“Absence of proof is not proof of absence” – Prof Noakes on Keys’ Cholesterol Con

In Part Thirteen of Professor Tim Noakes’ ‘Ancel Keys’ Cholesterol Con’ series, Noakes picked up with the invention of the Mediterranean Diet in 1993 and looked at studies that attempted to define what the Mediterranean diet is by comparing the diets of countries and communities of Northern Europeans and Mediterraneans. In Part Fourteen, the final instalment of the Cholesterol Con series, the stunningly comprehensive account of the greatest scam in the history of modern medicine continues with the findings of the Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) study reported in 1994. These results and subsequent instances detailed in the article show how the authors drew astonishing conclusions without any supporting evidence, the consequences of which remain deeply entrenched in today’s healthcare systems. – Nadya Swart

---

Ancel Keys Cholesterol Con. Part 14. 1994-2017

58. 1994. The findings of the Multinational Monitoring of Trends and Determinants in Cardiovascular Disease (MONICA) study are reported.

The MONICA study was initiated in 1974 with the goal of measuring trends in cardiovascular mortality and coronary heart disease and stroke morbidity in 21 countries and to assess the extent to which these trends could be explained by changes in known risk factors, daily living habits, health care, or major socioeconomic features measured at the same time in defined communities in those participating countries. In essence, the goal was to determine whether there truly had been a dramatic reduction in heart disease rates across the world and to determine if this reduction could be explained on the basis of changes in so-called coronary risk factors. 

Thus: “Where were the mortality changes real, and if so, how much was attributable to change in the incidence of coronary events…? Could the changes in coronary-event rates be related to population trends in known coronary risk factors of cigarette smoking, blood pressure, and serum cholesterol?” (1, p.675).

By the time the first results were reported 20 years later (1), the field had moved on significantly. In effect, the NHLBI had essentially decided that risk factor modification would reduce the risk of future cardiovascular and cerebrovascular morbidity and mortality. That is what the NHI National Consensus Conference (NCC) and the National Cholesterol Education Program (NCEP) had decreed not just to the US but also to the rest of the world. Especially with regard to the role of elevated blood cholesterol concentrations as the key driver of CHD.

The fact that the MRFIT findings and increasingly the findings from the Framingham Study did not support this conclusion was simply ignored, as is the case today. Following the NCC and the NCEP, the cholesterol stallion had long since bolted from the stable, and it was already too late in the 1990s to attempt to secure the stable door.  

The MONICA study, therefore, has real historical significance because it was initiated at a time when epidemiologists had great faith in the role of the “established” coronary risk factors, especially blood cholesterol concentrations, cigarette smoking and hypertension, as key predictors of future CHD and stroke risk (2). 

So results reported in 1994 (3) sought to explain changes in mortality statistics in different countries based on changes in the prevalence of the three most revered coronary risk factors – smoking, high blood pressure and high blood cholesterol concentrations. The null hypothesis being tested was that in these populations under study, “there is no relationship between 10-year trends in the major cardiovascular disease risk factors of serum cholesterol, blood pressure and cigarette smoking; and 10-year trends in incidence rate (fatal plus non-fatal attack rates) of CHD” (3). 

The first report of the study’s findings made the interesting observation that “mortality differences of the present study may very well be strongly influenced by genetic, cultural, climatic or other environmental factors in addition to the three risk factors considered” (3, p.512). In other words, the authors were simply re-stating the criticisms that Yerushalmy and Hilliboe (4) had raised in their critique of Keys’ original study (5). Only now, there was ample evidence to support that criticism.

The MONICA study found that whereas populations with higher blood pressure were at greater risk of CHD mortality, the same did not apply to populations with higher proportions of smokers or higher blood cholesterol concentrations. This led to the conclusion that the presence of confounders masked the true effects of the “major” CHD risk factors. In fact, according to the authors’ null hypothesis, they should have concluded that at least two of the three “major” CHD risk factors are not “major” risk factors at all. But given the global influence of the NCC and the NCEP, even if it was largely subconscious, it was no longer scientifically acceptable to suggest that conclusion. I would guess there was essentially zero possibility that an article presenting that conclusion would have been accepted for publication in 1994. 

The second report published in 2000 (1) also concluded that since the data did not agree with the certainty of the predictions from the NCC and the NCEP, it was, once more, not the theory that could possibly be wrong. The only reasonable explanation is that the data are wrong. 

Thus: “The apparent contribution of the classic risk factors in CHD over ten years…has been less precisely estimated than had been hoped” (1, p.685). 

But this is an astonishing conclusion. Where is the evidence showing this – that the classic risk factors were “less precisely estimated”? If data are missing because they were not measured, then it cannot be assumed that the evidence was there but was simply missed according to the convenient post-hoc explanation – the absence of proof is not proof of absence.

Maybe the missing evidence was not there in the first place, and so it was never missed? That is the proper conclusion the authors should have drawn. But the global influence of the NCC and the NCEP simply made it no longer acceptable to draw that conclusion.

So the authors concluded that perhaps 15% of the variance in coronary-event rates in women and 40% in men could be “explained,” by trends in the coronary risk factors so that “the larger changes in the incidence of CHD in many populations than expected from the risk-factor changes suggest that there is a broader range of interventions potentially available that may or may not be already identified” (1, p.685). 7

But the problem with those “broader range of interventions” is that they stretch beyond the limited scope of the pharmaceutical industry and so are unlikely ever to be tested in appropriate clinical trials. So they will remain forever “hidden”.   

Yet there was one interesting finding from the MONICA study that has predictably been conveniently ignored. A sub-sample of 354 men and 404 women of the Northern Sweden population of the MONICA study aged between 25-64 performed an oral glucose tolerance test to determine their levels of insulin resistance (6). The participants were then classified into four subgroups based on a grading of their insulin resistance and their fasting serum insulin concentrations. Seventeen percent of the males and 18% of the females had a combination of low insulin sensitivity and high fasting serum insulin concentrations. Persons of either sex with this combination were more likely to have higher body mass index, higher waist-hip ratios, high blood pressures, higher serum triglyceride and lower serum HDL cholesterol concentrations – all the markers of those with insulin resistance who are eating high carbohydrate diets. 

The authors concluded that this “combination of insulin resistance and high fasting insulin concentrations is associated with a marked clustering of cardiovascular risk factors and is present in one-sixth of the middle-aged population in the north of Sweden” (6, p.263). This in a population characterised by “high serum cholesterol levels, intermediate blood pressure levels, a relatively low prevalence of severe obesity and a high consumption of smokeless tobacco” (7, p.99). Yet insulin resistance could still be identified as a key risk factor amongst all these other potential confounders. 

One wonders if the WHO MONICA study might have made a greater contribution to our understanding if it had included data on diet and insulin resistance as predictors of future cardiovascular risk.

But it did not because that was not a research question allowed by those driving Keys’ Diet-Heart agenda.

59. 2003-2013. The Look AHEAD (Action for Health in Diabetes) Trial is planned.

In 2003 the Look AHEAD Research Group announced the planning of a major study of the effects of weight loss on the future risk of CHD in persons with T2DM, funded by the NIH (8). The reasoning was that whereas overweight and obesity are major contributors to both T2DM and cardiovascular disease (CVD), individuals with T2DM who are overweight or obese are at particularly high risk for CVD morbidity and mortality. The long-term consequences of intentional weight loss in overweight or obese individuals with T2DM had not been adequately examined in a large trial at that time. Thus the primary objective of the Look AHEAD clinical trial was to assess the long-term effects (up to 11.5 years) of an intensive weight loss program conducted over four years with overweight and obese individuals with T2DM. 

Approximately 5000 male and female participants with T2DM, aged 45-74 years, with a body mass index greater or equal to 25 kg/m2, were randomised into one of two groups – an intensive lifestyle intervention group or the control group. 

The goal of the intensive lifestyle intervention was to produce sustainable weight loss through decreased caloric intake and increased physical activity. The control group received diabetes support and education alone. The primary study outcome was the time of the development of the first major cardiovascular disease (CVD) event. Secondary outcomes of interest included components of CVD risk, cost and cost-effectiveness, diabetes control and complications, hospitalizations, intervention processes, and quality of life.

The first results were reported in 2014.

60. 2003-2013. The Prospective Urban Rural Epidemiology (PURE) study is initiated.

Between January 1st 2003, and March 31st 2013, the Prospective Urban Rural Epidemiology (PURE) study enrolled and finally studied 135 335 individuals aged 35-70 years living in 628 urban and rural communities in 18 countries on five continents (9). The study included three high-income (Canada, Sweden, and United Arab Emirates), 11 middle-income (Argentina, Brazil, Chile, China, Colombia, Iran, Malaysia, occupied Palestinian territory, Poland, South Africa, and Turkey) and four low-income countries (Bangladesh, India, Pakistan, and Zimbabwe). On entry to the study, participants completed standardised questionnaires that collected information about their socioeconomic status, lifestyle, health history, medication use and physical activity. Food intake was assessed using validated, country-specific food frequency questionnaires. Follow-ups occurred at 3, 6 and 9 years. 

Participants were followed for an average of 7.4 years through annual face-to-face contact or by telephone. During follow-up, the onset of any primary (total mortality and major cardiovascular events) or secondary outcomes (heart attack, stroke, cardiovascular disease mortality and non-cardiovascular disease mortality) were recorded. 

The initial goal of the study was to determine whether the burden of risk factors and the incidence of cardiovascular disease was “higher in low- and middle-income countries than in high-income countries, whether mortality after a cardiovascular event is higher in low- and middle-income countries than in high-income countries, or whether both are true” (9, p.818).

In time the study would evolve into the most thorough epidemiological evaluation of the Diet-Heart and Lipid hypotheses ever undertaken, dwarfing everything that had gone before and most especially Keys’ error-ridden Seven Countries Study (5,10). 

Like the Seven Countries Study, it would search for relationships between dietary fat and carbohydrate intakes in these populations and their total mortality and rates of cardiovascular disease events.  

The first results would be reported in 2017.

61. 2005. The inconvenient findings of the Women’s Health Initiative Randomised Controlled Dietary Modification Trial (WHIRCDMT) are conveniently hidden since they disprove Keys’ Diet-Heart hypothesis.   

The 8-year-long WHIRCDMT initiated in 1993 found that eating according to the USDA Dietary Goals did not reduce the risk of cancers of either the colo-rectum (11) or breast (12). This is predictable if both cancers are associated with insulin resistance (IR) and T2DM (13), which would be worsened by low-fat, high-carbohydrate diets (or perhaps those rich in omega-6 polyunsaturated fats (14) of the type promoted by the 1977 USDA Dietary Goal.  

Karin Michels and Walter Willett from Harvard Medical School postulated that preventing weight gain would be perhaps the best approach to prevent these cancers (15).  

Given that high-carbohydrate diets cause weight gain in persons with IR, it would have been more logical for Michaels and Willett to postulate that high-carbohydrate diets explain the co-existence of obesity with colorectal and breast cancers in persons with IR. 

But the 1984 National Consensus Conference, followed by the 1987 National Cholesterol Education Program that vilified saturated fat, declaring sugar and carbohydrates as the healthiest foods (16), ensured that that conclusion could never be expressed publicly. Especially not by the principal directors of the study, including Professor Jacques Rossouw, who were all employees of the NHI. The special contribution of Rossouw to the WHIRCDMT debacle will be exposed shortly. 

With regard to the effects of the prudent low-fat diet on cardiovascular outcomes, the major reported finding of the WHIRCDMT was that: “… a reduced total fat intake and increased intake of vegetables, fruits, and grains did not significantly reduce the risk of CHD, stroke, or CVD in postmenopausal women and achieved only modest effects on CVD risk factors” (17, p.655). As I will present shortly, the truth is that this study, published long after the USDA Dietary Guidelines were first released, found that the low-fat “heart healthy” eating guidelines harmed the health of these postmenopausal women. The concerns expressed by the experts in 1977, detailed previously, had all come to pass just as they had warned (18).  

In fact, the only significant finding in that study had apparently “escaped” the attention of the authors until quite recently, when it was finally brought to their attention (19). That evidence is presented on the seventh page of the published manuscript, where the following is stated: “The HR for the 3.4% of women with CVD at baseline was 1.26 (95% CI, 1.03-1.54)” (17, p.661). 

Properly interpreted, this finding indicates that women with established heart disease at the start of the trial had a 26% increased risk of developing further cardiac complications if they adopted the diet that followed the USDA Dietary Goals. By showing that postmenopausal women with heart disease were at lower risk of developing subsequent cardiac complications if they continued to eat more fat and fewer vegetables, fruits and grains, the study essentially disproves the diet-heart hypothesis. How can a diet designed to prevent heart disease be associated with a worsening of the condition in those who are the most vulnerable because they already have the disease? 

As I have described previously in more detail (19), this finding was not discussed further in the abstract, the discussion or the conclusions of that paper. In addition, a key line of text was “inadvertently” missing from the most important Table in the text. 

When challenged to explain these errors and omissions (19), the authors (20) dismissed the only significant finding in their study as “likely to be a chance finding” because “there is no biological basis for expecting a different outcome in this (ill) subgroup, as shown in cholesterol-lowering trials of women with prior disease” (p.882).  

Thus an inconvenient outcome that the authors heartily disliked was ignored because of their certainty that this adverse result has no (currently known) biological basis. But this explanation is unacceptable (21).

For example, the authors failed to reference the Estrogenic Replacement and Atherosclerosis (ERA) Trial, which found that coronary atherosclerosis progressed significantly more rapidly over a 3-year period in postmenopausal women eating the equivalent of the WHIRCDMT low-fat “prudent” diet (22, Figure 1 in 23). Higher carbohydrate intakes accelerated coronary artery disease progression, as did the substitution of dietary saturated fats with polyunsaturated fats.  

In contrast, postmenopausal women who ate the most saturated fat (and the least carbohydrates and polyunsaturated fats) showed no progression of coronary atherosclerosis, even though that group included a significantly higher proportion of current smokers. As expected, women eating the most saturated fat also had significantly higher serum HDL-cholesterol and lower serum triglyceride concentrations and lower Total cholesterol: HDL cholesterol ratios.

These findings, the subject of an accompanying Editorial (24), predict that the clinical manifestation of coronary heart disease should increase in participants in the WHIRCDMT eating the low-fat “prudent” diet that elevates blood triglyceride and lowers HDL-cholesterol concentrations. And this is what seems to have happened in that study.

The authors of the editorial reviewing the ERA study were fascinated by its paradoxical finding that “a high-fat, high-saturated fat diet is associated with diminished coronary artery disease progression in women with the metabolic syndrome, a condition that is epidemic in the United States. This paradox presents a challenge…”. (24, p.1103). They concluded that the study should be “hypothesis-generating”. 

Unfortunately, because of their dependence on the senior authors of these articles on the NIH (Rossouw) and the “vegetable” oil industry (Mozaffarian), neither of these scientists would ever be prepared to acknowledge that paradox. Nor, as heavily funded scientists supposedly searching for the truth, did they apparently appreciate their responsibility to generate novel hypotheses to unravel the paradox. Instead, they continue to pretend that the paradox exposed by their studies does not exist.

So when the WHIRCDMT reported their “paradoxical” finding (17), the responsibility of Rossouw and his co-workers was to explain why the conclusions from the ERA study were not relevant to their discovery. Instead, they ignored that research, choosing rather to advance their deceptive “biologically implausible” argument.  

Eminently plausible biological explanations for this inconvenient finding in the WHIRCDMT would include the favourable changes in blood HDL-cholesterol and triglyceride concentrations measured in the ERA trial, together with the evidence that a high-fat, low-carbohydrate diet reduces the blood concentration of small, dense LDL-cholesterol particles (25,26) which, when oxidised (27) or glycated (especially in those with T2DM) (28), are considered particularly atherogenic (29-34). 

Importantly for Professor Rossouw and his colleagues to consider is the evidence that low-fat, especially very-low-fat diets, increase the blood concentrations of small, dense LDL particles, the so-called Pattern B (35-38)) (figure 1), which is associated with increased risk of CHD. Conversely, a higher fat diet increases the percentage of subjects with a greater proportion of larger, less dense LDL particles (Pattern A). Pattern B is also associated with lower blood HDL-cholesterol and higher triglyceride concentrations (37,38; Figure 2). So Pattern B is the “bad” lipoprotein pattern, whereas Pattern A is considered “good” or healthy.

Legend to figure 1. There is an inverse linear relationship between the percentage of fat in the diet and the percentage of subjects eating that diet with a greater proportion of larger, less-dense LDL particles, the healthier Pattern A lipoprotein profile. Reproduced from figure 4 in reference 38. 

Figure 1 shows that as the percentage of fat in the diet is reduced from 50% to 10%, the percentage of subjects with the pathogenic Pattern B LDL-lipoprotein particle pattern (many small, “dense”, LDL-particles) increases, reaching >50% in those eating a diet with less than 20% fat. In contrast, less than 15% of those eating a 50% fat diet show that abnormality. Instead, they exhibit the healthy Pattern A profile (right side of figure 1).

Figure 2 shows that in persons who transition from a low- (20-24%) to a very low (10%) fat diet, those who maintain or increase the size of their LDL-particles – that is, they have a stable Pattern A lipoprotein profile – also maintain low blood triglyceride concentrations (left side of the graph). In contrast, those who increase their number of smaller, “denser” LDL particles and transition to the “bad” Pattern B lipoprotein profile also show marked increases in their blood triglyceride concentrations (right side of the graph).

Legend to figure 2. There is an inverse linear relationship between the change in LDL-particle size and the increase in blood triglyceride concentrations in people who are unable to maintain a stable Pattern A lipoprotein profile when they transition from a very low (10%) to a low- (20-24%) fat diet. Reproduced from figure 3 in reference 38.

For some reason, whilst this evidence was clearly available to the WHIRCDMT researchers when they tried to explain their unhappy findings, they simply ignored it. Importantly the senior author, Rossouw, believes that the condition of insulin resistance barely exists and its role in affecting human health can safely be ignored (39, p.71-72).

The WIRCDMT also found that although the higher carbohydrate intake of the intervention diet did not influence blood glucose control in women without diabetes, it caused a progressive worsening of control in those with T2DM (40). This finding “agrees with some, but not all, previous studies evaluating the effects of high- and low-carbohydrate diets in persons with diabetes” (40, p.83), forcing the authors to conclude that “caution should be exercised in recommending a reduction in overall dietary fat in women with diabetes unless accompanied by additional recommendations to guide carbohydrate intake” (40, p.84). That diets with a high glycemic load are associated with increased risk for the development of T2DM is well established but largely ignored (41-44).

In truth, the authors of these WHIRCDMT studies should have stated what is obviously true: Their findings indicate that persons with established heart disease or type 2 diabetes mellitus (T2DM) should be mandated to eat a higher, not lower, fat diet to limit the further progression of their diseases.  

But the “paradoxical” findings of the WHIRCDMT did not end there. 

First, the study found that postmenopausal women who took statins to lower their blood cholesterol concentrations were at a 39% increased risk of developing T2DM (45). 

Second, these women did not enjoy any long-term cardiovascular benefits from using statins (46). 

Neither of these points is ever mentioned by the proponents of the low-fat “heart-healthy” diet, like Professor Rossouw (47,48), since it so obviously disproves the lipid hypothesis, on the truth of which their academic credibility is absolutely dependent. 

Third, the postmenopausal women eating the low-fat ‘prudent’ diet gained more weight during the study regardless of whether they began the trial with a normal weight or were overweight or obese (49). The opposite occurred in those women who continued to eat more dietary fat. The authors concluded: “These findings suggest that a low-fat diet may promote weight gain whereas a reduced-carbohydrate diet may decrease the risk of postmenopausal weight gain” (49, p.1189).

As a result: “Consuming a reduced-carbohydrate diet, with moderate fat and high protein intake, may decrease the risk of weight gain in postmenopausal women. However, prevailing dietary recommendations call for limiting fat intake to promote optimal health and prevent chronic disease. Our findings, therefore, challenge prevailing dietary recommendations, suggesting instead that a low-fat (diet) may promote rather than prevent weight gain after menopause” (49, p.1196).

So, in the end, the NHI could have done much better if they had ignored the biased advice of the National Consensus Conference and the National Cholesterol Education Program and simply chosen to study a high-fat dietary intervention as part of the WHIRCDMT. Who knows, the findings might have been so remarkable that its authors might perhaps have won a coveted Nobel Prize.

In summary, the WHIRCDMT has definitively established that eating according to the USDA Dietary Goals is associated with an increased risk for the development of complications of heart disease and of T2DM. In addition, the diet is of no benefit in terms of weight control since it increases weight gain. These findings from the most expensive low-fat diet RCT ever undertaken should have destroyed any remaining credibility of Keys’ Diet-Heart hypothesis.

Even more importantly, by showing that harm was caused to those who reduced their dietary saturated fat intakes, the study establishes that it is, from now and forever henceforth, unethical to advise anyone to remove saturated fat from the diet. It is unethical because this advice is now known to cause harm. And the first ethical rule in medical practice is: First, do no harm.

The WHIRCDMT Project Director, Professor Jacques Rossouw, would explain away all these inconvenient findings with the following: “This study shows that just reducing total fat intake does not go far enough to have an impact on heart disease risk. While the participants’ overall change in LDL “bad” cholesterol was small, we saw trends towards greater reductions in cholesterol and heart disease risk in women eating less saturated and trans-fat” (50). 

The then Director of the NHLBI, Elizabeth G. Nabel, confirmed that she too has a limited understanding of the proper scientific method – the one that aims to discover the truth, not the one designed to bolster an already decided outcome: “The results of this study do not change established recommendations on disease prevention. Women should continue to…work with their doctors to reduce their risks for heart disease including following a diet low in saturated fat, trans fat and cholesterol” (50).

No, Professors Nabel and Rossouw. The WHIRCDMT proved the (null) hypothesis that replacing dietary saturated fat with PUFAs will have no beneficial and only adverse effect on CHD outcomes.  

More is the pity that your medical training did not teach you that if the null hypothesis is proven, it remains the “truth” until it is disproven by new evidence from another experiment testing a novel null hypothesis.

But following the path already well-worn by Keys in ignoring the problematic evidence from the Seven Countries Study, of Franz in ignoring the inconvenient evidence from the Minnesota Coronary Experiment, Dawber in ignoring the bothersome evidence from the Framingham Heart Study; of Stamler in ignoring the awkward evidence from the MRFIT, and others yet to be exposed in these columns, Nabel and Rossouw simply thought that by ignored all the annoying evidence that disagreed with their unyielding biases, that evidence would somehow disappear. 

But both are guilty of the wretched scientific crime described by Albert Einstein: “The right to search for the truth also implies a duty one must not conceal any part of what one has recognized to be true”.    

But even more unforgivably, the WHIRCDMT authors successfully buried the evidence that the prescription of a low-fat diet to those postmenopausal women who began the trial with established heart disease or with type 2 diabetes mellitus (T2DM), caused real harm as both conditions worsened in those eating the intervention low-fat diet (17,40). A rather critical finding that the authors still refuse to acknowledge more than a decade later and have done nothing (20) to rectify the errors (21) identified in their published manuscript (17).

So in the end, Rossouw, another modern Keys acolyte, simply ignored the negative findings and continued to promote the low fat “heart-healthy” “prudent” and “in moderation” dietary mantra (47,48) despite the mountains of contrary evidence (39,51). It is as if their proof of the null hypothesis never happened. And as if there was no evidence that removing saturated fat from the diet causes direct and measurable harm. 

But Karma lives on. Eventually, with the help of Dr Andreas Eenfeldt, we uncovered and were able to expose the hidden data in the WHIRCDMT trial (19). 

In summary, the key conclusion from the WHIRCDMT is that, forever more, it is unethical to advise persons to remove saturated fats from their diets. 

Perhaps, after all, this $700 million effort was worth it.

62. 2008. Publication of a study comparing Weight Loss on Low-Carbohydrate, Mediterranean, or Low-Fat Diets.

2008 saw the publication of one of the most important studies in modern nutrition literature (52). But because the study was relatively modest, not a colossal intervention trial like the WHIRCDMT or the MRFIT, it has largely escaped the attention of the nutrition sciences. Nina Teicholz is responsible for identifying the essential importance of this study (53, p.309-311).

For this two-year trial, 322 moderately obese subjects, with an average age of 52 years, were randomly assigned to one of three diets: 

1. Low-fat, restricted-calorie – 30% of calories from fat with 10% of calories from saturated fat; 300mg cholesterol per day.
2. Mediterranean, restricted-calorie – “rich in vegetables and low in red meat, with poultry and fish replacing beef and lamb”; no more than 35% of calories from fat, mainly from olive oil and nuts. This diet is the Willett “Mediterranean” diet (54,55). 
3. Low-carbohydrate, non-restricted-calorie – 25 g of carbohydrate/day for the first 2 months with a gradual increase to 120g/day. Intake of total calories, protein, and fat was not limited. This diet is based on the Atkins low-carbohydrate diet (56).  

Importantly, to ensure dietary compliance, subjects received eighteen 90-minute sessions with dietitians over the two years of the trial. Although all participants ate in the same canteen, the food options available to the three different groups were clearly identified with colour coding and full descriptions of their calorie and macronutrient contents.  

The rate of adherence to the different diets was exceptional – 95% at one year and 85% at two years. The most important results are depicted graphically in Figures 3-5.

Legend to figure 3. Changes in body weight over two years in the three diet groups in the study of Shai et al. (52). Reproduced from figure 2 in reference 52.

Figure 3 shows that weight loss in all three groups was greatest during the first 4-6 months of the trial, with some gradual weight regain thereafter. But weight loss was greatest in the low-carbohydrate group (blue line) and least in the low-fat group (red line). 

Recall that the low-carbohydrate group was advised to increase their carbohydrate intake after two months, and part of the weight regain in that group would have resulted from that change. Weight loss in the low-carbohydrate and Mediterranean groups was significantly greater than in the low-fat group.

Figure 4 shows changes in relevant blood lipid biomarkers during the early and late phases of the trial in the three different diet groups.

Legend to figure 4. Changes in blood lipid profiles during the early and late phases of the two-year trial in the three diet groups studied by Shai et al. (51). Reproduced from figure 3 in reference 52.

The top left panel of figure 4 shows that HDL cholesterol increased more in the low-carbohydrate group than in the other two groups; this difference was greatest after the first six months of the study when the low-carbohydrate group was eating the fewest carbohydrates.

Blood triglyceride concentrations (top right panel) fell the most in the low-carbohydrate group in the first six months of the trial; after two years, the reduction was equal in the low-carbohydrate and Mediterranean groups, but blood triglyceride concentrations were essentially unchanged in the low-fat group. 

LDL-cholesterol concentrations (bottom left panel) were unchanged in the low-fat group but fell marginally in the other two groups. As a result, the ratio of total cholesterol to HDL-cholesterol (bottom right panel) fell the most in the low-carbohydrate group and the least in the low-fat group, with the Mediterranean diet falling in between. This finding is paradoxical since all the evidence from short-term feeding studies is that low-fat diets cause a reduction in blood LDL-cholesterol concentrations. Predictably this finding is not ever discussed.

Figure 5 shows changes in other biomarkers in the three diet groups during the early and late phases of the trial.

Legend to figure 5. Changes in additional blood biomarkers during the early and late phases of the two-year trial in the three diet groups studied by Shai et al. (52). Reproduced from figure 4 in reference 52.

The top left panel in figure 5 shows that high-sensitivity C-reactive protein concentrations, a marker of whole-body inflammation, fell the most in the low-carbohydrate group and least in the low-fat group, but differences were not significant between groups. Similarly, adiponectin, a marker of improved metabolic health, increased most in the low-carbohydrate group, although the response of the different groups was not significantly different (top middle panel). Blood leptin concentration also fell equally in all groups, with most change occurring in the first six months.

Fasting blood glucose concentrations in non-diabetic (T2DM) subjects were not influenced by diet (left half of bottom left panel), but in patients with T2DM (right half of bottom left panel), fasting blood glucose concentrations increased in those eating the low-fat diet; were unchanged in those eating the low-carbohydrate diet but fell significantly in those on the Mediterranean diet. 

In those without T2DM, fasting insulin concentrations fell most in the low-carbohydrate group (left half of the middle panel). In those with T2DM (right half of middle panel), the reduction was greatest in those on the Mediterranean diet. 

Change in HOMA-IR (Homeostatic Model Assessment – Estimated Insulin Resistance), an indirect measure of an individual’s level of insulin resistance (right half of bottom right panel), were slightly more in those without diabetes in response to the low-carbohydrate diet (left half of the bottom right panel), but in diabetic patients, the improvement was greatest in those eating the Mediterranean diet (right half of the bottom right panel)). 

The authors concluded that the “Mediterranean and low-carbohydrate diets might be effective alternatives to low-fat diets. The more favourable effects on blood lipids (with the low-carbohydrate diet) and glycemic control (with the Mediterranean diet) suggest that personal preferences and metabolic considerations might inform individualised tailoring of dietary interventions” (p.229). 

This study established two irrefutable facts. 

First, the AHA-prescribed low-fat “heart-healthy” diet performed the worst of the three tested diets and was substantially worse than the Atkins low-carbohydrate diet. This has since been shown repeatedly, yet even today, it is as if the evidence simply does not exist.

This is important since, in defending their thoroughly discredited diet, the AHA has wasted much effort vilifying (57-59) the Atkins low-carbohydrate diet, which we now know to be more effective.

Second, the Mediterranean diet was no better and, in some cases, not as effective as the Atkins diet. This key fact has been rigorously suppressed (53, p.213-215). 

Thus Teicholz notes that: “Indeed, what (senior author Meir Stampfer MD) doesn’t like to advertise, and what the study report doesn’t emphasise, was the notable success of the third arm of the trial. This was the group eating a low-carbohydrate diet, relatively high in fat. The participants on this diet, it turned out, looked the healthiest of all. They lost even more weight (12 pounds), and their heart biomarkers looked even better: their triglycerides were lower, and their HDL-cholesterol was much higher than the other two groups. Only LDL-cholesterol looked better for the Mediterranean dieters, yet this biomarker has proven less reliable than previously thought. Therefore, although the finding has received no attention, there’s really no doubt that the low-carb diet performed better than both the low-fat and the Mediterranean diets” (53, p.214). 

But Stampfer could never acknowledge this since it conflicts so absolutely with the dogmatic position of himself, his colleagues and his employers at the Harvard School of Public Health, who rely unbendingly on epidemiological associational data as definitive “proof” that diets high in fat increase CHD risk (60). A randomised controlled trial that disproves their dogma will always be ignored. 

The key conclusion that Teicholz draws is that when the PREDIMED study (61) subsequently found that those eating a “Mediterranean” diet had fewer cardiovascular (actually cerebrovascular) events over five years than did those eating a standard low-fat AHA diet (that is not the standard Spanish diet because it was low in eggs, nuts, fatty fish, oils and high-fat foods of all kinds (53, p.215; 55)), the global conclusion was that finally, the Mediterranean diet had proved itself to be the ultimate diet for the prevention of CHD. 

But there were significant problems with the original publication of the PREDIMED study, so it has since been retracted and republished (62). The problems, not all of which were properly addressed in the retraction, included the following:

• The hubristic claim that the diet “reduced the incidence of major cardiovascular events” (61, p.1279) was based on the reduction in only one class of such events – strokes – and this occurred only in the first year of the trial. So that “when events occurring in the first year were excluded in a supplementary sensitivity analysis, this difference was no longer significant” (63, p. 674). In effect, the authors exaggerated the essentially non-existent benefits of the diet, just as did those involved in the Lipid Research Clinics Coronary Primary Prevention Trial (LRC CPPT) described in detail earlier (64). Recall that the authors of that study used the failed results of their LRC CPPT as the essential propaganda driving the 1987 National Cholesterol Education Program. Clearly, not much has changed in the past 30 years.
• Not every subject in the trial had been randomised individually to one of the three study groups. Instead, the study enrolled all members of the same households on the same dietary intervention instead of randomising individuals in the same household to different dietary interventions. The same applied to subjects at different study sites (65). 
• The personal characteristics of those “randomised” to the control group were not well matched to those “randomised” to the Mediterranean diet groups supplementing with either nuts or virgin oil. Thus the control group had significantly higher percentages of women; those with obesity; those using diuretic medications; and those using oral diabetic drugs (66).
• The control group was not eating the standard Spanish diet, as would be required if the group was truly a control group. Instead, they had been prescribed a low-fat diet “because that diet had been the international standard for so many decades. This low-fat group was advised to avoid eggs, nuts, fatty fish, oils, and high-fat foods of all kinds” (53, p.215). In other words, the control group was advised to follow the low-fat diet tested in the Women’s Health Initiative Randomised Controlled Dietary Modification Trial (WHIRCDMT). And the diet tested in the WHIRCDMT had worsened health outcomes in those with established heart disease or diabetes (55). Thus the control group was forced to eat a diet known to worsen cardiovascular outcomes. Even when measured against this inferior diet, the intervention “Mediterranean” diet achieved essentially no health benefits.
• But more fatally, for the first three years of the trial, the control group were essentially left to their own devices, having received only a leaflet summarising the recommendations to follow a low-fat diet. In contrast, the two dietary intervention groups – receiving the Mediterranean diet with either added nuts or extra-virgin olive oil – received both supplementary foods and regular counselling. After three years of the four-and-a-half year study, the researchers suddenly realised that “such a low-grade intervention might potentially represent a weakness of the trial and amended the protocol to include quarterly individual and group sessions with the delivery of food descriptions shopping lists, meal plans and recipes (adapted to the low-fat diet) in such a way that the intensity of the intervention was similar to that of the Mediterranean diet groups, except for the provision of supplemental foods for free” (61, Supplementary material p.10; 67, p.6). Recall that the sole differences in the trial occurred during this period when the control group was not receiving any support.

Teicholz concludes that all the PREDIMED study found was that the Mediterranean diet “was better than the low-fat diet”. Yet this was not a particularly important finding since the WHIRCDMT had already established that the low-fat diet increases, but it does not reduce CHD risk (55).

She concludes that one is left with a single clear possibility: The Mediterranean diet outperformed the standard AHA low-fat diet perhaps “simply because it delivered more dietary fat” since “the sole difference between the low-fat and the Mediterranean groups was the number of nuts and olive oil that they ate” (52, p.215). 

Teicholz continues: “It is perfectly possible that any national diet would look better when compared to the low-fat diet. Perhaps the traditional Chilean or Dutch diet, for example – or that of any country eating unrefined, traditional foods – would show fewer cardiovascular events compared to a diet low in fat. We don’t know because such experiments have not been done. Only the Mediterranean diet has been studied so thoroughly. It has monopolised the scientific landscape, with its many days in the Mediterranean diet” (53, p.215-216).

But what if the PREDIMED study had, like the Shai/Stampfer Study (52), included a fourth group, a group eating the Atkins higher-fat diet? 

Is it not possible that that diet group would have enjoyed a convincingly better outcome as was found in the Shai/Stampfer study (52)? 

The key evidence from the Shai/Stampfer study predicting a superior long-term outcome in the low-carbohydrate diet group was the superior total cholesterol to HDL-cholesterol ratio, the superior HDL-cholesterol to triglyceride ratio (not reported), and the substantially lower high sensitivity C-reactive protein concentrations. 

63. 2013.The Recovered Sydney Diet Heart Study is published.

The original findings of the Sydney Diet Heart Study (SDHS), which began in 1966, were reported in 1978 (68) as the following: “Survival was slightly better in the second (dietary intervention) group. Multivariate analysis showed that none of the dietary factors was significantly related to survival. The prognosis was determined largely by the extent of the coronary and myocardial disease as judged by the usual clinical parameters. Recreational physical activity strongly influenced survival when all other factors were kept constant” (68, p.115). 

But these conclusions were largely false as survival was clearly significantly superior in the group that continued to eat their normal diets and did not replace dietary saturated fats with linoleic acid from safflower oil and safflower oil polyunsaturated margarine (figure 6).

Legend to Figure 6: Five-year survival curves in control and experimental groups in the Sydney Diet Heart Study (SDHS) as reported in the original publication (68). Note that survival was significantly worse in the experimental group who replaced dietary saturated fat with omega-6 linoleic acid from safflower oil and safflower oil polyunsaturated margarine. 

In about 2010, Christopher Ramsden and Daisy Zamora of the National Institute on Alcohol Abuse and Alcoholism at the National Institutes of Health in Bethesda, USA, contacted Dr Boonseng Leelarthaepin, then retired but one of the investigators in the original SDHS (68). They requested and obtained permission to “recover, analyse, and interpret” the original SDHS data, which had been stored on a 9-track magnetic tape. Ramsden and his team converted those data into a usable format. To ensure accuracy, only variables that exactly matched published data were included after they had been verified by Dr Leelarthaepin. 

When the original data were “recovered” and subjected to independent analysis (69), a rather different result emerged. Thus: “The intervention group had (significantly) higher rates of death than controls” (69, p.1). Higher death rates occurred in all important categories – all causes, cardiovascular disease, and coronary heart disease (Figure 7) – in those in the experimental group. 

Legend to Figure 7: Data from the Recovered Sydney Diet Heart (RSDH) Study (69) show that cumulative death rates for all-cause (top panel), cardiovascular disease (middle panel) and coronary heart disease (bottom panel) were increased in the intervention group who replaced dietary saturated fat with the PUFA, linoleic acid. Note that the effect was apparent within a short time after the commencement of the dietary intervention. Redrawn from figure 2 in reference 69. 

Furthermore, when the authors included these new findings in a meta-analysis of studies reporting the effects of PUFA substitution for dietary saturated fats, they found no evidence of benefit. They concluded: “These findings could have important implications for worldwide dietary advice to substitute omega six linoleic acid, or polyunsaturated fats in general, for saturated fats” (69, p.1).

In reality, these findings simply confirm the conclusion from the WHIRCDMT. It is unethical to advise anyone to remove saturated fat from the diet.

Because the substitution of polyunsaturated fatty acids for saturated fats causes harm. 

Perhaps it is particularly interesting that the original investigators did not detect the overt evidence of harm shown in their original figure 6. My guess is that the pressure on them to find in favour of the substitution PUFA-rich diet was simply too great. 

They likely feared for their future careers should they come up with a study that disproved the Dietary Guidelines. 

64. 2014. The failed Look AHEAD trial is terminated prematurely.

The aim of the Look AHEAD trial (70) initiated in 2003 was to determine whether an intensive weight-loss intervention would improve health outcomes in those with obesity and T2DM. The nature of the weight-loss intervention was conventional. The goal was to produce a caloric deficit by increasing calorie expenditure and reducing calorie intake: “The intensive lifestyle intervention is designed to achieve and maintain weight loss through decreased caloric intake and increased physical activity” (70, p.610),

In October 2012, the planned 11.5-year study was terminated as “futile” after 9.6 years when it was established that these interventions were no more effective in slowing the progression of arterial damage than doing nothing (71,72).  

This confirms that T2DM will not be beaten by encouraging persons with T2DM to exercise more and to eat a diet rich in blood glucose and insulin-raising, obesogenic carbohydrates.

Fortunately, the Virta Health study, to be discussed subsequently, has finally succeeded where all other interventions for persons with T2DM have failed.

Remarkably, few ever refer to the failure of this definitive trial of the Calories In – Calories Out Model of weight loss. It simply confirmed the findings of the WHIRCDMT (49). 

65. 2016. The results of the Recovered Minnesota Coronary Experiment are published.

Buoyed by their success in re-analysing the SDHS study, researchers Ramsden and Zamora were handed another remarkable opportunity in the mid-2010s when the original data from the Minnesota Coronary Experiment (73) was gifted to them by Ivan Franz III MD, the son of Ivan Franz II MD, the lead investigator of the original MCE study. Whilst working through his recently deceased father’s academic papers, Franz III discovered the original MCE data files. Aware of the successful work of Ramsden and Zamora in their analysis of the recovered SDHS, Franz would have realised that Ramsden and Zamora were best placed to apply modern analytical techniques to his father’s data. 

Exactly as the Recovered SDHS study had found, the recovered MCE data (74) confirmed that removing saturated fat from the diet caused harm. 

Thus as described earlier (Figure 10 in 74; reproduced here as figure 8), no age group benefitted from the intervention. However, survival was significantly worsened in those over 65 who were placed on the “heart-healthy” prudent diet. In addition, lowering the blood cholesterol level with polyunsaturated fats proved harmful as “there was a 22% higher risk of death for each 30mg/dL (0.78mmol/L) reduction in serum cholesterol” (74, p.1).

Legend to figure 8: The Recovered Minnesota Coronary Experiment showed that survival of persons over 65 years of age was worse in the intervention diet group that replaced dietary saturated fat with “vegetable oils rich in linoleic acid” (74, p.1). Reproduced from figure 5 in reference 74.

Thus the evidence from the WHIRCDMT, the Recovered MCE and the Recovered SDHS are unequivocal. Removing saturated fat from the diet causes harm to both men and women.

It is, therefore, unethical to advise anyone to follow this dietary practice.

Perhaps even more compelling is that, had the results of this study been honestly analysed and reported in 1976 as they should have been, it would have been extremely difficult for the McGovern Senate Committee (18) to continue with its crusade to convert the world to eat a diet “rich” in cereals and grains and processed “vegetable” oils.

Perhaps if Dr Franz had understood the nature of science and the importance of hypothesis testing rather than performing studies to confirm one’s personal biases, the world might have been spared the resulting catastrophe of the obesity/T2DM epidemic, caused directly by the dietary guidelines produced by McGovern’s Senate Committee. 

66. 2017. A report of the association of food consumption, blood cholesterol concentrations and cardiovascular disease in 42 European countries 

A group of scientists from the Czech Republic used international statistics to search for associational relationships, which cannot be assumed to be causal, between nutritional factors and the prevalence of cardiovascular disease in 42 European countries. The mortality data were derived from the European Cardiovascular Disease Statistics, whereas the nutritional information came from the FAOSTAT website. The FAOSTAT reports the “total quantity of foodstuffs produced in a country added to the total quantity imported and adjusted to any change in stocks that may have occurred during the reference period” (76, p.2-3). The more interesting findings are depicted graphically in figure 9. 

Legend to figure 9. Associational relationships between markers of cardiovascular disease and different national consumption data for dietary fats, animal protein and carbohydrates in 42 European countries. 

Reproduced from graphical data in reference 76. 

The findings represented in the different panels (A-F) in figure 9 shows the following:

1. An inverse relationship between total dietary and animal fat intake and CVD mortality in women.
2. A linear relationship between %energy intake from carbohydrates and alcohol and CVD mortality in women.
3. An inverse relationship between CVD mortality and prevalence of raised blood cholesterol concentrations in women.
4. A linear relationship between animal fat and animal protein intake and prevalence of raised blood cholesterol concentrations in men.
5. An inverse relationship between % energy intake from potato and cereal and prevalence of raised blood cholesterol concentrations in men.
6. An inverse relationship between the prevalence of raised blood cholesterol concentrations and the prevalence of raised blood cholesterol concentrations in women. 

Importantly, identical relationships were found for men and women for all these different variables. For convenience, I have chosen to represent only one of the sexes to illustrate the different relationships that were detected. But identical relationships were found for both sexes.

The overall conclusion from these associational relationships (which cannot prove causation) might be that eating more animal fat and protein, and less carbohydrate is associated with high blood cholesterol concentrations and a lower prevalence of hypertension and cardiovascular disease. 

Whilst associational studies cannot prove causation, they can perhaps show what is unlikely to be true. Which, in this case, are Keys’ Twin Hypotheses.

The clearest evidence refuting Keys’ Lipid Hypothesis is the data shown in figure 10. That figure shows that in both men and women, there is an inverse relationship between actual CVD mortality rates in the 42 countries and the prevalence of raised blood cholesterol concentrations in those countries. 

Legend to figure 10. There is an inverse associational relationship between actual CVD mortality and the prevalence of raised blood cholesterol concentrations in both men (left panel) and women (right panel) in 42 European countries. Reproduced from figure 4 in reference 76 and from supplementary data provided with that publication. Note that the left panel has been drawn from the supplementary data. A line of identity, drawn by eye, has been added without statistical analysis or confirmation since that figure was (surprisingly) not included by the authors in their original publication. The statistical data analysis for the right panel is included. They show that this relationship is statistically significant.  

Taken together, figures 9 and 10 show that countries with a higher prevalence of elevated blood cholesterol concentrations have lower CVD mortality rates, the exact converse of Keys’ Lipid Hypothesis. 

Furthermore, the finding that blood cholesterol concentrations are raised in those eating more animal fat and protein (Panel D in figure 9) reverses the foundation pillar on which Keys’ Diet-Heart hypothesis is based. 

These data suggest that eating more animal fat and protein protects against CVD mortality (even though it may raise blood cholesterol concentrations). 

So how is it possible, according to Keys’ Twin hypotheses, for countries with the highest prevalence of elevated blood cholesterol concentrations to have the lowest rates of CHD? It is simply inexplicable, according to the Keys model. 

The authors’ conclusion was: “Our results do not support the association between cardiovascular disease (CVDs) and saturated fat, which is still contained in official dietary guidelines. Instead, they argue with data accumulated from recent studies that link CVD risk with the high glycemic index/load of carbohydrate-based diets. In the absence of scientific evidence connecting saturated fat with CVDs, these findings show that current dietary recommendations regarding CVDs should be seriously reconsidered” (76, p.1).

References:

1. Kuulasmaa K, Tunstall-Pedoe H, Dobson A, et al. Estimation of contribution of changes in classic risk factors to trends in coronary-event rates across the WHO MONICA Project populations. Lancet 2000;355: 675-687.
2. Pajak A. Geographical variation in the major risk factors of coronary heart disease in men and women aged 35-64 years. The WHO MONICA Project. Wld Hlth Statist Quart 1988;41:115-136.
3. Stewart AW, Kuulasmaa K, Beaglehole R. Ecological analysis of the association between mortality and major risk factors of cardiovascular disease. The World Health Organization MONICA Project. Int J Epidemiol 1994;23:505-516.
4. Yerushalmy J, Hilleboe HE. Fat in the diet and mortality from heart disease. A Methodological note. New Y State J Med 1957:57:2343-2354.  
5. Noakes TD. It’s the insulin resistance, stupid: Part 7. Crossfit Essentials. https://www.crossfit.com/essentials/its-the-insulin-resistance-stupid-part-7 
6. Lindah B, Asplund K, Hallmans G. High serum insulin, insulin resistance and their association with cardiovascular risk factors. The Northern Sweden MONICA population project. J Intern Med 1993;234:263-270.
7. Huhtasaari F, Asplund K, Wester PO. Cardiovascular risk factors in the Northern Sweden MONICA study. Acta Med Scand 1988;224:99-108.
8. Ryan DH, Espeland MA, Foster GD, et al. Look AHEAD (Action for Health in Diabetes): Design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes. Control Clin Trial 2003;24:610-628.
9. Yusuf S, Rangarajan S, Teo K, et al. Cardiovascular risk and events in 17 low-, middle-, and high-income countries. N Engl J Med 2014;371:818-827.
10. Noakes TD. Ancel Keys’ Cholesterol Con: Part 8. The Noakes Foundation. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-8-1970-1974 
11. Beresford SAA, Johnson KC, Ritenbaugh C, et al. Low-fat dietary pattern and risk of colorectal cancer. The Women’s Health Initiative Randomized Controlled Dietary Modification Trial. JAMA 2006;295:643-654.
12. Prentice RL, Caan B, Chlebowski RT, et al. Low-fat dietary pattern and risk of invasive breast cancer: the Women’s Health Initiative Randomized Controlled Dietary Modification Trial. JAMA 2006;295:629–642 
13. Giovannucci E, Habel LA, Harlan DM, et al. Diabetes and cancer. A consensus report. Diabetes Care 2010;33:1674-1685.
14. Bartsch H, Nair J, Owen RW. Dietary polyunsaturated fatty acids and cancers of the breast and colorectum: emerging evidence for their role as risk modifiers. Carcinogenesis 1999;20:2209-2218.
15. Michaels KB, Willett WC. The Women’s Health Initiative Randomized Controlled Dietary Modification Trial: a post-mortem. Breast Cancer Res Treat 2009;114:1-6.
16. Noakes TD. Ancel Keys’ Cholesterol Con: Part 12. The Noakes Foundation. https://thenoakesfoundation.org/news/the-ancel-keys-cholesterol-con-part-12-1984-1993
17. Howard BV, Van Horn L, Manson JE et al. Low-fat dietary pattern and risk of cardiovascular disease. The Women’s Health Initiative Randomized Controlled Dietary Modification Trial. JAMA 2006;295:655-666.
18. Noakes TD. Ancel Keys’ Cholesterol Con: Part 9. The Noakes Foundation. http://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-9-1976-1977
19. Noakes TD. The Women’s Health Initiative Randomized Controlled Dietary Modification Trial: An inconvenient finding and the diet-heart hypothesis. S Afr Med J 2013;103:824-825.
20. Rossouw JE, Howard BV. Noakes misses the point. S Afr Med J 2013;103:882
21. Noakes TD. WHIDMT. Rossouw and Howard blatantly miss the point. S Afr Med J 2014;104:261-262.
22. Mozaffarian D, Rimm EB, Herrington DM. Dietary fats, carbohydrates, and progression of coronary atherosclerosis in postmenopausal women. Am J Clin Nutr 2004;80:1175-1184.
23. Noakes TD. Ancel Keys’ Cholesterol Con: Part 6. The Noakes Foundation. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-6-1960-1967
24. Knopp RB, Retzlaff BM. Does saturated fat prevent coronary artery disease? An American paradox. Am J Clin Nutr 2004;80:1102-1103.
25. Volek JS, Fernandez ML, Feinman RD, et al. Dietary carbohydrate restriction induces a unique metabolic state positively affecting atherogenic dyslipidemia, fatty acid partitioning, and metabolic syndrome. Prog Lipid Res 2008;47:307-318.
26. Westman EC, Yancy WS, Jr., Olsen MK, et al. Effect of a low-carbohydrate, ketogenic diet program compared to a low-fat diet on fasting lipoprotein subclasses. Int J Cardiol 2006;110:212-216
27. De Graaf J, Hak-Lemmers HL, Hectors MP, et al. Enhanced susceptibility to in vitro oxidation of the dense low-density lipoprotein subfraction in healthy subjects. Arterioscler Thromb 1991;11:298-306.
28. Toma L, Stancu CS, Botez GM, et al. Irreversibly glycated LDL induce an oxidative and inflammatory state in human endothelial cells; added effect of high glucose. Biochem Biophys Res Commun 2009;390:877-882.
29. Austin MA, Breslow JL, Hennekens CH, et al. Low-density lipoprotein subclass patterns and risk of myocardial infarction. JAMA 1988;260:1917-1921.
30. Griffin BA, Freeman DJ, Tait GW, et al. Role of plasma triglyceride in the regulation of plasma low density lipoprotein (LDL) subfractions: relative contribution of small, dense LDL to coronary heart disease risk. Atherosclerosis 1994;106:241-253.
31. Steinberg D. Low-density lipoprotein oxidation and its pathobiological significance. J Biol Chem 1997;272:20963-20966.
32. Shoji T, Hatsuda S, Tsuchikura S, et al. Small, low-density lipoprotein cholesterol concentration and carotid atherosclerosis. Atherosclerosis 2009; 202:582-588.
33. Hoogeveen RC, Gaubatz JW, Sun W, et al. Small Dense Low-Density Lipoprotein-Cholesterol Concentrations Predict Risk for Coronary Heart Disease: The Atherosclerosis Risk in Communities (ARIC) Study. Arterioscler Thromb Vasc Biol 2014;34:1069-1077.
34. St-Pierre AC, Cantin B, Dagenais GR, et al. Low-density lipoprotein subfractions and the long-term risk of ischemic heart disease in men: 13-year follow-up data from the Quebec Cardiovascular Study. Arterioscler Thromb Vasc Biol 2005;25:553-559.
35. Dreon DM, Fernstrom HA, Miller B, et al. Low-density lipoprotein subclass patterns and lipoprotein response to a reduced-fat diet in men. FASEB J 1994;8:121-126.
36. Krauss RM, Dreon DM. Low-density-lipoprotein subclasses and response to a low-fat diet in healthy men. Am J Clin Nutr 1995;62:478S-487S.
37. Dreon DM, Fernstrom HA, Campos H, et al. Change in dietary saturated fat intake is correlated with change in mass of large low-density-lipoprotein particles in men. Am J Clin Nutr 1998;67:828-836.
38. Dreon DM, Fernstrom HA, Williams PT, et al. A very low-fat diet is not associated with improved lipoprotein profiles in men with a predominance of large, low-density lipoproteins. Am J Clin Nutr 1999;69:411-418.
39. Noakes TD, Sboros M. Real Food on Trial. Columbus Publishing, UK. 2019. 
40. Shikany JM, Margolis KL, Pettinger M, et al. Effects of a low-fat dietary intervention on glucose, insulin, and insulin resistance in the Women’s Health Initiative (WHI) Dietary Modification trial. Am J Clin Nutr 2011;94:75-85.
41. Salmeron J, Manson JE, Stampfer MJ, et al. Dietary fibre, glycemic load, and risk of non-insulin-dependent diabetes mellitus in women. JAMA 1997;277:472-477.
42. Salmeron J, Ascherio A, Rimm EB, et al. Dietary fibre, glycemic load, and risk of NIDDM in men. Diabetes Care 1997;20:545-550.
43. Schultze MB, Liu S, Rimm EB, et al. Glycemic index, glycemic load and dietary fibre intake and incidence of type 2 diabetes in younger and middle-aged women. Am J Clin Nutr 2004;80:348-356.
44. Halton TL, Liu S, Manson JE, et al. Low-carbohydrate-diet score and risk of type 2 diabetes in women. Am J Clin Nutr 2008;87:339-346.
45. Culver AL, Ockene IS, Balasubramanian R, et al. Statin use and risk of diabetes mellitus in postmenopausal women in the Women’s Health Initiative. Arch Intern Med 2012;172:144-152; Katsiki N, Banach M. Statin use and risk of diabetes mellitus in postmenopausal women. Clin Lipidol 2012;7:267-270.
46. Ma Y, Persuitte GM, Andrews C, et al. Impact of incident diabetes on atherosclerotic cardiovascular disease according to statin use history among postmenopausal women. Eur J Epidemiol 2016;31:747-761.
47. Rossouw JE. Serum cholesterol as a risk factor for coronary heart disease revisited. S Afr J Clin Nutr 2015;28:34-37.
48. Rossouw JE. The diet-heart hypothesis, obesity and diabetes. S Afr J Clin Nutr 2015;28:38-43.
49. Ford C, Chang S, Vitolins MZ, et al. Evaluation of diet pattern and weight gain in postmenopausal women enrolled in the Women’s Health Initiative Observational Study. Brit J Nutr 2017;117:1189-1197.
50. National Institutes of Health. News from the Women’s Health Initiative: reducing total fat intake may have a small effect on the risk of breast cancer, with no effect on the risk of colorectal cancer, heart disease, or stroke. February 7th 2006. Available at: https://www.nih.gov/news-events/news-releases/news-womens-health-initiative-reducing-total-fat-intake-may-have-small-effect-risk-breast-cancer-no-effect-risk-colorectal-cancer-heart-disease-or-stroke. 
51. 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 dietary recommendations do more harm than good”. S Afr J Clin Nutr 2015;28:19-33.
52. Shai I, Schwarzfuchs D, Henkin Y, et al. Weight loss with a low-carbohydrate, Mediterranean or low-fat diet. N Engl J Med 2008;359:229-241.
53. Teicholz N. The Big Fat Surprise. Why butter, meat and cheese belong in a healthy diet. Simon and Schuster, New York, NY. 2014. 
54. Willett WC. Eat, drink and be healthy. The Harvard Medical School Guide to Healthy Eating. Free Press, New York, NY. 2005.
55. Noakes TD. Ancel Keys’ Cholesterol Con: Part 13. The Noakes Foundation. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-13-1993-2005 
56. Atkins RC. Dr Atkins’ Diet Revolution: The High Caloric Way to Stay Thin Forever. New York: David McKay Company Inc.,1972.
57. Noakes TD. It’s the insulin resistance, stupid: Part 3. Crossfit Essentials. https://www.crossfit.com/essentials/its-the-insulin-resistance-stupid-part-3
58. Media Advisory. American Heart Association https://www.eurekalert.org/pub_releases/2002-11/aha-aha111902.php
59. White PL. A critique of low-carbohydrate ketogenic weight reduction regimens. A review of Dr Atkins’ Diet Revolution. JAMA 1973;224:1415-1419.
60. Hu FB, Stampfer MJ, Manson JE, et al. Dietary fat intake and the risk of coronary heart disease in women. N Engl J Med 1997;337:1491-1499.
61. Estruch R, Ros E, Salas-Salvado J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet. N Engl J Med 2013;368:1279-1290.
62. Estruch R, Ros E, Salas-Salvado J, et al. Primary prevention of cardiovascular disease with a Mediterranean diet supplemented with extra-virgin olive oil or nuts. N Engl J Med 2018;378:e34.
63. Hoermann R, Grossman M. Mediterranean diet for primary prevention of cardiovascular disease. N Engl J Med 2013;369:674
64. Noakes TD. Ancel Keys’ Cholesterol Con: Part 11. The Noakes Foundation. https://thenoakesfoundation.org/news/ancel-keys-cholesterol-con-part-11-1979-1984 
65. Anon. Harvard T.H. Chan School of Public Health. The Nutrition Source. PREDIMED study retraction and republication. What changed, what didn’t, and the big picture. Available at https://www.hsph.harvard.edu/nutritionsource/2018/06/22/predimed-retraction-republication/
66. Kopel E, Sidi Y, Kivity S. Mediterranean diet for primary prevention of cardiovascular disease. N Engl J Med 2013;369:672
67. Taubes G. Vegetable oils, (Francis) Bacon, Bing Crosby, and the American Heart Association. June 17th 2017. Available at http://garytaubes.com/blog/page/4/
68. Woodhill JM, Palmer AJ, Leelarthaepin B, et al. Low fat, low cholesterol diet in secondary prevention of coronary heart disease. Adv Exper Med Biol 1978; 109:317-31
69. Ramsden CE, Zamora D, Leelarthaepin B, et al. Use of dietary linoleic acid for secondary prevention of coronary heart disease and death. Evaluation of recovered data from the Sydney Diet Heart Study and updated meta-analysis. BMJ 2013 Feb 4;346:e8707.
70. Ryan DH, Espeland MA, Foster GD, et al. Look AHEAD (Action for Health in Diabetes): design and methods for a clinical trial of weight loss for the prevention of cardiovascular disease in type 2 diabetes. Control Clin Trials 2003;24:610-628.
71. Pi-Sunyer X. The Look AHEAD Trial: A review and discussion of its outcomes. Curr Nutr Rep 2014;3:387-391.
72. The Look AHEAD Research Group. Cardiovascular effects of intensive lifestyle intervention in type 2 diabetes. N Engl J Med 2013;369:145-154.
73. Frantz ID, Dawson EA, Ashman PL, et al. Test of effect of lipid lowering by diet on cardiovascular risk. The Minnesota Coronary Survey. Arteriosclerosis 1989;9:129-135.
74. Ramsden CE, Zamora D, Majchrzak-Hong S, et al. Re-evaluation of the traditional diet-heart hypothesis: analysis of recovered data from Minnesota Coronary Experiment (1968-73). BMJ 2016;353:i1246.
75. Noakes TD. It’s the insulin resistance, stupid: Part 8. Crossfit Essentials. https://www.crossfit.com/essentials/its-the-insulin-resistance-stupid-part-8
76. Grasgruber P, Sebera M, Hrazdira E, et al. Food consumption and the actual statistics of cardiovascular diseases: an epidemiological comparison of 42 European countries. Food Nutr Res 2016;60:31694.

Read Also:

• The invention of the Mediterranean Diet – Prof Noakes on Keys’ Cholesterol Con
• Institutional collusion and corruption runs DEEP – Prof Noakes on Keys’ Cholesterol Con
• “If the data doesn’t exist to support your hypothesis, you simply spin the results as if it did” – Prof. Noakes on Keys’ Cholesterol Con