A systemized review 1 analysed study reports containing information on physical activity and weight changes during long-term follow-up of adult subjects. The studies were either observational without weight loss intervention or included some type of intervention. Medline search was started from 1980 and produced slightly more than 1000 references of which 46 met the criteria determined prior to the search. The studies were classified into four groups.
Prospective population-based observation studies had to have a follow-up of at least 2 years. Sixteen studies were included in this group. Physical activity was assessed retrospectively with different questionnaires. Those studies that had assessed activity only at baseline were inconsistent when the effect of physical activity on weight changes was evaluated. In five studies activity was assessed at the end of the follow-up and these results were more consistent: if the amount of exercise at the end of the study was large, weight gain was smaller. In 11 studies physical activity had been assessed at several time points, and in seven of these studies increased activity was found to be associated with lower weight gain.
Dieting intervention studies had to include a follow-up of at least one year to be included in this review. The BMI of the subjects had to be at least 25 at the beginning of the study, and the result of dieting a reduction of 5% from baseline weight. These studies were divided into three subgroups, two of which contained a weight reduction period (with physical activity as one treatment mode either randomised or non-randomised into different groups) followed by an observation period. In twelve of the 17 non-randomised studies ample physical activity at the end of the follow-up was associated with a smaller weight gain after weight reduction. Four studies contained information on physical activity from several follow-up points and came to a similar conclusion. There were only eight randomised studies, four of which did not find exercising to improve weight loss, in three studies the result seemed to favour exercising and in only one study exercising increased weight loss induced by dieting.
The third group consisted of studies in which the weight reduction was followed by a separate weight control phase during which the subjects were randomised to exercise or diet treatment. There were only three studies of this type, and in two of them physical activity did not improve the maintenance of the weight loss.
Comment: In the intervention studies the effects of exercising on weight have thus been rather modest compared with epidemiological follow-up studies in which increased physical activity appears to reduce weight gain. There are very few well-performed dieting interventions with a long follow up. Dropping out from the study is a problem that is emphasized in exercise interventions. On the basis of this systematic review 1 weekly energy expenditure of 6.3-8.4 MJ (1 500-2 000 kcal) in epidemiological observations studies seemed to prevent weight gain. This is more than the suggested amount of physical activity in many intervention studies and much more than what most subjects could adhere to.
Another systematic review 2 compared a combined exercise and diet intervention to diet alone and included randomized controlled trials with at least 1-year follow-up after the active intervention phase. No weight reducing drugs were allowed in the intervention groups. The health status of the participants was variable (pregnant women excluded). After the literature search (up to year 2003) and examination of abstracts, 6 trials with a total of 407 subjects were included. The duration of therapy intervention was 10 to 52 weeks and the duration of follow-up was 1 to 2 years. Drop-out rates varied between 9% and 24%. In the analyses, the follow-up duration was set to one year.
The participants had BMIs over 25. At baseline the weight was on average 97 kg (BMI not reported). After the active intervention phase, weight reduction was approximately 20% more in the combined exercise and diet group compared to diet alone (-13.0 kg vs. -9.9 kg, p=0.063; -13% vs. -10 %). Effect size expressed as standardized mean difference (SMD) was -0.20 (95% CI -0.41 to -0.01). After one year follow-up, weight reduction from the baseline was approximately 20% more in the combined exercise and diet group compared to diet alone (-6.7 kg vs. -4.5 kg, p=0.058; -7% vs. -5 %). Effect size was -0.20 (95% CI -0.42 to 0.01). Weight gain after the active weight reduction phase was similar in both groups (50%).
Comment: The trials did not report the number of participants who had a clinically significant reduction in their overweight (e.g. 5% or 10% from baseline). It would be more reasonable to measure changes in the relative amount of fat tissue or at least changes in body mass index than changes in sole weight.
The quality of evidence is downgraded by limitations in study quality.
Primary/Secondary Keywords