Steroids in alcoholic hepatitis: Another salvo of data

Sixty-one of 65 original patients who were randomized to receive a month of prednisolone (50 mg/day) or placebo for acute alcoholic hepatitis are reported. Only those with clinically severe disease as indicated by the presence of hepatic encephalopathy, a discriminant function of at least 32, or both, were included. The 'discriminant function' is based on previous published work and on their own retrospective analysis which indicated that when [4.6(prothrombin time minus control)] plus [the serum bilirubin (in μmol/L) divided by 17] is 32 or greater, there is at least a 50% chance of dying. [This can also be expressed as 4.6 x prothrombin time (in seconds) plus the serum bilirubin (in mg/dl), using 93 as the cut-off discriminant function (1).] All patients had liver biopsy confirmation of disease, regardless of coagulation defects. Those with gastrointestinal bleeding or infection were included only if the complication was controllable within 48 h. The main finding was that at both 1 and 2 months after randomization, there was a statistically significantly improved rate of survival favoring the treated group. Eighty-eight percent were alive at both times, compared with only 62% and 45% survival at those times for the placebo-treated patients. Most patients met the discriminant function measure of severity, but were not encephalopathic; the benefit of corticosteroids was apparent for both those who were and those who were not encephalopathic. The dosage of prednisolone generally was well tolerated. Glucose intolerance was seen, and one patient dropped out because of mental changes induced by the prednisolone. On the other hand, infection and gastrointestinal bleeding were more common in the placebo-treated patients.