While everyone knows that exercise is a good idea, hard, scientific evidence about its benefits in the old and infirm has been surprisingly limited.

“For the first time, we have directly shown that exercise can effectively lessen or prevent the development of physical disability in a population of extremely vulnerable elderly people,” said Dr Marco Pahor, the director of the Institute on Aging at the University of Florida and the lead author of the study. Countless studies have found a strong correlation between physical activity in advanced age and a longer, healthier life. But such studies can’t prove that exercise improves older people’s health, only that healthy older people exercise.

For this latest study, the Lifestyle Interventions and Independence for Elders, or LIFE, trial, scientists at eight universities and research centres around the US began recr-uiting volunteers in 2010, using an unusual set of selection criteria. Unlike many exercise studies, which tend to be filled with people in relatively robust health who can easily exercise, this trial used volunteers who were sedentary and infirm, and on the cusp of frailty.

They recruited 1,635 sedentary men and women aged 70 to 89. Then the men and women were randomly assigned to either an exercise or an education group.

Those in the education assignment were asked to visit the research centre once a month to learn about nutrition, health care and other topics related to ageing. The exercise group received information about ageing but also started a programme of walking and light, lower-body weight training with ankle weights, going to the research centre twice a week for supervised group walks on a track, with the walks growing progressively longer. They were also asked to complete three or four more exercise sessions at home, aiming for a total of 150 minutes of walking and about three 10-minute sessions of weight-training exercises a week.

Every six months, researchers checked the physical functioning of all of the volunteers. The experiment continued for an average of 2.6 years, which is far longer than most exercise studies.

By the end of that time, the exercising volunteers were about 18 percent less likely to have experienced any episode of physical disability during the experiment. They were also about 28 percent less likely to have become persistently, possibly permanently disabled.

Most of the volunteers “tolerated the exercise program very well,” Dr Pahor said, but the results did raise some flags. More volunteers in the exercise group wound up hospitalised during the study than did the participants in the education group, possibly because their vital signs were checked far more often, the researchers say.

A subtler concern involves the surprisingly small difference, in absolute terms, in the number of people who became disabled in the two groups. About 35 percent of those in the education group had a period of physical disability during the study. But so did 30 percent of those in the exercise group.

“At first glance, those results are underwhelming,” said Dr Lewis Lipsitz, a professor of medicine at Harvard Medical School. “But then you have to look at the control group, which wasn’t really a control group at all.” That’s because in many cases the participants in the education group began to exercise, study data shows, although they were not asked to do so.

“It wouldn’t have been ethical” to keep them from exercise, Dr Lipsitz continued. But if the scientists in the LIFE study “had been able to use a control group of completely sedentary older people with poor eating habits, the differences between the groups would be much more pronounced,” he said.

Over all, Dr Lipsitz said, “it’s an important study because it focuses on an important outcome, which is the prevention of physical disability.”

In the coming months, Dr Pahor and his colleagues plan to mine their database of results for additional follow-up, including a cost-benefit analysis.

Dr Pahor cautioned that the LIFE study is not meant to prompt elderly people to begin solo, unsupervised exercise. “Medical supervision is important,” he said.

—©2014 The New York Times