Model for End‐Stage Liver Disease With Additional Criteria to Predict Short‐Term Mortality in Severe Flares of Chronic Hepatitis B

• Published in: Hepatology (Baltimore, Md.) Vol. 72; no. 3; pp. 818 - 828
• Main Authors: Fung, James, Mak, Lung‐Yi, Chan, Albert Chi‐Yan, Chok, Kenneth Siu‐Ho, Wong, Tiffany Cho‐Lam, Cheung, Tan‐To, Dai, Wing‐Chiu, Sin, Sui‐Ling, She, Wong‐Hoi, Ma, Ka‐Wing, Seto, Wai‐Kay, Lai, Ching‐Lung, Lo, Chung‐Mau, Yuen, Man‐Fung
• Format: Journal Article
• Language: English
• Published: United States Wolters Kluwer Health, Inc 01.09.2020
• Subjects: Acute-On-Chronic Liver Failure - diagnosis; Acute-On-Chronic Liver Failure - mortality; Alanine; Alanine transaminase; Antiretroviral drugs; Antiviral Agents - therapeutic use; Bilirubin; End Stage Liver Disease - diagnosis; End Stage Liver Disease - etiology; End Stage Liver Disease - mortality; Female; Guanine - analogs & derivatives; Guanine - therapeutic use; Hepatitis B; Hepatitis B e antigen; Hepatitis B, Chronic - diagnosis; Hepatitis B, Chronic - drug therapy; Hepatitis B, Chronic - epidemiology; Hepatitis B, Chronic - physiopathology; Hong Kong - epidemiology; Humans; Interferon; Liver diseases; Liver Function Tests - methods; Liver transplantation; Male; Medical prognosis; Middle Aged; Mortality; Patient Selection; Predictive Value of Tests; Prognosis; Proportional Hazards Models; Severity of Illness Index; Tenofovir; Tenofovir - therapeutic use; Hong Kong
• ISSN: 0270-9139; 1527-3350; 1527-3350
• DOI: 10.1002/hep.31086

Background and Aims The prognosis in severe acute flares of chronic hepatitis B (AFOCHB) is often unclear. The current study aimed to establish the predictive value using the Model for End‐Stage Liver Disease (MELD) score for short‐term mortality for severe AFOCHB. Approach and Results Patients with severe AFOCHB with bilirubin > 50 µmol/L, alanine aminotransferase > 10× upper limit of normal, and international normalized ratio > 1.5 were included. All patients were commenced on entecavir and/or tenofovir. Laboratory results and MELD scores were pooled to calculate mortality at four time points (days 7, 14, 21, and 28). A total of 240 patients were included. Median hepatitis B virus DNA was 7.77 log IU/mL (range, 4.11‐10.06), and 49 (20.4%) were hepatitis B e antigen–positive. The 7, 14, 21, and 28‐day survival was 96.7%, 88.5%, 79.5%, and 72.8%, respectively. Using pooled results derived from 4,201 blood samples, the area under the receiver operating curve for the MELD score to predict day 7, 14, 21, and 28 mortality was 0.909, 0.892, 0.883, and 0.871, respectively. For MELD ≤ 28, mortality at day 28 was low (<25%) compared with > 50% mortality for MELD ≥ 32. For MELD = 28‐32, higher day‐28 mortality was observed for four criteria: age ≥52 years, alanine aminotransferase > 217 U/L, platelets < 127, and abnormal baseline imaging (all P < 0.001). In this MELD bracket, the 28‐day mortality was 0%, 12.1%, 23.8%, 59.4%, and 78.8% for the presence of zero, one, two, three, and four criteria, respectively. Conclusions MELD score at any time points can accurately predict the short‐term mortality. Patients with MELD ≥ 28 should be worked up for liver transplantation, and those with MELD = 28‐32 with three to four at‐risk criteria, or MELD ≥ 32 should be listed.