Take the 2-minute tour ×
Skeptics Stack Exchange is a question and answer site for scientific skepticism. It's 100% free, no registration required.

According to Wikipedia,

Vegetarians tend to have lower body mass index, lower levels of cholesterol, lower blood pressure, and less incidence of heart disease, hypertension, type 2 diabetes, renal disease, osteoporosis, dementias, [and lower risks of] cancers of the esophagus, liver, colon, and the lungs. [...] A 2010 study [showed] vegetarians scored lower on depression tests and had better mood profiles.

Of course, correlation =/= causation - all of this can be explained by the fact that people who eat only (or just more) vegetables tend to be more health-conscious in other ways: eat better, sleep more, exercise more, less likely to smoke or drink alcohol/caffeine, etc.

Are there any studies which show positive causal effects of giving up meat, or is this all just correlational propaganda?


I always hear vegetarians talk about how when they gave up meat, they suddenly felt their body and mind working so much better. I've been eating a strict vegetarian diet for three weeks now, and the only thing that's working more are my bowels.

share|improve this question
7  
By way of comment, you'll find no end of anecdotes either way: this is one of the better writeups I've seen of someone ceasing a vegan diet for health reasons. My own attempt at veganism latest 12 months, followed by vegetarianism for a further 6 months. I now eat a small amount of meat products most days. Health effects (for me) of veganism included a general lack of energy and being mildly sick often (where previously I was very rarely). This is just my own experience, I must stress. –  ropable Apr 5 '11 at 6:42
1  
The hypothesis I would like to test would be that reducing meat/processed food has a beneficial effect, but total veganism is not warranted. (i.e. The benefits of veganism are negligible compared to a low meat diet, and the adverse nutritional and emotional effects would overcome any benefits.) Answered as comment because I really don't feel like researching this. If you have any evidence, go ahead an post it as an answer. –  Chris Cudmore Apr 5 '11 at 14:27
13  
This seems like a case of "if you can't prove what you wish to prove, prove something else and pretend they are the same". The assertion is that a vegan diet is more healthy; the evidence cited (even accepting that correlation is causation in this case) is that the incidence of certain diseases is lower. It doesn't show that health overall is better. After all, shooting yourself in the head would result in "lower risks of cancers of the oesophagus, liver, colon, and the lungs". –  Brian Hooper Apr 5 '11 at 20:58
2  
More healthy than eating ice cream and steak all day, or more healthy than a well-balanced diet? I bet you it's more healthy than the first, but less healthy than the second... –  Sklivvz Jan 8 '12 at 9:17
3  
@ChrisW yes, you're right. Cultural knowledge plays a big part. However, there is a problem of malnutrition in South Asia as also chronic diseases like cardio and diabetes. Not sure how much of this is an effect of vegetarianism, high milk consumption and sedentary lifestyle. –  phaedrus Aug 22 '13 at 12:43

2 Answers 2

There is a review titled "Nutrition concerns and health effects of vegetarian diets" published in the journal Nutrition in cinical practice. They summarize the research on nutrition concerns and possible health benefits associated with a vegetarian diet. One factor that makes a comparison hard is that there exists a large variety of eating patterns that fall under the label vegetarianism.

On the subject of cardiovascular disease they state

Compared with nonvegetarians, vegetarians (both lacto-ovo vegetarians and vegans) have a lower risk of death from ischemic heart disease, even after adjustment for body mass index (BMI) and smoking habits.

A vegetarian diet is associated with lower cholesterol levels, which might be a partial explanation of this observed effect. Also the lower average BMI of vegetarians seems to play a role.

One of the cited studies called "Hypertension and blood pressure among meat eaters, fish eaters, vegetarians and vegans in EPIC– Oxford" examined blood pressure in 11,000 men and women in the UK. They took different nutrient intake and lifestyle differences, e.g. exercise, smoking and alcohol consumption, into account.

A vegetarian diet also seems to be associated with a lower Body-Mass-Index

BMI values are reported to be higher in nonvegetarians compared with vegetarians for both men and women, and BMI values tend to increase as the frequency of meat consumption increases

On the subject of diabetes they state

Vegetarians have significantly lower rates of developing type 2 diabetes than do omnivores.

This can be partly explained by the lower average BMI for vegetarians.

There are also studies listed in the review that show a lower risk for certain cancers. For osteoporosis there were no significant differences found between vegetarians and omnivores.

But you also have to take possible nutrient deficiencies into account, especially for the more restrictive vegetarian diets like veganism. The authors of the review conclude

Appropriately planned vegetarian diets are healthful and nutritionally adequate and are beneficial in the prevention and treatment of certain chronic diseases. Poorly planned vegetarian diets can be deficient in vitamin B12, calcium, vitamin D, zinc, iron, and long-chain ω-3 fatty acids.

share|improve this answer
15  
But once again, all of this is just meaningless correlation; it is quite likely that many of the people observed are not healthier because they are vegetarians, but vegetarians because they are healthier - they also likely exercise more, sleep more, eat less junk-food, visit the doctor more, etc. etc. etc. A meaningful study would be to take a large number of people, group them into similar exercise habits/sleep habits/eating habits (other than meat-eating)/doctor visits/ethnicity/social class/place of residence/etc., and then compare (cont) –  BlueRaja - Danny Pflughoeft Apr 7 '11 at 19:51
1  
@BlueRaja Clinical trials usually take known confounding factors into account. That is obviously not perfect and does not work for unknown factors, but the evidence here is certainly more than just correlation. I've added a bit about one of the studies the review is based on. –  Fabian Apr 7 '11 at 20:10
2  
@BlueRaja If I read that correctly, the values are after correction for the mentioned factors. They mention in the conclusion that lifestyle factors seem to play a small role. But a large effect really seems to come from the BMI difference between vegetarians and omnivores. –  Fabian Apr 7 '11 at 21:01
2  
@BlueRaja, the data shows that the most important factor is by far the BMI, and that alcohol intake, exercise and macronutrients also plays a role. It shows that those other factors have an effect, but the BMI effect is larger. The conclusion of the study is that the lower blood pressure of vegetarians/vegans is mainly due to the difference in BMI. –  Fabian Apr 7 '11 at 21:24
3  
But BMI is not a cause itself, it's a result. Whatever causes a person to have a healthy-BMI is likely the cause of their healthy-everything-else. So what's the underlying factor? How much of their general healthiness is influenced by exercise and eating (in all other ways) well, and how much is influenced by not eating meat? –  BlueRaja - Danny Pflughoeft Apr 7 '11 at 21:28

From your comments, I am guessing that you will be a hardened skeptic. You've come to the right place.

Let me address a couple of points. First, I agree with you that correlation is not causation. But my guess is that you are not looking for causation (for example, the biochemical processes that causes a higher incidence of heart disease in meat eaters). After all, who knows what causes heart disease? There are experts who might guess, but I don't think any researcher has come up with a single, definitive cause.

(EDIT start: heightening the contrast between correlation and causation.) Let's turn our attention to the causation / correlation contrast for a different health-related matter: Smoking (perhaps unhealthy) verses Non-Smoking (perhaps healthier).

Here is a proposition, turned into a parallel of the vegan/healthier argument. Does non-smoking cause not-lung-cancer? OR is non-smoking correlated to not-lung-cancer? (I apologize for the Aristotelian flip here, but just think "absence of lung cancer" as a parallel to "healthier.)

To ask for causation is to ask too much. Some non-smokers get lung cancer. See http://carcin.oxfordjournals.org/content/24/5/875.short. Therefore, there is no causal link between not-smoking and absence of lung cancer.

In contrast, science has asserted that smoking is correlated with an increased incidence of lung cancer. Now that is a different question. Even a skeptic would say there is a correlation. Non-smoking correlates to lower lung-cancer incidence.

It would be scientific, but unethical, to cage humans and force one group to smoke and not have another group smoke. It would be scientific, but unethical, to cage humans and force-feed one group meat and force-feed another group vegetables. (EDIT ends.)

A scientifically accepted way to test humans is to do an epidemiological study. One type of epidemiological study would be to monitor many people over a long period of time.

It turns out that protein intake has been studied for both monozygotic (identical) twins and vegetarianism for an epidemiological survey of a homogeneous group.

The first study is called “Dietary protein and blood pressure in monozygotic twins” by Richard J. Havlik M.D, et. al, published in Preventive Medicine (Vol 19, Issue 1, pp. 31-35). It shows that

Using differences in monozygotic twins, a direct association of dietary protein intake and diastolic blood pressure was identified and persisted after adjustment for known covariates of blood pressure. Adjusting for known covariates and holding total calories constant, a 9-g difference in daily protein intake was directly associated with a 1 mm Hg difference in diastolic blood pressure.

Simply put, what they found was that after adjusting for differences other than genetics (because monozygotic twins have the identical genetic makeup), more protein correlates to higher blood pressure.

As you have noted, vegetarian diets tend to be poorer in protein. You might object that this study shows a relationship between protein and BP, not meat and BP. All that an inveterate skeptic would conclude is that this study shows a link between vegetarian diets and lower BP, but perhaps not a strong link. On to the second study.

The epidemiological study may be found in Am J Clin Nutr September 1988 vol. 48 no. 3 739-748, which you may read in full, online. It compares a large number (n=27,529) of Seventh-Day Adventists. The Adventists are prohibited from using tobacco, alcohol, and pork. They are discouraged from consuming other meats, fish and eggs. Because eating a vegetarian diet is optional, this made for an ideal epidemiological study in that the meat-eaters and the vegetarians have similar life styles and all live in the same US state (California).

Before I discuss the study’s conclusions, let me observe that it addresses almost all of your objections. It is a large study. It draws scientific conclusions, based on well known statistical methods. It minimizes lifestyle differences and geographic differences of the participants. It was published in a peer-reviewed journal. (The reviewers are, in a sense, skeptics-for-hire. Their job is to ferret out any design flaws, incorrect inferences, and confounding factors.)

The conclusion is:

Within this population, meat consumption was positively associated with mortality because of all causes of death combined (in males), coronary heart disease (in males and females), and diabetes (in males). Egg consumption was positively associated with mortality because of all causes combined (in females), and cancers of the colon (in males and females combined) and ovary. Milk consumption was positively associated only with prostate cancer mortality, and cheese consumption did not have a clear relationship with any cause of death.

“Positively associated with” means more x correlates with more y. (x = meat, eggs, milk; y=death for all causes, death by colon cancer, death by heart disease). If you delve into the article, you will see that the tests are very statistically significant. For the case of all-cause mortality in males, the significance (p < .0001) means that only in 1 case out of 1000 would you find that the results were due to chance.

This significance is much higher than that you would see for drug trials, for instance.

The author, Dr. Snowdon, has done his job of showing correlation. (Remember, the standard is to show correlation to a given significance, not to show causation.) He even has a section about the limitations of the study. (There were three: 1) Gathering a lot of data means a simpler survey; 2) Dietary habits may have changed since the study began; 3) He did not study what substituted for meat (e.g., fruits and vegetables substituted for meat may be the cause for lower mortality).)

His only grant was from the US National Cancer Institute, so the paper was not sponsored by PETA or the Dairy industry, for example.

If you compare the scientific studies of smoking correlating with lung cancer (for example, http://cancerres.aacrjournals.org/content/54/22/5801.short), you will see that for some studies, the statistical significance levels are still very high (P = 0.001), but not as high as the Adventist study. The relationship between eating meat and these major diseases is more strongly linked.

share|improve this answer
5  
This is a great answer. I really enjoyed reading it and noticed how much care you took in compiling the facts a providing background information. –  Ferdinand Beyer Jul 27 '11 at 14:26
    
Thank you, @Ferdinand. The OP was clearly seeking causation, so I needed to explain why medical journals go for correlation. I threw in a little statistics, as well, to show the strength of the Seventh Day Adventist study. –  rajah9 Jul 27 '11 at 14:42
1  
You should note that usually the proposed effects of diet to various diseases is fairly small. Compare this to smoking & lung cancer where the risk for smokers is 15-30 times compared to non-smokers. cdc.gov/cancer/lung/basic_info/risk_factors.htm –  Illotus Jan 10 '12 at 23:14
6  
"Some non-smokers get lung cancer. [...] Some smokers do not get lung cancer. Therefore, there is no causal link." I have seen this unconventional use of the word "causal" once before on Skeptics.SE, and I remain bewildered by it. By the same argument, being shot in the head doesn't cause death, it is only correlated. Huh? –  Oddthinking Jan 11 '12 at 5:20
3  
Yes, I agree that the OP is asking for a particularly difficult level of proof. If we took a sufficiently large sample of humans and randomly assigned them diets (and then locked them up in cages to ensure they didn't cheat) for their entire lives, we could get what he is asking for. –  Oddthinking Jan 13 '12 at 5:39

protected by Oddthinking Jan 13 '12 at 15:27

Thank you for your interest in this question. Because it has attracted low-quality answers, posting an answer now requires 10 reputation on this site.

Would you like to answer one of these unanswered questions instead?

Not the answer you're looking for? Browse other questions tagged or ask your own question.