Tuesday, December 16, 2014

Is Meat Unhealthy? Part V

In this post, I'll examine the possible relationship between meat intake and type 2 diabetes.  Type 2 diabetes is the most common form of diabetes, and it is strongly linked to lifestyle factors.

Non-industrial cultures

Non-industrial cultures have an extremely low prevalence of diabetes, whether they are near-vegan or near-carnivorous.  This is supported by blood glucose measurements in a variety of cultures, from the sweet potato farmers of the New Guinea highlands to the arctic Inuit hunters.  Here is what Otto Schaefer, director of the Northern Medical Research Unit at Charles Camsell hospital in Edmonton, Canada, had to say about the Inuit in the excellent book Western Diseases (Trowell and Burkitt, 1981):

In Canadian Eskimos no case of diabetes mellitus has yet been reported in the traditional-living central and Eastern arctic regions; but an increasing number of diabetics have been reported from the Western arctic, especially in the more acculturated Eskimos living in the Mackenzie delta.  Age-adjusted prevalence rates even in this area are less than one-third of those reported in most Western nations.  Comparable upward trends in prevalence have been reported in Alaska and Greenland Eskimos.  The prevalence of diabetes is reported to be rising slowly in Alaskan Eskimos (Mouratoff and Scott, 1973).  Recently a very high prevalence rate was reported in an Eskimoid population subjected to a long period of acculturation, the Aleuts of the Pribilof Islands (Dippe et al., 1976).
The low prevalence of diabetes in traditionally-living cultures is consistent with their leanness, high levels of physical activity, and younger mean age (although they have a very low prevalence of diabetes even when age is taken into account).  These three factors, plus genetics, explain the large majority of type 2 diabetes risk.

Observational studies

The Adventist Health Study examined the health of vegans, vegetarians, semi-vegetarians, pescetarians, and omnivores in a Seventh-Day Adventist community in California.  All non-omnivore diet groups had a reduced diabetes risk compared with omnivores, but vegans had the lowest risk (1).  After adjusting for BMI, physical activity and other measured confounding factors, the relationship was attenuated but vegetarians and vegans still had a lower diabetes risk than omnivores--  about half the risk of omnivores in the same community.  This suggests that the difference in risk is related to their relative leanness, as well as other factors related to diet and/or lifestyle.  After adjustment, vegans did not have a significant advantage over lacto-ovo vegetarians in the initial study, although they did in a follow-up study (2).

What about specific types of meat?  In general, unprocessed poultry and seafood consumption have neutral associations with diabetes risk (3, 4, 5, 6, 7, 8).  Unprocessed red and processed meat, on the other hand, tend to be associated with increased risk, although the association with unprocessed red meat is very weak (9, 10, 11).

While we're at it, I may as well cover dairy.  Dairy intake typically has neutral or favorable associations with diabetes risk (12, 13, 14).  In 2013, my colleagues and I published a review paper on the association between high-fat dairy consumption and obesity and cardiometabolic disease risk (15).  We conducted a comprehensive literature review, and found that the results on diabetes were mixed.  Some found harmful associations, others found protective associations.  This year, we updated the analysis, identifying seven new studies.  Of the seven, three reported protective associations, three reported no association, and one reported a harmful association (16, 17, 18, 19, 20, 21, 22).  The literature overall provides little support for the idea that dairy, including full-fat diary, contributes to diabetes risk.

As always, it's worth keeping in mind that these are observational studies, and it isn't necessarily easy to tease apart the health effects of meat from everything else that comes along with eating it.  This is particularly true for culturally distinct groups such as the Seventh Day Adventists.

Possible mechanisms

Body weight is probably one mechanism that links processed meat to diabetes risk.  Processed meat tends to be calorie-dense and highly palatable, so it lends itself to overconsumption.

The association between unprocessed red meat consumption and diabetes risk is so weak it's hard to take seriously, but there is actually a plausible mechanism by which red meat (processed and unprocessed) could promote type 2 diabetes in some people: excess iron.  Red meat is a particularly rich source of highly bioavailable dietary iron-- which is helpful in certain situations-- but excess iron can degrade metabolic control.

Several lines of evidence support this conclusion.  Serum ferritin is the best blood marker of iron status, and it's strongly correlated with the risk of developing diabetes and metabolic syndrome (23, 24, 25, 26).  Iron depletion by repeated blood removal improves glucose control in diabetics and non-diabetics (27, 28).  Iron overload can be a problem for men and post-menopausal women, while pre-menopausal women often benefit from higher iron intake.  Eating less red meat, avoiding iron supplements, and avoiding iron-fortified cereals are three effective ways of reducing iron intake.  The body does not have any natural way to dispose of excess iron once it's absorbed.  Blood donation can help reduce the body's iron load, but it must be done regularly for a prolonged period of time to be effective.

Vegan diet interventions

Neal Barnard has been involved in several low-fat vegan diet trials with type 2 diabetics (29, 30, 31).  These trials show that a very low-fat vegan diet causes weight loss and improves blood glucose control more effectively than commonly used diets for diabetics.

One might be tempted to conclude that avoiding animal foods improves glucose control, but when we consider the literature as a whole, that explanation seems unlikely.  Older studies show that very low-fat diets high in unrefined carbohydrate improve glucose control in diabetics at least as well as a low-fat vegan diet, and although they were low in animal foods, they were not vegan (32, 33, 34).  These older studies found that many patients are able to discontinue insulin therapy after 1-2 months on the diet, due in part to increased insulin sensitivity.  Furthermore, the Paleo diet has been shown to improve glucose control more effectively than standard diabetes diets as well, and lean meat is a key component of these trials (35, 36, 37).

Although low-fat vegan diets can improve glucose control, there is currently little evidence that the effect is directly related to meat avoidance.  The effect may instead be attributable to 1) weight loss, and 2) the insulin-sensitizing effect of very low-fat diets high in unrefined carbohydrate.

Synthesis and conclusions

Poultry, seafood, and dairy consumption don't appear to be linked to type 2 diabetes risk, nor is there any clear mechanism linking them to diabetes risk (to my knowledge).  Processed meat may increase diabetes risk, and I speculate that this is mostly due to its ability to promote overconsumption and weight gain.  Unprocessed red meat may contribute to diabetes in some people due to its ability to increase body iron stores.

However, the potential effect of red meat appears to be very small, and it's important to keep it in perspective.  Obesity, age, physical activity, and genetics are the dominant risk factors for type 2 diabetes.  Everything else is probably peanuts.

18 comments:

Reijo said...

Good review of evidence. Perhaps it would be good to restate what meat products are actuallly classified as processed meat as there is such a big difference in risk ratios vs red meat?

Because you mention paleo diet as a healthy example of high meat diet in T2D, I cannot resist reminding that another dietary pattern diet, Mediterranean diet, reduced the risk of incident type 2 diabetes by more than 50% in randomized Predimed-Reus study (Diabetes Care 2011). Mediterranean diet is rather low both in red and processed meat.

Stephan Guyenet said...

Hi Reijo,

Good points about the Mediterranean diet and defining processed meat. I'll modify the post if I have a chance.

Gretchen said...

What are commonly used diets for diabetics? Even the ADA admits these days that no one diet works best for everyone with diabetes.

If you're trying to sell a diet, one good way is to show that it works better for people with diabetes than what they were eating before, which is often high-GI Food Pyramid food, especially in older studies when the Food Pyramid ruled.

Unknown said...

Despite Neal Barnard's claim that a vegan diet can cure diabetes, looking at the largest and most recent of his own studies (Reference 30), there was no statistically significant reduction in either Body Weight or HgB A1C. His own study directly refutes his own claim.

As you point out, "Obesity, age, physical activity, and genetics are the dominant risk factors for type 2 diabetes." However, the increase in diabetes rates only dates to the 1980s. Neither age, physical activity nor genetics has changed in that time (rates are age-adjusted). That only leaves the well known connection between diabetes and obesity.

This leaves us with the question - what causes type 2 diabetes, or elevated insulin resistance? Dr. Roy Taylor in the Banting Memorial Lecture in 2012 published an interesting hypothesis that points to fatty liver or non alcoholic steatohepatitis (NASH) as the underlying cause. Fatty liver, through the export of newly created lipids via De Novo Lipogenesis may lead to fat deposition in the pancreas, causing the beta cell 'burnout' phase that precedes clinical diagnosis.

Unknown said...

As far as type 2 diabetes, high meat consumption could be partially to blame. But it think it's more due to the fact that high calorie meat foods take up a higher percentage of the diet and overcrowd protective foods like legumes and other whole starches and their fibre content.

Aegirsson said...

Re Mediterranean diet. Doesn't it mean that it is low in very bio-available iron due to its being low in red meat (unprocessed as much as processed) ? In this case, we get the iron vs diabetes correlation again.

Unknown said...

Lars said...
At a quick glance I didn´t notice carbo hydrates as a possible risk factor. From a Swedish perspective that would certainly be high on the list...

Newbie said...

Re - ferritin as a marker.....
Thank you for providing references (23, 24, 25, 26), I couldn't see that any of them answered the question I will pose - although the JAMA article did somewhat address my query, it utilized transferrin receptor assays, which are not readily available.
Since ferritin is an acute phase reactant, elevated levels may reflect processes other than iron overload. Are you aware of any research that incorporates TIBC, % transferrin saturation and serum iron as the markers? The point I am getting at - is it the elevation of ferritin per se, even in the face of normal iron studies, that predicts an elevated risk of progression to T2dm?

RLL said...

I think we need a better set of terms for 'processed red meats"

1. They are ground. This is perhaps the most problematic, as surface bacteria are spread throughout all of the meat. But many would not label them as processed

2 Spices or other flavorings are added to the ground meat. Again, are they processed

3 That meat is fermented or smoked, most would call it processed.

4 Or red meat is air dried in after heavier surface salting - jerky beef, old fashion hams. Probably called processed.

5 Meat is injected with a brine - modern hams. Processed?

6 Some consider Nitrates/nitrites the issue, but there also is doubt about harm, and obvious benefits.

It is fair to ask in these various cases: What are the specific dangers of each these processed meats?

Stephan Guyenet said...

Hi Newbie,

Great point-- I forgot to consider that in my post. I'll have to look back at the studies and see if they addressed that concern.

Unknown said...

Hi Stephan,

I am an avid reader of your post.

I would like to know your views on Ma-pi 2 diet and how it can help in diabetes.

All along there has been this talk of switching to low carb/paleo for reduction in Diabets and general improvement of health. But the Ma-pi2 diet reduces diabetes with whole grains and vegetables..

Please let me know your thoughts.

Sincerely,
Sudarsan.D

Anonymous said...

Well we went over heart disease and diabetes, so i hope cancer is next because that certainly is the area on which vegans focus most. More specifically I'm interested what you find on methionine and the high correlation between a high methionine diets and cancer. (Casein in the case of dairy). Also, the supposed balance between methionine and glycine and how this ration determines inflammation (glycine higher good, methionine higher bad). Arachidonic acid is also supposed to be a boogyman these days (poultry and pork have high amounts), leading to inflammation. Sorry I'm jumping ahead. The suspense is killing me!

Dan said...

@Stephan, I'm also curious about the point RLL brought up. What exactly is processed meat and what is harmful about it? Is it the fact that it's ground, the fact that it's got spices, etc.

Jin said...

http://www.irondisorders.org/chronic-diseases-affected-by-iron

Newbie, curious if you'll find your answer in the work of Eugene Weinburg (Indiana University) or at the web site above.

Is the iron in the diet causing the problem, or making an underlying problem worse?

tomR said...

@lian johnston - about fiber contents - there was some information on freetheanimal on the animal prebiotics like glycans, or just animal parts like hair, feathers, skin etc.

The research on coprolites from the paleolithic suggests that animals were eaten as a whole, whic necessairly would include eating this animal prebiotics:

http://vegetablepharm.blogspot.com.au/2014/12/the-history-of-gut-bugs-part-3-of-3.html

"Animal bones were found in 97% of coprolites examined, some were tiny fish and lizard bones, but many were fairly large[12]! In one coprolite, a squirrel’s leg bone 1.2” long was found. Rodent bones were the most common, such as rabbits, rats, and mice. Almost half the coprolites had hair, skulls, teeth and jaws of mice and rats indicating these critters were eaten whole!"

Richard Nikoley said...

In Re processed meats, I'd simply point out that there is a world of difference between a traditional French, Italian, or Spanish charcuterie, and Oscar Myer.

I lived in France for two years and ate courtesy of the French Navy (petit dejeuner and dejeuner in port, those plus dîner at sea). We often had charcuterie as an app for dejeuner—along with baguette, butter, cornichons). But I never, ever had processed or cured meats as a meal. Never.

This goes in part to Stephan's palatability/reward argument, part to sane cultural practices.

Stephan, please do a part comparing 1st world countries and their food vs all the bad outcomes. Hint hint. :)

Lindsey Lewandowski said...

I have heard multiple things about this so this was a very interesting to read! I eat meat in moderation, but I am veggie lover at heart.

Linda said...

Thanks for sharing this post. I am not obese, but I stumbled upon your blog and decided to comment and to thank you for sharing all the information here. I was diagnosed with osteoarthritis in my early 30's (I am 58 now), and processed meats are something I no longer eat, as well as beef. I am on Allopurinol to prevent gout, as I have had many attacks and beef seems to be my enemy so I stay away from it. Not only does it seem to cause attacks of gout but I don't digest it well. Happy New Year and warm greetings from Montreal, Canada.