How genetic is obesity?

The NYTimes ran an excerpt of a book called Rethinking Thin: The New Science of Weight Loss -- and the Myths and Realities of Dieting by Gina Kolata. Having read the excerpt -- I haven't read the whole book -- I take issue with how Kolata frames the issue of the genetics of obesity.

My problem with most articles and books discussing genetics -- particularly with respect to behavior -- is that they don't emphasize the concept of an environmental X genetic interaction in determining outcome. It's either all genetics or all environment.

Before, I talk about the genetics of obesity and the problems I have with this book, here is my main point: Whether or not you are obese is determined by the interaction of your genes and your environment. With respect to obesity, genes are permissive, not deterministic. They confer risk, not destiny.

Let's take a look at the excerpt. The excerpt summarizes two general pieces of evidence for obesity being genetic. The first is that when obese people lose weight they showed dramatically lower metabolisms when compared to skinny individuals of the same weight:

It seemed straightforward. Dr. Hirsch found eight people who had been fat since childhood or adolescence and who agreed to live at the Rockefeller University Hospital for eight months while scientists would control their diets, make them lose weight and then examine their fat cells.

The study was rigorous and demanding. It began with an agonizing four weeks of a maintenance diet that assessed the subjects' metabolism and caloric needs. Then the diet began. The only food permitted was a liquid formula providing 600 calories a day, a regimen that guaranteed they would lose weight. Finally, the subjects spent another four weeks on a diet that maintained them at their new weights, 100 pounds lower than their initial weights, on average.

Dr. Hirsch answered his original question -- the subjects' fat cells had shrunk and were now normal in size. And everyone, including Dr. Hirsch, assumed that the subjects would leave the hospital permanently thinner.

That did not happen. Instead, Dr. Hirsch says, "they all regained." He was horrified. The study subjects certainly wanted to be thin, so what went wrong? Maybe, he thought, they had some deep-seated psychological need to be fat.

So Dr. Hirsch and his colleagues, including Dr. Rudolph L. Leibel, who is now at Columbia University, repeated the experiment and repeated it again. Every time the result was the same. The weight, so painstakingly lost, came right back. But since this was a research study, the investigators were also measuring metabolic changes, psychiatric conditions, body temperature and pulse. And that led them to a surprising conclusion: fat people who lost large amounts of weight might look like someone who was never fat, but they were very different. In fact, by every metabolic measurement, they seemed like people who were starving.

Before the diet began, the fat subjects' metabolism was normal -- the number of calories burned per square meter of body surface was no different from that of people who had never been fat. But when they lost weight, they were burning as much as 24 percent fewer calories per square meter of their surface area than the calories consumed by those who were naturally thin.

This piece of evidence in and of itself is not evidence of genetics, however. It is evidence of physiological effects of environment. Basically, if you become obese than your chances of not being obese decrease because of a persistent physiological change in your body.

To make the genetic argument, the Kolata compares the behavior in these individuals to skinny individuals who deliberately gained weight and then tried to lose it:

It began with studies that were the inspiration of Dr. Ethan Sims at the University of Vermont, who asked what would happen if thin people who had never had a weight problem deliberately got fat.

His subjects were prisoners at a nearby state prison who volunteered to gain weight. With great difficulty, they succeeded, increasing their weight by 20 percent to 25 percent. But it took them four to six months, eating as much as they could every day. Some consumed 10,000 calories a day, an amount so incredible that it would be hard to believe, were it not for the fact that there were attendants present at each meal who dutifully recorded everything the men ate.

Once the men were fat, their metabolisms increased by 50 percent. They needed more than 2,700 calories per square meter of their body surface to stay fat but needed just 1,800 calories per square meter to maintain their normal weight.

When the study ended, the prisoners had no trouble losing weight. Within months, they were back to normal and effortlessly stayed there.

When we compare the results of these two experiments it is easy to draw the following conclusion: fat people had trouble losing weight because they are genetically fat, and the skinny had people could gain and lose weight easily because they are genetically skinny.

However, there are problems with that interpretation. (I will ignore for the moment the issue of sample sizes and whether you can really say that inmates lost weight willingly or under duress.)

Both experiments involved crash dieting or radical increases in weight. What if changes in the "set-point" of weight -- the weight your body naturally tends to stay at -- change more slowly? If this were true, we could expect the "genetically fat" people who remained skinny to have lots of trouble initially, but they would gradually change their physiology to match their lower weight. Likewise, with the "genetically skinny" inmates. If they stayed fat, their physiology could adjust to try and stay there.

This is fundamentally speculation, but I do think it is a reasonable criticism of using this work to argue for a genetic cause of obesity. There are examples of individuals who have maintained weight loss for long periods -- for example, the National Weight Control Registry.

The next argument the article discusses for the genetics of obesity are twin studies:

[Stunkard]...found the perfect tool for investigating the nature-nurture question -- a Danish registry of adoptees developed to understand whether schizophrenia was inherited. It included meticulous medical records of every Danish adoption between 1927 and 1947, including the names of the adoptees' biological parents, and the heights and weights of the adoptees, their biological parents and their adoptive parents.

Dr. Stunkard ended up with 540 adults whose average age was 40. They had been adopted when they were very young -- 55 percent had been adopted in the first month of life and 90 percent were adopted in the first year of life. His conclusions, published in The New England Journal of Medicine in 1986, were unequivocal. The adoptees were as fat as their biological parents, and how fat they were had no relation to how fat their adoptive parents were.

The scientists summarized it in their paper: "The two major findings of this study were that there was a clear relation between the body-mass index of biologic parents and the weight class of adoptees, suggesting that genetic influences are important determinants of body fatness; and that there was no relation between the body-mass index of adoptive parents and the weight class of adoptees, suggesting that childhood family environment alone has little or no effect."

In other words, being fat was an inherited condition.

Dr. Stunkard also pointed out the implications: "Current efforts to prevent obesity are directed toward all children (and their parents) almost indiscriminately. Yet if family environment alone has no role in obesity, efforts now directed toward persons with little genetic risk of the disorder could be refocused on the smaller number who are more vulnerable. Such persons can already be identified with some assurance: 80 percent of the offspring of two obese parents become obese, as compared with no more than 14 percent of the offspring of two parents of normal weight."

A few years later, in 1990, Dr. Stunkard published another study in The New England Journal of Medicine, using another classic method of geneticists: investigating twins. This time, he used the Swedish Twin Registry, studying its 93 pairs of identical twins who were reared apart, 154 pairs of identical twins who were reared together, 218 pairs of fraternal twins who were reared apart, and 208 pairs of fraternal twins who were reared together.

The identical twins had nearly identical body mass indexes, whether they had been reared apart or together. There was more variation in the body mass indexes of the fraternal twins, who, like any siblings, share some, but not all, genes.

The researchers concluded that 70 percent of the variation in peoples' weights may be accounted for by inheritance, a figure that means that weight is more strongly inherited than nearly any other condition, including mental illness, breast cancer or heart disease. (Emphasis mine.)

The abstract for Dr. Stunkard's 1986.study is here. The abstract for the 1990 study is here.

First, with respect to this evidence let me say that both of these studies for the heritability of obesity resulted in heritability rates -- about 70% -- that have been well supported by later research. I don't dispute their findings.

Second, the idea that obesity does have genetic risks is also widely accepted in the literature. To digress briefly, scientist have identified three types of obesity with respect to genetics. (Summarized in this review.)

• 1) Monogenic obesity -- These constitute a group of rare disorders where a single mutation causes obesity. Examples of this group are genes such as leptin and leptin receptor. However, monogenic obesity constitutes a relatively small fraction of the known cases.
• 2) Syndromic obesity -- These are syndromes of which obesity is a large component but not the only one. Examples are Prader-Willi syndrome (PWS) and Bardet-Biedl syndrome (BBS). These also constitute a small fraction of the known cases.
• 3) Polygenic obesity -- This is the overwhelming majority of obesity where many, many genes are involved and causality cannot be attributed to any one of them individually. In humans, over 127 genes have been identified with Single Nucleotide Polymorphisms (SNPs) that confer risk for obesity. (Although only 22 of these genes have been confirmed in more than 5 studies.)

With respect to this evidence, what I object to is not that obesity has genetic risks. That proposition has been overwhelmingly supported. What I object to is the notion that genetics is the sole determinant of weight -- which is what Kolata appears to be arguing for.

In order to make this argument Kolata ignores the huge environmental changes that have occurred in Western society over the past 100 years.

For example, take the Danish Study. What has changed in the world between now and 1947? Primarily what has changed is our diets. The amount of calories in our diets have increased dramatically since the beginning of the last century, and this has caused a commensurate increase in obesity rates.

The heritability of obesity may have been very high in that study, but that says nothing about the overall obesity rates, and those obesity rates have increased over time because of poor diet in Western society.

I am not the only one who has made this point. Emily Bazelon in her review of the book in the NYTimes writes:

She's surely right to push back against bafflement and intolerance, and her argument that we've tilted too far toward blaming fat people's bad habits for their weight is convincing. But Kolata goes so far in arguing for biological predestination that she sometimes seems to completely dismiss the other part of the fat equation -- what we eat. In all likelihood, the obesity rate has doubled in the United States since 1980 for all the familiar reasons: fattening food has never been so cheap, convenient and cunningly marketed. "The genes that make people fat need an environment in which food is cheap and plentiful," she writes. It's in a world of giant muffins and bowls of office candy that Americans need wider movie seats and larger coffins.

Kolata knows this. She touches on reasons that poor people are more likely than rich people to be overweight, all of them environmental. But she treats childhood obesity as virtually inevitable. In addition to the twin and adoption studies, she cites research showing that teaching kids to eat right in school, and serving them leaner lunches, has no effect on their weights. The researchers concluded that the intervention was too limited -- the children's diets needed to change at home as well as at school. But Kolata scoffs at the "popular solution," which is "not to question the premise but rather to increase the intensity of the intervention."

Fundamentally, I don't think that Kolata is making a scientific argument so much as defending a political proposition that obese people have been unjustly blamed for their condition, that it is a disease, and that they deserve treatment not questioning of their will power.

I am sympathetic to this position. Our society does not make it easy to eat healthy. (If you would like a full explication of the complete, sordid history of how our diets became so poor, I recommend reading Fat Land by Greg Critser.) Furthermore, there is a strong social stigma attached to obese people in this society, as if their failure to achieve a BMI less than 17 is a failure of nerve rather than the failure to beat near insurmountable odds.

However, just because the odds are stacked against them does not suggest that we live in a world of genetic determinism. Likewise, social trauma is not in my opinion sufficient reason to deny scientific evidence. To emphasize the genetic over the environmental may will not spur efforts any further to find pharmaceutical fixes to obesity -- those are going along quite well on their own. What it will do is stifle any attempt at lifestyle changes on the part of obese people by lulling them into a false sense of complacency that nothing can be done.

Something can indeed be done, and doing it requires first that we acknowledge to complexity of the illness we are dealing with.

Mutch and Clement summarize this complexity well in their review of obesity genetics:

The complex interactions underlying polygenic obesity demonstrate that genetic, social, behavioral, and environmental factors are all capable of influencing the obese phenotype.