Is Meat Unhealthy? Part VI

In this post, I'll examine the possible relationship between meat consumption and cancer risk.

Is cancer risk even modifiable?

Cancer is caused by the uncontrolled division of a population of rogue cells in the body.  These cells essentially evolve by natural selection to escape the body's multiple anti-cancer mechanisms.  

To a large extent, cancer appears to be a numbers game.  The human body contains about 37 trillion cells.  To get cancer, all you need is one cell that develops key mutations that allow it to shed its built-in restrictions on cell division.  The older you are, the more time you have to accumulate mutations, explaining why cancer risk rises sharply with age.

Unlike other common non-communicable diseases, we don't know to what extent cancer is caused by modifiable diet and lifestyle factors vs. bad luck that's completely outside our control.  Some cancers, such as lung cancer, are typically linked to lifestyle factors like cigarette smoking-- yet the majority of cancers aren't so easily understood.

There is an ongoing scientific debate about how much of cancer risk is modifiable.  Although there is clearly a certain amount of luck involved in cancer risk, the various stages of cancer development could theoretically be influenced by diet and lifestyle factors-- and this has often been supported by research.  Observational studies have suggested that at least 35-40 percent of cancers are preventable, although these estimates rely on major assumptions (1, 2).  In the end, I doubt there is any diet or lifestyle that can eliminate cancer risk, but overall cancer risk is probably modifiable enough that it's worth trying (particularly for the obvious risk factors like cigarette smoking).

Non-industrial cultures

A number of field physicians have reported that cancer is rare among traditionally-living cultures (Cancer: Disease of Civilization? Stefansson. 1960).  However, cancer rates are difficult to measure accurately in these cultures, due to their shorter mean lifespans, and the fact that many cancers are internal.  These reports are intriguing, but it's difficult to draw firm conclusions from them.

We do know that cancer existed in ancient times, because we can occasionally observe cancerous lesions in bones that are thousands of years old (3).  It's difficult to know what the age-specific cancer risk might have been among our distant ancestors, but we do know that their absolute cancer incidence would have been low simply because few of them survived to old age.

If we look at cancer incidence in different parts of the world (adjusted for age of course), the rates tend to be lower in less industrially developed areas such as West Africa (4).  Cancer mortality rates are lower as well, although the difference is less striking than the difference in incidence.  While some of these differences could be due to differences in diagnosis, these figures are at least consistent with the idea that cancer risk is modifiable by diet and/or lifestyle.

The risk of specific cancers can vary dramatically between cultures.  In particular, industrialization and affluence are strongly associated with a higher risk of colon/rectal, breast, lung, and prostate cancer (Western Diseases. Trowell and Burkitt. 1981).  This is thought to relate to changes in reproductive patterns, tobacco use, and diet.

Observational studies

Let's start with the Adventist Health Study, which examined the diet, lifestyle, and health of a religious vegetarian community in California.  Lacto-ovo vegetarian Seventh-Day Adventists (SDAs) in this study had a cancer risk similar to their omnivorous peers, while vegans had a slightly lower risk (5).

There are a number of other observational studies that have compared the cancer risk of vegetarians to omnivores-- too many for me to discuss here-- so I'll rely on meta-analyses (studies of studies).  The results are mixed, although the most recent one I could find did report an 18 percent lower risk among vegetarians (6, 7).  It's a small difference, but it was statistically significant.  

How about specific types of meat?  In general, poultry and fish consumption are associated with neutral or lower risk of common cancers such as breast, prostate, and colon/rectal cancer (8, 9, 10, 11, 12, 13, 14, 15, 16).  However, there may be an association between poultry intake and kidney cancer (17).

In contrast, red meat and processed meat are often associated with a small but statistically significant increase in total cancer risk (18, 19, 20), and fairly consistently associated with a larger increase of colon/rectal cancer risk specifically (21, 22, 23, 24, 25, 26, 27, 28).

It's worth keeping in mind the usual limitations of observational studies, such as mediocre diet assessment methods, and possible confounding by unexpected or unmeasured factors.  Sadly, we don't have any human randomized controlled trials to back up these hypotheses, and we probably never will.

What are processed and red meats?

Reijo Laatikainen and others suggested that I define what I mean by processed and red meats in this series, which is an excellent idea.  

For our purposes, the most useful definitions of processed and red meat are the ones that are most similar to the definitions used in the studies I cite.  In general, processed meats are those that have been preserved in some way to extend their shelf life, for example, hot dogs, bacon, sausage, ham, and cold cuts.  

Unprocessed red meats are those that are no more protected from spoilage than fresh red meat, such as all fresh whole cuts of beef and lamb, and ground meats.  Pork is also typically considered a red meat, despite being similar in color to chicken, probably because it has some nutritional similarities to beef.

Possible mechanisms

There are several known mechanisms by which red and processed meat might increase cancer risk-- buttressing the case made by observational studies.

When meat is cooked at high temperature, such as by grilling or frying, it creates potent carcinogens including heterocyclic amines (HCAs; 29).  HCAs promote the development of multiple cancer types in animal models, including the types of cancer that are most common in humans (29). Consistent with the animal studies, human observational studies report an association between the consumption of high-heat-cooked meat and several cancers including colon/rectal and breast cancer (29, 30).  This applies to all meat, not just red and processed meat.  It may be best to limit high-heat cooked food, and instead focus on lower-heat cooking methods such as steaming, pressure cooking*, poaching, and boiling.

As I discussed in the post on meat and diabetes risk, free (unbound) iron is a powerful driver of free radical reactions, and red meat is a rich source of loosely bound (heme) iron.  The potential resulting damage to the gut's epithelium could in theory contribute to an increased risk of colon/rectal cancer-- and this is supported by both animal and human studies (31, 32, 33).

Insulin-like growth factor 1 (IGF-1) is a hormone related to insulin that has received attention as a possible contributor to cancer risk.  IGF-1 is a growth factor, meaning that it instructs tissues to grow by increasing cell division.  IGF-1 levels are increased by high-quality protein, such as that provided by animal foods, and dairy seems particularly effective at increasing it (34, 35).  Animal foods seem to put the body into more of an anabolic (building up rather than taking down) state, which we'll come back to in later posts.  Theoretically, this could increase the rate of cell division of cancerous cells.  

Vegans have slightly lower IGF-1 levels than omnivores and vegetarians (36).  However, human studies have found inconsistent associations between IGF-1 levels and overall cancer risk, with the exception of pre-menopausal breast cancer and prostate cancer (37).  The association between circulating IGF-1 and prostate cancer risk is consistent with the (small but significant) association between dairy consumption and prostate cancer risk (38).  The meat-IGF-1 hypothesis is definitely intriguing, and it could be involved in specific cancers, but the link with overall cancer risk remains unclear to me.

Processed meat is often preserved using sodium nitrite to prevent spoilage.  Although nitrites and nitrates are a healthful component of many vegetables, nitrite can form potent carcinogens called nitrosamines when heated in the presence of protein such as meat.  Nitrosamine intake has been linked to digestive cancers in observational studies (39, 40).  Even processed meats that say "nitrite/nitrate-free" typically contain celery juice or another "natural" source of nitrite/nitrate.  

The newest arrival onto the scene is a sugar called Neu5Gc.  Unlike the energy-yielding roles of the sugars glucose and fructose, Neu5Gc plays a role in cell signaling as part of the complex sugar chains on the exterior of cells.  It's part of a family of such sugars called sialic acids.  During our evolutionary history, humans lost the ability to make Neu5Gc, but most other mammals still make it and their tissues contain it.  When we eat the flesh of mammals such as cows, sheep, and pigs, we absorb Neu5Gc and incorporate it into our tissues.  Since it's a foreign molecule however, our tissues mount an immune response against Neu5Gc, producing antibodies and possibly activating an inflammatory response.  A previous paper that developed this mechanism suggested that this inflammatory response could be involved in cancer risk (41).

Until recently, there was little evidence to support the hypothesis that Neu5Gc increases cancer risk.  However, a very recent paper has moved the needle somewhat toward plausibility (42).  First, they measured the Neu5Gc content of a variety of foods, showing that it's moderate to high in beef, lamb, pork, cheese, and caviar, but very low in poultry, fish, eggs, milk, and vegetables.  Then, starting with a strain of mice that is naturally susceptible to liver cancers, they knocked out a protein that is required for Neu5Gc synthesis, making the mice similar to humans in that regard.  They subsequently immunized the mice against Neu5Gc so that they would mount an immune response to it, similar to what is observed in humans.  Then, they fed the mice Neu5Gc, and found that it caused a large increase in the development of liver cancer, suggesting that the human-like immune response mounted by the mice increases cancer risk.

This study needs to be interpreted in the proper context.  First, it was conducted in mice.  Second, the amount of Neu5Gc in the diet was equivalent to a diet composed entirely of beef-- and not just any beef, but the highest-Neu5Gc beef measured in the study (beef Neu5Gc ranges 10-fold).  The exposure to dietary Neu5Gc was therefore some 40-fold higher than what most red meat eaters would experience.  Still, the study outlines a plausible mechanism for a link between red meat and cancer, and that helps increase our confidence in the observational findings.  Hopefully, follow-up studies in humans will paint a clearer picture of the importance of this mechanism to human health.

Synthesis and conclusions

Vegetarians and vegans have a modestly reduced overall risk of cancer relative to omnivores, although we don't know how much of that relates to meat avoidance specifically.  

There is little evidence that poultry or seafood intake contribute to cancer risk, however, I'm fairly convinced that red and processed meat contribute to cancers of the digestive tract such as stomach, colon, and rectal cancers (and possibly at other sites as well).  The observational studies, animal studies, and mechanisms are relatively consistent with that hypothesis.  We will probably never have randomized controlled trials in humans to test it.

Red meat is a highly nutritious food that our ancestors have been eating for millions of years.  I don't think we need to stop eating it entirely.  Still, from a cancer standpoint it's probably best to favor poultry and fish, and when we do eat red meat, to cook it gently.

* Pressure cooking involves somewhat higher temperatures than boiling, but the temperatures are still much lower than frying, grilling, and roasting.