So on my return to regular Scienceblogging, I see that Mike the Mad Biologist and Razib are taking exception to a point made by Megan McArdle in the Atlantic. McArdle observes that the heritability of weight is quite high - almost as high as the heritability of height:
Twin studies and adoptive studies show that the overwhelming determinant of your weight is not your willpower; it's your genes. The heritability of weight is between .75 and .85. The heritability of height is between .9 and .95. And the older you are, the more heritable weight is.
Okay: how you take that statement depends on how you define willpower (as RT astutely commented on Razib's thread). As a neuroscientist, I think nothing in the brain is independent of genes, including willpower. But I also recognize that most people think of willpower/self-control as a moral strength which is independent of physiology/genes, and that's how McArdle's audience will probably take these lines. The distinction here is not really between willpower and genes, but between psychology and physiology. . . and the fact that there is no clean distinction there makes this hard to discuss in precise terms.
But in general, Megan is trying to make the point that weight depends strongly on genes - more strongly than most people think. Mike and Razib argue that the concept of heritability doesn't mean quite what Megan implies. Heritability is a misunderstood concept, they say. That's true. Like many genetics concepts, heritability gives me a headache, because there is math involved. Horrors! But basically, heritability is a measure of how much of the variability in a trait in a population (like different weights in a group of people) is accounted for by variants in the genes of the members of a population.
Heritability matters immensely to evolutionary biologists because a trait that is not inherited is not one you'd expect evolution to select for (or against) in a population. Natural selection acts on genes, and can't act on something that isn't genetic in the first place. Razib and Mike do excellent jobs of discussing heritability in this context, though I think both their posts are a bit technical for general audiences. (This stuff is hard.)
On the other hand, heritability matters to health professionals for a less esoteric reason: because genetics help account for individual differences in health, and in responses to treatment. And that's why I think McArdle's piece makes some important points.
Consider my mom, who is resistant to Novocain. As a child, dentistry was excruciating for her, because her dentist and parents thought that Novocain worked equally well on everyone. They assumed she was being a crybaby, and she had a lot of teeth drilled under insufficient anesthetic. Today, though, it is recognized that individuals have different responses to anesthetics - and my mom's dentist gives her another class of drug. Knowing that the patient has a genetic resistance or susceptibility to certain environmental factors can help the dentist or doctor make a better decision for the patient - because not all patients are the same.
The thing that makes me angry about the "weight genes" debate is that it shouldn't be about placing blame on either genes or people. It should be about improving health. Knowing that genes do strongly influence weight - that weight is strongly heritable - can be a very good thing. It helps explain why a given diet works for one person yet not for another. Some of us have a propensity to put on weight despite eating less and being more active than our rail-thin friends. Some of us are hungry all the time. Some of us work out three times weekly and gain weight instead of losing it. Knowing these different outcomes have a real biological basis can help us redirect our frustration at the unfairness of it all toward finding the solution that works for us - recognizing that the best solution may be a totally different diet or exercise plan than the one that works for somebody else.
Life is unfair. Evolution is (perhaps by definition?) unfair. Fat is unfair. And it's also unfair that our society judges people for being overweight even when some of them are making herculean efforts to fight it, and when many thin people make no effort at all to be thin. I think weight has such a powerful stigma largely because there appear to be so many solutions out there - diets, exercise plans, surgeries, supplements, traditional "willpower". But none of these solutions works for everyone, most of them are expensive, and when someone tries hard yet fails to keep the weight off - which the overwhelming majority of dieters do - it's another nail in the coffin. Why should they keep trying to lose weight, only to fail over and over? (This is the theme of Gina Kolata's excellent book, Rethinking Thin.)
The answer to that - the reason we should keep trying - lies in Megan's point about heritability. We are all genetically different, and just as we need different sizes of shoes and individualized study habits, we need personalized diet/exercise solutions. It may take a while to find the right one. I've struggled with my own weight my entire life, but I only got it under control when I had a double epiphany: I have a PhD in biology and I know my body better than anyone else, so why on earth was I taking advice from rail-thin dieticians and musclebound trainers? I came up with an admittedly draconian plan; my naturally thin friends feared I had become anorexic and harassed me for eating so little. But I knew I have a really slow metabolism. I had never eaten enough calories to sustain my BMI, as calculated by conventional dietary equations, and a quick-and-dirty oxygen consumption test in college placed me an order of magnitude lower than my thinner classmates. My body's rules are different - and more annoying - than most people's. Once I accepted that and ignored other people's well-meaning attempts to make me use *their* miracle diet plans (some of which advocated more calories than I was eating before I started dieting!), I lost a lot of weight. After two years, I was running five miles a day and in the best shape of my life. Unfortunately, I got mono last year and gained a lot of weight back, so I have to start over. But I know that I can do it - I simply have to play by my body's rules. That makes an incredible difference.
Anyway, I think the point Megan McArdle makes in the Atlantic piece is sorely needed. Yeah, it's a flawed analogy. But so what? She isn't saying that high heritability means weight is *only genetic.* Height isn't *only genetic* either. Plus, the end of her article summarizes a list of candidate hypotheses for why Americans are increasingly obese, and virtually all are about diet. What McArdle is trying to do with her imperfect weight/height analogy is force people to see weight from a new angle - as something that individuals do not have complete control over. It's extremely easy to judge fat people if you've never felt gnawing, dizzying hunger for weeks on end, because your brain isn't wired to send those signals. But just because you don't feel them doesn't mean no one else does - any more than the fact that Novocain works on you means it will work on my mom. What works for one person will not necessarily work for another. It's that simple.
When it comes to placing blame, obesity is remarkably similar to drug addiction. How many of us have friends and family who want desperately to stop smoking, are shamed by social taboos against it, financially strapped by the cost, can feel it affecting their lungs, and yet keep smoking anyway? Heritability estimates (based largely on twin studies) for addiction are at least 0.5 - in some studies, the heritability estimates are comparable with those for obesity. Genetic factors do influence whether people who try drugs like nicotine will become addicted to them. And this makes intuitive sense: we've all seen friends start smoking, don't like it, and quit, while others end up smoking a pack a day and can't stop. We know addiction is not one size fits all, because we've seen it firsthand. Yet people still use anecdotes to justify global policy assertions about addiction.
Consider those vocal people who got over their own addictions to alcohol or drugs cold turkey, and think that anyone who can't do the same is weak and worthless. Anecdotes do not translate into a global prescription. Good for you that you quit cold turkey, but most addicts do need treatment and help to quit. While they did make the choice to take drugs in the first place, it's not their fault that their bodies and brains reacted differently to the drugs than yours did. On the other hand, we also have people who say that marijuana is not addictive or harmful, because they personally have been smoking for twenty years and aren't addicted or harmed. Well, good for you - but some people do become addicted to marijuana, and any plan to legalize marijuana should take the need to treat those people into account, not deny they exist. People are different. (Duh.)
I'd argue that McArdle's piece, and those like it, do more good than harm. We need more voices saying that yes, people do indeed respond differently to diet and exercise, through no fault of their own. That doesn't make being unhealthy okay - we have to avoid the trap of genetic determinism. And whether or not current BMI standards are realistic (or valid) is a different argument entirely. But we should give more credit to people struggling with their weight, because some of them are carrying a heavier load than others - literally. And the sooner we recognize that as a society, the sooner we'll be able to help them.
- Log in to post comments
Bless you.
If you are happy, you are beautiful - regardless of size or shape.
Since height is a key determinant of weight, and height is highly heritable, wouldn't it naturally follow that weight is highly heritable as well?
Jessica,
I don't disagree with you; genetic variation is important. But there are several areas where I think McArdle is really wrong (and I'm leaving aside her anti-health insurance reform agenda; "We can't 'fix' obese people, so preventative medicine is foolish"):
1) Heritability measures populations, not individuals. If that's her point, then she should discuss the norm of reaction literature, not heritability--it's not germane to her argument.
2) We can have an environmentally-determined general increase in BMI which could lead to bad health outcomes, while still having significant individual-to-individual variation. We can do something about the general increase.
3) Related to #2, changes in environment can affect various people differently. A 'heavier' environment can do very little to one person's BMI (and potentially lead to adverse outcomes), but really change another's BMI. But without being able to identify which genomes are more susceptible, we need to encourage healthy behaviors, whether or not they lower BMI.
4) I view BMI and 'overweight' and 'obesity' as proxies for a 'healthy' state--they are correlated with bad health outcomes. Obviously, some people run heavy, and some run 'light', but these are still reasonable proxies. Again, the issue isn't 'fatness', but healthy diet and exercise, which in some cases, will result in 'slimness' and in others have little effect.
5) There is a larger issue, which is that heritability is also misused in behavioral and IQ studies, which just bugs me (although I'm easily bugged....).
Jon, we're being sloppy here with terminology - "weight" is being used in the sense of body mass, usually measured as BMI, which compensates for height already.
Mike:
1) Heritability measures populations, not individuals. Yes. But I don't understand your objection. Are you saying that it doesn't matter to an individual whether the disorder they suffer from has been shown to have a genetic contribution? I'd say that it matters a lot. It's true that you must not equate ".85 heritability" to ".85 of any individual's weight is genetic" - but I don't think McArdle did make that generalization.
"A 'heavier' environment can do very little to one person's BMI (and potentially lead to adverse outcomes), but really change another's BMI:" that is exactly my point about genetics: we're not inbred lab rats that respond in sync to our diets. When you say "we need to encourage healthy behaviors," we may disagree if you think all healthy behaviors are equally effective for all people. While a generic "more exercise!" is almost always good advice, some healthy behaviors are far more effective for certain individuals than for others. If we do not have the genetic markers to predict which will work, at least we can admit when a strategy is not working and try something else. Frankly, I don't see a lot of accounting for such individual variability in the treatment of obesity. I think that is a real problem.
"Again, the issue isn't 'fatness', but healthy diet and exercise, which in some cases, will result in 'slimness' and in others have little effect." Agreed, except that unfortunately society judges people on their "fat" or "slim" appearance - not their resting heart rate, ability to run five miles, or cholesterol levels. "Fatness" is an issue, insofar as it is a justification for social stigma, even when it is not the best indicator of health.
"heritability is also misused in behavioral and IQ studies." Yes, and I understand your sensitivity to it. But I just don't think McArdle is doing the same thing in this article. I don't see anything more sinister than an effort to discuss genetic variability in plain language - something that is very hard to do.
Heritability would measure how much genes contribute to
a trait varying from the mean (average). For obesity, the
mean itself has shifted, in this country, in the last few
decades. This seems to invalidate heritability as a useful
guide to figure out how much genes make one obese. The shift
in the mean is more significant than the individual
variation from it.
The mean seems to be set by genes and the environment
combined. Changing the environment should make a difference.
Some of that change can possibly be willed, in the common
sense of that term - say by avoiding eating with gluttons.
Heritability assumes genes and the environment influence
a trait linearly, that is the value of the trait variation
goes up directly proportional to the gene or environment
value.
This is a problem with height and weight. Weight is
proportional roughly to the square of the height. This means
genes can contribute linearly to either height or weight,
but not both. If it is linearly related to height, then it
is related quadratically to weight. If linearly related to
weight, then related by a square-root to height. So
heritability can be used only for height or weight but not
both. Height is the more likely candidate.
Using the logarithm of how much the weight varies from the
mean makes the value related linearly to the genetic
contribution. But this makes the numbers very small and
imprecise. People do not typically weight multiples of
other's weights.
Ajoy, there are plenty of influences on weight other than height - fat deposition, muscle mass, metabolism, hormone levels, etc. It's hard to tease out exactly which factors contribute because height and weight are complex traits influenced by many different factors, both genetic and environmental, many of which interrelate. But that doesn't mean you can't estimate heritability for weight (as measured in BMI) while controlling for height in the analysis. People do that sort of analysis all the time.
Jessica, thank you for this thoughtful post. I just returned from a conference of an organization I belong to called the Association for Size Diversity and Health (ASDAH). What you've said above is very consistent with the message of ASDAH, and I hope you and your readers will check out the website (http://www.sizediversityandhealth.org/index.asp). Basically our message is that obesity and BMI alone are not sufficient indicators of health, and that by incorporating healthy habits such as mindful eating of healthy foods, listening to your body's signals and exercising in a manner you can enjoy will improve anyone's health. It has been definitively shown, time and again, that the vast majority of dieters gain back their weight and then some, so there has to be another way of getting healthy, even if one is fat. Gina Kolata's book "Rethinking Thin" is excellent, as you mentioned above. I would also recommend Linda Bacon's book "Health at Every Size."
Ajoy, we know that nothing in physiology is linear. Sometimes a linear model may be useful, but we know that it is wrong.
Twins share an in utero environment, so there can be plenty of shared epigenetic modifications to their DNA that program the adult phenotype but which are not inherited from the genomic DNA. Many disorders such as obesity, diabetes, heart disease, hypertension, are affected by in utero epigenetic programming.
That changing epigenetic environment is very likely the âcauseâ of the apparent increased heritability of height in the past. The genetic heritability of height has not changed in a generation. The gene pool of humans simply hasnât changed very much in a generation to have a dramatic effect on such a polygenetic trait as height or weight. That also goes for every other polygenetic trait, which includes essentially all of the chronic diseases and disorders including obesity, diabetes, heart disease and hypertension.
If the in utero environment changed in a characteristic way that epigenetically programmed increased adult size that would show up as a decreased heritability of adult size because there is a skewing of the average size higher.
We know that the size, mass and growth rate of farm animals can be increased and their sexual maturity accelerated by giving them antibiotics in their feed. The mechanism for this change is unknown, but we know it canât be genetic. We know it isnât composition of diet either, other than the presence or absence of antibiotics. Many of the effects of antibiotics in animal feed mimic the changes observed in humans in the past 150 years, where the age of menarche declined from ~17 in 1850 to about 12 today. Increased height and weight and increased efficiency in converting food to biomass is observed in humans and in farm animals fed antibiotics.
The antibiotic effect is likely not a chemical effect of antibiotics because diverse antibiotics which work via multiple mechanisms have the same effect. Antibiotics donât increase the height and weight of bacteria-free animals. The changes in humans started to occur long before antibiotics were developed.
I want to make two comments, while I agree with the comments about hereditability, thin people are as a rule not thin because they have more self control, the fact is there is an obesity epidemic which did not exist 40 years ago.
My first comment is this: completely annecdotal, I travel between South Africa and the US frequently. Obese people in the US have different fat deposits to obese people in South Africa. There seems to be fat deposits around the neck and groin that you just don't see in South Africa ( I am talking specifically about the Caucasian population in South Africa). Obesity rates are very similar, but there is one significant difference in the food that is consumed. In South Africa junk food is sweetened with cane sugar (readily available and cheap), not glucose syrup. This is changing, would be interesting if the pattern of fat deposits also change.
The second comment I want to make is that if I go onto an SSRI to control depression, I gain 5-10 pounds, when I discontinue the medication because of this weight gain, I eventually revert to my previous setpoint. What is the difference? I consistently run 20 miles a week. The difference is appetite, SSRI's stimulate my appetite for Carbohydrates. It really is the difference between feeling sated on a modest diet, and feeling as if I am starving all the time.
Is the widespread use of SSRI's not also partly responsible.
I can't comment on how sound the basis of genetics and weight isâit's not my patchâbut I think the general point you make about that knowing that genes do (strongly) influence a health issue can be a good thing is a good point. I was talking recently with a colleague whose wife is doctor and he commented that there may be a shift to preventative medicine; with a knowledge of a person's genetics, there is an opportunity to learn what a person is susceptible to and perhaps from that help them in a preventative way. (Your own approach to you weight loss seems to be founded on a similar approach but using physiological measurements rather than genetics.)
This was an amazing read for me.
One of my favorite posts so far, even if it's not the typical art & science that I have grown to adore you for.
Not only was I shaking my fist as I was reading this and saying "YES!" victoriously over and over again, but I've *never* gotten the opportunity to be exposed to anyone else who has discovered the same thing that I have about simply listening to one's body and realizing the the caloric intake recommended is just not reasonable.
I had an almost identical experience philosophically- I was clinically obese, and finally decided to face this problem head on. After looking at many "diets' and discovering that their caloric recommendations were for more that I had been eating while I was *gaining* weight, I realized they were, at least for me, unusable.
I have a fantastically slow metabolic rate, and even with exercise, still burn more than most people who sit on their bums all day.
Anyway, my methods were also "draconian" (excellent choice of words) and I was chastised for under nourishing- but I managed to drop to a healthy weight and remain healthy and strong and content.
It's amazing what simply acknowledging our variations can do.
Anyway, thanks- as always, you rock my nerdy socks
Thanks Esmeralda! So glad to hear your story - I love to hear from someone else who bucked the popular wisdom and did her own thing to get where she needed to be. w00t! :D
I've always been fat, hopefully I manage to find something as successful as you have.