Complications with hypertension, diabetes and other diseases obscure statistics.
The rising level of obesity, alongside diabetes, cardiovascular disease and other related diseases, is predicted to slow down or reverse the decline in mortality seen in most Western countries.
But closer examination shows the picture is more complicated than that.
Obesity rates have been increasing in developed countries since the 1950s (albeit slowly at first). And since the 1960s, deaths from disease have steadily decreased in most developed countries while life expectancy has consistently increased.
It seems obesity's negative impact on mortality may be outweighed by other factors favourably influencing life expectancy. In the United States, for instance, cholesterol, smoking, and physical activity levels have all improved in recent years. Indeed, the rate of decline in mortality might have been even faster if it weren't for the increasing prevalence of diabetes.
But some researchers predict that the negative effects of increasing levels of obesity will soon outweigh the benefits from reductions in smoking.
Improved medical treatment of some of the pathways linking obesity to mortality may have also blunted obesity's negative impact. Hypertension, for example, has been better managed in recent years.
The impact of obesity may itself have been overestimated because its main adverse effects are experienced by a minority of the population. The best estimates of the association between body mass index (BMI) and mortality suggest that the mortality risk from excess body weight increases from a BMI of 25 but isn't substantial until BMI exceeds 32 or 35.
Between 15 per cent and 25 per cent of the US population have BMIs in this range. While this is a significant proportion, it is nevertheless a minority. And the relationship between obesity and health appears to reverse in old age.
In old age, people who have low body weight are at higher risk of disability and mortality. But this reversal may be because of weight loss in old age through disease.
Indeed, body weight may not be a significant risk factor for mortality in itself. Instead it might simply be a surrogate marker for a particular lifestyle, or a particular diet, physical activity level, and genetic factors. If this were so, obese individuals would represent a heterogeneous group of people with high body weight for different reasons, some of which may not be strongly related to morbidity or mortality.
But the negative impact of obesity trends, including longer duration of living with obesity, may not have yet affected life expectancy through the lag time between the onset of obesity and disease. And the adverse impact of obesity may be because of both its severity and duration.
As extreme obesity is becoming increasingly common and people are becoming overweight and obese earlier in life, it may be with more recent cohorts that we observe the deleterious impact of obesity on life expectancy.
Mortality statistics demonstrate, for the first time in over 25 years, a slight decline in US life expectancy. But diabetes and cardiovascular mortality rates have maintained a continual decline.
Studies of obesity's effect on cardiovascular disease and mortality, and of the positive effects from weight loss after bariatric surgery, suggest relatively short lag times of 10 years and less (two years for bariatric surgery).
Then there's the "obesity paradox" in which survival from acute events such as myocardial infarction, heart failure, and dialysis for renal failure is better in patients who are overweight or obese.
The reasons for this are unknown, but its unlikely to be simply that very unwell patients lose weight because of the severity of their illness. It's also possible that diseases such as hypertension and diabetes may be more aggressive and of a different origin in lean individuals than when associated with obesity.
Considerations such as lag time, duration of exposure to obesity, and increasing proportions of the population with severe obesity suggest that obesity may in the future have a considerable impact on life expectancy.
But there are also important reasons why obesity may not be strongly linked to life expectancy, except at the extremes of the weight distribution.
Several studies suggest that obesity's main adverse impact is the risk of becoming disabled, rather than on life expectancy itself.
It's also possible that health and life expectancy gains could be even greater if it weren't for the increasing prevalence of extreme obesity.
Helen Walls is a research fellow at the national centre for epidemiology and population health at Australian National University.