The choice for treating patients with established and diagnosed cases of severe AH is corticosteroids[1
]. There have been 12 randomized placebo controlled trials (RCT) to assess the benefit of corticosteroids in AH patients (Table ). Results from these RCTs, conducted during the last 40 years, have varied with the sample size, inclusion/exclusion criteria, disease severity, end-points, type of corticosteroid used and treatment duration. These studies have shown conflicting data on the benefit of steroids with only five studies showing a survival benefit (Table ).
Randomized studies to assess corticosteroids for treatment of acute alcoholic hepatitis
Meta-analyses of RCTs provide the best evidence for efficacy. To date, four meta-analyses have been published on the efficacy of steroids in AH[2
]. The latest Cochrane analysis concluded that there is no clear evidence that steroids are effective in the management of AH. The potential for bias is due to heterogeneous data[5
]. However, the same meta-analyses concluded that steroids do have survival benefit for patients with severe AH (discriminant function index, DFI ≥ 32)[5
One of the means to tackle the issue of heterogeneity is to perform meta-analysis on the individual patient data from each study[6
]. This had been performed earlier by Mathurin and colleagues from France where they analyzed individual patient data from 3 RCTs. The results of this meta-analysis showed that steroids have survival advantage for severe AH with 28 d survival of 85% among treated patients and 65% for patients receiving placebo (P
= 0.001). This was also associated with improvement of liver function starting within the first week of starting the steroids[4
Corticosteroids act by reducing inflammatory cytokines such as tumor necrosis factor-α (TNF-α), intercellular adhesion molecule 1, interlukin (IL)-6 and IL-8[7
]. Inflammation is a major component of AH pathogenesis. In fact, in one study, peripheral white blood cell count > 5500/cm and the amount of polymorphonuclear leucocytic infiltration on the liver biopsy specimen were independent predictors for response and survival on steroid treatment[9
Although, many agents have been used across different studies, prednisolone is preferred (but not demonstrated to be better) over prednisone as the latter requires conversion within the liver to its active form, prednisolone. The drug is given orally in a dose of 40-60 mg/d for a total duration of 4 wk. The treatment is then tapered over next 2-3 wk. If the patient is unable to take it orally due to nausea, vomiting or altered sensorium, an intravenous preparation such as methylprednisolone may be used until the patient is capable to take medication by mouth.
It is prudent to screen patients for any contraindication prior to starting steroids. One of the most important contraindications is the presence of infection which is fairly common among patients with severe AH. This used to be considered an absolute contraindication for steroids[10
]. However, the latest data from France have shown that if a patient is adequately treated for an established infection, steroids can be safely started and even improve the outcome in these patients. In this study, all 246 patients studied prospectively were treated with steroids. Patients with infection (25% of the group) were treated adequately with antibiotics prior to starting steroids. Survival with steroids at 2 mo was similar, irrespective of the presence of infection prior to starting steroids (71% vs
Other contra-indications are an active gastrointestinal bleeding, renal failure, acute pancreatitis, active tuberculosis, uncontrolled diabetes and psychosis. Patients should be assessed for response to steroids. It has been shown that a decrease in bilirubin at 1 wk (early change in bilirubin, ECBL) is a reliable and specific marker for response. Patients who achieved ECBL had a better survival at 6 mo compared to patients who did not achieve ECBL (98.3% vs
]. Based on ECBL and other variables, French workers have derived a score (Lille score) based on the patient’s age, serum albumin, ECBL, renal insufficiency and prothrombin time. Patients with a Lille score of ≥ 0.45 are defined as non-responders to steroids (NRS). This score, with a cut off at 0.45, has an accuracy of 75% in predicting death at 3-6 mo[13
]. Patients should also be screened for infective complications while on steroids. Occurrence of sepsis and infective complications while the patient is on steroids is a poor prognostic sign. A total of 57 patients developed infection after starting steroids which occurred more frequently among NRS than responders to steroids (42% vs
]. Lille score was an independent predictor for the occurrence of infection after starting steroids.