Development and validation of a new prognostic score for hepatitis B virus-related acute-on-chronic liver failure

Jiaqi Li,Xi Liang,Shaoli You,Tingting Feng,Xin Zhou,Bing Zhu,Jinjin Luo,Jiaojiao Xin,Jing Jiang,Dongyan Shi,Yingyan Lu,Keke Ren,Tianzhou Wu,Lingling Yang,Jiang Li,Tan Li,Qun Cai,Suwan Sun,Beibei Guo,Xingping Zhou,Jiaxian Chen,Lulu He,Peng Li,Hui Yang,Wen Hu,Zhanglu An,Xiaojun Jin,Jin Tian,Baoju Wang,Xin Chen,Shaojie Xin,Jun Li,Chinese Group on the Study of Severe Hepatitis B
DOI: https://doi.org/10.1016/j.jhep.2021.05.026
IF: 25.7
2021-11-01
Journal of Hepatology
Abstract:BACKGROUND & AIMS: <AbstractText Label="BACKGROUND & AIMS">Early determination of the prognosis of patients with hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) is important to guide clinical management and decrease mortality. The aim of this study was to develop a new simplified prognostic score to accurately predict outcomes in patients with HBV-ACLF.</AbstractText>METHODS: <AbstractText Label="METHODS">Prospective clinical data from 2,409 hospitalized patients with acute deterioration of HBV-related chronic liver disease were used to develop a new prognostic score that was validated in an external group.</AbstractText>RESULTS: <AbstractText Label="RESULTS">A total of 954 enrolled patients with HBV-ACLF were diagnosed based on the Chinese Group on the Study of Severe Hepatitis B-ACLF (COSSH-ACLF) criteria. Six predictive factors were significantly related to 28-day mortality and constituted a new prognostic score (=1.649×ln(international normalized ratio)+0.457×hepatic encephalopathy score+0.425×ln(neutrophil)+0.396×ln(total bilirubin)+0.576×ln(serum urea)+0.033×age). The C-indices of the new score for 28-/90-day mortality (0.826/0.809) were significantly higher than those of 4 other scores (COSSH-ACLF, 0.793/0.784; CLIF-C ACLF, 0.792/0.770; MELD, 0.731/0.727; MELD-Na, 0.730/0.726; all p <0.05). The prediction error rates of the new score for 28-day mortality were significantly lower than those of the 4 other scores: COSSH-ACLF (15.9%), CLIF-C ACLF (16.3%), MELD (35.3%) and MELD-Na (35.6%). The probability density function evaluation and risk stratification of the new score also showed the highest predictive values for mortality. These results were then validated in an external cohort.</AbstractText>CONCLUSION: <AbstractText Label="CONCLUSION">A new prognostic score based on 6 predictors, without an assessment of organ failure, can accurately predict short-term mortality in patients with HBV-ACLF and might be used to guide clinical management.</AbstractText>LAY SUMMARY: <AbstractText Label="LAY SUMMARY">Hepatitis B virus-related acute-on-chronic liver failure (HBV-ACLF) is a complex syndrome that is associated with a high short-term mortality rate. We developed a simplified prognostic score for patients suffering from this condition based on a prospective multicentre cohort. This new score had better predictive ability than 4 other commonly used scores.</AbstractText>
gastroenterology & hepatology
What problem does this paper attempt to address?
This paper aims to solve the early prognosis problem of patients with hepatitis B virus - related acute - on - chronic liver failure (HBV - ACLF). Specifically, the research objective is to develop a new simplified prognostic scoring system to accurately predict the short - term mortality of these patients and perform risk stratification on them. Through this new prognostic scoring system, high - risk patients can be identified earlier, so that more aggressive treatment measures can be taken to reduce the high mortality rate. The paper mentions that although some existing prognostic scoring systems (such as COSSH - ACLFs, CLIF - C ACLFs, MELDs and MELD - Nas) can predict the prognosis of patients to a certain extent, their evaluation processes are relatively complex and their accuracy needs to be improved. Therefore, based on the standards of the Chinese Severe Hepatitis B Research Group (COSSH), this study screened out six independent risk factors significantly related to 28 - day mortality from multiple clinical factors and constructed a new prognostic scoring system (COSSH - ACLF IIs). These six factors are international normalized ratio (INR), hepatic encephalopathy score (HE score), neutrophil count, total bilirubin (TB), serum urea nitrogen level and age respectively. The formula of the new scoring system is as follows: \[ \text{COSSH - ACLF IIs} = 1.649\times\ln(\text{INR})+ 0.457\times\text{HE score}+0.425\times\ln(\text{neutrophil})+0.396\times\ln(\text{TB})+0.576\times\ln(\text{serum urea})+0.033\times\text{age} \] The research results show that the new prognostic scoring system performs excellently in distinguishing between surviving and non - surviving patients. Its C - indices are 0.826 (28 - day mortality) and 0.809 (90 - day mortality) respectively, which are significantly higher than those of the other four scoring systems. In addition, the new scoring system also performs well in terms of prediction error rate, probability density function analysis and calibration performance, further verifying its advantages in the prognosis evaluation of HBV - ACLF patients.