Tuesday, 19 February 2013

Diet-Heart: The Role of Vegetarian Diets in the Hypothesis

A recent publication from the EPIC-Oxford cohort with 15,000 vegetarians and 30,000 non-vegetarians found that the vegetarians had a 32% lower risk of hospitalization or death from coronary heart disease.1 These findings are consistent with a previous meta-analysis of 5 cohort studies with 48,000 non-vegetarians and 28,000 vegetarians which found that lacto-ovo vegetarians had a 34% lower risk of fatal coronary heart disease compared to regular meat eaters.2 These findings remained significant even after adjusting for non-dietary factors and alcohol intake. In addition, in each of these 6 cohort studies, vegetarians and non-vegetarians shared a similar interest in healthy lifestyles or were of a similar religious background, therefore limiting the number of potential confounders that could have obscured these findings.

This review will focus on the evidence from randomized controlled trials and long-term prospective cohort studies addressing the influence of vegetarian dietary patterns on the risk of coronary heart disease, and how these findings have contributed to the current understanding of the diet-heart hypothesis. This review will also consider the question as to whether the simple definition of a vegetarian diet is meaningful in the context of a healthy diet to reduce the risk of coronary heart disease. Regarding cohort studies, this review will primarily consider the influence of lacto-ovo vegetarian diets on the risk of coronary heart disease due to limited evidence from these studies addressing the long-term adherence to other types of vegetarian diets. A more informative analysis maybe possible after a longer follow-up of the on-going and largest cohort of vegetarians, the Adventist Health Study 2, which has observed more favorable cardiovascular risk factors within different vegetarian subgroups, particularly vegans.3

Skeptics of the diet-heart hypothesis often suggest that there are no plausible mechanisms in which a vegetarian dietary pattern can lower the risk of coronary heart disease, and often ascribe the observed benefits of vegetarianism to factors other than the avoidance of animal foods. Typically either ignored or downplayed by these skeptics is the convincing evidence that vegetarian dietary patterns can lower LDL cholesterol, which is an established risk factor for coronary heart disease.4 5 6

Establishing Causation

In the 6 cohorts described, a sizable portion of the non-vegetarians consumed significantly less meat than the general population. For example, in the EPIC-Oxford cohort, most participants were either occasional meat eaters, or affiliated with vegetarians or with vegetarian societies. Also, a potential problem in these cohorts is that measurement error of usual dietary intake of meat may have resulted in misclassifying a sizable portion of non-vegetarians as vegetarians. For example, in the Health Food Shoppers Study included among these cohorts, a validity assessment of the survey used to classify the participants vegetarian status suggested that 34% of the participants classified as vegetarians actually consumed meat. This data strongly suggests a much smaller than otherwise expected difference in dietary intake between the groups classified as vegetarians and non-vegetarians, potentially masking a stronger protective effect of a vegetarian dietary pattern.7

Another potential problem in these cohorts is the possibility that a sizable portion of participants classified as vegetarians stopped consuming meat or other animal foods in response to deteriorating health or unfavorable risk factors that would ultimately become life-threatening. This has been referred to as the ‘sick quitter effect’, which is known to mask the protective effect of smoking cessation in studies due to participants quitting in response to deteriorating health.8 In regards to diet, it has been documented that people tend to lower intake of saturated fat and cholesterol in response to unfavorable serum cholesterol levels, which has actually been shown to bias the association between diet and serum cholesterol in the opposite direction than expected [reviewed previously]. This bias is known as reverse causation, and may explain why in the Adventist studies that participants with short-term adherence (less than 5 years) to a vegetarian diet experienced an increased risk of mortality, while participants with long-term adherence (more than 17 years) to a vegetarian diet experienced a significantly lower risk of mortality compared to non-vegetarians (Fig. 1).8

Figure 1. Life expediencies for long-term vegetarians and short-term vegetarians in the Adventist Health Study and Adventist Mortality Study*

These factors should be taken into account when testing for causality as failing to do so may mask a protective effect of a vegetarian dietary pattern. One of the most important factors in order to establish causality is to address whether the association is biologically plausible, which in this case requires examining how vegetarian dietary patterns can influence cardiovascular risk factors.

Serum Lipids

In 1922, de Langen published what was perhaps the first study that provided strong evidence that a vegetarian dietary pattern favorably effects serum cholesterol when he placed five native Indonesians consuming a rice-based vegetarian diet into a metabolic ward and shifted the diet to one rich in meat, butter and egg fats, resulting in significant elevations in serum cholesterol [reviewed previously]. In 1954, Hardinge and Stare published what was perhaps the first observational study comparing the serum lipids of vegetarians to non-vegetarians in an affluent population. Lacto-ovo vegetarians and especially vegans had significantly lower serum cholesterol concentrations despite relatively high intakes of saturated fat.9 10 In 2009, Ferdowsian and Barnard published a systematic review of 27 randomized controlled trials and observational studies on either vegetarian or predominantly plant-based diets, and found that certain plant-based dietary patterns can lower LDL cholesterol by up to 35%, independent of changes to body weight (Figs 2, 3).4

Figure 2. Effects of plant-based diets in normolipidemic individuals: Randomized controlled trials*

Figure 3. Effects of plant-based diets in hyperlipidemic individuals: randomized controlled trials*

In the Lifestyle Heart Trial lead by Dr. Dean Ornish, intensive lifestyle changes including a vegetarian diet that allowed a small amount of non-fat dairy foods successfully reduced LDL by 37.2%, angina episodes by 91% and regressed coronary atherosclerosis in the experimental group after 1 year. In both the experimental and control group LDL and total cholesterol was correlated with changes in coronary atherosclerosis.11 

A recent meta-analysis of statin based randomized controlled trials found that lowering LDL cholesterol to less than 100 mg/dl was associated with regression of coronary atherosclerosis in participants with coronary heart disease.12 Similarly, a recent mendelian randomization study of over 100,000 individuals found that genetically-predicted higher LDL cholesterol was associated with greater carotid atherosclerosis, but there was no causal association for HDL cholesterol and triglycerides.13 Consistent with these lines of evidence, it has been consistently demonstrated in experiments on non-human primates that coronary atherosclerosis induced by feeding of dietary cholesterol and saturated fat can be reversed by a cholesterol lowering diet [reviewed previously]. Therefore the preponderance of evidence strongly suggests that the findings from the Lifestyle Heart Trial of a correlation between lower LDL cholesterol and regression of coronary atherosclerosis was causal, and can at least partly be explained by the intervention of a cholesterol lowering vegetarian diet. 

In the meta-analysis of 5 cohorts it was found that in a sample of participants from 3 of the cohorts that serum cholesterol ranged from between 13 mg/dl to 23 mg/dl lower in the vegetarians compared to non-vegetarians. The researchers suggested that the difference in serum cholesterol could have largely explained the difference in fatal coronary heart disease between these groups.2 In the EPIC-Oxford cohort, serum lipids and blood cholesterol were measured in a sample of the participants. Non-HDL cholesterol was 17 mg/dl lower and systolic blood pressure was 3.3 mmHg lower in the vegetarians compared to the non-vegetarians. The researchers calculated that the differences between these two risk factors alone would expect to lower the risk of coronary heart disease by 24%, which is less than the observed 32% lower risk.1 

The researchers from the EPIC-Oxford cohort suggested that the high ratio of polyunsaturated fat to saturated fat largely explained the difference in non-HDL cholesterol between groups, but failed to mention that a number of other plant based nutrients may have also contributed to this difference.1 It has been repeatedly demonstrated in randomized controlled trials that intake of plant protein, particularly from soy, plant sterols, and dietary fiber can also lower LDL cholesterol.14 15 16 In fact in many of the interventions with the greatest diet induced decrease in LDL cholesterol, the decrease could not be explained by changes in dietary fat and cholesterol intake alone, but also likely due to the additive effects of a number of plant based nutrients.17 18 19

It is clear that the LDL cholesterol levels of the vegetarians in these cohort studies far exceeded optimal levels, likely due to a diet deficient in whole plant foods and still relatively rich in animal foods. If these vegetarians had adhered to a much more phytonutrient rich cholesterol lowering diet such as that used in the aggressive dietary experiments, an even significantly lower risk of coronary heart disease may have been observed. Plant Positive recently referred to this informative statement made by Michael Brown and Joseph Goldstein the year before they were awarded the Nobel Prize for their research on the LDL-receptor:20
If the LDL-receptor hypothesis is correct, the human receptor system is designed to function in the presence of an exceedingly low LDL level. The kind of diet necessary to maintain such a level would be markedly different from the customary diet in Western industrial countries (and much more stringent than moderate low-cholesterol diets of the kind recommended by the American Heart Association). It would call for the total elimination of dairy products as well as eggs, and for a severely limited intake of meat and other sources of saturated fat.
Evidence from over one hundred randomized controlled trials has proven beyond plausible doubt that changing from a diet rich in animal foods to a dieter richer in certain whole plant foods significantly lowers LDL cholesterol.4 14 15 16 21 22 Similarly, evidence from over one hundred randomized controlled trials has proven beyond plausible doubt that lowering LDL cholesterol decreases the risk of coronary heart disease and all-cause mortality.5 6 Therefore consistent with the diet-heart hypothesis, there is convincing evidence that an appropriately designed vegetarian diet would reduce the risk of coronary heart disease, and that this reduction can at least be partly explained by lower LDL cholesterol.

Beyond Cholesterol

There are likely a number of dietary related factors that contribute to the lower risk of coronary heart disease observed in people with vegetarian dietary patterns that cannot entirely be explained by lower LDL cholesterol. For example, it has been shown in randomized controlled trials that a number of plant based nutrients can lower blood pressure, which may explain the lower blood pressure observed in vegetarians in a number of observation and intervention studies [reviewed previously]. Furthermore, appropriately designed vegetarian diets likely reduce the risk of being overweight and developing type II diabetes.23 24 25 26 27 Other factors such as reduced oxidation of LDL and changes in blood clotting have also been suggested as explanations for the lower risk of coronary heart disease observed in vegetarians.28 29

Perhaps the main concern with an inappropriately designed vegetarian diet is that it may result in elevated homocysteine due to an inadequate intake of vitamin B12, suggested to be a risk factor for coronary heart disease. Although deficiency of vitamin B12 is rarely observed in some populations in the developed world consuming a predominantly plant based diet, perhaps due to regular contact with vitamin B12 producing bacteria, health authorities strongly recommended that vegetarians diets be supplemented regularly with a bioavailable source of vitamin B12.30 31 Jack Norris, RD regularly posts informative reviews on the latest research on vitamin B12 intake and homocysteine, and updates his recommendations for vitamin B12 supplementation in response to new findings.

In all of the cohort studies, and perhaps most intervention studies carried out on vegetarians, there is little doubt that only very few vegetarians were actually consuming a diet predominantly based on whole plant foods, and as expected although these vegetarians experienced a significantly lower risk of coronary heart disease than their omnivorous counterparts, they still experienced a substantial residual risk of coronary heart disease.32 In Dr. Caldwell Esselstyn’s more recent decade long study (pending publication) of around 200 patients that were advised to consume a whole foods, plant-based diet, it was found that recurrent cardiac events only occurred in 0.5% of adherent participants. This is an approximately 40 fold lower risk than achieved in other dietary or statin based trials, strongly suggesting that these results can only partially be explained by the use of LDL cholesterol lowering medication [reviewed previously]. This is an excellent example of how a whole foods, plant-based diet can confer significant benefit over-and-above favorable changes to traditional risk factors.

Caldwell Esselstyn on making heart attacks history

The definition of a vegetarian diet typically only defines which type of animal foods are restricted and not the quantity and quality of plant foods consumed. As all vegetarian diets are not created equal, studies on vegetarians may only provide limited information of the influence a more nutrient dense vegetarian dietary pattern on the risk of coronary heart disease.33 The restriction of certain animal foods however may encourage at least a modest increase of intake of high quality plant matter, including fruits, vegetables, whole grains, legumes and nuts in order to make up for calories and certain nutrients otherwise consumed from animal foods. Nevertheless, even the studies examining less than optimal vegetarian diets may contribute more to the knowledge of optimal dietary patterns than many studies on homogenous populations due to greater differences in intake of specific foods and nutrients. Vegetarian diets should be designed according to not only which animal foods are restrict, but also the quality of plant foods consumed in order to minimize and preferably eliminate the risk of developing coronary heart disease. There is very strong evidence that such a diet would also lower the risk of numerous other chronic and degenerative diseases.

Diet-Heart Posts

Part I - Diet-Heart: A Problematic Revisit
Part II - Diet-Heart: Saturated Fat and Blood Cholesterol
Part IV - Cracking Down on Eggs and Cholesterol
Part V - Cracking Down on Eggs and Cholesterol: Part II

Please post any comments in the Discussion Thread.