RESUMO
BACKGROUND: Treating severe alcoholic hepatitis involves the exposure of patients to corticosteroids for 7 days to assess "response". AIM: To assess the prognostic and therapeutic implications of baseline neutrophil-to-lymphocyte ratio (NLR) in patients with severe alcoholic hepatitis. METHODS: Patients recruited to the STOPAH trial and an independent validation group were analysed retrospectively. Area under the receiver operating curve (AUC) analysis was performed. Kaplan-Meier analysis was used to assess survival. Log-rank test and odds ratio (OR) were used for comparative analysis. RESULTS: Baseline NLR was available for 789 STOPAH patients. The AUC for NLR was modest for 90-day outcome (0.660), but was associated with infection, acute kidney injury (AKI) and severity of alcoholic hepatitis. Ninety-day survival was not affected by prednisolone treatment if NLR < 5 or > 8 but mortality was reduced with prednisolone treatment when the NLR was 5-8 (21.0% cf. 34.5%; P = 0.012). Prednisolone treatment increased the chance of Lille response if the NLR was ≥ 5 (56.5% cf. 41.1%: P = 0.01; OR 1.86) but increased the risk of day 7 infection (17.3% cf. 7.4%: P = 0.006; OR 2.60) and AKI (20.8% cf. 7.0%: P = 0.008; OR 3.46) if the NLR was > 8. Incorporation of NLR into a modified Glasgow alcoholic hepatitis score (mGAHS) improved the AUC to 0.783 and 0.739 for 28-day and 90-day outcome, respectively. CONCLUSION: The NLR is associated with AKI and infection in severe alcoholic hepatitis. The NLR identifies those most likely to benefit from corticosteroids at baseline (NLR 5-8). The mGAHS has a good predictive value for 28- and 90-day outcomes.