- 1970. Report of the Inter-Society Commission for Heart Disease Resources relating to the Prevention of Cardiovascular Disease and the Primary Prevention of the atherosclerotic process.
At the same time that the NHLBI Task Force on Atherosclerosis was meeting under the chairmanship of Fredrickson, Keys’ principal acolyte Jeremiah Stamler was convening another Commission to address the primary prevention of atherosclerosis. The report of the committee was released in 1970 (1). Even as the NHLBI Task Force was deciding that a test of the Diet-Heart Hypothesis was not feasible, yet Stamler’s Commission was acting as if the evidence was already in hand. And that it was definitive. Recall that this was 1970, still 7 years shy of the Committee of Senator McGovern’s Dietary Goals for Americans (2).
Thus amongst the Commission’s recommendations (1) were a number of novel dietary guidelines squarely based on Keys’ model of how atherosclerosis develops.
- Eat to maintain optimum weight.
- Eat less than 300mg of cholesterol per day. This despite clear evidence, as espoused by Keys himself, that dietary cholesterol has no influence on blood cholesterol concentrations (3). For as Keys and his colleagues had written in 1956: “It is concluded that in adult men the serum cholesterol level is essentially independent of the cholesterol intake over the whole range of natural human diets. It is probable that infants, children and women are similar” (3, p.56).
- Consume about 30% of total daily energy from fat, equally divided between saturated, polyunsaturated and monounsaturated fats. This essentially established an upper limit of 10% of total calories for daily saturated fat intake. Predictably there was no experimental evidence to support this advice.
- Foods to be labelled for their contents of calories, of cholesterol and saturated, polyunsaturated and monounsaturated fats.
- Substitution of margarine for butter.
Fredrickson, in his position then as Director of Intramural Research at the NHLBI was not entirely convinced. In a lecture presented at the College of Physicians and Surgeons in London on 18th December 1970 (4), he voiced his concerns about Stamler’s Commission’s proposals:
“In the light of what is actually known, the injunctions on consumption of cholesterol and fats seem too radical as they stand. What evidence do we have that an egg yolk a day spells jeopardy for all Americans? Do we have enough information about marginal hyperglyceridaemia or incipient diabetes to advise everyone to eat a diet which will tend to provide more than half of the calories as carbohydrates? What of sucklings and older infants? The Commission’s report leads to an inference that a third of their calories from fat should also be polyunsaturated. Are we convinced of the safety of a diet containing 10% of polyunsaturates to the extent that we want to insist on this in baby’s formula?”
“Finally, are we certain enough of the efficacy of such sweeping changes that we, as physicians, can convincingly follow them ourselves?”
“Observing that, despite a mass of suggestive and encouraging evidence, there is no conclusive proof of the preventive value of the dietary change it advocates, the Commission urges the Government to get on with large-scale trials of the effect of its diet on the incidence of ischaemic heart disease. It wants, however, its recommendations concerning diet, and others aimed at wide spread treatment of mild hypertension and a phasing out of the cigarette industry, to be implemented at the same time that trials to determine their effects are being planned”.
“Here the Commission seems to find itself somewhat at cross-purposes. If its recommendations were wholeheartedly adopted tomorrow would a field trial then be possible in the general population? Surely no one could accept seriously the results of a study in which the control group would necessarily consist solely of non-adherers”.
“The hard truth is that, without its conversion to public policy – with the official acceptance of the legitimacy of the diet prescription, the mass education, and the Federal supports to the endangered sectors of the food industry this implies – the momentum of a drive to lower the mean cholesterol enough to lower convincingly the collective incidence of ischaemic heart disease is likely to fail. And one looks vainly in the Commission Report for something new and strong enough to change the present governmental non-position concerning dietary fat consumption” (4, p.187-188).
In short Fredrickson, the father of the Fredrickson Classification of Hyperlipidemias, was utterly unconvinced that in 1970 there was sufficient evidence to force humans to remove foods from their diet – like eggs and butter – that we had eaten for eons. And to replace those with “vegetable” oils that have been in the food chain only since the early 1900s.
He also raised the inconvenient question about “sucklings and older infants”: What would be the effects on their long-term health of removing fat from their diets? And what about the addition of so much dietary carbohydrate to everyone’s diet? What long-term consequences would that have? This inconvenient question continues to be ignored even today.
Nor was he convinced that these dietary changes would ever produce a sufficient lowering of blood cholesterol concentrations to reduce CHD rates. Did this suggest some uncertainty in his mind of the real role of cholesterol in the causation of CHD?
A subsequent editorial by Paul Stolley MD (5) added additional warnings which become important in any retrospective analysis of exactly what were the long-term consequences of this document.
Like Fredrickson, Stolley from the Johns Hopkins School of Hygiene and Public Health was concerned that the Committee, directed by Keys’ premier acolyte, Jeremiah Stamler MD, was acting ahead of the evidence. He noted that: “The Committee is convinced (my added emphasis) that a diet high in saturated fat and cholesterol, hypertension and cigarette smoking are widespread major risk factors leading to the development of atherosclerotic disease” (5, p. 661).
As a result of this conviction: “There are those who believe that the evidence for the dietary hypothesis are so overwhelming that awaiting results of such field trials (of Keys’ Hypotheses) would be unwarranted….The Commission anticipates (my added emphasis) the results of the experiment and would start us on a national program of dietary modification while we are still conducting field trials. Even if the results of these trials were then unconvincing, their recommendations, if implemented would still have improved the quality of much of our packaged food supply by substituting protein for fat. Accordingly, the Commission recommends that the food industry be encouraged to make available leaner meats, low-fat dairy products, and so on, and to substitute protein or low-saturated fat in many prepared foods that are highly saturated – in effect, put more meat than fat in our hot dogs and cold cuts and reduce the effort to produce fat cattle” (p. 662)
In other words, the Commission had essentially adopted the position of Keys, Stamler and the other acolytes, specifically that there really was no need for any further science to be performed. Keys was correct – everyone of consequence “knew” that to be the truth. As a result it was fully justified to be convinced that the anticipated outcomes would be achieved regardless of the results that any clinical trials might produce. For even if those results were not favorable, the advice to replace dietary saturated fat with protein would improve the quality of many “prepared foods”.
Recall that this happened in 1970, long before any longer term randomized controlled trials of this type of dietary advice had even been contemplated.
But Stolley persisted and asked three additional questions:
- Does the knowledge exist for disease prevention?
- What are the costs involved in applying the knowledge?
- What benefits in terms of lives saved, disability prevented, and other economic and social losses averted, can be achieved?
He also made that astute observation that this campaign would “suddenly challenge the nutritional precepts most housewives were taught in school – serve plenty of red meat, dairy foods, organ meats, and shellfish – and it is not yet proved that it would prevent heart disease” (p.662). This, incidentally, is the dietary advice on which my mother raised me in the ‘fifties.
Stolley’s advice was rather to favour “awaiting the results of a national dietary study before this step (of advocating a radical overhaul of the US diet)” (p.662) especially since when working from inadequate evidence, it is “not yet known if the decisions of the Commission are right…” (p.662).
With the benefit of half a century of hindsight, we now know that the attitude that the Commission promoted – acting before the evidence was available on the basis of anticipation, because to act otherwise would be harmful – would be the exact approach adopted by the disastrous George McGovern US Senate Select Committee on Nutrition and Human Needs when it produced its Dietary Goals for the United States in 1977 (2).
We also know that those guidelines would prove catastrophic because they heralded the beginning of the global obesity/diabetes epidemic that started immediately those novel 1977 dietary guidelines were promoted (6).
And a key reason for that development was because of a dietary change that the 1970 Commission neither advised nor foresaw. Whereas the Commission presumed that the US diet would be improved by substituting dietary protein for saturated fat, what actually happened after 1977 was that dietary saturated fats were replaced by sugar, carbohydrates and “vegetable oils”.
First in the US diet. And then, by osmosis, in the rest of the world.
- 1971. The World Health Organization convenes an international meeting to initiate international trials into risk factor modification for CHD prevention.
In 1971 the WHO convened an international meeting to discuss the implementation of large intervention studies of CHD risk factor modification to determine whether such intervention could indeed reduce the growing burden of CHD in developed countries.
The Americans and Norwegians reported that they were already planning the MRFIT and Oslo Primary Intervention Trials (7) respectively. The UK delegation led by Professor G Rose proposed a multifactorial trial in industry. So was born the European Collaborative Trial of Multifactorial Prevention of Coronary Heart Disease.
For the trial, 60 881 men aged between 40-59 years were recruited from 80 factories in the UK, Belgium, Italy and Poland. One half received preventive advice on a cholesterol-lowering diet, on smoking cessation, on control of body weight and of blood pressure and on regular physical activity (8,9). The other half formed the control group, 10% of whom were examined for coronary risk factors at the start of the trial. At the time it was the largest randomized trial of CHD prevention ever undertaken.
The first results began to be reported in 1980 (9).
- 1971. The Multiple Risk Factor Intervention Trial (MRFIT) is initiated.
The MRFIT is one of the most intensive risk factor modification trials yet undertaken. It will never be repeated. As Moore (10) describes, the study took 10 years, involved 28 medical centres across the USA, and occupied the time of 250 researchers, all at a cost of $115 million.
The goal was to find 12 888 persons at high risk for future CHD. This required the initial screening of 361,662 men. The selected cohort was then divided into two groups – a Usual Care Group whose health care continued to be managed by their own physicians. And a Special Investigation Group (SIG) that received an intensive educational and behavioural intervention, unmatched in its scope before or since. The aim of these interventions was to change the health behaviours of the SIG for the better. Or, at least, those behaviours that the researchers believed would substantially reduce their risks for developing CHD.
The study was under the control of Jeremiah Stamler, one of Keys’ most loyal lieutenants, and the person most certain that the Keys’ Twin hypotheses were divinely correct and did not require further proof (1). The man whose career owed everything to the closeness he maintained with Keys. Recall that when Keys needed to convince the AHA to back his unproven hypothesis in 1961, he ensured that Stamler was placed with him on the crucial committee that would make that decision (11). With his faithful lieutenant in tow, Keys successfully achieved that crucial objective.
Stamler had also inherited control of the Western Electric Study which found that “the amount of saturated fat in the diet was not significantly associated with risk of death form CHD (12, p.68)”. One might have thought that this finding would have spiked Stamler’s interest. Instead throughout that article and especially in the abstract, he carefully towed the party line: “The results support the conclusion that lipid composition of the diet affects serum cholesterol concentration (True but only in controlled laboratory experiments, not in free-living populations – my addition) and risk of coronary death in middle-aged American men (False for this study – my addition)” (12, p.65).
Actually no, Dr Stamler. They support an utterly opposite conclusion. That dietary fat is unrelated to risk of CHD. As we will discuss subsequently, this has been found rather too frequently to be repeatedly ignored, over and over again.
In her book, Big Fat Surprise, Nina Teicholz describes how she interviewed Stamler and asked him to explain this finding from the Western Electric Study. He yelled: “It (saturated fat) had no INDEPENDENT effect (on CHD risk)” (13, p. 97). Which meant to her that Stamler knew that the Western Electric study disproved Keys’ Diet-Heart hypothesis since it found no evidence that saturated fat predicted future risk of CHD.
This was the man responsible for overseeing the study the results of which would be reported for the first time in September 1982.
A man who would continue to believe that Keys was correct even when he turned 100 in 2019 (14). Clearly his own health had not suffered as a result of that belief.
Keys too would live for more than 100 years.
- 1971-1998. The Helsinki Policemen Study is initiated.
Between 1966 and 1967, 1326 men over 30 years of age and free of CHD who were working for either the Police Department of the city of Helsinki or who were members of the National Police Force stationed in Helsinki were examined medically. In 1971, a total of 1259 of these men were re-examined.
From that group 970 men who were aged between 34-64 years and free of CHD or other clinically significant cardiovascular disease, became subjects for the Helsinki Policemen Study (15).
The study would become one of the first prospective studies examining hyperinsulinemia as a risk factor for the development of CHD.
The findings from the study would be reported over the next 30 years.
- 1972. The National Cooperative Pooling Project data are published.
In 1972, the American Heart Association began a process to collect what, at that time, was the largest body of data measuring the (associational) relationship of coronary “risk factors” and the subsequent development of CHD – the National Cooperative Pooling Project of the AHA (16,17). It was hypothesized that “pooling data from several studies would yield analytical results giving increase assurance concerning the relationship of key risk factors to CHD, and more precise quantitative estimates of their importance singly and in combination than the individual studies permitted” (16, p.2). Which again sounds as if the authors and the AHA were fairly certain of what they would (have to) find.
In the end, data from five US studies including the Western Electric Study, the Framingham Study, Keys’ Minneapolis Professional and Business Men Study, and the Tecumseh Community Health Study were included in the data base. Importantly all the key players promoting Keys’ Diet-Heart hypothesis were involved including Jeremiah Stamler, William Kannel, Thomas Dauber, William Castelli, Ancel Keys himself, Henry Blackburn and Frederick Epstein.
The main conclusion from the study was that “the relationship between the major coronary risk factors and susceptibility to coronary events is a strong one” (17, p.260). This was true for blood pressure, smoking, relative weight and the presence of ECG abnormalities.
With regard to blood cholesterol concentrations – “Serum cholesterol once again is revealed to be consistently and strongly related to coronary heart disease. Over the range of values many clinical laboratories designate as within the normal limits (e.g. 220-300mg/dL; 5.7 – 7.8mmol/L), risk is proportional to cholesterol level. An optimal value cannot be readily identified based on this experience of middle-aged American men because of the paucity of data in the lower two quintiles, but it is somewhat well under 200mg/dL (5.1 mmol/L)” (17, p.260).
Thus the findings were very conveniently, entirely compatible with Keys’ Diet-Heart hypothesis.
But when these findings for blood cholesterol concentrations are presented graphically as Reisser (18) has done, the practical relevance of the finding becomes somewhat more perplexing.
Figures 1 and 2 depict the relationship between blood cholesterol concentrations and rates of first major coronary event, of sudden death, of all CHD deaths and of all deaths as reported in the Pooling Project.
Legend to figure 1. The relationship between blood cholesterol concentrations and subsequent 10-year risk for first major coronary event (left panel) and sudden death (right panel). Whilst the rate of first major coronary event rises with increased blood cholesterol concentrations (left panel), the magnitude of the increase is too trivial to be meaningful. Reproduced from figure 1 in reference 18.
Figure 1 shows that rate of first major coronary event rises with increasing blood cholesterol concentrations (left panel) as does the rate of sudden death (right panel). But the real question is: By how much did the risk increase for any given increase in the blood cholesterol concentration? In other words: How helpful is knowing the blood cholesterol concentration in predicting future risk of a dire cardiovascular event, such as a first heart attack or sudden death?
The data show that the increase in rates of first major coronary event rose from 45 per 1000 per 10 years to 67 per 1000 per 10 years as the blood cholesterol rose from <175mg/dL (4.5mmol/L) to 249mg/dL (6.4mmol/L). In effect this was an increase from 0.45/100/year to 0.67/100/year. In other words for the individual, what is now considered an extremely high blood cholesterol concentration increased the risk of an adverse cardiovascular event from 0.0045 to 0.0067 per annum; an infinitesimal increase in risk that was “too trivial to be considered” (18, p.868).
Even comparing rates of first major coronary events in those with the lowest and highest blood cholesterol concentrations, rates rose from 0.45/100/year at a blood cholesterol concentration of <175mg/dL (4.5mmol/L) to 1.62/100/year at a blood cholesterol concentration >330mg/dL (8.5mmol/L). Whilst this is indeed a 3.6-fold (360%) increase, in absolute terms the increase in risk is irrelevant.
Similarly rates of all CHD deaths and all deaths (Figure 2) also increased insignificantly with increasing serum cholesterol concentrations.
Legend to figure 2. The relationship between blood cholesterol concentrations and subsequent 10-year risk for All CHD Deaths (left panel) and All Deaths (right panel). Whilst the rates of All CHD deaths and of All Deaths rises somewhat with increasing blood cholesterol concentrations, the magnitude of the increase is too trivial to be meaningful. Reproduced from figure 1 in reference 18.
Thus the rates of all CHD deaths rose from 24/1000/10 years at the lowest blood cholesterol concentrations to 58/1000/10years at a blood cholesterol concentrations >299mg/dL (7.8mmol/L) indicating an annual predicted increase in risk from 0.24/100 persons 0.58/100 persons. Deaths from all causes rose from 0.62 to 0.90/100 per annum for the same increase in blood cholesterol concentrations.
Reisser concluded that: “Such data do not support the theory that risk increases with each increment in serum cholesterol concentration” (18, p.868).
Reisser emphasized his point by quoting iconic Texas heart surgeon Dr Michael Debakey who reported that only 30-40% of persons with atherosclerotic heart disease have elevated blood cholesterol concentrations (19): “If you say that that’s (elevated blood cholesterol concentration) the cause, how do you explain the other 60 to 70%?” (18, p.870).
DeBakey and his colleagues, who daily treated persons with CHD, concluded their article: “..studies which focus on determination of basal serum cholesterol levels to the exclusion of other possible determinants of atherosclerosis carry only the remote likelihood of contributing additionally or significantly to solution of the problem of the etiology of occlusive and aneurysmal disease” (19, p.659).
The answer is that there is another biological step, missing in Keys’ Diet-Heart Hypothesis, that progresses coronary atherosclerosis into acute myocardial infarction or sudden death (20,21) (figure 4 in 20) – factors influencing coronary artery plaque rupture and blood clotting leading to coronary thrombosis.
And those factors are largely absent from any discussions of Keys’ Twin Hypotheses.
But perhaps the truly important point is that anyone examining these graphs in 1972 when the data were published could only have come to one conclusion:
That the theory that an elevated blood cholesterol concentration is the key driver of CHD simply cannot be correct. There has to be more. Much more.
- 1973 Lipid Research Clinics (LRC) Coronary Primary Prevention Trial (CPPT) is planned
This trial like the MCE, involved a number of key players whose future careers would be significantly impacted by any research they produced which found that lowering the blood cholesterol concentration, either by diet or drugs, produced favourable outcomes in terms of cardiovascular health.
The trial was under the directorship of Dr Basil Rifkind and Dr Robert Levy then working in the Program Office of the NHLBI. The directors of three of the 12 collaborating laboratories, Dr Antonio Gotto at Baylor College of Medicine, Houston, Dr Daniel Steinberg of the University of California at San Diego, and Dr Ivan Franz of the University of Minnesota in Minneapolis would subsequently feature prominently in the drive to have the Diet-Heart and Lipid Hypotheses accepted as the consensus opinion.
Thanks to this interest, Dr Gotto especially would become particularly close to a number of pharmaceutical companies producing cholesterol-lowering drugs. So enamoured did he become with the prescription of those drugs that, in time, he would predict that statins “will have conquered” atherosclerosis as early as 2000 (22, p.58).
Which of course would be true if cholesterol is the singular cause, not one of many reasons (23,24) for the development of coronary atherosclerosis. The fact that Dr Gotto’s prediction has failed to materialize confirms that cholesterol cannot then be the singular cause of atherosclerosis (23).
Gotto continues: “Statins have had a greater impact on heart disease than anything that’s come out in the past 50 years,” “They are very safe drugs. I think their downside is very low” (25, p.1). Gotto would describe himself as an “educational consultant for several pharmaceutical companies that produce statins” (25, p.1). One wonders exactly on what topics the pharmaceutical companies felt an academic cardiology could educate them. Perhaps something to do with increasing their sales?
Many of us are still waiting for the evidence that statins make any difference at all to CHD outcomes.
As was the case with the MCE, the outcome would help define the careers of all its senior researchers. A positive finding would boost their careers; an indifferent result would mean they would have to re-direct their future careers along some other research avenue. But would that research direction be as well funded? And would they be invited to serve on the expert review committees that oversaw where future monies should be spent? For these were just of few of the insider advantages that their obedience to Ancel Keys and the AHA/NHI alliance provided for these acolytes.
The point is that the senior scientists charged with conducting the LRC CPPT were not disinterested participants. Instead they were careerists whose future academic prospects would be enhanced by any findings that supported the Diet-Heart and Lipid hypotheses. And all were firmly convinced that Keys’ Twin Hypothesis were divinely ordained.
The work of the CPPT was provided by 12 selected lipid research clinics that agreed beforehand to all the testing protocols. They wished to ensure that when the results were finally analysed, all 12 collaborating laboratories would be comfortable with the results, whatever the outcome. Perhaps the unspoken assumption was that the study would convincingly prove the Lipid Hypothesis sending a clear and unambiguous message to the world. And they would be remembered as the brilliant scientists who fashioned this medical breakthrough.
The goal of the intervention was to use the drug, cholestyramine, to lower blood cholesterol concentrations by about 25%. Cholestyramine binds with bile, secreted by the liver, in the intestine. Since bile is produced in the liver from cholesterol, its increased removal from the body lowers the blood cholesterol concentration somewhat. The researchers believed that the drug would lower the blood cholesterol concentration by a minimum of 25% with a corresponding reduction in CHD events by at least 50% in the seven years of the trial.
So certain where the researchers that the result would be unambiguous that in their original description of the research protocol, they indicated they would use the most rigorous statistical methods available. As they described it, their method of analysis would ensure that there would be a “99 percent certainty that the results were not a statistical fluke” (10, p.49).
“Since the time, magnitude, and cost of this study make it unlikely that it could ever be repeated, it was essential to be sure that any observed beneficial effect of lower cholesterol was a real one” (10, p.49). As we shall see, once the results of the study failed to show that the benefit of lowering the blood cholesterol concentration was “a real one”, they reneged on their statistical purity, proving that what they truly desired was a positive outcome, regardless of the manipulations it would take to produce that result.
As was the case with the MRFIT trial, the researchers had to find a population of men at the highest possible risk of CHD. Since the presumption was that the highest risk would occur in those with the highest blood cholesterol concentrations, the researchers needed to find a group of 3,810 men with the very highest blood cholesterol concentrations, higher than 95% of the US male population. To find those elusive men, the researchers would need to screen 480,000 middle-aged US men, a process that required 4 long consultations with each potential participant spread over 16 weeks. The trial would cost $142 million.
The 3,810 men were split into a control and intervention group and both were placed on a cholesterol-lowering diet with fewer eggs, leaner meat, and lower dairy intake than the national average. This would reduce any potential effect of subjects in either group choosing to change their diets during the 3-year trial. But as Taubes points out, the ethics of this was unclear since “the LRC investigators had no proof that such a diet would benefit the subjects, rather than harm them” (22, p.57). Recall the ethical dilemma – First do no harm.
However, importantly, the trial was not a diet trial – it was a drug trial. Yet, as I describe subsequently, the results would be presented to the world as if they came from a drug trial.
- 1973. The first results of the NiHonSan (Nagasaki Hiroshima Honolulu San Francisco Japanese ancestry) study are reported.
When studied as part of Keys’ Seven Countries Study, it was found that persons of Japanese ancestry living in Japan have very low rates of coronary heart disease – about one-fifth of US rates – associated with very much lower average blood cholesterol concentrations (4.25 vs 6.20mmol/L) than US citizens (26). This finding would provide pivotal support for Keys’ Diet-Heart Hypothesis since the Japanese data anchored the bottom left corner of his iconic slide (figure 1 in reference 27).
Yet these Japanese data are paradoxical for at least two reasons. First, despite much lower historical blood cholesterol concentrations, the Japanese have very much higher rates of cerebral strokes than do Europeans (28) and US citizens. According to the Keys Hypothesis, this would be because the Japanese must have more severe atherosclerosis in their cerebral arteries. But this is paradoxical if it is argued that low blood cholesterol concentrations protect the Japanese from suffering heart attacks. How do the same low blood cholesterol concentrations put the Japanese at increased risk for strokes but protect them for heart attacks?
Second, the intake of dietary saturated fat by the Japanese has tripled between 1960 and 1990 (29) with an increased consumption of fat and protein from animals with a striking reduction in carbohydrate consumption from a peak of about 420 grams/day in 1950 to 190 grams/day in 1990 (Figure 3).
Legend to Figure 3: Changes in macronutrient intakes from 1946 to 1990 in Japanese living in Japan. Reproduced from figure 3 in reference 29.
Figure 3 shows that since the end of World War II in 1945, there has been a dramatic change in the Japanese diet. Whereas in 1946 81% of calories in the Japanese diet came from carbohydrates (lowest panel in figure 14), this had fallen to 59% in 1990. Although protein intake has changed somewhat – from 12% to 16% of total calories over the same time period – the major change has been in the contribution of dietary fat to total daily caloric intake. This has risen from 7% in 1946 to 25% of total daily calories in 1990. This reflects an increase from 18 to 58g/person/day (third panel from top in figure 3).
The second and third panels from the top in figure 3 show that animal sources have provided the greater part of the increase in protein and fat consumed by the Japanese.
According to Keys’ Twin Hypotheses and to those, like Harvard Professors Walter Willett and Frank Hu who promote a vegetarian-type “Mediterranean Diet” (30), the consequences for Japanese health must have been dire. Such a change, supposedly identical to what the McGovern Commission (incorrectly) claimed had occurred in the US after 1920, precipitating the US CHD epidemic (2), must have caused a steep increase in blood cholesterol concentrations and a dramatic rise in CHD.
Indeed average blood cholesterol concentration has risen quite steeply in the Japanese men and women since (31) whereas it has fallen progressively in seven other countries in Europe, North America and Australasia (figure 4).
Legend to figure 4. There has been a progressive increase in the average blood cholesterol concentrations in both Japanese men (blue line – left panel) and women (blue line – right panel) from 1980 to 2000 associated with increased consumption of animal products (figure 3). In contrast, blood cholesterol concentrations have fallen linearly in the same time period in the United Kingdom, Australia, Sweden, United States, Canada, Spain and France. Reproduced from figure 1 in reference 31.
Despite these rising blood cholesterol concentrations, mortality from CHD and stroke in both Japanese men and women has been in decline since 1970 (31,32). Today despite increased dietary fat intake including from animal sources and substantially higher blood cholesterol concentrations, at birth, the Japanese have the longest life expectancy in the world (33). Interestingly these differences are not explained by lesser degrees of atherosclerosis in Japanese males than, for example, in white males in the US (34). Thus the relative immunity of the Japanese to CHD is not because they are less prone to atherosclerosis. Instead they must be more resistant to atherosclerotic plaque rupture with the development of thrombosis in the coronary arteries (11).
Predictably these findings cannot be explained by any part of Keys’ Twin Hypotheses. Thus the findings are paradoxical and require explanation.
Between 1973 and 1975, the results from one of the first studies attempting to address this paradox were published. The study would become known as the NiHonSan study (35-41). The acronym comes from the four cities in which persons of Japanese ancestry were studied – Nagasaki, Hiroshima, Honolulu and San Francisco.
The initial goal of the study was to try to explain the paradoxical finding that whilst Keys’ Seven Countries Study (SCS) found a relationship between saturated fat intake and blood cholesterol concentrations in different populations (actually this is a common error, because Keys’s did not find that in the SCS data – see references 27 and 42), this relationship could not be found within members of a common population. The extended goal of the NiHonSan study was to “determine the risk factors of cardiovascular disease among subjects having essentially the same genetic background, but different dietary, and possibly other, environmental patterns” (37, p.373). Thus persons of Japanese ancestry living in either Japan or California or Hawaii were studied.
The initial study published in 1973 (36) found that blood cholesterol concentrations rose progressively in the three populations of Japanese; it was lowest in Japanese living in Japan and highest in those living in California. Interestingly blood triglyceride concentrations changed in the same direction but with one ignored difference. Non-fasting blood triglyceride concentrations rose 75% across the three population groups whereas the rise in blood cholesterol concentrations was only 26%. Thus dietary and other environmental factors had a much greater effect on blood triglyceride that on blood cholesterol concentrations. The result was that Japanese living in California and Hawaii had not only higher blood cholesterol concentrations but they had markedly higher blood triglyceride concentrations. The latter is a marker of carbohydrate-sensitive hypertriglyceridemia.
The study found, not unexpectedly, that Japanese living in North America ate more total fat, more cholesterol, more animal protein, much more saturated fat and more simple carbohydrate but less total carbohydrate, less vegetable protein and less complex carbohydrate.
As a result there was a positive relationship between dietary saturated fat intake and blood cholesterol concentrations in all three populations and an inverse relationship between dietary intake of complex carbohydrates and blood cholesterol concentrations. No dietary nutrient was associated with non-fasting blood triglyceride concentrations perhaps because carbohydrate intakes in all three groups was high (>250 g/day).
Subsequent publications showed that total all-cause mortality rates and stroke rates were highest in all age groups in persons living in Japan and were lower in those Japanese living in Honolulu and San Francisco (39). The opposite pattern was present in All CHD mortality rates. These were least in those living in Japan and higher in those living in California and Hawaii.
Remarkably the very next publication in this series begins with the correct statement that “As shown by …. Worth at al. (39) there is a gradient in mortality from coronary heart disease among men of Japanese ancestry living in Japan, Hawaii and California” (40, p.491). But the paper makes no reference to the fact that all-cause mortality rates and stroke rates showed an exactly opposite pattern.
The paper reported blood cholesterol, non-fasting triglyceride, uric acid and glucose one hour after ingestion of a 50g glucose load in the three different Japanese populations. The study confirmed the earlier finding (36) that blood cholesterol and triglyceride concentration were higher in Japanese living in the US than in Japanese living in Japan, but added the new evidence that blood uric acid as well as 1-hour blood glucose concentrations matched that same trend.
Marmot et al. (41) next examined 11 900 men, aged 45-69, from this cohort to determine the rates of electrocardiographically-proven definitive CHD as well as rates of possible CHD defined as presence of a history of angina pectoris, or of pain of possible heart attack (myocardial infarction), both determined by questionnaire. They showed that the prevalence of definitive and possible CHD were least in Japanese men living in Japan; they were somewhat higher in Hawaiian Japanese and were highest in Californian Japanese.
Exactly the same trends were seen for blood cholesterol, triglyceride and blood glucose concentrations, 1-hr post-glucose ingestion (in a glucose tolerance test). All these blood values were least in Japanese living in Japan but rose substantially in those living in California or Hawaii. But the prevalence of raised blood pressures was the reverse and was highest in the Japanese living in Japan.
Figure 5 compares the prevalence of all CHD events in the three populations at three different levels of risk for the three most commonly accepted CHD “risk” factors.
Legend to figure 5. Prevalence of all CHD events at three different levels of risk for blood pressure (top panel), serum cholesterol concentrations (middle panel) and smoking (bottom panel) in persons of Japanese origin living in Japan, Hawaii or the US. Note that at any level of “risk”, there is a progressive increase in CHD events from Japanese living in Japan, to those living in Hawaii, to those living in the US. Reproduced from figures 1-3 in reference 41.
The point these graphs make is that at any level of coronary risk be it blood pressure, blood cholesterol concentration, or smoking, there is a gradient of increasing risk from Japan, to Hawaii to California. Thus the same level of a particular risk factor produces quite different outcomes depending on where in the world those of Japanese origin have spent most of their lives.
Thus 20 years after Keys had used the low rates of heart disease, low blood cholesterol concentrations and low dietary saturated fat intakes in Japan to anchor the left hand side of his iconic graph (figure 1 in reference 27), the authors would indirectly call him out for his overly simplistic interpretation.
Thus: “The very low levels of serum cholesterol in Japan, where the prevalence of CHD is very low, are consistent with the established role of hypercholesterolemia as a risk factor. However, at equivalent levels of serum cholesterol, California Japanese still have higher CHD prevalence. Similarly, at equivalent levels of blood pressure, the high California prevalence persists. To make definite statements about relationship of risk factors to disease from prevalence figures would be inappropriate, but the cautious interpretation from these simple prevalence comparisons is that other factors interact with blood pressure and cholesterol in the etiology of CHD (my added emphasis)” (41, p.523). If incidence data confirm that cholesterol, blood pressure and smoking differences do not completely explain these gradients in cardiovascular disease, then we must look elsewhere for explanations” (p. 524).
“The Japanese in America lead very different lives from the Japanese in the home country. Among other things, they eat different diets, they have different patterns of occupation and they live in a different social and cultural milieu. The relationship between the changes in these factors and CHD, both within and between study cohorts, will be the subject of later reports” (p.524).
A subsequent study of the NiHonSan populations found that the incidence of stroke, recorded as either cerebral haemorrhage (bleeding) or cerebral infarction (death of brain tissue), was about three times as high in Japan as it was in Hawaii (43). Elevated blood pressure (hypertension) was the key risk factor for stroke in both Japan and Hawaii. Yet the prevalence of hypertension was not different in Japanese living in either Hawaii or Japan so other factors must also be important.
Intriguingly that study also found that animal protein intake was inversely related to total stroke incidence in Hawaiian Japanese. The authors suggested that “animal protein and saturated fat intake which is inversely associated with stroke incidence, is much greater in Hawaii than in Japan. This explanation would support epidemiological and experimental studies in Japan which suggest that dietary animal protein and fat exert an inhibitory effect on the incidence of stroke” (43, p.15).
For completeness, another study found that higher rates of arterial hypertension in Japanese living in Japan explain their higher rates of stroke, recorded as either cerebral haemorrhage or cerebral infarction (44). Importantly low blood cholesterol concentrations increased the risk for cerebral haemorrhage but not for cerebral infarction providing a possible mechanism whereby a higher intake of animal produce might reduce risk of cerebral haemorrhage.
The clear take-away message from the NiHonSan study must be that differences in the health of Japanese living in different parts of the world cannot be explained solely by lower or higher intakes of saturated fat and lower or higher blood cholesterol concentrations. Instead the key measure of overall health – total mortality – is not different between any of these groups despite many lifestyle differences.
In short, the NiHonSan study is just another detailed investigation that has failed to support Keys’ Twin Hypotheses.
Two other studies are relevant to the Japanese Paradox.
Ueshima and colleagues from Osaka, Japan reported that their nationwide study found a significant negative correlation between the mean serum cholesterol concentration and the incidence rate of stroke, both hemorrhagic and thrombotic (45). This effect began at blood total cholesterol concentrations lower than 190mg/dL (4.9mmol/L). The authors concluded that: “Such low levels of total serum cholesterol may have derived mainly from the traditional Japanese diet, i.e., poor intake of animal fat and protein” (p.105).
“In order to prevent stroke, especially cerebral haemorrhage, it may be better to keep the level of serum cholesterol ‘not too low’. Our results suggest that the desirable level of total serum cholesterol in men would be somewhere between 180 and 200mg/dL, where incidence rates of both CHD and stroke are low” (p.105).
Not unexpectedly this blasphemy drew sharp responses from a number of Keys’ acolytes. William Kannel MD, Director of the Framingham study, made the interesting comment that “a relationship (of stroke) to lipids has never been firmly established” (46, p.106) – a reality that is seldom mentioned in the modern day when cholesterol-lowering drugs are routinely prescribed for stroke survivors.
Robert I. Levy MD of the NHLBI warned that – “Therapeutic recommendations based solely on apparent epidemiological associations are open to criticism” (47, p.107). Which is certainly true but which applies equally to Keys original study in which the Japanese data were absolutely critical for establishing his iconic linear relationship between dietary fat and CHD rates in six countries (42).
Henry Blackburn, Ancel Keys and David Jacobs argued that a key weakness of the study was because “it is difficult to conceive of mechanisms for a direct causal connection between very low cholesterol in the blood and cardiovascular disease itself” (48, p.110). In other words, the validity of an observation can be rejected if there is not already an established proof explaining what causes that observation. Yet, almost from the outset, Keys’ hypothesis (42) had been accepted by these co-authors as proven fact, even in the absence of a proven biological mechanism.
The reality is that there is a large body of evidence showing associational relationships between low blood cholesterol concentrations and a variety of medical conditions (20).
Second, a study of Japanese centenarians (49) and of other elderly Japanese followed prospectively, established that increased intakes of animal protein and animal foods such as eggs, milk, fish and meat were associated with greater longevity.
Clearly all these findings are in direct opposition to Keys’ Twin Hypotheses.
- 1974-1992. The Helsinki Businessmen Study is initiated.
Business executives living in Helsinki and born between 1919 to 1934 participated in a series of health check-ups in the late 1960s. In 1974, 1222 of these men who were clinically healthy but who had established risk factors for CHD were entered into a primary prevention trial (50). Six hundred and twelve men were randomized to the intervention group and the remainder to the control group. The study lasted 5 years during which the intervention group received multiple interventions aimed at lowering their risk for future development of CHD. In particular those in the intervention group were advised to stop smoking; to achieve a normal body weight with the adoption of a prescribed diet and to moderate their alcohol intakes. In addition those with hypertension or hypercholesterolemia were treated with appropriate prescription drugs.
The results of the intervention were first reported in detail in 1991.
References.
- Stamler J, Beard RR, Connor WE, et al. Report of Inter-Society Commission for Heart Disease Resources. Prevention of Cardiovascular Disease. Primary prevention of the atherosclerotic diseases. Circulation 1970;42:A55-A95.
- Noakes TD. Ancel Keys’ Cholesterol Con: Part 9. The Noakes Foundation.
- Keys A, Anderson JT, Mickelsen O, et al. Diet and serum cholesterol in man: Lack of effect on dietary cholesterol. J Nutr 1956;59:39-56.
- Fredrickson DS. Mutants, hyperlipoproteinaemia, and coronary artery disease. BMJ 1971;1:187-192.
- Stolley PD. The Primary Prevention of Atherosclerotic Disease: A reqport of the Inter-society Commission for Heart Disease Resources. Ann Intern Med 1972;76:661-113
- Noakes TD. The 2012 University of Cape Town Faculty of Health Sciences centenary debate: Cholesterol is not an important risk factor for heart disease, and the current dietary recommendations do more harm than good. S Afr J Clin Nutr 2015;28:19-33.
- Hjerman I. A randomized primary preventive trial in coronary heart disease: The Oslo Study. Prevent Med 1983;12:181-184.
- World Health Organization European Collaborative Group. An international controlled trial in the multifactorial prevention of coronary heart disease. Int J Epidemiol 1974;3:219-224.
- World Health Organization European Collaborative Group. Multifactorial trial in the prevention of coronary heart disease: 1. Recruitment and initial findings. Europ Heart J 1980;1:73-79.
- Moore TJ. The cholesterol myth. The Atlantic. 1989;264(September):37; Moore TJ. Heart Failure: A Critical Inquiry into American Medicine and the Revolution in Heart Care. New York, NY: Simon and Schuster, 1989.
- Noakes TD. Ancel Keys’ Cholesterol Con: Part 6. The Noakes Foundation. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-6-1960-1967
- Shekelle RB, Shryock AM, Paul O, et al. Diet, serum cholesterol, and death from coronary heart disease. The Western Electric Study. N Engl J Med 1981;304:65-70.
- Teicholz N. The Big Fat Surprise. Why butter, meat and cheese belong in a heathy diet. Simon and Schuster, New York, NY. 2014.
- Dampier C. At 100, the father of preventive medicine is still going strong – as living proof that he was right all along. Chicago Tribune December 16th 2019. https://www.chicagotribune.com/lifestyles/ct-life-100-year-old-scientist-stamler-20191226-jeprzoeqazha7nuvs2prfcndwy-story.html
- Pyorala K. Relationship of glucose tolerance and plasma insulin to the incidence of coronary heart disease: Results from two population studies in Finland. Diab Care 1979;2:131-141.
- National Collaborative Pooling Project of the American Heart Association. Circulation 42, December 1970, Revised April 1972. Cited by Reisser R; reference 48.
- The Pooling Project Research Group. Relationship of blood pressure, serum cholesterol, smoking habit, relative weight and ECG abnormalities to incidence of major coronary events: Final report of the pooling project. J Chron Dis 1978;31:201-306.
- Reisser R. Oversimplification of diet:coronary heart disease relationships and exaggerated diet recommendations. Am J Clin Nutr 1978;31:865-875.
- Garrett HE, Horning EC, Creech BG, et al. Serum cholesterol values in patients treated surgically for atherosclerosis. JAMA 1964;189:655-659.
- Noakes TD. Ancel Keys’ Cholesterol Con: Part 4. Crossfit Health. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-4-by-prof-tim-noakes
- Ahrens EH, Hirsch J, Insull W, et al. Dietary control of serum lipids in relation to atherosclerosis. JAMA 1957;164:1905-1911.
- Taubes G. Good calories bad calories. Fats, carbs, and the controversial science of diet and health. Anchor Books, New York, NY. 2008.
- Stehbens WE. Basic precepts and the lipid hypothesis of atherogenesis. Med Hypoth 1990;31:105-113.
- Stehbens WE. Causality in medical science with particular reference to heart disease and atherosclerosis. Persp Biol Med 1992;36:97-119.
- Jauhar S. Weighing benefits and risks of statins. New York Times September 5th 2000.
- Verschuren WMM, Jacobs Dr, Bloemberg BPM, et al. Serum total cholesterol and long-term coronary heart disease mortality in different cultures. Twenty-five-year follow-up of the Seven Countries Study. JAMA 1995;274:131-136.
- Noakes TD. It’s the insulin resistance, stupid: Part 7. Crossfit Essentials. https://thenoakesfoundation.org/news/blog/ancel-keys-cholesterol-con-part-7-1968-1970
- Furuta Y, Ninomiya T. [Epidemiology of stroke in Japan and comparision with the world] (article in Japanese). Nihon Rinsho 2016;74:549-553.
- Yoshiike N, Matsumura Y, Iwaya M, et al. National nutrition survey in Japan. J Epidemiol 1996;6:S189-S200.
- Noakes TD. Ancel Keys’ Cholesterol Con: Part 12. Crossfit Health.
- Sekikawa A, Miyamoto Y, Miura K, et al. Continuous decline in mortality from coronary heart disease in Japan despite a continuous and marked rise in total cholesterol: Japanese experience after the Seven Countries Study. Int J Epidemiol 2015;1614-1624.
- Ueshima H. Explanation for the Japanese Paradox: Prevention of increase in coronary heart disease and reduction in stroke. J Atheros Thromb 2007;14;278-286.
- Ikedo N, Saito E, Inove M, et al. What has made the population of Japan healthy? Lancet 2011;17-28. Published OnLine September 1, 2011.
- Sekikawa A, Curb JD, Ueshima H, et al. Marine-derived n-3 fatty acids and atherosclerosis in Japanese, Japanese-American, and White Men. J Am Coll Cardiol 2008;52:417-424.
- Tillotson JL, Kato H, Nichaman MZ, et al. Epidemiology of coronary heart disease and stroke in Japanese men living in Japan, Hawaii, and California: methodology for comparison diet. Am J Clin Nutr 1973;26:177-184.
- Kato H, Tillotson J, Nichaman MZ, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Serum lipids and diet. Am J Epidemiol 1973;97:372-385.
- Kagan A, Harris BR, Winkelstein W, et al. Epidemiologic studies of coronary heart disease and stroke in Japanese men lining in Japan, Hawaii and California: demographic, physical, dietary and biochemical characteristics. J Chron Dis 1974;27:345-364.
- Syme SL, Marmot MG, Kagan A, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Introduction. Am J Epidemiol 1975;102:477-480.
- Worth RM, Kato H, Rhoads GG, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Mortality. Am J Epidemiol 1975;102:481-490.
- Nichaman MZ, Hamilton HB, Kagan A, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Distribution of biochemical risk factors. Am J Epidemiol 1975;102:491-501.
- Marmot MG, Syme SL, Kagan A, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Prevalence of coronary and hypertensive heart disease and associated risk factors. Am J Epidemiol 1975;102:514-525.
- Noakes TD. Ancel Keys’ Cholesterol Con: Part 7. Crossfit Health. https://thenoakesfoundation.org/news/blog/ancel-keys-cholesterol-con-part-7-1968-1970
- Takeya Y, Popper JS, Shimizu Y, et al. Epidemiological studies of coronary heart disease and stroke in Japanese men living in Japan, Hawaii and California. Incidence of stroke in Japan and Hawaii. Stroke 1984;15:15-23.
- Tanaka H, Ueda Y, Hayashi M, et al. Risk factors for cerebral haemorrhage and cerebral infarction in a Japanese rural community. Stroke 1982;13: 62-73.
- Ueshima H, Iida M, Komachi Y. Is it desireable to reduce total serum cholesterol level as low as possible. Prev Med 1979;8:104-105.
- Kannel WB. In search of an optimal total serum cholesterol. Prev Med 1979;8:106-107.
- Levy RI. In search of an optimal total serum cholesterol. Prev Med 1979;8:107-109.
- Blackburn H, Keys A, Jacobs DR. In search of an optimal total serum cholesterol. Prev Med 1979;8:109-111.
- Shibata H, Nagai H, Haga H, et al. Nutrition for the Japanese elderly. Nutr Health 1992;8:165-175.
- Miettinen TA, Huttunen JK, Naukkarinen V, et al. Multifactorial primary prevention of cardiovascular disease in middle-aged men. Risk factor changes, incidence and mortality. JAMA 1985;254;2097-2102.
About the Author
Professor Tim Noakes has dedicated his life to the pursuit of knowledge and undoing the last 50 years of ‘bad’ nutritional science. His aim is to fix the future outlook of human health, by changing the way people eat and the food policies to enable the change.
Prof. Noakes has published more than 750 scientific books and articles. He has been cited more than 19 000 times in scientific literature, has an H-index of 71 and has been rated an A1 scientist by the National Research Foundation of South Africa for a second 5-year term. He has won numerous awards over the years and made himself available on many editorial boards.