Obesity has been trumpeted as the biggest health risk of modern times. But is it really? Jerome Burne reports on startling new research

Obesity is the modern health obsession. Every week seems to bring a new warning about the health "timebomb" that threatens to cut short the lives of future generations unless radical action is taken to reduce Westerners' weight. Earlier this month, the British Medical Journal warned in stark terms that this epidemic needed a politically driven public health policy to curb it.

But could the dangers have been overstated? Recent authoritative research suggests that the obesity epidemic is more of a mirage conjured up by various statistical glitches. In fact, if you're officially classified as overweight, you may have a lower risk of premature death than those deemed to be a healthy weight. And that's not the only major obesity rethink underway at the moment. Long term, not dieting may be better for your health than trying to lose weight.

Only a year ago, a widely quoted research paper claimed that obesity in America was responsible for 400,000 extra deaths every year - almost on a par with the toll exacted by smoking. This figure has just been radically revised downwards by a paper in the Journal of the American Medical Association, which suggests that "it is still far from certain whether there is any measurable mortality toll at all among overweight and obese Americans as a group." The author, Dr Katherine Flegal, a researcher at the prestigious Centre for Disease Control and Prevention (known as the CDC) in Atlanta, found that "even severe obesity failed to show up as a statistically significant mortality risk".

This is a remarkable turnaround and, unsurprisingly, now the topic of a fierce debate. However, several academics see it as a powerful vindication of what they have been saying for years. A detailed account of the doubter's case is put forward in the June edition of the Scientific American. "We are creating a disease simply by labelling it as such," says Paul Campos, professor of law at the University of Colorado and author of The Obesity Myth.

What he means is that, at the moment, people are labelled as overweight at an arbitrary point on the BMI (Body Mass Index) scale, which relates your weight to your height. Score between 19 and 24, and you are officially "healthy", but between 25 and 30 you immediately become overweight, while anything over that makes you obese. More and more people are falling into the "overweight" and "obese" categories, but this is only cause for alarm if being overweight automatically means a rise in mortality. It's this assumption that is being put under the microscope.

There are more than 100 independent risk factors for heart disease, such as poor diet, lack of physical fitness, high stress levels, being poor and certain gene variations. But most studies linking obesity to heart disease lump them all in with being obese.

However, when Dr Flegal separated them out a quite different picture emerged: the number of deaths linked with being "moderately and severely obese" ranged from between "122,000 more to 7,000 fewer". Her analysis is controversial and there is obviously a wide margin of error, but it does suggest that warnings about epidemics may be rather premature. The implication is that if you are overweight, smoke, eat poor food, don't exercise and so on then you are certainly more at risk from illness. But if you are overweight - or even obese - but have a healthy lifestyle, then the risk becomes less certain.

"By the same criteria we are blaming obesity for heart disease, we could accuse smelly clothes, yellow teeth or bad breath for lung cancer," says Eric Oliver, a political scientist at the University of Chicago and author of Obesity: the Making of an American Epidemic, due out in the autumn.

Oliver and his fellow contributors to the Scientific American article have other arguments, too. For instance, over the last two decades the average BMI of Americans has risen remorselessly. However, their average levels of risk factors for heart disease, such as blood pressure and cholesterol, have gone in the opposite direction. Far from rising, as you would expect if they were being driven by obesity, they have declined by about 50 per cent. What's more, the increase in diabetes in the 1990s was not nearly as sharp as the rise in BMI.

Exactly what practical effect this major rethink will have isn't at all clear. Dr Flegal has been accused of giving comfort to the junk food lobby by downplaying the dangers of being overweight. But the new analysis doesn't offer a refuge to pizza-scoffing couch potatoes. Obesity itself may not be such a killer as was previously thought, but that is probably partly because of improved treatments and because more people classed as overweight by their BMI are exercising, eating a whole food diet and avoiding obvious risky behaviours.

The critics not only believe that the dangers of obesity are overstated, but also question the value of dieting to reduce it. As many as 75 per cent of Americans are trying to lose weight or keep it off at any one time, but what is the effect in the long run? A recent survey by CDC of people over 65 found that only 6 per cent of officially non-obese older adults had been obese a decade earlier. In other words, even if the weight comes off it very rarely stays off.

Campos, for example, points out that there is a strong genetic component to the amount of fat you store, which means that there is no safe and practical way for losing more than about 5 per cent body weight long-term. As a result, advice to stay within BMI range is "literally impossible" for many. It is data like this that has led a group at the University of California, Davis to try a new approach called "Health at Every Size".

This doesn't involve dieting; instead, people considered to be overweight are taught how to pay more attention to the internal body cues that signal hunger and fullness, and encouraged to be more accepting of their large size. Last month, the group reported on a two-year trial involving overweight and obese women, half of whom went on the new course and half of whom were on a regular diet.

Those on the course were not only happier but also healthier than the dieters. They were four times more likely to be engaged in physical activity, had significantly improved blood pressure and cholesterol levels, and were less depressed.

It seems that a campaign that concentrates on boosting people's health, rather than demonising them for being the wrong size, could be more rewarding - and more effective.