Elsevier

Journal of Hepatology

Volume 36, Issue 4, April 2002, Pages 480-487
Journal of Hepatology

Corticosteroids improve short-term survival in patients with severe alcoholic hepatitis (AH): individual data analysis of the last three randomized placebo controlled double blind trials of corticosteroids in severe AH

https://doi.org/10.1016/S0168-8278(01)00289-6Get rights and content

Abstract

Background/Aims: Controversy surrounding the efficacy of corticosteroids in severe alcoholic hepatitis (AH) persists. The aims of our study were: (a) to analyze individual data of patients with severe AH discriminant function (DF) ≥32 from the last three randomized controlled trials; and (b) to identify the independent prognostic factors associated with short-term survival.

Methods: Individual data were collected from the three principal investigators. Survival analysis was performed at 28 days using the Kaplan–Meier method and log-rank test. The independent prognostic values were assessed by the proportional hazards regression model.

Results: About 102 placebo and 113 corticosteroid patients with DF≥32 were analyzed. At 28 days, corticosteroid patients had significantly higher survival: 84.6±3.4% vs. 65.1±4.8%, P=0.001. In univariate analysis, corticosteroid treatment, age, DF, albumin, creatinine and encephalopathy were prognostic factors. In multivariate analysis, age (P=0.0001), serum creatinine (P<0.002) and corticosteroid treatment (P=0.002) were independent prognostic variables. A more dramatic decrease of median serum bilirubin values (μmol/l) was observed at 7 and 14 days in corticosteroid patients (P<0.05) : −76.5 vs. −35 and −105 vs. −45.

Conclusions: Corticosteroids improved short-term survival of patients with severe AH. Age and serum creatinine are independent prognostic factors. Corticosteroids are recommended for patients with severe AH.

Introduction

Thirteen randomized controlled trials that evaluated corticosteroids in patients with alcoholic hepatitis (AH) yielded inconsistent results [1], [2], [3], [4], [5], [6], [7], [8], [9], [10], [11], [12], [13].

Three meta-analyses concluded that corticosteroids improved short-term survival of patients with a severe form of AH [14], [15], [16]. However, the latest meta-analysis which attempted to adjust for prognostic factors, questioned the efficacy of corticosteroids in AH [17]. This meta-analysis used a different weighting method than previous meta-analyses that may overweight Mendenhall's RCT.

No accurate criteria for assessing disease severity in AH were available until Maddrey et al. described a discriminant function (DF) [6] and later modified it [12] to its more widely used form. This modified DF is calculated as follows: 4.6 (prothrombin time-control time [in seconds])+serum bilirubin (in micromoles per liter)/17.

Recently, representatives of the American College of Gastroenterology recommended glucocorticosteroids for patients with severe AH as defined by the Maddrey DF [18]. However, Christensen et al. suggested that the proposed guidelines were not sufficiently evidence-based [19]. They encouraged the authors of the largest trial (Mendenhall et al.) to make their data available for more elaborate analysis as this study preceded the Maddrey DF for assessment of disease severity and based severity only on the serum bilirubin level (i.e.=5 mg/dl) [10].

To determine whether corticosteroids are effective in severe AH, the most accurate approach would be to combine the individual data of patients with DF≥32 from previous randomized placebo controlled trials. In the present study, individual data from the three most recent randomized controlled trials were combined [10], [12], [13]. Thus, using the individual data on patients with DF≥32, we aimed to assess the efficacy of corticosteroid treatment and to identify the independent prognostic factors associated with short-term survival.

Section snippets

Patients

The three studies were conducted in a randomized, double-blind fashion. Randomized treatments consisted of prednisolone or placebo.

Inclusion and exclusion criteria in each study

Carithers' and Ramond studies included only patients with severe AH defined by DF≥32 or spontaneous hepatic encephalopathy. The three studies excluded patients with active peptic ulcer, neoplasms, non-alcoholic liver disease, active infection or positive test for hepatitis B surface antigen. Gastrointestinal bleeding was an exclusion criterion in Carithers' and

Survival data of each trial regardless of DF function

In Mendenhall's study the 28-day-survival of placebo (n=88) and corticosteroids (n=91) were not significantly different: 80.5±4% vs. 83±4%. In Carithers' and Ramond' studies that included only patients with severe form of AH (DF≥32 or presence of encephalopathy), corticosteroids patients had higher survival than placebo patients: 94±4% vs. 63±8.9% (P<0.01) and 87.5±5.8% vs. 62±9% (P=0.03).

Identification of patients with a DF≥32 in the three trials

We retrospectively identified the randomized patients with a DF≥32 at admission or at the first day of

Discussion

Meta-analyses evaluating corticosteroids in AH yielded controversial results and included trials too heterogeneous [14], [15], [16], [17]. The authors of the last meta-analysis used multivariate statistics to adjust for confounding variables between corticosteroid and control groups [17]. In a recent correspondence, Imperiale et al. found this method to be problematic [20]. They pointed out that the ability of this method to detect a treatment effect of corticosteroids is low because of the

Acknowledgments

The authors thank Sylvain Lavergne for his contribution in the data capture for the statistical analysis. This work is supported by a grant from the Association pour la Recherche contre le Cancer, France.

References (26)

  • B.L Blitzer et al.

    Adrenocorticosteroid therapy in alcoholic hepatitis: a prospective, double-blind randomized study

    Am J Dig Dis

    (1977)
  • J.B Shumaker et al.

    A controlled trial of 6-methylprednisolone in acute alcoholic hepatitis

    Am J Gastroenterol

    (1978)
  • A Theodossi et al.

    Controlled trial of methylprednisolone therapy in severe acute alcoholic hepatitis

    Gut

    (1982)
  • Cited by (303)

    • Use of immunosuppression in non-transplant hepatology

      2021, Best Practice and Research: Clinical Gastroenterology
      Citation Excerpt :

      The study failed in reaching its primary endpoint, a six-month survival improvement, and it showed just a trend in reducing the 28-days mortality (13.8% vs 18%, p = 0.06) [61]. Rambaldi et al.([63]), and Mathurin et al.([64]), published two meta-analyses in 2008, reporting more encouraging results. Both studies showed a positive effect for the 1 month-mortality in patients with severe acute hepatitis.

    • Long-term outcomes in patients with decompensated alcohol-related liver disease, steatohepatitis and Maddrey's discriminant function &lt;32

      2020, Journal of Hepatology
      Citation Excerpt :

      Independent predictive factors of 5-year mortality in the control group were age (HR 1.096; 95% CI 1.021–1.176; p = 0.011) and presence of encephalopathy (HR 3.463; 1.092–10.983; p = 0.035). While the short-term prognosis for patients with AH and an mDF <32 is favorable,3 data regarding the long-term outcomes for these patients are scarce, emphasizing the need for this study. The present study confirms that the risk of death at 1 and 3 months is about 5%, which is comparable to the limited data available in the literature, including a recent meta-analysis demonstrating the paucity of high-quality studies in patients with non-severe AH.10–13

    View all citing articles on Scopus
    View full text