Liver failure is a severe complication in which the liver fails to perform basic functions due to severe or massive liver injury, which also disrupts the vital functions of other organs and systems. It typically manifests as jaundice, hepatic encephalopathy, and coagulopathy (1, 2). The mortality rate of liver failure may be as high as 70% (3). Hepatologists must be able to assess and predict the outcomes of liver failure accurately and early in patients, especially to determine if and when liver transplantation is required. Dynamic changes in the conditions of a patient should also be closely monitored and followed, as timely prevention and treatment of complications significantly impact the clinical outcomes. The models for end-stage liver disease (MELD), MELD-Na, and Child-Turcotte-Pugh (CTP) scores are three established systems used to numerally grade the severity of liver failure in patients, which can reveal the probability for favorable or non-favorable outcomes (4, 5). All three model systems were established by assessing patients with liver failure mainly caused by hepatitis C virus (HCV) infection and alcoholic liver disease. However, liver failure caused by hepatitis B virus (HBV) infection is characterized by acute onset, severe liver injury, and high mortality, with distinct pathogenesis and clinical characteristics from those caused by HCV and alcohol. The three model systems can be expanded or modified to better grade HBV liver failure. We previously established a new scoring system known as the Chinese scoring system and its performance and utility were evaluated for patients with hepatitis B liver failure in a retrospective study (6–9). To further investigate the utility of this new scoring system in grading the severity of hepatitis B liver failure and predicting outcomes, we conducted this prospective study at seven liver disease centers across China.

MELD, MELD-Na, and CTP scores have been established to address the need to accurately identify patients with liver failure who require liver transplants. All three systems are widely accepted in the clinical setting. However, these models were constructed using clinical data largely from patients with liver failure in Europe and the United States where the major etiologies for liver failure are alcohol, drugs, hepatitis C, and cholestasis. In contrast, hepatitis B represents the dominant etiology for liver failure in China and is responsible for 85% of liver failure cases in the country (11). Therefore, these three models (12, 13) may not completely fit Chinese patients with liver failure. As noted by Lin Xianfeng et al. (14), the MELD score failed to accurately predict the prognosis of 327 Chinese patients with hepatitis B liver failure. One factor that may have affected this result is that patients with HBV liver failure do not respond well to comprehensive treatment and often need an artificial liver, liver transplant, or other treatment urgently. The other factor is that chronic HBV infection may exhaust the immunity of patients, leading to severe complications. A new system that considers these factors is required (15).

Child-Turcotte-Pugh (CTP) score has been widely employed in managing patients with liver diseases because of its simplicity and ease of calculation. However, it is often hindered by a lack of consistency as CTP is mainly based on albumin level and prothrombin time, both of which can be masked if exogenous albumin and plasma products are infused. In addition, hepatic encephalopathy and ascites measurement can be subjective and variable. CTP only consists of three grades with a narrow score range of 5–15, which limits the ability to differentiate. Patients with the same score may differ greatly in the severity of the liver disease.

Although the MELD and MELD-Na scoring systems are recognized as easy to use, reproducible, and accurate in clinical applications, they bear an inherent limitation because they only include bilirubin, coagulation function, renal function, and etiology, while excluding the complications of liver failure (infection, hemorrhage, hepatic encephalopathy, brain edema, or ascites). The Na concentration in the MELD-Na scoring system can be easily affected by insufficient intake, diuretics, or sodium pump failure (12, 16–20). A study conducted in China (21) showed that the AUC of MELD and MELD-Na scores for the 12-week outcomes of patients with hepatitis B-related liver failure was only 0.731, 0.735, and 0.773, respectively, and the sensitivity was below 0.7.

This prospective study conducted at seven large liver disease centers across China showed that the Chinese scoring system better distinguished patients that survived from those who did not, in comparison to MELD (P < 0.001). The area under the ROC curve for the Chinese scoring system was 0.954 in the peak disease stage, suggesting a stronger ability to grade the severity of the liver failure. The AUC for the MELD scoring system was 0.896, which was relatively high, but significantly lower than that for the new scoring system (P < 0.001). We also found that the efficiency in predicting outcomes on the 30th, 90th, and 180th days after discharge with the Chinese scoring system was significantly higher than that with MELD, MELD-Na, and CTP (P < 0.001).

Massive or submassive liver injury following a flare-up of HBV replication or HBV reactivation compromises synthetic, metabolic, and detoxification functions in the liver, leading to reduced albumin levels and ascites, hyperbilirubinemia, decreased PTA, increased international standard ratio of prothrombin (INR), and hepatic encephalopathy. In addition, a severely injured liver may also hurt Kupffer cell function and reduce complement levels that may limit the anti-infective capacity. The accumulation of toxins and a decrease in renal blood flow may facilitate hyperbilirubinemia, renal function insufficiency, or even renal failure. In theory, the severity of liver failure among different individuals can be indicated by biomarkers that reflect critical alterations in the pathophysiology of liver failure. Thus, we formulated this new scoring system to improve severity grading and outcome prediction. A total of seven clinical indicators of hepatitis B liver failure are included in the Chinese scoring system, consisting of not only objective indicators such as Cr, total bilirubin, PTA, and liver size but also the complications of hepatic encephalopathy, ascites with pleural effusion, and pulmonary infection, which reflect the pathophysiological changes resulting from hepatitis B liver failure. The score boundary point was confirmed using the interactive chi-squared test, guided by the principle that emphasizes simplicity and clarity. Each index is graded between 0 and 4 points, allowing a high degree of differentiation that increases prediction accuracy. This new scoring system utilizes common clinical indicators, which can be routinely ordered to allow easy and wide applications.

This new scoring system demonstrated superior performance in grading the severity and predicting the outcomes of HBV liver failure in this cohort compared to the three established scoring systems, representing progress in improving the management of HBV liver failure. However, our findings will need further verification in larger cohorts that include not only patients with HBV liver failure but also those with liver failure with other etiologies.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Clinical Trials registry (registration number: NCT01961440, https://clinicaltrials.gov/). The patients/participants provided their written informed consent to participate in this study.

WZ-b performed the case summary and statistical analysis and wrote the manuscript. LB-l and PL were responsible for statistical analysis. CZ, ZX-x, TD-m, RW-h, WK, and YX-b were responsible for case enrollment and data collection in other subcenters. KW-m designed this scoring system. ZY-b was responsible for the whole quality of the study. GZ-l was responsible for research design and the whole quality of the study. All authors contributed to the article and approved the submitted version.

This work was supported by the National Science and Technology Major Project (2018ZX10302204-002) and the National Natural Science Foundation of China (81672701).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.