Do diets work? The great majority of doctors think so. Experience tells us to expect an enormous failure rate, yet most of us continue to hold out hope that the diets we prescribe will result in lost weight or other health benefits. Why do we keep believing? First of all, because dieting really ought to work—here's the standard explanation. Calories are energy. The body stores excess energy (beyond what we need for motion and the chemical processes of life) in fat. If we cut down our caloric intake (or increase our energy output) below the amount of energy the body uses, our weight will fall.
This should happen in a predictable and mechanistic way, just as we were taught in pre-med science courses. Thus, dieting failure is supposed to reflect failure of adherence, which we've assumed to be the result of moral failure—lack of self-discipline and control. That probably explains why for years "obesity" wasn't recognized as a diagnosis by Medicare, Medicaid, and private insurers. But this approach hasn't worked well: Over the last 25 years, the obesity rate in the United States has jumped from 15 percent to about a third of the population, with another third deemed simply overweight. Insurers have come to understand that patients with untreated obesity develop more expensive medical problems in the future. Now they'll pay for dieting as a treatment—and we are forced to confront the question of what that's likely to accomplish.
There is considerable evidence that in the short term, calorie restriction generally leads to an average loss of about 5 percent to 10 percent of body weight. But the more important questions are these: Is the weight loss sustained over time? And does it have any long-term health benefits? These questions have recently been rigorously addressed in an exceptionally fine paper (PDF) by a group of researchers working with Traci Mann in UCLA's psychology department. These scientists systematically reviewed many studies of the long-term outcomes of dieting. And their results were hardly encouraging. Though the chemistry and physics of dieting seem straightforward, the biology and psychology may not be.
Mann and her colleagues examined three different kinds of research on the long-term outcome of dieting. First, they concentrated on the kind of studies that science likes best: random control trials. In these studies, patients interested in weight loss were randomly assigned to one of two groups. One group was given a calorie-restricted diet to follow; the other group was not. Many, many studies of this kind have been done over a relatively short period of time (often several months). But Mann and her group found only seven studies that followed their subjects over an extended period of more than two and a half years. And only a single "pure" long-term study randomly assigned participants to follow a diet or not, and simply measured weight change.
In this study, the experimental subjects were kept on their low-calorie diets for 18 months without other weight-loss interventions (like medications, exercise, or even pep talks) and their weight was measured again a year after the diet ended. A comparison of the dieting subjects with the control group (who were, instead, placed on a waiting list for a diet) showed some weight loss a year post-diet, but it was disappointingly small: an average decrease of 3.75 pounds. The other six long-term studies (which were muddier because they included other interventions besides diet) showed similarly unimpressive results: The average dieting weight loss maintained over extended time (between 2.5 and 10.5 years) was less than 2.5 pounds.
Mann and her colleagues also examined other kinds of studies on the long-term effects of dieting. In 14 studies that lacked control groups, but followed dieters for at least four years after a prescribed reduced-calorie diet, the average early weight loss after dieting was almost 31 pounds. But by the end of the follow-up period, on average the dieters gained back more than 24 of the pounds they had lost. In 10 studies in which nutritional scientists tracked the weight of people who put themselves on any diet of their choosing, the results were even worse. Of the 10 reports, only one described lasting weight loss, two showed no long-term effect, and the remaining seven studies found that dieting led to weight gain in the long run.
That result, of course, is particularly disturbing because it suggests that dieting is somehow monkeying with the body—perhaps by resetting the thermostat that controls how efficiently we utilize food—so as to make the struggle to lose weight all the more difficult. For instance, an excellent 2003 study of almost 15,000 preteens and teens followed for three years clearly showed that the kids who dieted gained more weight than the nondieters. This excess gain could not be explained by initial chunkiness, differences in caloric intake, or in the amount of energy contained in dietary fat.