200 mg/dL as “at risk” and called for mass cholesterol screening, even though the most ardent supporters of the lipid hypothesis had surmised in print that 240 should be the magic cutoff point. Such screening would, in fact, need to be carried out on a massive scale as the federal medical bureaucracy, by picking the number 200, had defined the vast majority of the American adult population as “at risk.” The report resurrected the ghost of Norman Jolliffe and his Prudent Diet by suggesting the avoidance of saturated fat and cholesterol for all Americans now defined as “at risk,” and specifically advised the replacement of butter with margarine.
The Consensus Conference also provided a launching pad for the nationwide National Cholesterol Education Program, which had the stated goal of “changing physicians’ attitudes.” NHLBI-funded studies had determined that while the general population had bought into the lipid hypotheses, and was dutifully using margarine and buying low-cholesterol foods, the medical profession remained skeptical. A large “Physicians Kit” was sent to all doctors in America, compiled in part by the American Pharmaceutical Association, whose representatives served on the NCEP coordinating committee. Doctors were taught the importance of cholesterol screening, the advantages of cholesterol-lowering drugs and the unique benefits of the Prudent Diet. NCEP materials told every doctor in America to recommend the use of margarine rather than butter.