The skins crawled, and the meat that was stuck on the spits bellowed,
both roast and raw, and the noise was like the lowing of cattle.
– Homer
The first question is: healthy compared to what? To a lacto-ovo-vegetarian diet? or to a standard Western diet? or to the mean world diet, if that is even a coherent concept? The papers I will cover in this post don't all answer the same question, so I'll try to be clear in pointing out what they are comparing to. They also differ in which non-carnivorous diets they are looking at. There are many varieties of vegetarian & even vegan diets. Since there aren't that many studies on vegan diets specifically, I'll also be looking at studies of vegetarian diets. Why? Because I think it's likely (though not guaranteed) that any positive or negative effects in vegetarian diets are also present in vegan diets (though there might be positive or negative effects from vegan diets that aren't produced by vegetarian diets). I'll make these distinctions, too, where relevant.
The other thing I'll note before we get underway is that vegans & vegetarians are strongly self-selected groups that differ significantly from the general population. That means there are a lot of possible confounds here. If we find that vegans are on average healthier than non-vegans, we can't infer that it is the vegan diet that makes them healthier, because correlation is not causation. It could just be that vegans exercise more or smoke less, for example, & that it is the exercise or the lack of smoking, not the diet, that makes them healthier. What that means is that we must look at observational studies with a critical eye. Instead, wherever possible, I will be citing randomised controlled trials (RCTs), where the variable being studied (in this case, a particular diet) is isolated so that causality can be inferred.
Just one final word of warning. I am myself a vegetarian leaning vegan, my wife is vegan & I believe that eating vegan is morally right, at least for the vast majority of humans. So it would be good for me if vegan diets were healthy, or at least risk-free. That said, I'll do my best to be fair-minded & to let the evidence speak for itself.
Cardiovascular Disease
If the Four Horsemen of the Apocalypse in the Christian tradition were Death, Famine, War & Conquest, today they would have had to have been named Heart Problem, Cancer, Respiratory Disease & Diabetes. But Heart Problem would be the ruler of them all, which is probably why there are so many studies on veganism/vegetarianism & cardiovascular disease.
High body mass index, high blood fat, high blood pressure & diabetes all increase the risk of heart problems, stroke & so on. In a meta-study from 2018, Benatar & Stewart looked at these risk factors among vegans. They included 40 observational studies, mostly done in the West but with a considerable number also from Taiwan. In what will be a bit of a recurring theme, they found different results in the West & in Asia. In Western studies, vegans had lower body mass index, lesser waist circumference, reduced blood sugar levels, lower low-density lipoprotein (protein that transfer lipids around the body; fats are a kind of lipid) cholesterol (a type of lipid), less body fat & lower blood pressure than omnivores. But in the Taiwanese studies these differences were much smaller or vanished completely. Other observational studies seem to produce similarly significant results when comparing with Western diets.
But again, these are not RCTs. I don't think these studies even control for confounds, or at least the authors of the meta-studies don't mention that. They just do some subpopulation analyses. But as mentioned before, if self-selected vegans tend to exercise more, smoke less, eat less junk food & afford healthier lifestyles generally, that might be enough to explain the difference.
Fortunately, there are a bunch of RCTs made on veganism/vegetarianism & cardiovascular disease.
Yokoyama et al. did a meta-study (2014) of seven controlled trials (of which six were RCTs) & 32 cross-sectional (examining population differences at a single point in time) studies on vegetarian diets & blood pressure. High blood pressure is a risk factor for heart failure & ischemic heart disease among other things. Heart failure is where the heart's not able to pump blood as effectively as needed. Ischemic heart disease is when deposits build up in the arteries of the heart, reducing blood flow & causing heart attack & other issues. They found that vegetarian diets were associated with decreased blood pressure both in the controlled trials & in the cross-sectional studies. A 2017 meta-study by some of the same authors, reviewing 30 observational studies & 19 clinical trials, found that a vegan diet was associated with & caused a reduction in three out of four markers for blood lipids.
The cross-sectional studies in the earlier Yokoyama et al. meta-study were diverse geographically, but the controlled trials were all done in the West. In fact, most of the studies I've found use different varieties of Western diets as control group, not because of researcher prejudice, but simply because most of the relevant research has been made in the West, particularly in the United States. We'll be seeing indications that results in the West don't generalise, so I will keep pointing this out.
López et al. did a meta-study (2019) of 11 RCTs, also looking at blood pressure. They found evidence that vegan diets don't result in decreased blood pressure. As the authors note, this is surprising given that the Yokoyama et al. study showed decreased blood pressure with a vegetarian diet. However, most of these studies had a high (50+) mean age & many were of diabetics or dyslipidemics (people who have a lot of lipids in their blood), meaning the result may not generalise. What's more, some of the studies covered here use various "good" omnivorous diets as the baseline case, e.g. diets recommended by medical societies & so on. The authors write: "Throughout the studies, all participants were advised to reduce their red meat intake, increase their vegetable intake, and lose weight."
In a 2020 meta-study of 15 RCTs, mostly from the West, Lee et al. found that vegetarian diets reduced blood pressure. They also found, in a subgroup analysis, that vegan diets were more effective in reducing blood pressure than were vegetarian diets. (I don't think they correct for multiple comparisons, but this result was significant even after I made a Bonferroni correction.) Once again the effect was stronger in U.S.-based studies (which makes sense given the pretty abysmal diet of the median American).
Viguiliouk et al. conducted a meta-study (2018) of 9 RCTs, mainly done in the West, all of which looked specifically at diabetics. The control diets were a pretty even mixture of normal diets & conventional diabetes diets. They found that vegan diets were really good on body weight & possibly better on blood fats & blood sugar level. But there was no effect at all on blood pressure. That makes me think that the Lòpez et al. meta-study, which found that vegan diets don't result in decreased blood pressure, found no improvement precisely because six of its 11 included studies were of diabetics. Actually, having just now gone back to that study, I see the authors commenting that "because major clinical trials and a recent meta-analysis have noted differences in hypertension outcomes between diabetic and non-diabetic patients, this could have influenced our findings".
Verdict: I'm pretty confident that vegan diets are better for the heart than common Western diets, though the difference seems to be smaller for diabetics & may not exist at all between vegan diets & common non-Western diets or healthy omnivore diets generally.
Cancer
If death worldwide was an Olympic sport, cancer would be a silver medallist. (It's worth pointing out, though, that cancer is not a leading cause of death in much of the developing world, where infectious disease is more common.) Reviewing the health effects of a vegan diet, Craig (2009) notes that fruit, vegetables, legumes, fibre, soy & vitamin C are all protective against cancer. He also writes:
Vegan diets are usually higher in dietary fiber, magnesium, folic acid, vitamins C and E, iron, and phytochemicals, and they tend to be lower in calories, saturated fat and cholesterol, long-chain n–3 (omega-3) fatty acids, vitamin D, calcium, zinc, and vitamin B12.
There's apparently some evidence that vitamin D shortage, which is common among vegans, is associated with increased risk of cancer. But in theory the vegan diet has a lot going for it here. Does the evidence bear that out?
I wasn't able to find many studies on population-level epidemiology here, maybe because it seems pretty vain to assign a bunch of people vegan diets, wait ten weeks or so & see how many people got cancer in that time. I suppose maybe there's no reliable biomarker for cancer, like blood pressure is for cardiovascular disease. That said, Craig writes that vegetarians have lower risk of prostate & colorectal cancer & that, because obesity increases the risk of cancer, vegans' lower body mass index may be protective.
A meta-study from 2017 found in ten prospective (selecting subjects before their conditions are known) cohort (following subjects over time) studies that vegetarianism & especially veganism were associated with a lower overall risk of cancer. In a 2015 review of long-term health effects of vegans/vegetarians, looking at eight large prospective cohort studies, all done in the West, Appleby & Key found that "[o]verall cancer rates may be slightly lower in vegetarians, but the data are inconclusive for most common individual cancers". And of course, though the nutritional facts give us some hints, it's not clear that vegan diets actually cause these small effects.
Verdict: Inconclusive, but it seems probable that veganism isn't harmful & might even be somewhat beneficial?
Inflammatory Disease, Including Diabetes
Chronic inflammatory diseases apparently include things like type 2 diabetes, heart disease & cancer. There's some evidence that inflammatory biomarkers are linked or contribute to these chronic diseases. A biomarker is just something in the body that you can measure. Menzel et al., in a 2020 meta-study of 21 cross-sectional studies, look at associations between veganism/vegetarianism & abnormal levels in these inflammatory biomarkers. They found that vegan & vegetarian diets are associated with lower C-reactive protein levels (this being a major inflammation marker). They found no differences in the many other biomarkers studied, though they caution that the important ones (except for C-reactive protein) haven't been studied much yet.
Of course, correlation does not equal causation. And the included studies were mainly carried out in Asian countries. As of now, there are no RCTs on veganism & inflammatory biomarkers. If you are in the business, make one!
Nor are there any RCTs on vegan/vegetarian diets & diabetes – we had seen many studies of diabetics, but so far we haven't looked at the risk of getting diabetes. Lee & Park performed a meta-study in 2017 of 12 cross-sectional & two cohort studies. They found that vegetarianism was associated with not having diabetes, though not in the Asian studies; this result held even after conditioning on body mass index. Though these were all observational studies, they note that other studies indicate that whole grains, fruit & vegetables help prevent diabetes whereas higher consumption of red & processed meat is positively correlated with diabetes risk.
Verdict: There's some evidence that veganism reduces risk of diabetes & other inflammatory disease, though it's hardly overwhelming. Without RCTs, I am uncertain.
Weight Reduction
As we've seen, obesity is associated with a number of diseases, including diabetes but also cardiovascular disease & others. Huang et al. reviewed 12 RCTs in a 2015 meta-study on veganism/vegetarianism & weight reduction. They found that, though vegetarian diets did better than the omnivore diets, vegan diets beat them by a wider margin. Again, this makes sense, because vegan diets are high on whole grains, fruit & vegetables. However, the improvement was attenuated over time. So it's hard to say whether this is a long-term effect (though given that observational studies point in the same direction, it seems probable).
Judging purely from the author names, I assume most of these 12 RCTs were carried out in the West, but the control subjects here were assigned low-fat diets, anti-diabetes diets, weight-reducing diets & so on. So the results carry not only against the standard Western diet but seemingly also against healthier omnivore diets.
Verdict: Veganism seems a good bet to reduce body weight & reduced body weight can be good for many reasons.
Bone Health
There are two main ways of measuring bone health. One is by looking at the bone mineral density; the lower the bone mineral density, the higher the risk of bone fracture. The other is by simply counting bone fractures. Iguacel et al., in a 2018 meta-study looking at 20 observational studies (of which 15 studied bone mineral densities & 5 fracture rates), found that vegans had both lower bone mineral density & higher fracture rates than both vegetarians & omnivores. These studies mostly included women subjects, elderly women being at particular risk of fracture, & were pretty evenly split between Western & Asian countries. They are observational studies, of course, so they are not going to prove causation. However, since vegan diets are known to be low on calcium & vitamin D, & since adequate calcium & vitamin D intakes are the most important ways of preventing low bone mineral density, we can make guesses.
Ho-Pham et al., in a 2009 meta-study, find lower bone mineral densities among vegans, but note that the difference is "clinically insignificant". They don't mention anything about controlled trials, so I suppose these are observational studies.
Craig writes that there's evidence of low bone mineral density among some vegans, specifically those with low protein & calcium intakes. And in the previously mentioned 2015 review of long-term health effects of veganism/vegetarianism, Appleby & Key found that "[b]one fracture rates [...] are higher in vegans [than in non-vegetarians] if they have inadequate intakes of calcium". But bone health also depends on some nutrients, like vitamin D, vitamin K, potassium & magnesium, as well as foods like soy, fruit & vegetables, most of which vegans do well at. Craig mentions several RCTs of menopausal women in particular that showed increased soy intake as improving bone mineral density.
Verdict: There seems to be some risk of vegans having worse bone health than do omnivores & vegetarians, likely because they don't get enough calcium & vitamin D. This can be solved by taking supplements or eating foods fortified with these.
Nutritional Deficiencies
Let me quote Craig again on the nutritional aspect:
Vegan diets are usually higher in dietary fiber, magnesium, folic acid, vitamins C and E, iron, and phytochemicals, and they tend to be lower in calories, saturated fat and cholesterol, long-chain n–3 (omega-3) fatty acids, vitamin D, calcium, zinc, and vitamin B12.
It's pretty well known that cutting out all animal products from your diet puts you at risk of some nutritional deficiencies, iff you don't eat foods fortified with these or take supplements. The main problems are with vitamin D & vitamin B12. Vitamin D deficiency is especially a risk for those vegans who also have other risk factors, like for example living in Lapland, being Black, wearing a burka or niqab & so on – anything that reduces sun exposure.
Vegans are more likely to have vitamin B12 deficiency than both vegetarians & omnivores. A 2013 meta-study found a significant association with veganism/vegetarianism & vitamin B12 depletion & deficiency. It looked only at observational studies, but given that vitamin B12 is found almost exclusively in animal products, I think I can go out on a limb here & guess that the culprit is the diet. A cross-sectional study from 2017 of over 90,000 participants (about three percent of whom were vegetarian or vegan) looked at nutritional intake among self-reported vegans/vegetarians:
[V]egans exhibited a higher estimated prevalence of inadequacies for some nutrients, in particular vitamin B12 (69.9% in men and 83.4% in women <55 years of age), compared to meat-eaters. [...] Our study highlighted that, overall, self-reported vegetarians and vegans may meet nutritional recommendations.
Sebastiani et al. reviewed (2019) the evidence for effects of vegan diets on mothers & offspring during pregnancy. Nutrition is obviously essential for pregnant women, as deficiency can lead to chronic disease & other lifelong health issues in the child. Many of the studies reviewed were done on relatively poor populations in Asia & Africa & may not generalise; few studies were RCTs. Their opinion seems to be that vegan diets are fine, so long as one gets enough of the necessary nutrients (proteins, vitamin B12, vitamin D, calcium, omega-3 & iron are those that vegans risk not getting enough of). If one does get the right nutrients, there's some evidence that vegan diets can even reduce the risk of complications in pregnancy.
Verdict: Vegans run the risk of not getting enough of some nutrients, especially vitamin D & vitamin B12. This can be solved by taking supplements or eating foods fortified with these.
Fatigue & Cognitive Function
So much for the stuff that kills you. One of the more common complaints I hear anecdotally about veganism is that makes you tired or enervated or unable to focus. This seems less important than not dying to me, but sure, it's no fun not having any energy.
Unfortunately, I haven't been able to find many studies on this. Medawar et al. reviewed (2019) the evidence for effects of vegan/vegetarian diets on cognitive function among other things. They cover a few observational studies the results of which I suspect can be explained by selection effects, concluding that:
[T]here is an overall robust support for beneficial effects of a plant-based diet on metabolic measures in health and disease. However, the evidence for cognitive and mental effects of a plant-based diet is still inconclusive.
Verdict: No idea. Please write me if you come across any good studies on this.
Conclusion
Overall, it seems highly probable that vegan diets are healthier than common Western diets, & perhaps also somewhat healthier than good Western & non-Western omnivore diets. However, that is assuming that the vegan gets all the nutrients that they need, especially vitamin D & vitamin B12, of which two vitamins many vegans don't have enough. These can be gained through supplements or by eating foods fortified with them.
Thanks for this post! I'm currently transitioning to a vegetarian diet, and am considering doing or trying a vegan diet in the future. So the title of this post was something I've thought about. I'm glad to see someone compile a literature review on this question.
I guess one question though that health-conscious meat-eaters might still ask after reading this is, "Okay a vegan diet is healthy, but I want to have the healthiest diet. Wouldn't the healthiest diet for humans be one that is not vegan?"
It would probably be hard to find out what is the healthiest diet for humans from existing literature, but I've seen someone argue in this forum that the healthiest diet is probably not vegetarian:
I'm curious to hear if you have any thoughts on this!
Ah, good question. Like the author of your quote, I'm also not a nutritionist, nor am I a medical doctor. That said, I wouldn't be surprised if the healthiest diet did include some animal products. That's because vegan/vegetarian diets optimise for something else – they optimise for the removal of meat & animal products. It shouldn't be surprising that a diet optimising purely for health might be better than (& different from) one that optimises for something else entirely.
I suppose in the end one has to sort out one's motivations in choosing a diet. How much importance do I place on my health versus, say, animal suffering? (Or, in more deontological terms, how do I reconcile the duties I have to myself with those I have to other creatures?) Personally, I would strive to eat vegan/vegetarian even if I learned that it was relatively unhealthy. But I'm well aware that not everyone would do that!
Yeah what you said makes sense. I agree that I think practically everyone who can should go for a vegetarian/vegan diet even if it is not the healthiest one. It's already good enough that it's likely a healthier diet compared to most people's diets. Thanks for your thoughts!
You can optimise for health on a vegan diet as well. The only difference would be any nutrients that are exclusively found in animal products. But, as I stated in my other comment below, I think there are good reasons to believe that it's unlikely there are any such nutrients with non-negligible health benefits, other than those that we are aware of and can be supplemented. The main reason is all the knowledge we collectively have about the nutrients in the human diet (and the compounds which are important for metabolism in the human body).
I know this is very late, but I felt obliged to reply since I disagree with a lot of the points made in the post you quoted.
Unprocessed red meat is still classified as a type 2A carcinogen by the WHO's IARC (International Agency for Research on Cancer), meaning it is 'probably carcinogenic'. Also, fish and other seafood are known to contain relatively high levels of mercury, meaning there are health issues associated with consuming them too much.
Additionally, saturated and trans fats are almost ubiquitous in animal products, often at relatively high levels. Both of these are widely considered to be unhealthy. Indeed, the NAS (Nutritional Academy of Sciences) in the US has stated that there is no safe level of trans fat consumption.
Levels of cadmium (another toxic metal) seem to be typically higher in offal, like liver, than in plant foods (source). This is somewhat unsurprising as heavy metals tend to accumulate in the liver when animals are trying to metabolise them. Also, it seems like if consuming a decent amount of liver would lead to getting far too much of some vitamins and minerals, which is generally not good (e.g. see these composition data).
Humans are very capable of converting ALA (the predominant omega-3 fatty acid found in plant foods) to EPA and DHA (source). It is not hard to get the amount of ALA required for sufficient EPA and DHA, even assuming a somewhat low conversion rate. This amount would typically be a few grams per day. Also, the conversion rate is taken into account in all nutritional guidelines.
Taurine is synthesised in the human body from cysteine. Carnosine is synthesised from histidine, cytosine and uracil. Cysteine and histidine are both found in high amounts in plant proteins (source), while cytosine and uracil, being nucleotide bases, are quite ubiquitous in the body. I think it is reasonable to assume that the human body will produce the amount of taurine and carnosine required for optimal health, even without obtaining them through a diet, and I would place a high prior probability on that. Additionally, the evidence so far suggests that there are not health benefits from having dietary intake or higher levels of these compounds.
For creatine the evidence is a little more mixed, although the majority of evidence still suggests that dietary intake does not provide health benefits. Nonetheless, one may want to supplement it out of an abundance of caution.
I don't think supplementing generally takes more than ~5 minutes extra per day. And generally I find that supplements are a fairly negligible expense. (This is especially the case because some brands have very high amounts of the nutrients in their products and personally I often break up the tablets into smaller proportions.) I can add some more information about costs here if anyone is interested.
I think anyone who can maintain a typical lifestyle is good enough at keeping habits to utilise supplementation. (Not criticising the OP - maybe they just needed a bit more time to get used to it.) Nonetheless, foods fortified with nutrients like vitamin B12 and vitamin D are widely available, so one could simply use those instead. If this also isn't an option for some reason, you could just take larger amounts of the supplements once every few days or even once a week, depending on the elimination half-life of the nutrient. To my knowledge, this would work for vitamin B12 and vitamin D (although not for creatine apparently).
Preservatives and flavourings (and additives in general) are not automatically unhealthy, and indeed most of them are perfectly healthy, as they have to pass rigorous safety standards before being approved (and they are generally studied extensively even after they're approved). As such, store-bought bread and fortified cereals are not unhealthy either, and I would say they're actually good components of a diet. However, if anyone wants to share evidence to the contrary, feel free.
I find that vegetarian or vegan products are often healthier than the alternatives. This is probably because there seems to be a frequent association between plant-based foods and healthiness in marketing, e.g. because companies think the target demographics overlap significantly (which I could imagine being true).
I think this is the strongest argument in the post. However I want to point out that nutrition as a field has existed for a long time and by now we have characterised and studied the majority of nutrients that are present in typical human diets. It seems unlikely that there are nutrients we are unaware of that are exclusively present in animal products and have non-negligible health benefits (from dietary intake). Additionally, there are cohort studies which measure a large number of health outcomes for omnivores and vegetarians and/or vegans, and they haven't identified any particular negative effects from either of the latter (indeed it's usually the other way around). The main disadvantage I can think of there is that these studies don't measure every possible health outcome that may be relevant.
I can understand that there may be a small negative expected value from nutrients lacking in a vegan diet that we are currently unaware of. However I think this is plausibly outweighed by the negative expected value from some of the health effects (e.g. those above) that are associated with most animal products, given that these are more probable and seemingly more serious.
Or, to put this another way, your prior for the best diet containing some animal products might initially be quite high, but in light of the evidence against the healthiness of many animal products, I think the probability becomes quite low.
That being said, I disagree that it takes more time and money to create an optimal (or just reasonably good) vegan diet than an optimal omnivorous diet, for reasons listed above, and because I think the latter is significantly more difficult than one might intuitively believe.