Looking at individual diseases is informative, but it can cause us to become myopic, making broad health-related decisions based on narrow information. It can cause us to miss the forest for the trees. In this case, the "trees" are individual diseases and the "forest" is total mortality: the overall risk of dying from any cause. Does eating meat increase total mortality, shortening our lifespans?
Traditionally-living cultures such as hunter-gatherers and non-industrial agriculturalists are not the best way to answer this question, because their mean lifespans tend to be short regardless of diet. This is due to ~30 percent infant mortality, which drags down the average, as well as a high risk of death in adulthood from infectious disease, accidents, and homicide/warfare. It can also be difficult to accurately measure the age of such people, although there are reasonably good methods available.
However, there are semi-industrialized cultures that can help us answer this question, because they feature a somewhat traditional diet and lifestyle, combined with modern medicine and the rule of law. The so-called Blue Zones, areas of exceptional health and longevity, fall into this category. These include Sardinia, Italy; Okinawa, Japan; Loma Linda, California; Nicoya Peninsula, Costa Rica; and Icaria, Greece.
What do people eat in the Blue Zones? With the exception of one mostly vegetarian Blue Zone in Loma Linda, California, all of them eat meat, and most eat dairy and eggs. However, they tend to eat much less meat than people in affluent nations (though not necessarily less dairy and eggs). Blue Zone diets tend to be starchy, often based on grains, starchy tubers, and legumes.
It's worth keeping in mind that these populations are relatively isolated and may carry genetic variants that promote longevity. Also, they exhibit a complex of healthy behaviors that extends beyond diet, including regular physical activity and social interaction. However, these cultures provide us with a dietary template that is, at the very least, compatible with a long, healthy life.
Let's begin again with the Adventist Health Study of vegetarian Seventh-Day Adventists in California. In the first round of study, male but not female vegetarians had a lower mortality than omnivores (1, 2). In the second round of study, male and female vegetarians had a 12 percent lower mortality rate than omnivores (3). However, lacto-ovo vegetarians, vegans, pescetarians, and semi-vegetarians were all approximately tied, which is not consistent with the idea that eliminating meat is necessary for the survival benefit. In absolute terms, the longest lived were the fish eaters, although the error bars are too large to conclude that the various subgroups are different from one another. As usual, keep in mind that this is an atypical population with a cluster of healthy behaviors, particularly among the most observant, vegetarian Seventh-Day Adventists.
There have been many other observational studies on vegetarianism and mortality, so I'll rely on meta-analyses that compile and analyze these individual studies.
One of the first meta-analyses was published in 1999 (4). Although it's older and only included five studies, it's interesting because it reported results for subgroups of vegetarian and semi-vegetarian diets, including vegans and pescetarians. Overall, being vegetarian was not associated with a survival advantage. However, subgroup analysis suggested that lacto-ovo vegetarians, semi-vegetarians, and pescetarians had a survival advantage, while vegans had exactly the same mortality risk as omnivores.
Consistent with this early meta-analysis, a recent meta-analysis including seven studies also reported no survival advantage for vegetarians, although there was a modest trend toward lower mortality (5).
What about specific types of meat? Some studies have shown an association between red and processed meat consumption and higher mortality risk, most notably the large observational studies from the Harvard School of Public Health (6). A highly publicized study on elderly Americans came to a similar conclusion (7). The two most recent meta-analyses, which considered a broader swath of studies, arrived at a different conclusion. Neither one found a significant association between red meat consumption and total mortality, although there was a modest trend toward higher risk (8, 9). Both confirmed the association between processed meat and mortality.
In general, poultry and fish consumption have neutral or beneficial associations with total mortality (10, 11, 12, 13, 14, 15), and dairy consumption is not associated with mortality (16).
I'll summarize before moving on. Being vegetarian is not associated with a clear survival advantage, nor is being vegan. Some studies have suggested that fresh red meat consumption is associated with increased mortality risk, but the overall literature does not strongly support that conclusion. Processed meat consumption is fairly consistently associated with higher mortality risk. There is no evidence that eating fish, poultry, or dairy is associated with a higher mortality risk. In fact, people who eat fish seem to die a bit less often than anyone else.
It's very unlikely that we'll ever have long-term randomized controlled trials to test these questions.
The most obvious mechanisms by which meat might shorten lifespan are by increasing the risk of the primary killers we've already discussed, including cardiovascular disease, diabetes, and cancer. I won't repeat what I've already written on those topics, except to say that the negative potential of meat appears to be linked specifically to processed and red meat.
Yet there is a larger mechanism that some have suggested may link meat to shorter lifespans, and that is by increasing the activity of an enzyme called mTOR (mammalian target of rapamycin). mTOR is one of two key sensors of cellular energy status, the other being AMPK (AMP-activated protein kinase). My own research has touched on mTOR, as my graduate work involved basic aging research (17).
One of the interesting things about mTOR is that it senses more than just energy abundance. It is also highly sensitive to amino acids, the building blocks of protein. When the body has abundant energy and amino acids, such as after eating a large steak and potatoes, mTOR activity is high. When the body has low energy and amino acids, such as during calorie restriction, mTOR activity is low. As you might predict, when mTOR is activated, it tells the cell to fire on all cylinders because its energy and protein balance sheets are in the black. This favors anabolic processes such as cell division, RNA and protein synthesis.
This brings us to the big hypothesis: firing on all cylinders effectively accelerates cellular aging.
There is some pretty compelling evidence to support this general biological principle. In brewer's yeast-- a surprisingly productive organism for aging research-- removing or inhibiting the yeast equivalent of mTOR extends lifespan* (18). This is thought to be similar to what happens during calorie restriction, which inhibits TOR and extends lifespan in yeast and some other organisms. One of the ways to inhibit TOR is by using a drug called rapamycin.
What about in mammals? At least in certain species under specific conditions, we know that calorie restriction can extend lifespan. Does this relate to lower mTOR activity? As it turns out, rapamycin can be used in mice, as long as it's administered carefully. And it does indeed extend lifespan (19). What else extends lifespan in mice? Low-protein, high-carbohydrate diets (20).
Together, this suggests that dietary protein may be able to increase mTOR activity and accelerate aging-- at least in rodents under lab conditions. Meat is the most concentrated, highest-quality source of protein available. Plant proteins tend to be less concentrated, less completely digested/absorbed, and have a less complete amino acid profile, leading them to be less anabolic.
It's a beautiful theory, but real life is complicated. Beautiful theories are no substitute for empirical evidence. What we need is actual evidence that eating meat increases mortality risk-- then we can say that the mTOR theory might explain what we've observed. Unfortunately (fortunately?), we don't have such evidence. While there is some hint that red meat intake may shorten lifespan, there is no evidence that poultry, fish, or dairy do the same. Since these are all concentrated sources of high-quality protein, I believe our hypothesis has run aground. Perhaps someday it will receive more convincing support, but for the time being it's a little too speculative for my tastes.
Synthesis and conclusions
When we consider individual diseases, specific types of meat appear to have substantial impacts. However, when we zoom out and consider health as a whole by examining total mortality, the picture becomes quite a bit fuzzier. If your overall goal is to live a long, healthy life, there appears to be little to gain from avoiding all meat.
Most of the healthiest, longest-lived cultures eat meat in modest quantity, and there is no evidence that poultry or seafood consumption increase overall mortality risk. There is not strong evidence that red meat intake is associated with mortality, but there is fairly consistent evidence that processed meat does.
Limited evidence suggests that vegans do not share the modest survival advantage that lacto-ovo vegetarians and pescetarians have over the general omnivorous population. There are no examples of long-lived vegan cultures to reassure us that this dietary pattern sustains long-term health and longevity. So while a vegan diet may protect against specific diseases such as cardiovascular disease, it may not be the safest strategy for overall health and longevity.
*It's surprisingly complicated to measure lifespan in yeast. The best way is to measure "replicative lifespan", or how many times a cell can bud (divide) before becoming senescent (too "old" to bud). Genes that affect replicative lifespan in yeast often seem to relate to lifespan (or lifespan-related factors such as diseases of aging) in mammals as well. mTOR (TOR1/2 in yeast), AMPK (SNF1 in yeast), and SirT1 (Sir2 in yeast) are good examples of this. It's pretty remarkable that a single-celled fungus we use to brew beer, which is separated from us by a 600+ million year evolutionary chasm, has such similar pathways for sensing cellular energy status and regulating lifespan. Evolution doesn't fix what isn't broken.