A fortuitously controlled study of steroid therapy in acute viral hepatitis: I. Acute disease

Abstract

An analysis was made of the course of clinical manifestations and liver function tests from the data incorporated in the charts of patients with hepatitis admitted to either of two hospitals in which treatment was comparable except that steroids were used frequently (89.5 per cent) at one and much less often (16.5 per cent) at the other. Both sexes and a wide spectrum of ages were represented. Hospital populations were compared rather than treatment groups in order to avoid the bias that exceptions to routine therapy would impose. Admission parameters were distributed in a compensating fashion which permitted use of the entire hospital populations in evaluating steroid effects. A special grouping method of covariance analysis was employed. Steroid treatment was associated with more rapid resolution of elevated serum levels of bilirubin throughout hospitalization and of transaminases and alkaline phosphatase during the first two weeks. The frequency of intercurrent illness, chiefly peptic ulcer and pyogenic infections, was twice as high at the hospital using steroids routinely. The relapse rate was three times that of the control hospital patients. These significant findings (p < 0.01) were partially offset by less convincing evidence that steroid treatment may prevent hepatic coma. Three per cent of the patients at the “steroid hospital” died of peptic ulcer complications. Neither hospitalization nor time lost from work was shortened by steroid therapy. Subclassification of patients according to prognostic attributes such as age or serum bilirubin levels on admission did not identify any category in which steroid therapy was more advantageous.