The use of corticosteroids in severe alcohol hepatitis: We need to look beyond this controversy
Article Outline
The efficacy of corticosteroids in alcoholic hepatitis remains controversial for some authors. Three meta-analyses of the literature, from three different groups, concluded that the survival effect of corticosteroids was restricted to patients with severe disease. Conversely, two other meta-analyses of the literature from the same group questioned the efficacy of corticosteroids in alcoholic hepatitis, regardless of disease severity, although their latest meta-analysis observed that corticosteroids reduced mortality significantly in the subgroup of trials in patients with Maddrey discriminant function ⩾32 or hepatic encephalopathy [1]. Unexpectedly, their latest meta-analysis included some inaccuracies: (i) the authors considered Richardet’s study to be a trial comparing corticosteroids with no treatment, whereas both arms were treated with corticosteroids (information provided by Dr. Richardet); (ii) mortality data in the short-term were mixed with long-term mortality data; and (iii) the authors included a study (Galambos) quoted by another author (Conn) that has never been published even in abstract form.
In fact, the main limit of meta-analysis of the literature in alcoholic hepatitis is that evaluation of any treatment effect on short-term survival is achievable only in the subgroup of patients with a significant risk of death in the short-term. For that purpose, the use of reproducible tools is mandatory but this requirement was not met in most studies published before the Maddrey discriminant function (DF) became available. In numerous studies from different groups aiming to predict short-term mortality, the AUROC curve of DF is around 0.8 in patients with severe and non-severe forms [2], [3]. The prognostic values of DF was confirmed by several independent groups showing that spontaneous short-term survival of patients with a DF ⩾32 fluctuated between 50% and 65% [2], [3], [4], [5] whereas 28-day survival of patients with a DF <32 is close to 90% [6].
It is not a reasonable approach to evaluate treatment impact in patients with a negligible risk of death during exposure to the drug. However, meta-analysis of the literature cannot pool the results restricted to patients with DF ⩾32 as most of the previous RCTs did not supply the specific survival data of this subgroup. Christensen and his colleagues stated that the variability of severity is not an issue as, in their meta-analysis, most of the trials had similar or higher group mortality [1]. In fact, this inaccurate statement is related to the fact that they combined long-term survival data from some studies with short-term survival in others. How can clinicians consider long-term survival data as a relevant endpoint to analyze the impact of just 28
days of exposure to any compound in an acute process? How can a comparison be performed using different mortality endpoints? As an example, the exact data on short-term mortality for the corticosteroid group is 15/90 instead of 55/90 in the Mendenhall study [6] and 1/24 instead of 5/24 in Bories’ study [7]. The prognostic value of DF is confirmed by the fact that, in Mendenhall’s study, approximately 85% of deaths were observed in the subgroup of patients with DF ⩾32. Thus, therapeutic intervention for short-term outcome must focus only on patients with DF ⩾32.
The use of meta-analyses combining individual data is warranted in order to overcome the limitations associated with using the literature for comparative purposes. The need for such an approach is confirmed by the fact that, 20-year truth survival of conclusions derived from meta-analysis from the literature was lower (57%) than that from non-randomized studies (87%) (p<0.001) or randomized trials (85%) [8]. Whenever possible, a meta-analysis of individual patient data should be done because this provides the least biased and most reliable means of addressing questions that have not been resolved satisfactorily by individual clinical trials. In the specific setting of the evaluation of corticosteroids in patients with severe alcoholic hepatitis only a meta-analysis combining individual data is able to obtain detailed information from each individual with DF ⩾32 or hepatic encephalopathy.
Therapeutic intervention which improves liver function is liable to improve short-term survival and studies focusing on a strategy integrating the early impact of treatment on liver function are warranted [9]. Early improvement in liver function is very useful clinically in predicting short-term survival of patients with severe alcoholic hepatitis whether or not they are treated with corticosteroids [3], [10]. A recent international study combined individual data of patients with DF ⩾32 or encephalopathy from the five most recent randomized controlled trials comparing corticosteroids to placebo or any inefficient treatment [11]. Approximately 400 patients were included. The 28-day survival was significantly higher in corticosteroid treated than in non-corticosteroid treated patients (79% vs 64%, p=0.0005). Corticosteroid treated patients had an early and greater improvement of liver function and a better response to the assigned therapy assessed by the Lille model, a score integrating early improvement in its formula. In a multivariate analysis, DF, and the Lille model, encephalopathy and corticosteroid treatment were associated independently with short-term survival [11]. This study should end the controversy surrounding the short-term efficacy of corticosteroids in severe alcoholic hepatitis.
However, the current therapy for alcoholic hepatitis (i.e. corticosteroids and/or pentoxifylline) is not effective in all patients. Early identification of responders with a substantial improvement in hepatic function following treatment with corticosteroids constitutes an advance in the management of severe alcoholic hepatitis [10]. Conversely, in non-responders without an early improvement in hepatic function, novel therapies are needed urgently.
For all of these reasons, the recent American practice guidelines recommend the use of corticosteroids or pentoxifylline for patients with severe alcoholic hepatitis as defined by the Maddrey DF [12].
Conflicts of interest
The author declared that he does not have anything to disclose regarding funding or conflict of interest with respect to this manuscript.
References
- . Systematic review: glucocorticosteroids for alcoholic hepatitis – a Cochrane Hepato-Biliary Group systematic review with meta-analyses and trial sequential analyses of randomized clinical trials. Aliment Pharmacol Ther. 2008;27:1167–1178
- MELD accurately predicts mortality in patients with alcoholic hepatitis. Hepatology. 2005;41:353–358
- . Meld is a better prognostic model than Child–Turcotte–Pugh score or discriminant function score in patients with alcoholic hepatitis. J Hepatol. 2005;42:700–706
- Methylprednisolone therapy in patients with severe alcoholic hepatitis: a randomized multicenter trial. Ann Intern Med. 1989;110:685–690
- . Antioxidants versus corticosteroids in the treatment of severe alcoholic hepatitis – a randomised clinical trial. J Hepatol. 2006;44:784–790
- 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. J Hepatol. 2002;36:480–487
- . Traitement de l’hépatite alcoolique aiguë par la prednisolone. Presse Med. 1987;16:769–772
- Truth survival in clinical research: an evidence based requiem?. Ann Intern Med. 2002;136:888–895
- . Alcoholic hepatitis. N Engl J Med. 2009;360:2758–2769
- The Lille model: a new tool for therapeutic strategy in patients with severe alcoholic hepatitis treated with steroids. Hepatology. 2007;45:1348–1354
- Corticosteroids improve 28-day survival in patients with severe alcoholic hepatitis: individual data analysis of the last 5 randomized controlled trials. Hepatology. 2008;48:170A
- . Alcoholic liver disease. Hepatology. 2010;51:307–328
PII: S0168-8278(10)00385-5
doi:10.1016/j.jhep.2010.03.014
© 2010 European Association for the Study of the Liver. Published by Elsevier Inc. All rights reserved.
