Friday, 17 August 2012

Forks Over Knives and Healthy Longevity: Perhaps the Science is Legit After All

This is the second part of a series of posts that addresses the science regarding plant based diets and the documentary Forks Over Knives and the very serious inaccuracies and omissions that compromise the critiques authored by the cholesterol skeptics, in particular Denise Minger.

Part I: Forks Over Knives: A Missed Opportunity for the Cholesterol Skeptics

Denying the Preponderance of Evidence

Large systematic reviews of the nutritional literature authored by major international health authorities and by panels of leading nutrition researchers, many of which have disclosed ties to livestock industry, have consistently come to the conclusion that diets should be predominately plant based.1 2 3 4 The documentary Forks Over Knives features a number of doctors who have come to the conclusion that the allowance of animal foods in 'small-to-modest' amounts (made by researchers who often have financial ties to the livestock industry) are too permissive and that an optimal diet should be almost entirely composed of minimally refined plant foods. They also conclude that many major chronic and degenerative diseases that affluent populations succumb to can be prevented, and in many cases even reversed by consuming a whole-foods plant based diet. For the vast majority of nutritional researchers the question is no longer as to whether a plant based diet or an animal based diet is more optimal, but as to what the upper tolerable intake is in an optimal diet for foods not derived from minimally refined plant foods.

Many leading nutritional researchers and prominent health authorities actually do agree that the medical literature supports many of the dietary recommendations made by the doctors in Forks Over Knives, but are often unable to make similar recommendations to the public, in part due to socioeconomic factors. For example, Eric Rimm from the Department of Nutrition, Harvard said to Reuters in regards to a major health report produced by the National Academy of Science, which he was an author of that:
We can’t tell people to stop eating all meat and all dairy produces. Well, we could tell people to become vegetarians... If we were truly basing this on science we would, but it is a bit extreme.
Similarly, Walter Willett, the Chair of the Department of Nutrition, Harvard previously said in regards to findings on cancer that:
If you step back and look at the data, the optimum amount of red meat you eat should be zero.
Diethelm et al. published an excellent review addressing the five characteristics of denialism. The first characteristic describes how the cholesterol skeptics attempt to downplay the scientific consensus regarding the disease promoting effects of elevated LDL cholesterol and animal based diets rich in saturated fat:
The first is the identification of conspiracies. When the overwhelming body of scientific opinion believes that something is true, it is argued that this is not because those scientists have independently studied the evidence and reached the same conclusion. It is because they have engaged in a complex and secretive conspiracy.
The cholesterol skeptics will also attempt to downplay the scientific consensus often by insisting that scientists are ignoring certain studies, studies which these denialists fail to mention are compromised by a number of very serious flaws and omissions.5 6 7 Diethelm et al. also explains the motivations behind denialism:
Denialists are driven by a range of motivations. For some it is greed, lured by the corporate largesse of the oil and tobacco industries. For others it is ideology or faith, causing them to reject anything incompatible with their fundamental beliefs. Finally there is eccentricity and idiosyncrasy, sometimes encouraged by the celebrity status conferred on the maverick by the media.
Perhaps the cholesterol skeptics persistent denialism can be explained by conflicts of interest associated with the sales of merchandise or the desire for celebrity status on the internetBrownell et al. reminds us how serious and real conflicts of interest are, describing the tactics used by the tobacco industry, who for decades attempted to dismiss the 'junk' science linking smoking to lung cancer and other associated diseases, whose personal gain from this caused millions of people to perish. They asserted:8
A striking event occurred in 1994 when the CEOs of every major tobacco company in America stood before Congress and, under oath, denied believing that smoking caused lung cancer and that nicotine was addictive, despite countless studies (some by their own scientists) showing the opposite.
This merits exploration as to whether the cholesterol skeptics motivations are any different than these other denialists, and whether many of the hundreds of peer-reviewed papers they also dismiss as 'junk' science are actually informative and contain potentially life-saving findings.

Hormone Free, Pasteurised Animal Foods and Primitive Populations

Cholesterol skeptics will often claim that the results of any study suggesting harmful effects of animal foods were obscured due to participants consuming unnatural foods produced by intensive farming practices. The cholesterol skeptics however provide scant evidence regarding the perpetrated health benefits of replacing whole plant foods with naturally produced animal foods. Studies cited throughout both Part I of this review and this current post demonstrate that the association between replacing minimally refined plant foods with animal foods and poorer health expectancy can largely be explained by the fact that animal foods are typically naturally rich in methionine, dietary heme, saturated fat, dietary cholesterol, ruminant trans-fat, and hormones, and deficient in dietary fiber, antioxidants, carotnoids and phytochemicals, just to name a few.

A number of studies that have actually found some of the strongest associations between animal food intake and an increased risk of chronic diseases were actually carried out in populations where livestock is primarily grass fed and administration of hormones to livestock are banned by law, such as in Uruguay.9 10 11 12 These studies cannot simply be explained as exceptions as they are consistent with evidence from before the widespread use of intensive farming practices that produce unnaturally raised livestock.

In 1892 the renowned French geographer, Reclus noted that:13
...cancer is most frequent among those branches of the human race where carnivorous habits prevail.
In 1908, Williams published an extensive review of the medical literature and documentation from a large number of populations around the world in regards to the causation of cancer, and came to conclusions that were consistent with Reclus's findings:14
Careful study of the life-history of centenarians and of persons of advanced age who, as we have seen, are very rarely the victims of cancer shows that they are generally of spare figure, medium height, and that they eat frugally, taking but little meat and alcohol....In this connexion, it may be well to recall the fact, that although cancer is remarkably rare in vegetarian communities, yet complete exemption cannot be claimed for such ; and the like is true of herbivorous, as compared with carnivorous animals. In spite of these facts, which indeed are only such as might have been expected from the essential nature of the problem, there cannot be the slightest doubt in face of the overwhelming evidence I have adduced in the course of this work that the incidence of cancer is largely conditioned by nutrition.
In 1925, Kuczynski described the poor health of a population who subsisted on a diet based predominantly on organic pasteurized animal foods. As later described by Stamler:15
Kuczynski (1925) reported on an Asian population at the opposite end of the dietary spectrum - nomadic Kirghiz plainsmen who habitually consumed large amounts of meat and milk. He noted high incidence of obesity, premature extensive atherosclerosis, contracted kidney, apoplexy, and arcus senilis. Their urbanized kinsmen, subsisting on more varied fare, did not exhibit such severe vascular disease.
In 1932, Raab noted in regards to the distribution of atherosclerosis, that:
…the relative rarity of atherosclerosis and hypertension among the chiefly vegetable-consuming inhabitants of China, Africa, Dutch East India, and the enormous frequency of arteriosclerosis and hypertension among the peoples of Europe and North America who consume large quantities of eggs, butter...
In 1934, Rosenthal reviewed 28 papers from observations carried out around the world, and reached a conclusion that was consistent with Raab’s findings, noting that:16
…in no race for which a high cholesterol intake (in the form of eggs, butter and milk) and fat intake are recorded is atherosclerosis absent...
In 1940, based on years of clinical practice and reviewing medical reports, Bertelsen who is considered the father of Greenland epidemiology stated in regards to the mortality patterns amongst the Greenland Inuit that:17 
...arteriosclerosis and degeneration of the myocardium are quite common conditions among the Inuit, in particular considering the low mean age of the population. 
In 1904, Bertelsen proved the existence of cancer in the native Inuit, diagnosing a case of breast cancer. During the following decades researchers documented that the existence of cancer was exceedingly common among the Inuit despite their relatively short life expectancy.18 Consistent with Bertelsen’s findings, an Inuit predating western contact who was mummified in approximately 1475, 450km north of the Arctic Circle was shown to have evidence of cancer, likely of the breast.19 It has also been documented that numerous preserved pre-contact Inuit who were mummified dating all the way back to 1,500 years ago had a severe degree of atherosclerosis, osteoporosis, and osteoarthritis, consistent with studies of Inuit living in the 20th century.20 21 22 23 Other evidence of poor health among the pre-contact Inuit includes iron deficiency anemia, trauma, infection, dental pathology, and children with downs syndrome and Perthes disease.24 25

A large number of the examined mummies from ancient Egypt have also provided clear evidence of atherosclerosis in ancient civilizations. Unfortunately some researchers have previously confused the diets of the mummified elites of ancient Egypt who exhibited atherosclerosis with the plant based diets of the lower classes of Egypt.26 More recent research on the interpretations of the ancient Egyptian hieroglyphs and isotope analysis of hair samples from the mummies has provided strong evidence that the elites of Egypt, being those primarily mummified consumed a diet rich in meat and saturated animal fat.27 28 29 The researchers asserted that:
It is important to point out that there was a marked difference between the mainly vegetarian diet most Egyptians ate and that of royalty and priests and their family members whose daily intake would have included these high levels of saturated fat. Mummification was practised by the elite groups in society, ensuring that their remains have survived to provide clear indications of atherosclerosis; by contrast, there is a lack of evidence that the condition existed among the less well-preserved remains of the [mainly vegetarian] lower classes.
The following videos (Videos 1-6) are from the very inspirational Primitive Nutrition Series produced by Plant Positive, providing further details that address the very serious flaws and omissions that compromise the claims from advocates of animal based diets regarding primitive population studies.

Video 1. Primitive Nutrition 27: The Eskimo Model, Part I

Video 2. Primitive Nutrition 28: The Eskimo Model, Part II

Video 3. Primitive Nutrition 29: The Masai Model, Part I

Video 4. Primitive Nutrition 30: The Masai Model, Part II

Video 5. Primitive Nutrition 31: The Native Australian Model

Video 6. Primitive Nutrition 32: Ancient and out of Fashion

Taking Unfounded Wheat Claims with a Grain of Salt

At least eight large prospective studies consisting of a total of over 1,125,000 participants found that intake of foods containing whole grains or grain fiber were associated with a significantly lower risk of all-cause mortality, even in studies where the primary consumed grain was wheat or other gluten containing grains.30 31 32 33 34 35 36 37 38 These results are very impressive in light of the fact that the definition of whole grains in these studies were typically considered as foods containing as little as 25% whole grain, and that grain fiber intake in most developed nations is largely derived from refined grains, potentially underestimating the benefits of actual whole grain intake.39

A meta-analysis of 7 prospective studies found that intake of foods containing whole grain was associated with a 21% decrease of cardiovascular disease, and a pooled analysis of 11 prospective studies that used dietary validation methods found a 10 g/d increment of grain fiber was associated with a 25% decrease risk of death from coronary heart disease.40 41 Meta-analyses and systematic reviews have also found that foods containing whole grain and grain fiber significantly decreases the risk of colorectal cancer, type II diabetes, hypertension, and obesity.42 43 44 45 Furthermore, multiple studies have found that replacing animal foods with whole grains significantly decreases the risk of type II diabetes.46 47

Denise Minger has been criticised for failing to carry out multivariate analyses in order to adjust for multiple confounding variables simultaneously in her critiques of the China Study. A Swedish blogger carried out a multivariate analyses with multiple response variables from the raw China Study data, finding contrasting results to that of Minger s, primarily that animal protein and fat were associated with an increased risk of ischemic heart disease, whereas green vegetables, dietary fiber and grains, including wheat were associated with a decreased risk (Fig. 1). (48 49 50 English version from Google Translate)  These findings certainly raise questions as to whether Minger's oversimplified analysis holds up to her claims that 'wheat is murder'. When compared to Minger’s analyses, these findings are far more consistent with the findings from major cross-country studies, with the exception of vegetable fat (from PUFA and MUFA), which is typically associated with a decreased risk after controlling for saturated fat.51 52 53 54 These findings are also consistent with the observation that the large majority of the world’s population were largely free from coronary heart disease while consuming their traditional diet centered on tubers, legumes, and grains including wheat prior to adopting a western dietary pattern.55 56 57

Figure 1. Various foods and nutrients and risk of ischaemic heart disease in a multivariate regression analysis with multiple response variables for ages 35-69 in China Study II

Often overlooked by the promoters of wheat free diets is that well fermented wheat eliminates the great majority of gluten and has shown to be tolerable by celiac patients.58 Furthermore both epidemiological studies and randomized controlled trials have found that whole grain wheat does not  promote weight gain or impair satiety and is likely beneficial.59 60 61 62 63 Lastly, the fact that many of the authors of some of the most popular weight loss diet books that advocate the restriction of wheat and grain intake in favor of animal foods have remained over-weight despite apparently following their advocated diets for decades raises further questions as to whether such diets are the most optimal for weight loss and healthy longevity (Video 7). 

Refer here for a very informative review that documents the very serious inaccuracies, omissions, and oversimplifications presented in the book Wheat Belly.

Video 7. Low Carb vs. Plant-Based

Lowering Serum Cholesterol on a Plant Based Diet

Evidence from over 100 randomized controlled trials, large meta-analyses of mendelian randomization studies, and prospective studies consisting of several million individuals have firmly established a causal relationship between lowering LDL and non-HDL cholesterol and a decreased risk of cardiovascular disease and all-causes mortality, without evidence of a threshold beyond which a lower concentration does not provide additional benefit.64 65 66 67 68 69 70 71 

Regardless of the overwhelming evidence, Denise Minger suggested that the cholesterol levels among Dr. Esselstyn’s patients (serum cholesterol, >100mg/dl; LDL, >48mg/dl) are unhealthy and 'super-low'. It has already been established for decades that these 'super-low' cholesterol levels are actually well above levels required in order to support normal growth and development, and are the typical levels that were likely maintained for tens of millions of years throughout human evolution.72 73 These are also the typical concentrations found in free-ranging non-human primates, and among these primates consuming only food founds in their natural habitat higher cholesterol concentrations have been associated with atherosclerosis despite having cholesterol concentrations that Minger may define as 'super-low'.74 These lines of evidence refute the claims suggesting that foods found in nature promote optimal health regardless of their effect on LDL and serum cholesterol. Furthermore, individuals born with extremely rare conditions that cause life-long LDL levels of <15 mg/dl display normal growth and actually experience increased longevity.75 Michael S. Brown and Joseph L. Goldstein who were awarded a Nobel Prize for their research on the metabolism of LDL cholesterol elaborated on in this topic in their Nobel Prize lecture:76
In view of the 10 to 1 gradient between concentrations of LDL in plasma and interstitial fluid, a level of LDL-cholesterol in plasma of 25 mg/dl would be sufficient to nourish body cells with cholesterol. This is roughly one-fifth of the level usually seen in Western societies.  Several lines of evidence suggest that plasma levels of LDL-cholesterol in the range of 25-60 mg/dl (total plasma cholesterol of 110 to 150 mg/dl) might indeed be physiologic for human beings. First, in other mammalian species that do not develop atherosclerosis, the plasma LDL-cholesterol level is generally less than 80 mg/dl. In these animals the affinity of the LDL receptor for their own LDL is roughly the same as the affinity of the human LDL receptor for human LDL, implying that these species are designed by evolution to have similar plasma LDL levels. Second, the LDL level in newborn humans is approximately 30 mg/dl, well within the range that seems to be appropriate for receptor binding. Third, when humans are raised on a low fat diet, the plasma LDL-cholesterol tends to stay in the range of 50 to 80 mg/dl. It only reaches levels above 100 mg/dl in individuals who consume a diet rich in saturated animal fats and cholesterol that is customarily ingested in Western societies.
Minger made an unreferenced suggestion that Dr. Esselstyn’s patients while having no additional coronary events over several decades while adhering to the prescribed diet, may have somehow achieved better results if they consumed a diet that raised their HDL and lowered their triglycerides. She also made an unreferenced suggestion, perhaps referring to epidemiological studies that lowering cholesterol can increase the risk of developing a number of diseases, including cancer and neurological disorders. As is typical with the cholesterol skeptics, Minger will often dismiss any epidemiological study with findings suggesting adverse health effects of animal based foods as being largely uninformative, insisting that 'correlation isn’t causation'. Therefore this merits exploration into how epidemiological studies compare to randomized controlled trials used to prove 'causation' relating to Minger's concerning comments about serum lipids.

As already mentioned in the first post, a meta-analysis of 108 randomized controlled trials of various medical and dietary based lipid modifying interventions found that lowering LDL cholesterol significantly decreased the risk of coronary heart disease and all-cause mortality, while modifying HDL or triglycerides provided no clear benefit after controlling for LDL cholesterol.77

More recently a meta-analysis of mendelian randomization study of 170,000 participants found that inheriting genetic variants that are associated with life-long elevated HDL do not affect the risk of coronary heart disease.78 Another recent meta-analysis of mendelian randomized study with over 312,000 participants found that inheriting any of the nine studied genetic variants associated with life-long reduced LDL, but do not alter other known risk factors equally predicted a three-fold greater decreased risk of coronary heart disease per unit lower of LDL than statins do when started later in life (Video 8).79 Furthermore, this study and others found that individuals who inherit a variant of the statin drug targeted HMGCR gene that is associated with life-long reduced LDL, have an equal degree lower risk of coronary heart disease as individuals who inherited any of the other 8 studied gene variants.80 81 

These studies provide convincing evidence that the primary mechanism in which statins lower coronary heart disease can be explained by it effects on lowering LDL cholesterol. These studies also provide convincing evidence that the benefit of lowering LDL depends on both the timing and the magnitude of the LDL reduction, and that the benefits associated with lower LDL are largely independent of the mechanism in which LDL is lowered. This in-turn supports the benefits of a plant based diet combined with exercise, the safest way to significantly lower LDL cholesterol, beginning as early in life as possible (Video 8).

Video 8. Long Term Reduction in Low-Density Lipoprotein Cholesterol Beginning Early in Life 

In contrast to prospective epidemiological studies that focused on only base-line HDL, the results from prospective studies that tested the effect of HDL modification and the risk of cardiovascular disease have not been so consistent, for which the largest study failed to find any association after controlling for other risk factors.82

If raising HDL can actually modify cardiovascular risk, then this data provides convincing evidence that it is important to consider the mechanism of HDL rather than the concentration alone. For example it has been shown that diets high in saturated fat impairs the anti-inflammatory properties of HDL compared to polyunsaturated fat, and that high fiber low-fat diets convert HDL from pro to anti-inflammatory HDL.83 84 85 In addition animal fat contains dietary cholesterol and ruminant trans-fat which has been shown to raise the LDL/HDL ratio in randomized controlled trials.86 87 88 Therefore the preponderance of evidence demonstrates that attempting to raise HDL in the presence of an increased saturated animal fat intake as advocated by cholesterol skeptics will only increase the risk of cardiovascular disease.

It is well established that carbohydrate foods stripped of their nutritional value elevate triglycerides, but also that nutrient dense carbohydrate foods do not produce such an elevation, especially when the carbohydrate is largely derived from dietary fiber and resistant starch.89 90 91 92 93 94 This evidence strongly suggests that the elevated levels of triglycerides in a number of Dr. Esselstyn’s participants was a marker of excessive intake of refined carbohydrates that Minger claimed that they had eliminated from their diet. Minger appears to have confused Dr. Esselstyn's recommendations in his recent book with those provided to his initial set of patients well over two decades ago. Dr. Esselstyn did not specify in his papers that his initial set of patients were asked to eliminate refined carbohydrate foods, which may have played a large role in their diets considering that carbohydrate intake was increased to approx. 80% of total energy. The success seen among Dr. Esselstyn's initial set of patients therefore can probably not be explained by a reduced intake of nutrient poor carbohydrate foods.95

In regards to the cholesterol skeptic's typical claims about low serum cholesterol and an excess risk of cancer, an editorial authored by the American Cancer Society attributed this association to reverse causation, refuting the suggestion of a causal relationship:96
Many epidemiologic studies published in the 1980s documented an association between low circulating cholesterol and higher overall cancer incidence and mortality. This association has been attributed to reverse causation, that is, undiagnosed cancer causing a reduction in cholesterol levels. Reverse causation is strongly supported by observations that cholesterol levels decline before cancer diagnosis and that associations between low cholesterol and cancer incidence and mortality weaken when the first few years of study follow-up are excluded. In addition, a meta-analysis of randomized trials of cholesterol-lowering statins found no effect on risk of cancer, although only short-term effects could be addressed due to the short duration of most trials.
More recently several mendelian randomization studies have demonstrated that individuals who inherit genetic variants associated with life-long reduced LDL do not have an excess risk of cancer, compensating for the relatively short-term cholesterol-lowering trials.97 98 99

Minger’s own words, 'Yikes! Did we slip and fall back into the ’80s?' more accurately describes her own misleading statements than that of the doctors in Forks Over Knives. Several recent prospective studies with up to 37 and 40 years follow-up, sufficient to potentially eliminate the possibility of reverse causation found that elevated serum cholesterol was associated with an increased risk of aggressive prostate cancer or prostate cancer death.100 101

In regards to low cholesterol and neurological disorders, a mendelian randomization study found that individuals who inherit genetic variants associated with life-long reduced LDL do not have increased depressive symptoms.102 Also, randomized controlled trials have found that vegetarian diets which are associated with lower cholesterol are also associated with improved moods compared to omnivorous diets.103 104 105 Other randomized controlled trials found that psychological symptoms including depression, hostility and anger improved significantly on a cholesterol lowering, complex-carbohydrate rich diet compared to baseline or carbohydrate restricted diets.106 107 Conversely, systematic reviews and meta-analyses of epidemiologic studies have found an association between dairy intake and an increased risk of Parkinson’s disease, and saturated fat and an increased risk of dementia.108 109 110 111

For the evidence regarding diet and serum lipids and the risk of stroke please refer to Part I and Part II of my review addressing this subject.

The World Wars Revisited

During the British blockade in World War I, the food imports that Denmark heavily relied upon were cut off and the population was forced to sell a large portion of their livestock due to the inefficient conversion of livestock feed into meat. Following a transition towards more of a plant based diet the Danish capital experienced a significant decrease in mortality from chronic disease and all-causes during the period of significant regulation towards the end of the war.112 113

As previously discussed in Part I of this review, in World War II a number of Scandinavian and Low Countries of Europe experienced deprivations of animal foods together with a significant decline in cardiovascular mortality. In contrast, in the United States heart disease mortality increased together with an increased intake of dairy and eggs. In regards to the scarcity of war time tobacco as a potential confounding variable, smoking was very rare among Scandinavian women prior to the war yet cardiovascular mortality decreased similarly in both men and women. Furthermore in nations such as Denmark and Great Britain where there was a similar scarcity of tobacco as other Scandinavian countries but intake of cholesterol rich food was not significantly altered, there was no significant change in cardiovascular mortality. It was also observed that the decline of cardiovascular mortality was in the order of the nations that experienced shortages of animal fats earlier in the war.114 115 116 Another valuable finding was the observation that children born in Norway and other parts of Scandinavia during or shortly after the war experienced lower than expected rates of a number of cancers during the following decades, suggesting the importance of a plant based maternal diet.117 118

During the food shortages in West Germany under the Allied occupation from 1945-48 there was an observed absence of advanced forms of atherosclerosis, diabetic complications and cardiovascular mortality compared to the pre and post occupation period, as well as compared with the parts of Southern Germany that did not experience severe food shortages. This was attributed primarily to a cholesterol lowering diet that was low in energy, animal fat and animal protein and rich in foods high in dietary fiber. Other major modifiable risk factors were unable to explain the changes in cardiovascular disease as these were modified only slightly between 1948-50 when the sharp increase in heart attacks was observed.119  

In regards to dietary intake of Norway during the war, Minger cited evidence from a chart of food intake of men from 50 families in Oslo which included foods obtained outside of rationing from illegal sources. As the authors pointed out, these illegally obtained foods were 'mostly of bread and flour', suggesting that the data for animal food intake cited by Strom and Jensen in Part I of this review was not compromised by the exclusion of illegally obtained foods. Nevertheless, Minger referred to the almost two fold increase of flour, meal and groats during the war as being only a 'slight' increase. The chart suggests that among these men, compared to the 1936-37 pre-war period, intake of meat, eggs and added fats had significantly decreased by at least early 1941 and declined throughout the year. This provides further evidence that refutes Minger s claims that intake of animal foods did not decline until late 1941. However, compared to 1936-37 total dairy intake among these men was observed to be slightly higher in 1941 but then declined from early 1942, primarily from high fat sources.120

Table 1. Dietary intake of food groups for Oslo men from 50 families, 1936-45

In regards to dairy intake in Norway, Denise Minger claimed that:
There’s no doubt about it: In 1941, when cardiovascular disease started plummeting, Norwegians were eating more total dairy (light blue line) than they were before the war, when the death rate was higher. 
Minger derived this data from these same Oslo men from 30 to 50 families (Table 1) and graphed the intake of dairy intake using the per-war years of 1936-37 and the war years of 1941-45. Regarding the dietary intake of these Oslo men Minger clearly contradicted herself by first stating that 'it’s hard to say how accurately this represents the food intake of Norway’s whole population', and then stating 'There’s no doubt about it', as if she is certain that this data accurately reflected the intake of the entire Norwegian population. When examining the observed dietary changes of these Oslo men and cardiovascular disease statistics, it may be more informative to provide statistics specifically for Oslo men, rather than the entire nation as Minger has done, as the observed rates of cardiovascular mortality were significantly different between Oslo men and Oslo women, as well as between Oslo and the entire nation.121 122

In Oslo the observed heart attack rate of men was actually slightly higher in 1941 compared to 1936-37, but then plummeted in 1942 simultaneously with the decreased intake of dairy, particularly from high fat sources. The observed increased heart attack rate in Oslo men in 1945 with an increased intake of diary in early to mid-1945 among these Oslo men (Figs. 2, 3).

Figure 2. Analysis of hospitalized cases of myocardial infarction (heart attack) in a sub-group that was estimated to much more accurately reflect that of the general population in Oslo, Norway 1935-49

Figure 3. Dairy intake and myocardial infarction in Oslo men, 1936-45

Although Minger correctly points out that Dr. Esselstyn did not mention the significant increased intake of fish in Norway during the war years, her claim that the Norwegian diet was 'marine based' appears somewhat fishy. According to the tables of food intake for the Oslo men that Minger focused on, intake of the major starch groups, including flour, meal and groats, bread, potatoes, and roots and vegetables was approximately 1 kg/day, roughly 4 times that of fish intake. Fish intake can not explain the large decline of cardiovascular morality in Scandinavia and the low countries of Europe. In Finland where the observed decline in arteriosclerosis mortality was even greater than in Norway, availability of fish was actually lower throughout most of the period where wartime rates of mortality were plummeting, and also declined throughout the late 1980s and 1990s when mortality was again plummeting.123 124 125 126

Several recent meta-analyses of randomized controlled trials, including the highest quality double-blinded, placebo-controlled trials failed to show that fish oil decreased the risk of cardiovascular events, including subgroup analyses when compared to oleic or n-6 rich vegetable and mixed oils.127 128 129 Furthermore, large prospective studies have found that benefits of fish are only apparent when displacing less healthy foods such as red meat, and that further benefit was found when fish was replaced with whole plant foods (Fig. 1).130 131 132 133 The lack of cardiovascular benefits associated with the consumption of fish oil maybe in part explained by the fact that the majority of marketed fish and fish oils regardless of labelled claims contain high levels of environmental pollutants, such as mercury and PCBs.134 135 136 137 Several studies have found an association between mercury exposure and an increased risk of all-cause mortality, and a recent review elaborated on the numerous detrimental effects of mercury exposure and risk of vascular diseases:138 139
The overall vascular effects of mercury include increased oxidative stress and inflammation, reduced oxidative defense, thrombosis, vascular smooth muscle dysfunction, endothelial dysfunction, dyslipidemia, and immune and mitochondrial dysfunction. The clinical consequences of mercury toxicity include hypertension, coronary heart disease, myocardial infarction, cardiac arrhythmias, reduced heart rate ariability, increased carotid intima-media thickness and carotid artery obstruction, cerebrovascular accident, generalized atherosclerosis, and renal dysfunction, insufficiency, and proteinuria. 
A number of studies have also found evidence of ill-effects of prenatal exposure to mercury from fish on child neurodevelopment.140 Nevertheless, there is suggestive evidence of benefits from long chain omega 3 fatty acids, particularly for those with low circulating levels which could be obtained more efficiently through the consumption of microalgae oil to minimise the risk of exposure to harmful toxins.141

The Primitive Nutrition Series

The following videos (Videos 9-12) are from the Primitive Nutrition Series produced by Plant Positive that address some of the very serious flaws and omissions in Denise Minger's interpretation of the China Study.

Video 9. Primitive Nutrition 62: China Studies, Part I

Video 10. Primitive Nutrition 63: China Studies, Part II

Video 11. Primitive Nutrition 64: China Studies, Part III

Video 12. Primitive Nutrition 65: China Studies, Part IV

The following videos (Videos 13-19) were created as a response to Denise Minger's statements about the Primitive Nutrition Series. In regards to video 16 and the plausibility of removing countries from cross-country comparisons due to the time lag between the war-time decrease in coronary heart disease mortality and the data from 1950 obscuring the diet-heart relationship, Minger correctly pointed out that Denmark did not experience a reduction of heart disease mortality during the war. As previously explained, unlike the other Scandinavian countries, there was no significant decrease in cholesterol rich foods in Denmark, nor was there a decrease in cardiovascular mortality.142

Although the explanation of time-lag may have weakened the diet-heart relationship, the large distribution of mortality rates among countries with high fat intake can largely be explained by differences in availability of saturated fats, which were lower in Denmark, Norway and Sweden. When researchers specifically looked at availability of saturated fat, there were no longer any significant exceptions in the data from the 1950s, with saturated fat explaining almost 70% of the variance of heart disease between countries (Fig. 4).143 Hegsted et al. used a multiple-regression equation using later data and found that almost all variance of coronary heart disease mortality between countries (r=0.92) could be explained by a combination of saturated fat which was positively associated, and polyunsaturated fat and alcohol which were inversely associated with risk (Fig. 5).144

Figure 4. Association of international death rates of coronary heart disease with percentage intakes of total and saturated-types of fat

Figure 5. Mortality predicted by consumption of saturated and polyunsatured fat and all sources of alcohol versus observed coronary heart disease mortality

In regards to the data from the 22 countries, Minger cautioned her readers about interpreting the results of diet and the risk of coronary heart disease and all-cause mortality due to the significant differences in socioeconomic factors between the countries. Less favorable socioeconomic factors are associated with an increased coronary heart disease and all-cause mortality, and would have potentially biased the results towards finding a positive association between plant based nutrients and mortality as it was the developing nations that typically consumed more plant based foods.145 146 147 148 Minger also stated that the GDP per capita for each nation is a 'pretty good way to estimate standard of living'. However, many of the countries with moderate intakes of saturated fat rather than total fat, such as Denmark and Sweden, and higher intakes, such as the U.K and Canada had similar high levels of medical care and GDP per capita, yet the countries with higher intakes of saturated fat had significantly higher mortality rates of coronary heart disease (Fig. 4).149 150

It is not the least bit surprising that while criticizing Ancel Key’s findings in regards to the association between animal based nutrients and all-cause mortality, Minger failed to inform her readers that in the Seven Countries Study prospective study that was not confounded by such significant differences in socioeconomic factors, saturated fat was associated with a significantly increased risk of all-cause mortality in the 10, 15 and 25 year follow-ups.151 152 153

Video 13. "Vegan Propaganda"

Video 14. Response to Denise Minger 1: Scrupulous

Video 15. Response to Denise Minger 2: Not Benefiting from Hindsight

Video 16. Response to Denise Minger 3: Cherry Picking

Video 17. Response to Denise Minger 4: China Revisited

Video 18. Response to Denise Minger 5: Wheat and Carbs

Video 19. Response to Denise Minger 6: Number Needed to Treat

The Fictional Independent Correlation

The cholesterol skeptics frequently cite research that they claim shows no independent association between consuming higher intakes of animal foods and chronic and degenerative diseases by examining these specific foods in isolation, while ignoring what sources of energy these are typically substituted for. These observations are misleading because they ignore the law of thermodynamics where in the general population a decrease of one source of energy will generally lead to an increase in other sources in order to maintain energy balance. In order to provide a more informative analysis of the association between diet and risk of developing diseases, it is therefore essential to study effect of substituting different foods on health outcomes. This was elaborated on in a large research panel organised by Walter Willett:154 155

For example, it may not be useful, as is usually done, to compare a specific food to all other sources of energy, which are usually mainly refined starches, sugars, red meat, and fat-rich dairy products in typical Western diets.
Not surprisingly meta-analyses and systematic reviews that do not compare foods with appropriate alternatives often fail to even find a relationship between refined grains and the risk of cardiovascular disease, weight gain and all-cause mortality, even though it is well documented that replacement with whole grains reduces risk.156 157 While such an over-simplistic approach is apparently sufficient for the cholesterol skeptics to claim that animal foods are safe to consume in almost unlimited quantities, they demonstrate very little interest to judge the health properties of other foods, such as refined grains using the same methods, as doing so would provide little rational for their unfounded claims. Unfortunately this enormous loop-hole has invited a very unwelcome opening for the industries and advocates of fad diets to exploit the medical literature in order to promote the consumption of disease promoting foods. Meta-analyses funded by the associated industries have used this loop-hole to downplay the damaging effects of saturated fat and soft drink consumption, but have nevertheless been refuted by other researchers who demonstrated that the studies that used better methodology found an increased risk of disease.158 159

Studies that focus on substituting food are far more informative and provide increasing evidence for replacing animal foods, particularly red meat and saturated fat rich foods with minimally processed plant foods (Fig. 6).160 161 162 163 164 Note that the definition of whole grains in the studies represented in Figure 6. are grains with at least 25% whole-grain or bran content by weight, permitting up to 75% of the grains as being essentially refined.165
Figure 6. Type II diabetes associated with replacement of other food groups for red meat from repeated dietary data surveys from over 200,000 men and women in the Health Professionals Follow-Up Study and Nurses’ Health Study I & II

The results of many of the referenced epidemiological studies are likely conservative estimates due to imprecise dietary measurement methods that are prone to attenuation bias (regression dilution bias). For example, it is well documented that observational studies significantly underestimate the strong relationship between diet and serum lipids that has been firmly established by hundreds of tightly controlled feeding studies.166 167 168 It was estimated over four decades ago that in order to estimate within 20% of the actual dietary intake, there is a requirement of at least 22 days of 24-hour dietary recalls for a number of macronutrients including saturated fat, and at least 45 days for dietary cholesterol.169 Inaccurately measuring intraindividual variation has been shown to lead to a miss-classification of subjects into ranges of usual dietary intakes, and biasing correlation coefficients towards null.170 Despite excessively referring to epidemiological studies that use some of the most imprecise dietary measurement methods such as single 24-hour dietary recalls as being of 'high quality' or 'good science', the cholesterol skeptics constantly dismiss any epidemiological studies that produce results contradicting their claims as being flawed and entirely uninformative rather than recognising that the results were likely conservative estimates.171 172 

In a recent post Denise Minger discussed a prospective study which found a significant association between higher-protein, lower-carbohydrate diets and an excess risk of cardiovascular disease among Swedish women. She attempted to downplay the study by pointing out the limitation that dietary intake was only measured by a dietary survey focusing on the 6 month period prior to the 16 year follow-up. She suggested that dietary and lifestyle changes throughout the follow-up period may have obscured the results towards finding this excess risk, ignoring the evidence presented by the researchers indicating the exact opposite:173
The long interval between exposure and outcome is a source of concern, because certain participants may change their dietary habits during the intervening period. However, this is more likely to generate non-differential misclassification and, thus, attenuate the evaluated association. In fact, we saw a tendency for the incidence rate ratios to decline with longer follow-up.
Minger also failed to mention the researchers discussion within the paper explaining how the findings from this study were consistent with numerous other prospective studies, including the Nurses's Health Study (NHS) and the Health Professional's Follow-up Study (HPFS) which measured diet up to six times throughout the follow-up period.174 175 176 177 178

The NHS and HPFS are well known for the use of virtually the highest quality dietary measurement methods used in any large scale epidemiological study, including detailed food frequency questionnaires (FFQ) that have been updated every 4 years for over 20 years and validated against multiple-day food records of at least 14 days in a year. A number of cholesterol skeptics have attempted to downplay the quality of the NHS and HPFS, particularly when a recent paper was published describing a significant association between red meat intake and an increased risk of all-cause mortality. They criticised this paper in part due to the variance of dietary intake between the FFQ and the multiple-day food records, clearly ignoring the fact that the researchers found that the association between red meat and excess mortality was actually significantly strengthened in a sensitivity analysis using data from multiple FFQ against the multiple-day food records in order to account for measurement error. Not surprisingly the researchers also found that when compared to multiple FFQ, the association was weakened when using data only from only the single base-line FFQ. Furthermore, no other foods associated with red meat intake could explain this association with excess mortality, and the only nutrients that partly explained this association were those that are naturally found in red meat, including saturated fat, dietary cholesterol and dietary heme. This strongly suggests that the excess mortality was associated with constituents naturally found in red meat, and refutes the claims that the results were obscured by other unhealthy foods associated with red meat intake.179 

Another important factor is that it has been observed that participants often change their dietary and lifestyle habits, including refraining from consuming meat and saturated fat due to illnesses or unfavourable risk factors that ultimately become life threatening.180 This frequently causes researchers to observe participants with a high risk of developing diseases as having lifestyles and consuming diets that are portrayed as being healthy, a phenomenon known to bias the benefits of smoking cessation towards null.181 It is also important not to neglect that many of the recommendations to avoid consuming a number of toxic substances, including tobacco are not based off results from controlled clinical trials investigating hard disease end points, but often purely based off epidemiologic, metabolic and laboratory studies. Ignoring the totality of the evidence simply due to a lack of large clinical trials can result in absurd consequences, including a missed opportunity for healthy longevity.

A Bid Farewell to the Confusionist's

Denise Minger’s biases have become exceptionally clear throughout her posts compromised by very serious and evidently intentional inaccuracies and omissions regarding the evidence supporting the health benefits of plant-based diets. She has also demonstrated her biases through descriptive statements such as ‘How to Win an Argument with a Vegetarian’ and ‘wheat is murder’, as well as with the posting of completely irrelevant photos of vegans that she appears to be ridiculing (Video 13). 

Minger constantly attempts to refute that she is bias in favor of promoting animal foods by claiming that this cannot be true simply because she was previously a vegan, and still eats a primarily plant based diet. Such claims should be considered with extreme caution in light of the fact that a number of other confusionists, including several honorary members of the Weston A. Price Foundation, such as William Douglas and Barry Groves who were opposed to smoking tobacco until they apparently discovered as they have described it, that the consensus that smoking is disease promoting was derived from ‘prejudices based on false science and government propaganda’, that there is a ‘broad spectrum of therapeutic and preventive applications of tobacco smoking for human medicine’, that ‘tobacco smoke contains no carcinogens’, and even that nicotine can ‘Help you live longer’.182 183 184 Perhaps the real reason Minger consumes a plant based diet is simply because she is well aware of the health benefits over a diet rich in animal foods.

Lastly, it is not scientifically justified to assume that someone's diet is optimal just because they perceive themselves as being healthier after a dietary change. Humans simply do not have the ability to feel the initial stages of development of atherosclerosis and cancer, nor feel the cancerous N-nitroso compounds form in their digestive tract and their serum cholesterol rise after digesting an animal rich meal. There is usually a significant time lag of up to several decades between dangerous lifestyle changes and the maximum risk of developing the associated diseases (Video 7).185 186 Furthermore, sudden cardiac death which has been associated with high protein and ketogenic diets is a leading cause of death in the developed world, a disease where there are typically no associated symptoms up until one hour prior to death.187 188 189 190 Dr. Michael Gregor outlines sudden cardiac death in a video regarding the remarkable findings from the China Study explaining how no one needs to succumb to this (Video 20).

Video 20. China Study on sudden cardiac death

Please post any comments in the Discussion Thread.