Current issues of ACP Journal Club are published in Annals of Internal Medicine


Prednisolone improved short-term survival in patients with severe alcoholic hepatitis

ACP J Club. 1992 Sept-Oct;117:36. doi:10.7326/ACPJC-1992-117-2-036

Related Content in the Archives
Glucocorticosteroids are probably ineffective in alcoholic hepatitis

Source Citation

Ramond MJ, Poynard T, Rueff B, et al. A randomized trial of prednisolone in patients with severe alcoholic hepatitis. N Engl J Med. 1992 Feb 20;326:507-12.



To evaluate the effectiveness of prednisolone in improving short-term survival in patients with severe alcoholic hepatitis.


Randomized, double-blind, placebo-controlled trial of 2 months duration.


2 hospitals in France.


65 patients hospitalized between March 1987 and June 1990 with biopsy-proven alcoholic hepatitis (characterized by hyaline necrosis and infiltration of polymorphonuclear leukocytes) and with spontaneous hepatic encephalopathy, a discriminant-function value > 32, or both, were included. The discriminant function used was 4.6 (prothrombin time - control time [in seconds]) + serum bilirubin (in µmol/L)/17. Exclusion criteria were gastrointestinal bleeding or bacterial infection that could not be effectively treated within 48 hours; gastric or duodenal ulcer or ulcerated esophagitis; neoplastic disease; presence of hepatitis B antigen; presence of HIV antibodies; or anticoagulation therapy. 4 patients subsequently found to be ineligible were excluded from the analyses. 1 patient was lost to follow-up.


Prednisolone, 40 mg (n = 32), or identical placebo (n = 29) was given in a single dose of 2 pills each morning for 28 days. Patients unable to take oral medication received intravenous infusions of prednisolone or placebo. All patients were provided with a 3000-kcal diet containing 1 gram of protein per kilogram of body weight.

Main outcome measure

2-month survival.

Main results

Intention-to-treat analysis was used. Within 66 days of randomization, 16 of the 29 placebo recipients died compared with 4 of the 32 prednisolone recipients (55% vs 13%, P = 0.001). {This absolute risk reduction of 42% means that 3 patients would need to be treated with prednisolone (rather than placebo) to prevent 1 additional death, 95% CI 2 to 5; the relative risk reduction was 77%, CI 45% to 91%.}* The mean cumulative rates of survival at 1 and 2 months were 88% {CI 78% to 98%}* for both periods in the prednisolone group and 62% {CI 44% to 80%}* and 45% {CI 29% to 61%}*, respectively, in the placebo group. The survival advantage of prednisolone persisted after stratification according to hospital (P = 0.001) and the presence or absence of encephalopathy (P = 0.002), and after adjustment for prognostic factors in a Cox proportional-hazards model. Drug therapy was discontinued in 4 patients, 3 in the prednisolone group (2 because of drug toxicity) and 1 in the placebo group.


Prednisolone improved short-term survival in patients with biopsy-proven severe alcoholic hepatitis.

Source of funding: Not stated.

Address for article reprint: Dr. B. Rueff, Service de Traitement Ambulatoire des Maladies Alcooliques, Hôpital Beaujon, 92118 Clichy, France.

*Numbers calculated from data in article.


In a carefully conducted, randomized controlled trial, Ramond and colleagues report that treatment with prednisolone improves the short-term survival of patients with severe biopsy-proven alcoholic hepatitis. The treatment benefit was clinically impressive and persisted after adjustment for differences in baseline prognostic factors. The study protocol was generally well followed, although the analysis included 3 deaths in the placebo group that occurred shortly after the 2-month follow-up time, which was defined as the primary end point.

The finding that corticosteroids reduce early mortality in patients with acute alcoholic hepatitis is consistent with the conclusions from a previous meta-analysis (1) in which a protective effect was found in studies that excluded patients with acute gastrointestinal bleeding (as was done in the study by Ramond and colleagues). Unlike the meta-analysis, which concluded that steroids were of benefit for patients with hepatic encephalopathy only, patients with and without encephalopathy were helped by steroids in this trial. A possible reason for this discrepancy is that patients without encephalopathy had severe alcoholic hepatitis as measured by a discriminant-function score that included the prothrombin time and serum bilirubin.

Clinicians should be cautious about accepting the authors' recommendation to initiate corticosteroid therapy in severely ill patients before the results of the liver biopsy are available and to discontinue treatment if the diagnosis is not histologically confirmed. If this recommendation is adopted, patients in clinical practice would be treated differently than were the patients in the trial. Specifically, patients in practice would be treated earlier than trial patients (on average, 15 days after hospital admission), and many without biopsy evidence of alcoholic hepatitis would be exposed to the risks of treatment without expectation of benefit. Physicians in practice need to decide whether the authors' recommendations represent prudent clinical strategy.

Ralph I. Horwitz, MD
Yale University School of MedicineNew Haven, Connecticut, USA


1. Imperiale TF, McCollough AJ. Do corticosteroids reduce mortality from alcoholic hepatitis? Ann Intern Med. 1990;113:299-307.