In an Australian study, men and women age 70−75 were less likely to die over a ten year period if they were slightly overweight.

Much of the study's results aren't unusual: weighing too little raised the risk of dying and so did weighing too much. What was unusual was that the lowest risk was centered in the overweight range, not the normal weight range.

This doesn't mean that the elderly can totally ignore their weight. It may mean that current definitions of normal weight and overweight for the elderly are a few pounds low.

There has been an idea in the general population for some time that a few extra pounds can be beneficial to elderly people, serving as a reserve for stressful times. For instance, during a bout of the flu, the extra pounds can serve as an energy reserve and prevent a person from ending up underweight after the flu has passed. While the Australian study's results agree with this idea, the researchers take a slightly different meaning from them. They suggest that these are not extra pounds: when it comes to classifying the elderly as normal weight or overweight by body mass index (BMI), the bar is set a bit too low.

The use of BMI as a weight standard became common in the 1980's. BMI assigns a normal weight for a person based on their height. It recognizes four main weight classes: underweight, normal weight, overweight and obese (very overweight). But the actual weight range chosen for each class was set primarily from studies of mortality and disease on younger and middle aged people. Very little information came from studies on the elderly. It is possible that BMI charts and tables are slightly low in their classification of normal weight and overweight in the elderly.

While being overweight has consistently been shown to lower life expectancy in younger and middle aged people, a few studies have suggested that a BMI in the overweight range does not raise mortality in the elderly. The present study sought mainly to answer one question: what BMI weight class is associated with the lowest mortality rate in the elderly? The study answer was the overweight range.

The study began in 1996 and followed over 9,200 men and women, aged 70−75, for 10 years or until their death. The risk of death during this period was about 13% lower for overweight subjects than for normal weight subjects. This benefit was not seen in obese subjects.

A weakness of the study was that subjects' weights were given only once, at the start of the study, and were self−reported, not measured. No information was kept on how a person's weight fluctuated during the 10 years. A strength of the study was that it looked separately at healthy and ill subjects. Older people who become unwell for any reason often lose weight before death; this could give the appearance of lower weight increasing mortality when it was only one effect of serious illness. However, the study results were essentially the same for healthy and ill participants.

While the exact meaning of these findings is sure to be debated, one message seems to be that being slightly overweight is better than being slightly underweight around the age of 70. And a finding of the study that's certainly not controversial is that being sedentary (lack of activity) increased the risk of death for all weight classes.

The results of the study were published in the February 2010 issue of the Journal of the American Geriatrics Society. The article is freely available. Looking at the shape and positioning of the mortality graphs may help bring the study results home in a more visual manner than words can.