Wednesday, December 3, 2014

Is Meat Unhealthy? Part IV

In this post, I'll address the question: does eating meat contribute to weight gain?

Non-industrial cultures

I'll get right to the point: humans living in a non-industrialized setting tend to be lean, regardless of how much meat they eat.  This applies equally to hunter-gatherers, herders, and farmers.

One of the leanest populations I've encountered in my reading is the 1960s Papua New Guinea highland farmers of Tukisenta.  They ate a nearly vegan diet composed almost exclusively of sweet potatoes, occasionally punctuated by feasts including large amounts of pork.  On average, they ate very little animal food.  Visiting researchers noted that the residents of Tukisenta were "muscular and mostly very lean", and did not gain fat with age (1, Western Diseases, Trowell and Burkitt, 1981).

!Kung man gathering mongongo fruit/nuts.
From The !Kung San, by Richard B. Lee.
Another remarkably lean hunter-gatherer population is the !Kung San foragers of the Kalahari desert.  The !Kung San are so lean that many of them would be considered underweight on the standard body mass index scale (BMI less than 18.5).  Average BMI doesn't exceed 20 in any age category (The !Kung San, Richard Lee, 1979).  Is this simply because they're starving?  It is true that they don't always get as much food as they'd like, but on most days, they have the ability to gather more food than they need.  The fact that they are able to reproduce normally suggests that they aren't starving.  Richard Lee's detailed work with the !Kung San indicates that approximately 40 percent of their calories came from animal foods during his study period in the 1960s.  This was mostly meat, with occasional eggs when available.

Traditionally, the arctic Inuit ate more meat than almost any other population on Earth, due to the scarcity of edible plant foods in their arctic ecosystem.  We often perceive the Inuit as fat, but in fact this is a misconception caused by their voluminous clothing and broad faces.  Accounts of traditionally-living Inuit suggest that they were in fact lean underneath all that fur.  Arctic explorer and anthropologist Vilhjalmur Stefansson spent a number of years with the Inuit of Alaska and Canada in the early days of European contact, reporting that they were characteristically lean and fit (My Life with the Eskimo, Vilhjalmur Stefansson, 1912).

Consistent with Stefanssons's account, Mouratoff and colleagues reported that in Western Alaska in 1962, only four percent of Inuit men, and ten percent of Inuit women, were overweight (2).  Among middle-aged and older men and women, even fewer were overweight, in contrast to the gradual weight gain that characterizes affluent industrialized societies today.  By 1972, the prevalence of overweight had increased substantially, as diets and lifestyles continued to modernize in the area.

Meat eating doesn't seem to have much of an impact on body weight in traditionally-living societies, suggesting that it may not be a major determinant of body weight.  However, those societies differ from ours in many ways.  What is the relationship between meat eating and body weight in affluent societies today?

Observational studies

Vegans don't eat any animal foods.  The stereotypical vegan body type is extremely lean with low fat and muscle mass-- and this stereotype is fairly accurate.  Although it's possible to be vegan and obese, most vegans are indeed quite a bit leaner than the typical omnivore, and vegetarians are somewhere in between.  Data from the 38,000-person EPIC-Oxford study illustrate this well, using body mass index as a measure of fatness (3):

Omnivores are the fattest, pescetarians (fish eaters) and vegetarians are tied, and vegans are the leanest.  The study suggests that these differences in body weight were linked with both dietary and non-dietary factors (e.g., physical activity level).  This study also illustrates the difficulty of disentangling the effect of meat eating from other diet and lifestyle factors, since the vegan group exercised much more, ate more fiber, smoked fewer cigarettes, and was much less likely to be married and have children.

In 2010, Bernstein and colleagues published a paper based on the massive Nurses' Health Study, reporting the association between the intake of different types of meat and cardiovascular risk (4).  Although the study wasn't designed to investigate body fatness, the paper does report the association between meat intake and BMI at baseline (Table 1).  In this cohort, people who ate the most red meat or poultry did have a somewhat higher BMI than those who ate the least.  There was little association between fish intake and BMI.  Surprisingly, the strongest association was actually with poultry.  As usual, meat intake was associated with a number of other diet and lifestyle factors.

In 2011, Mozaffarian and colleagues published a paper on diet and weight changes that received a huge amount of attention (5).  Drawing from three large US cohorts collectively representing 120,877 people, they reported the association between changes in dietary habits and changes in body weight over 4-year intervals.  They reported a number of associations*, including for unprocessed red meat and processed meat, both of which were associated with weight gain.  The associations were among the stronger of those that they identified.  Unfortunately, they didn't report data for poultry or fish.

These observational studies use relatively inaccurate dietary measurement tools, and suffer from a number of potentially serious confounding factors, so we have to be careful in interpreting them.  Still, vegetarians and vegans clearly tend to be leaner than omnivores, and I believe meat is part of the explanation.  Here's why: people eat fewer calories when they eliminate a food category from the table.  Also, certain meats, such as fatty and cured meats, are calorie-dense and highly palatable, potentially contributing to overeating.  For what it's worth, low-fat vegan diet advocates are consistently among the leanest diet advocates.

Now, let's move beyond observational studies.  What happens to body weight when omnivores stop eating animal foods?

Vegan diet interventions

Physician, researcher and vegan diet activist Neal Barnard has conducted a number of vegan diet studies.  In 2007, he co-authored a two-year randomized weight loss study in overweight post-menopausal women "comparing a vegan diet to a more moderate low-fat diet" (6).  Unfortunately for our purposes, the study didn't isolate the effects of animal foods specifically, because the vegan diet was also low-fat.

However, the low-fat vegan diet did cause weight loss.  At one year, vegan participants had lost 11 pounds (5 kg), and they managed to keep most of it off for two years.  As expected, those who adhered closely to the diet did better than those who were less adherent.  This degree of weight loss is similar to what is reported from low-carbohydrate dietary interventions, although we don't have body composition data so we don't know how much of the loss came out of muscle tissue.

Vegetarian diet interventions seem to have a more limited effect on body weight.  A 2008 randomized trial comparing weight loss diets with or without meat found that both diets produced a similar degree of weight loss over 18 months (7).

Together, these studies suggest that avoiding meat doesn't enhance weight loss, but avoiding all animal foods might.  We don't yet have a very good sense of how much of this weight loss comes from fat vs. muscle tissue.

High-protein diet interventions

Low-fat vegan diets cause weight loss, but guess what else does too?  High-meat diets!  Or, more precisely, high-protein diets in which the protein comes mostly from meat.

Protein is the most satiating macronutrient: more filling than carbohydrate or fat, per unit calorie (8).  Lean meat is the most concentrated and convenient source of protein.  A number of studies have shown that high-protein diets cause people to eat fewer calories per day, and lose fat, without experiencing increased hunger (9, 10).  

Not only do high-protein diets reduce calorie intake and promote fat loss, they attenuate the loss of muscle mass and reduction of metabolic rate that occurs during dieting (11, 10).  Not surprisingly, this makes protein one of the most effective ways to support long-term fat loss maintenance (12, 13), along with physical activity.  Protein is a powerful tool in the fat loss toolbox, which is why it's a central part of the Ideal Weight Program.  We designed the Ideal Weight Program to accommodate omnivores, vegetarians, and everything in between, but meat is certainly an easy and effective way to meet your daily protein goal.

Some researchers believe that increased protein intake, rather than reduced carbohydrate intake per se, accounts for the fat loss effectiveness of low-carbohydrate diets (14, 10).  The case isn't closed yet, but there is some evidence to support that view.

Together, the evidence suggests that meat, particularly fresh lean meat, is a powerful fat loss tool.

Synthesis and conclusion

So, does eating meat contribute to weight gain?  Yes and no!

Meat intake has a complex relationship with body weight.  Vegetarian and vegan diets are associated with leanness, but at the same time, high-meat diets are one of the most effective fat loss tools we have.

I believe the effects of meat on body weight may be explained by three simple principles:
  1. Taking an entire food category off the table tends to reduce total calorie intake.
  2. Fatty and cured meats are calorie-dense and highly palatable, and can therefore contribute to overeating.
  3. Fresh lean meat is rich in protein, which tends to increase satiety and suppress calorie intake.

* This was the study that caused the media to proclaim that potatoes are fattening.  The reason has to do with a quirk of one of the figures: in Figure 1, the most widely reproduced figure, the authors lumped the data for French fries together with baked, boiled, and mashed potatoes!  Since French fries had by far the strongest association with weight gain of any of the factors they studied (implausibly strong), lumping it together with non-fried potatoes made it appear as if all potatoes share the same strong association.  It also had the added benefit of making Figure 1 look nice-- if they hadn't lumped them together, the figure would have been dominated by the massive, implausible, French fry data.  The association they found for baked, boiled and mashed potatoes is much less impressive.  People who increased intake of baked, boiled and mashed potatoes gained 0.6 lbs over 4 years, compared with 3.4 lbs for French fries, 1.7 lbs for potato chips, and 1 lb for sugar-sweetened beverages, unprocessed red meat, and processed meat.  Also, keep in mind that people slather all kinds of calorie-dense toppings on their potatoes, so the effects of the potatoes themselves will be difficult to disentangle from the toppings.  When is the last time you saw someone eat a plain potato?


Tucker Goodrich said...

Interesting. Both poultry and French fries are high in linoleic acid, while beef and potatoes are not...

Pork is, as well: I wonder if that signal would exist if you could segregate processed meats into pork and beef?

rantyscientist said...

Maybe I'm off on this, but I would think that the amount of calories consumed from meat would determine the amount of weight gain. I wouldn't expect any gain from eating small amounts of meat, and I wouldn't expect any weight gain if the diet consisted of only meat but remained less than the quantity of calories required to maintain a given weight.

Seems to me that it is difficult to determine the healthfulness of a food by looking at weight loss. One could also lose weight on a "whiskey and cigarette" diet.

Unknown said...

I know you already know this but BMI is a really bad measure. The studies aren't much use for determining health. Is the difference in weight because of fat mass or muscle mass?

When I think of a vegan I think of them having the figure of a long distance runner vs a meat eater having the figure of a muscular sprinter. This is probably confirmation bias since I'm not vegan.

Unknown said...

Neal Barnard ... wasn't he the guy about whom you said recently that he looked like he was starving ??

I concur about the BMI comment: a very muscular individual would have a rather high BMI. I'd be more worried about the lower BMI figures from vegans ...

Patrick Rochon said...

BMI as little to do with body fatness ...

Most bodybuilders / anaerobic athletes are obese by BMI standard

Stephan Guyenet said...

Hi folks,

Since there is so much interest in BMI, I'll address the point. The primary virtue of BMI is that it's easy to measure-- all you need is height and weight. BMI is actually quite well correlated with fat mass when it's used to characterize groups of people. It is therefore very useful for research purposes where groups of people are compared.

It's useful on an individual level too, but it needs to be applied with some common sense. If a person is unusually muscular, the BMI scale will overestimate that person's fatness. If they don't have much muscle, it will underestimate fatness. I think BMI is a useful tool when combined with another useful measure-- looking in the mirror.

frank r said...

Body fat % is also very easy to measure, using the Navy method. All you need is height, neck circumference and waist circumference for men, with an additional measure of hip circumference for women. This Navy method of body fat % estimation is supposedly much more accurate than caliper or impedance measures, though not so accurate as water displacement. For a population study, it should be quite accurate.

Weight is actually a fairly high-tech thing to measure: you need a scale and that scale can be miscalibrated. Whereas a tape measure can fit in your pocket and is very unlikely to be miscalibrated.

David L said...

I looked in several of the meat-related articles you posted and found it difficult to garner certain data relating to diet composition. To be more specific about protein:

Does lean meat include fish or chicken? Can eggs count towards lean meat? Is the diet ruined if meat is not lean? What about cheese?

Galina L. said...

One of wide-spread cliches is "moderation is the key". Your well-balances series on meat contradict the moderation approach to a weight-loss, and probably to the everyday diet as well.

David L said...

Please add nuts to my previous comment, also.

Chuck Steak said...

"[High protein diets] attenuate the loss of muscle mass and reduction of metabolic rate that occurs during dieting". I don't doubt they prevent muscle loss to some extent, but i find the second part, that they attenuate the reduction of metabolic rate, very hard to believe. One of the links you provided attributed this to the preservation of lean tissue, but that is a minor factor. In weight loss conditions, metabolic rate is not determined by the amount of lean tissue, but is controlled hormonally, as the body hits the brakes and starts lowering expenditure and increasing fat storage. This would occur with any diet, but a low-carb diet is even worse because it trashes your thyroid, and we all know low thyroid means very bad things, including low metabolic rate.

RichD said...

I would add that the composition of a vegan diet matters for leanness. I was a vegan for six years. My weight varied from 135 - 155 over that time period. (I'm male, 5'9".) When did I get heavier? When I started eating a lot of white pasta and desserts. (I almost certainly increased my calorie intake in that period.) As most vegans are health conscious (as I was most of the time), they probably eat more whole foods and less sugary and fatty desserts, even though the constraints of veganism permit otherwise. I know that Stephan (and most readers here) know this, but I thought it was good to point out.

For what it's worth, I happen to trying a low-carb/high meat diet currently. I find it quite satiating, and easy to keep calories low for weight loss.

9ff77380-7bbd-11e4-b80b-83b791ea25f0 said...

Too much focus on BMI without evaluating actual health outcomes. BMI is a biomarker and not an endpoint. That should be noted in this article here somewhere, lest some layman will merely think low BMI, high protein (atkins) = good.

Laura said...

Thanks for the note on the study with potatoes! I had read the study earlier and had difficulty believing that the plain potatoes ought to be lumped with french fries and potato chips (both quite difficult for me to stop eating). I don't have any trouble eating plain potatoes!

Stephan Guyenet said...

Hi Frank,

Thanks for sharing that method. There has been a tremendous amount of debate in the scientific community over what measures of body fatness are most appropriate. I suspect that if the Navy method offered a significant advantage, researchers would be using it.

Any measure that involves measuring circumference with a tape measure will introduce error, because there is clothing in the way and each person will pull it to a different degree of tightness. Those sources of error can be minimized w training, but weight and height measures are more accurate and consistent.

As much as it's criticized, on a population level BMI is actually closely correlated with fat mass, and it also predicts obesity-related health outcomes as well as any other measure of fat mass we have.

Hi David,

Lean meat includes most fish and many cuts of poultry. It doesn't include eggs or nuts because those aren't meat.

Hi Chuck,

The reduction of energy expenditure happens because of two processes: loss of lean mass, and reduction of energy expenditure per unit lean mass. High-protein diets attenuate both processes, leading to a smaller reduction in energy expenditure and more successful weight maintenance. High-protein diets can supply plenty of carbohydrate.

Hi 9ff,

If you read the rest of the series, you will see that it's mostly focused on health rather than weight. This particular post was specifically about weight. BMI is a well-validated measure of body fatness for population studies such as those I cited.

Sam said...

Stephan, in your reading have you seen anything that tracked changes in lean mass over age?

Chuck Steak said...

"...and reduction of energy expenditure per unit lean mass. High-protein diets attenuate both processes..." This is what I'm not getting. How does a high-protein diet attenuate energy expenditure (EE) per unit of lean mass? Through what mechanism? Are there any links i can read? I thought EE was controlled hormonally like a thermostat, as is discussed in body fat set point theory, a theory of which you're a proponent of if I'm not mistaken. The body detects the loss of adipose tissue (not lean tissue), and starts adjusting hormone levels to make you move less and eat more, and store more fat, in an attempt to get you back to your "real" weight--your set point. So how would high protein counter this? In my personal experience it certainly hasn't.

Stephan Guyenet said...

Hi Chuck,

Excellent question. The reduction of energy expenditure per unit lean mass is mediated by the energy homeostasis system, as you mentioned. It's essentially part of the "starvation response".

I believe high protein attenuates that response by lowering the setpoint. We know from rodent research that amino acids can act on energy homeostasis circuits in the hypothalamus to affect the setpoint. That was thought to be mediated by mTOR activation in the hypothalamus, but recent research by Chris Morrison is suggesting an alternative mechanism involving FGF21.

I don't know your diet and I won't make any assumptions about what you're eating, but I do know that people often underestimate the amount of protein they're eating when they try a high-protein diet. That's why we designed the Protein Unit system for the Ideal Weight Program-- to allow people to intuitively track protein intake and meet their optimal daily goal that's individualized for their own needs.

Billy Oblivion said...

The total difference between Vegans and Omnivores is 2 percentage points.

Given the inherent inaccuracy of BMI I'm not really so sure that's a significant difference.

I'm also not really certain that even if it was 2 percent body fat is a significant health risk.

Gretchen said...

Re BMI, are there corrections for advanced age? Let's say someone is 6 feet tall when young. When 80, that person has shrunk to 5 feet 8.

So the BMI would go up without any change in weight. And a person who was borderline overweight/obese would be pushed into another category.

Stephan Guyenet said...

Hi Billy,

Not two percent-- two BMI points. That's the equivalent of me gaining 13 lbs, for example. It's not a huge difference, but it is meaningful.

Hi Gretchen,

Ideally it should be corrected for age if you're trying to estimate BF% based on BMI. However, in these studies, they control for age so it's essentially the same in both groups and BMI measures can be compared without worrying about that variable.

Chuck Steak said...

I thought I had heard it all, but I never knew about this. You learn something every day :)
As to my diet, I am currently on the seafood diet (I see food and I eat it, ha-ha), but I have done Atkins and Paleo (of the high meat, Loren Cordain variety) in the past, and they both led to the common pattern of initial fat loss, followed by starvation response and hitting that plateau wall.
So why didn't FGF21 help me out? I believe I spotted a strong clue in the link you provided. In paragraph two of the introduction, the authors noted that "Plasma levels of FGF21 are paradoxically increased in the settings of obesity and insulin resistance." That sounds exactly like how obese people paradoxically have high levels of Leptin. So maybe we're looking at FGF21 resistance?

Reijo said...

High protein intake seems to be good in weight loss but it was not associated with better weight management in a systematic review of weight management RCTs and cohorts.

Meat particularly was associated with weight gain in 8 long term studies. Only one study lasting longer than 6 months found that meat was protective from weight gain. 4 showed no association. It's hard to believe after this data eating more meat would be an essential tool in prevention of weight gain. Protein per se did not associate with better weight management either but fiber did. This review did not find any association between potato consumption and weight gain but interestingly dairy, whole grains and nuts were associated with less weight gain.

Ref: Fogelholm M, Anderssen S, Gunnarsdottir I, Lahti-Koski M. Dietary macronutrients and food consumption as determinants of long-term weight change in adult populations: a systematic literature review. Food Nutr Res. 2012;56

Anyway, as you say "meat intake has a complex relationship with body weight". I would not deem meat as fattening based on that systematic review only.

Stephan Guyenet said...

Hi Reijo,

I think there are questions that can be asked about the effects of protein on weight gain prevention vs. weight loss. It's possible that the effects are different.

I think it's important to recognize that the high-protein diets that are effective for weight loss include more protein than what is commonly consumed in the general population. Few people spontaneously eat that much protein, so I'm not sure how well an observational study would be able to investigate the question. For example, in the Weigle RCT showing robust decrease in spontaneous kcal intake, they were eating 30% protein-- twice the average US % protein intake.

Thanks for the reference, I hadn't seen that review paper. Their findings on meat are similar to what I presented in my post, which is reassuring. However, I do find it troubling that they make sweeping conclusions about meat consumption in general, when the studies they cited did not consistently suggest that meat in general is associated with weight gain. Some of the studies compared vegetarians vs. non-vegetarians, other studies only found that specific types of meat are associated with weight gain, while other types were not. It seems difficult to summarize all this as suggesting a "probable" role for meat intake in weight gain. It's also troubling that they use language throughout that suggests causality.

TheGiantess said...

Oddly, I don't know any Vegans that are lean. Most Vegans I know are so because of ideals and not health and so rely heavily on processed foods. Lots of extra fat, bad skin, etc.

It seems that traditional cultures eat either one or the other.. that is they eat either starch or meat/fat, but rarely both. So, that's the secret? Eat a boring, monotonous diet to be lean?

tomR said...

A vegan - dr. Greger - demonized leucine as some kind of an aging agent. That is present in animal products.
"The lifespan extension associated with dietary restriction may be due less to a reduction in calories, and more to a reduction in animal protein (particularly the amino acid leucine, which may accelerate aging via the enzyme TOR)."

But leucine is THE bodybuilder's amino acid. The one legal, and approved by sport federation (unlike testosterone) supplement that gives you the most muscles of such natural supplements.

tomR said...

So the McDougal vegans are like in their 60s, while the new generation doesn't reproduce? Are there going to be vegans in the future?

Alex Kokolis said...

Hi Stephan,

Love the meat series, thanks for offerring a deep backdrop for this discussion. As I keep reading about all the groups across the world that have extremely unique and varied diets, Masai vs Inuit vs more vegan groups vs etc...and still show remarkable Heath with their own unique diet, I can't help but think of the one common denominator.

I feel it's that they are secluded groups that have lived in that area eating the same things for many, many generations. Meaning, the ones who could not handle the very specific diet just died off. Also, since these groups have been there forever, they have a perfect match of gut biome to local food availability. This is opposed to other areas of the world where we are basically transient and have insane diversity in our diets. We basically never get fully adapted due to he lack of deep generational roots to the environment and the crazy availability.

Anyway, just curious if that concept has been explored. It makes me feel findin that perfect diet for yourself is so important and should be determined based on your personal ancestral path and you should limit diversity from what works.

Thanks again for all of your insight,

tomR said...

About that muscle/anabolic topic: at least one prominent former vegan wasn't able to do heavy jobs while on a vegan diet:

"Veganism, when done correctly, is undoubtedly the best diet for a sustainable planet, but personally I've found it impossible to follow as someone who works on a ranch and engages in a fair amount of physical work each day. For me, powering my work takes a small amount of fish, fresh farm eggs, some [...] omega-3s and other fish oils [...] Overall, my diet is probably 95% plant based. No dairy."

George Henderson said...

Another reason that meat-eating is tied to higher BMI comes down to simple nutritional adequacy of the diet.
The B vitamins, vitamin A, iron and other micronutrients are growth factors. Meat is the richest source of these factors in the usual diet, eggs and fish are also better sources than vegetables.
You can't lay down fat, or build muscle, if you're vitamin deficient, no matter how many calories are available to you (you will tend to loose appetite).

Richard Nikoley said...

Well, I'm a bit mystified as to why so many seem to want to pick at nits. Part of the point of a SERIES is that no particular part is all encompasing.

Beyond that, thanks Stephan for convincing me that for the purpose of these sorts of studies, BMI is a reasonable, albeit imperfect tool. I recall from some years ago people dismissing ALL findings (baby, bathwater, so OUT!) because a study had 3 letters together and Michael Jordan (an extreme outlier) is fat, according to BMI.

Excellent, fair and balanced work.