Physical activity and erectile dysfunction: meta-analysis of population-based studies.

We searched for population-based cross-sectional studies, cohort studies and randomized controlled trials (RCTs) on erectile dysfunction (ED) through Medline, PubMed, PsychInfo and scanned though reference lists. Studies that did not include adjusted odds ratios (OR) of physical activity were excluded. Seven cross-sectional studies were suitable for meta-analysis, and the results from one cross-section study, two cohort studies and one RCT were summarized. Pooling the ORs using random effects models, we derived summary estimate for adjusted OR of physical activity in those with ED compared with those without ED, which was 0.53 (0.31, 0.91). Moderate and high physical activities were associated with a lower risk of ED, with ORs at 0.63 (0.43, 0.93) and 0.42 (0.22, 0.82), respectively. Funnel plot by visual inspection, and Begg's test and Egger's test did not detect significant publication bias. Sensitivity analyses revealed that the summary estimate from the random effects model was robust to changes in study sample size and level of statistical adjustment, but not so robust to changes in ED definition, although the summary estimate for each ED definition did not differ significantly. Although causality cannot be demonstrated from cross-sectional studies, the apparent 'protective' effect of physical activity on ED should be further investigated using large-scale cohort studies or RCTs.