Much ado is made about vegetarians having a lower all-cause mortality or CHD mortality than non-vegetarians. Suffice it to say, this is all unfounded. No randomized controlled trial has ever been conducted comparing meat-eaters with vegetarians, so on what grounds do people conclude that meat is causally associated with disease? There is none, any argument for the superiority of a vegetarian dietary regimen is unsubstantiated. In this article, I systemically assess the literature we have so far and debunk the much propagated myth that vegetarians live longer than omnivores .
Let's start with CHD mortality. Take for example this meta-analysis published in 1999 (1). It looked at 5 long-term prospective studies which are some the best we have involving vegetarian and non-vegetarian diets.
For those who don't know, these were observational studies which can NEVER prove causation because variables are never manipulated.
Firstly, this study says itself that there is no significant difference in all-cause mortality between vegetarian and non-vegetarian diets. What was found though was vegetarians had a 24% lower mortality from CHD than non-vegetarians. And they did adjust for the usual confounding factors. Oh no! It's game over, meat increases your risk of developing CHD! Hardly. It's generally impossible to account for everything that could be influencing health status. In observational studies, there's a problem called "residual confounding" where diet and lifestyle factors (which are impossible to measure perfectly) add up and can significantly sway the results of a study, even when researchers try their best to adjust for the variables with statistics.
And here's a good example of how hard it is to adjust for confounding with meat consumption. There was a well-publicized study called "Meat Intake and Mortality" (2) a few years ago that followed over 500,000 people for a decade, and found that red and processed meat was associated with increased risk of heart disease and cancer even after all the common confounders were adjusted for. But the full text reveals something quite surprising. Even in the most sophisticated statistical models, red meat had a stronger association with diseases like the flu, hepatitis B, liver disease, HIV, pneumonia, etc. than it did with heart disease or cancer! Theoretically, those diseases should have a neutral (no correlation) relationship with meat intake, if the statistical models had truly adjusted for all the confounders. The fact that meat had a stronger relationship with infectious diseases and liver disease (from alcoholism or hepatitis) than it did with chronic disease indicates the interference of things like socioeconomic status, work environment, living conditions, etc. that increase disease risk. This shows us that the researchers couldn't completely account for confounding, even though they made adjustments for things like exercise, smoking, drinking, etc.
The other problem with observational studies in general is that the food-frequency questionnaires they use to gather diet data are notoriously inaccurate; by some estimates, meat intake is only adequately gauged about 20% of the time, because of how people report (or fail to report) the things they've eaten.
No doubt vegetarians are healthier than the general population, but what has to be determined is whether that health boost is related to eschewing meat. The only way to study this would be with a RCT, which as I've said, has not been done yet.
So now let's look at all-cause mortality. I'll review in adequate detail every single long-term study comparing vegetarians with non-vegetarians.
Before we do this though, it's imperative to note that each study has the same problem; vegetarians also have much more healthful dietary habits; they tend not to smoke and limit their consumption of alcohol, participate in regular physical exercise and also tend to be less stressed (3). What this equals is a nightmare of confounding factors and covariate imbalances for researchers to adjust (which can never be done perfectly, leaving residual confounding factors which when aggregated can sway results significantly).
Let's start with the Adventist Health studies (4). The study found that vegetarians had lower risks of hypertension, diabetes, arthritis, colon cancer, prostate cancer, fatal CHD in males, and death from all-causes. But as previously discussed, non-vegetarians exhibit a wide range of other healthy lifestyle and dietary changes unrelated to meat intake. And no doubt this was exactly the case in this study, where many of the vegetarian Adventists practices these healthful habits that were not seen in their omnivorous brethren. Vegetarians also consumed more nuts and fruit, they drank less coffee and ate less donuts. Moreover, non-vegetarian Adventists consumed alcoholic beverages twenty-times more than their vegetarian counterparts (5), (6).
And next, The Health Foods Shoppers Study (7). There was no difference found in all-cause death rate among vegetarians and omnivores after 17 years.
Along to the Oxford Vegetarian Study (8) which found a 20% reduction in overall mortality in vegetarians after 12 years. A more recent follow-up study found the reduction in overall mortality had disappeared. The only significant difference remaining for any cause of death was seen for mental and neurological diseases, which were 2.5 times higher in the vegetarian group (9).
And a slightly more recent one, the EPIC-Oxford Study (10). No difference in overall mortality between vegetarians and omnivores after 6 years was seen. Worth noting is that vegetarians displayed slightly higher mortality from all cancers and stroke.
Another study, this time by the German Cancer Research Center (11). They followed 1900 vegetarians, and after 11 years, 225 had died. Because 470 deaths would have been expected in a normal (omnivorous) German population, this study is usually cited ad nauseum as support for vegetarianism. However once again, meat was not to blame. For example, when the researchers contrasted mortality among strict vegetarians, those who only occasionally ate meat, there was similar CVD, cancer and all-cause mortality rates between the two groups. Moreover, only about 4% and 3% of males and females respectively, were smokers in the study, whereas the corresponding figures for the rest of Germany were 41% and 26% respectively. After you take all this into consideration, its no wonder that vegetarians has such a lower mortality. And whats more, when authors examined the effect of various confounding factors they found that the far strongest was the level of physical activity (12).
So what happens when confounding variables are kept to a total minimum? Consider a female Bhuddist population in Taiwan (13) Same situation with the Adventists, except health-consciousness is much the same between vegetarians and non-vegetarians. The researchers state, "Most western vegetarians include fresh vegetables and fruits as their main source of nutrition and energy, based on health benefits of the foods. In contrast, most Taiwanese vegetarians choose a vegetarian diet because of their Buddhist religion, which teaches a policy of “no killing.” Buddhists in Taiwan have a dietary pattern similar to that of most Taiwanese in terms of meal patterns and cooking methods, except that they do not include any meat, fish, or poultry in their meals." So what did they find? They found the vegetarians have lower serum lipoprotein-associated phospholipase A2 (Lp-PLA2) which is a marker for vascular inflammation, but higher levels of C-Reactive Protein (CRP), homocysteine (HcY) and triglycerides (all serious risk factors for heart disease). So when confounders are eliminated or kept to an absolute minimum, vegetarians do not fare better than non-vegetarians.
Another showing the same thing, this time in China (14). Researchers looked at long-term vegetarians and compared heart disease rates against an omnivorous control group. The results were surprising. Vegetarians had considerably thicker arterial walls, higher systolic blood pressure, higher triglycerides and higher HcY. The authors concluded "In summary, contrary to common belief, vegetarians, at least in the Chinese, might have accelerated atherosclerosis and abnormal arterial endothelial function, compared with omnivore control subjects."
And another looking at an omnivorous Californian population who were healthier-than-usual (15). They found a 53% lower cancer mortality, a 48% reduction in cardiovascular deaths, a 53% lower all-cause mortality than the rest of the Californian population. For those who did not smoke, exercised regularly, and got proper sleep, the reductions were even more impressive. Cancer, cardiovascular and all-cause mortality were reduced by 66%, 86% and 78%, respectively.
- Nemo
Let's start with the Adventist Health studies (4). The study found that vegetarians had lower risks of hypertension, diabetes, arthritis, colon cancer, prostate cancer, fatal CHD in males, and death from all-causes. But as previously discussed, non-vegetarians exhibit a wide range of other healthy lifestyle and dietary changes unrelated to meat intake. And no doubt this was exactly the case in this study, where many of the vegetarian Adventists practices these healthful habits that were not seen in their omnivorous brethren. Vegetarians also consumed more nuts and fruit, they drank less coffee and ate less donuts. Moreover, non-vegetarian Adventists consumed alcoholic beverages twenty-times more than their vegetarian counterparts (5), (6).
And next, The Health Foods Shoppers Study (7). There was no difference found in all-cause death rate among vegetarians and omnivores after 17 years.
Along to the Oxford Vegetarian Study (8) which found a 20% reduction in overall mortality in vegetarians after 12 years. A more recent follow-up study found the reduction in overall mortality had disappeared. The only significant difference remaining for any cause of death was seen for mental and neurological diseases, which were 2.5 times higher in the vegetarian group (9).
And a slightly more recent one, the EPIC-Oxford Study (10). No difference in overall mortality between vegetarians and omnivores after 6 years was seen. Worth noting is that vegetarians displayed slightly higher mortality from all cancers and stroke.
Another study, this time by the German Cancer Research Center (11). They followed 1900 vegetarians, and after 11 years, 225 had died. Because 470 deaths would have been expected in a normal (omnivorous) German population, this study is usually cited ad nauseum as support for vegetarianism. However once again, meat was not to blame. For example, when the researchers contrasted mortality among strict vegetarians, those who only occasionally ate meat, there was similar CVD, cancer and all-cause mortality rates between the two groups. Moreover, only about 4% and 3% of males and females respectively, were smokers in the study, whereas the corresponding figures for the rest of Germany were 41% and 26% respectively. After you take all this into consideration, its no wonder that vegetarians has such a lower mortality. And whats more, when authors examined the effect of various confounding factors they found that the far strongest was the level of physical activity (12).
So what happens when confounding variables are kept to a total minimum? Consider a female Bhuddist population in Taiwan (13) Same situation with the Adventists, except health-consciousness is much the same between vegetarians and non-vegetarians. The researchers state, "Most western vegetarians include fresh vegetables and fruits as their main source of nutrition and energy, based on health benefits of the foods. In contrast, most Taiwanese vegetarians choose a vegetarian diet because of their Buddhist religion, which teaches a policy of “no killing.” Buddhists in Taiwan have a dietary pattern similar to that of most Taiwanese in terms of meal patterns and cooking methods, except that they do not include any meat, fish, or poultry in their meals." So what did they find? They found the vegetarians have lower serum lipoprotein-associated phospholipase A2 (Lp-PLA2) which is a marker for vascular inflammation, but higher levels of C-Reactive Protein (CRP), homocysteine (HcY) and triglycerides (all serious risk factors for heart disease). So when confounders are eliminated or kept to an absolute minimum, vegetarians do not fare better than non-vegetarians.
Another showing the same thing, this time in China (14). Researchers looked at long-term vegetarians and compared heart disease rates against an omnivorous control group. The results were surprising. Vegetarians had considerably thicker arterial walls, higher systolic blood pressure, higher triglycerides and higher HcY. The authors concluded "In summary, contrary to common belief, vegetarians, at least in the Chinese, might have accelerated atherosclerosis and abnormal arterial endothelial function, compared with omnivore control subjects."
And another looking at an omnivorous Californian population who were healthier-than-usual (15). They found a 53% lower cancer mortality, a 48% reduction in cardiovascular deaths, a 53% lower all-cause mortality than the rest of the Californian population. For those who did not smoke, exercised regularly, and got proper sleep, the reductions were even more impressive. Cancer, cardiovascular and all-cause mortality were reduced by 66%, 86% and 78%, respectively.
So there you have it, every single long-term study done on vegetarians and non-vegetarian mortality. Do vegetarians have more favorable mortality outcomes? No. Any claim made for it is false and I have just laid bare the science to prove it. What we need now is a RCT to definitively tell us whether meat is causally associated with disease. Until then, don't believe any of this vegetarian hogwash.
- Nemo
http://www.nature.com/bjc/journal/v101/n1/pdf/6605098a.pdf
ReplyDelete