Diet, Diabetes, and Doubt: Is Preventive Medicine Lost in Space?

The whole story on the federally-funded Look AHEAD trial


A large federal trial, looking at lifestyle—diet and exercise—for the treatment of diabetes was just terminated because, after 11 years, it wasn't working as intended. The Look AHEAD study was stopped early because it was not reducing the rate of heart attack and stroke in the intervention group relative to the control. The termination was reported in a press release by the National Institutes of Health, and picked up by mainstream media. The findings suggest that diet and exercise are not effective for reducing the cardiovascular complications of diabetes.

And so, AHEAD, or at least the media coverage of it, is inviting us to look back, and doubt what we thought we knew about diet as the best medicine we've got—for diabetes, at least. We thought we knew that lifestyle was among the most powerful determinants of health outcomes. We thought we knew that diet and exercise together could prevent heart attacks in high-risk people. Participants in the AHEAD intervention lost 8 percent of their body weight by the end of the first year of the trial and were still down 5 percent from their baseline weight at the four-year mark. We thought we knew that diet, exercise, and weight management like this exerted important influences on the course of diabetes. Now, the AHEAD findings suggest we were wrong. Right? Not so fast.

For one thing, the trial did generate many noteworthy benefits. Prior papers in the Archives of Internal Medicine and the New England Journal of Medicine have reported significant benefits of lifestyle intervention related to weight loss, fitness, blood glucose levels, blood pressure, cholesterol levels, and mobility. The study was terminated for failing to prevent heart attacks and strokes, but it did reduce medication use, and conferred other benefits—such as a significant reduction in sleep apnea.

The AHEAD methodology also helps account for the putatively disappointing results of the long-term study. Diabetes requires treatment—so all patients in AHEAD were treated. Those in the lifestyle intervention group reduced their reliance on medication, while those in the control group took more. But since failing to treat diabetes with state-of-the-art medication is unethical, everyone was provided that. The study was actually comparing feet and forks to pharmacotherapy. When both intervention and control groups are being treated, differences between them diminish, an occurrence known in research as "bias toward the null." This exerted a profound effect in the AHEAD trial, making the positive findings more noteworthy still.

But to the extent that the negative results, with regard to cardiovascular event prevention, remain both surprising and disappointing, there is a fundamental explanation for them: too little, too late. What works for prevention may not always work nearly as well for treatment.

Need an image to help that notion really resonate? Let's talk about jumping out of an airplane. A parachute is great for preventing a high-velocity collision with the ground,but it's of no use at all if opened after the landing. Sometimes, timing is everything.

The Diabetes Prevention Program (DPP)—the precursor to AHEAD, in fact, and based on the very same lifestyle intervention—showed that diet, exercise, and modest weight loss could prevent the development of diabetes in 58 percent of high-risk adults. The best drug we've got, metformin, was only half that good. And let's be clear: complications of diabetes don't happen when the diabetes doesn't happen.

The DPP was not a warm-up band for AHEAD. It was a huge federal trial in its own right, run by many of the same people who ran AHEAD. It enrolled thousands of pre-diabetic adults and was supported with a budget of $174 million. The DPP administered the lifestyle intervention adapted for AHEAD. We can't toss out the original DPP results just because AHEAD didn't serve up a repeat. The results of prior trials don't vanish just because new results come along. Whatever we need to learn from AHEAD needs to be reconciled with what we learned before.

We have long had evidence that a comprehensive lifestyle intervention can shrink plaque in coronary arteries. We have evidence that it can prevent heart attacks in high-risk individuals. And we have evidence that it can even change gene expression, and potentially reduce the risk of cancer occurrence, recurrence, and progression.

How do we reconcile such findings with the latest from AHEAD? As a case of a bit too little, quite a bit too late.

Consider that famous space mission analogy: A mission to outer space is running just a bit off course. If you fix it now, it's a minor adjustment—a small effort for an excellent outcome. The longer you wait, the more magnified that small diversion from the planned course. The later the correction, the larger the correction needs to be—and the less likely it is to be enough. Once you are in the wrong solar system, the show is over.

Preventive medicine is like that. We like to say it's never too late to benefit from health promotion—and that's mostly true. If health is heading off course, it's relatively easy to fix right away, possible to fix early, and hard if not impossible to fix late.

The DPP and AHEAD placed their bets on a balanced, prudent, moderate lifestyle intervention that is much more conducive to population-wide adoption, perhaps, than the intensive interventions that have shrunk coronary plaque and changed gene expression—but it's also a lesser dose of lifestyle as medicine. In the case of AHEAD, it appears to have been too little. The message here is that it takes a lesser dose of lifestyle to prevent diabetes than to prevent complications once diabetes is established. With that space mission in mind, this is no great surprise.

We have seen examples before of the disadvantages of intervening late. Studies of nutrients we have long thought could do us good—like vitamin E, vitamin C, and other antioxidants—have failed to prove beneficial, or even done harm, when applied to those with established disease. We have evidence that there is slight, net harm from hormone replacement therapy after menopause when it's provided late—and net benefit when it's provided early. Studies of fiber supplementation have failed to show benefit in those with a history of colonic polyps. But it may be that once we already have colonic polyps, the injury to our intestines that fiber could have helped prevent is already a fixture.

Lastly, diet and exercise have the potential to reverse type 2 diabetes. If that did not happen in AHEAD, it implies that the study subjects had relatively advanced diabetes that was already somewhat resistant to the effects of diet, exercise, and weight loss.

We like to say it's never too late—but that, of course, isn't quite true. It just stands to reason that the acute administration of broccoli and blueberries would do little good during a cardiac arrest. That does not belie all the claims made about the benefits of eating more of them over a lifetime. The law of diminishing returns pertains in preventive medicine as elsewhere. Early and good investments in lifestyle as medicine yield huge returns. Later, smaller investments yield smaller returns.

The results of any contrary study must be interpreted in context, and appended to what we already knew or hoped we did. We don't benefit when we simply replace our understanding with a new one each time an alternative view is published. That's like looking around in the middle of the night and concluding the earth is dark, then standing in daylight and concluding it's bright. It is, of course, both, at different times. And in the case of this contrary study, time is of the essence. Lifestyle is truly powerful medicine. But sometimes it is too late, and sometimes it is too little.

Preventive medicine, and lifestyle as medicine, are by no means lost in space. But they may have much in common with that space mission. Decisive corrections made early produce powerful effects. Too little done too late may be better than nothing, yet still leave you among the wrong stars.

Lifestyle remains good medicine once diabetes has developed; overall, the AHEAD findings confirm that. But we've known for years that the AHEAD intervention can prevent diabetes outright when administered earlier—so let's apply it accordingly. The fault lies not with the fundamentals of prevention, but with our cultural proclivity for procrastination.

Hungry for more? Write to with your questions, concerns, and feedback.

David L. Katz, MD, MPH, FACPM, FACP, is a specialist in internal medicine and preventive medicine, with particular expertise in nutrition, weight management, and chronic-disease prevention. He is the founding director of Yale University's Prevention Research Center, and principal inventor of the NuVal nutrition guidance system. Katz was named editor-in-chief of Childhood Obesity in 2011, and is president-elect of the American College of Lifestyle Medicine.