The association between corticosteroid therapy and subsequent infections was calculated by pooling data from 71 controlled clinical trials. The overall rate of infectious complications was 12.7% in the 2,111 patients randomly allocated to systemic corticosteroids and 8.0% in the 2,087 controls (relative risk [RR], 1.6; 95% confidence interval [CI], 1.3-1.9; P less than .001). The risk of infection was particularly high in patients with neurologic diseases (RR, 2.8; 95% CI, 1.9-4.3; P less than .001) and less pronounced in patients with intestinal (RR, 1.4; 95% CI, 1.1-1.7; P = .02), hepatic (RR, 1.4; 95% CI, 0.9-2.3; P = .25), and renal (RR greater than 1; P = .03) diseases. The rate was not increased in patients given a daily dose of less than 10 mg or a cumulative dose of less than 700 mg of prednisone. With increasing doses the rate of occurrence of infectious complications increased in patients given corticosteroids as well as in patients given placebo, a finding suggesting that not only the corticosteroid but also the underlying disease state account for the steroid-associated infectious complications observed in clinical practice.