Alcoholic Hepatitis Severity Scores; Which Score Is Better? Evidence From Meta-analysis of 5 Different Scoring Systems

Abstract

Introduction: We performed a systematic review and a meta-analysis of studies comparing various scoring systems to predict severity of and short-term mortality from alcoholic hepatitis. Methods: To be included in the current meta-analysis articles had to (1) report mortality in AH patients, (2) report AH severity by 2 or more scoring models, and (3) report area under curve (AUC) of each model. AUC from different studies were pooled using a MedCalc meta-analysis program for the following severity scoring systems (1) Model for End Stage Liver Disease (MELD), (2) Maddrey discriminant function (mDF), (3) Glasgow Alcoholic Hepatitis Score (GAHS), (4) Child-Turcotte-Pugh (CTP), and (5) Age, Bilirubin, International Normalization Ratio and Creatinine (ABIC). The analysis was stratified according to 28 and 90 day (D) mortality. Results: On 28 d mortality, AUC comparing MELD and DF from 9 studies did not show difference, 0.837 (0.779-0.895) vs. 0.803 (0.747-0.860), P=0.41. Comparisons were also non-significant on 28d mortality for DF vs. CTP from 3 studies: 0.774 (0.697-0.850) vs. 0.711 (0.559-0.864), P= 0.47), for ABIC vs. GAHS from 5 studies: 0.768 (0.714-0.823) vs. 0.772 (0.721-0.824), P= 0.92, and for MELD vs. GAHS from 4 studies: 0.768 ( 0.714-0.823) vs. 0.780 (0.713-0.0.846), P=0.78. Analysis on 90 d mortality showed similar results comparing AUC using MELD and DF from 7 studies: 0.802 (0.758-0.846) vs. 0.790 (0.782-0.797), P=0.6, comparing ABIC vs. GAHS from 7 studies: 0.765 (0.716-0.815) vs. 0.758 (0.711-0.805), P=0.84, and comparing MELD vs. GAHS from 6 studies:0.779 (0.725-0.832) vs. 0.746 (0.695-0.797), P= 0.38. Conclusion: Existing severity scores used in clinical practice are comparable in predicting severity and prognosis of alcoholic hepatitis. However, none of the scoring systems is ideal. Prospective studies are needed as basis for deriving scoring systems and/or biomarkers for better stratification of severity and prognosis of patients with alcoholic hepatiits.