Artificial Liver Support System Improves Short-Term Outcomes of Patients with HBV-Associated Acute-on-Chronic Liver Failure: A Propensity Score Analysis

Background Hepatitis B virus-associated acute-on-chronic liver failure (HBV-ALCF) is a complicated syndrome with extremely high short-term mortality. The artificial liver support system (ALSS) may improve the liver function for patients with HBV-ACLF, but the data on its short-term outcomes are insufficient in China. Methods We recruited HBV-ACLF patients in this nationwide, multicenter, retrospective study. Patients with HBV-ACLF were diagnosed by the COSSH-ACLF criteria. Propensity score matching (PSM) analysis was used to generate compared pairs. The short-term (28/90 days) survival rates between the standard medical therapy (SMT) group and ALSS group were calculated using a Kaplan–Meier graph. Result In total, 790 patients with HBV-ACLF were included in this retrospective study; 412 patients received SMT only (SMT group), and 378 patients received SMT and ALSS treatment (ALSS group). PSM generated 310 pairs and eliminated the baseline differences between the two groups (p > 0.05 for all baseline variables). The probabilities of survival on day 28 were 65.2% (205/310) in the ALSS group and 59.0% (185/310) in the SMT group; on day 90, they were 51.0% (163/310) and 42.3% (136/310). The short-term (28/90 days) survival rates of the ALSS group were significantly higher than those of the SMT group (p=0.0452 and p=0.0187, respectively). Compared to receiving SMT alone, treatment with ALSS was associated with a significant reduction in serum bilirubin levels and the model for end-stage liver disease (MELD) scores at day 7 and day 28. Multivariate logistic regression analysis revealed that older age, high total bilirubin (T-Bil), low albumin, high ALT, high MELD scores, and high COSSH-ACLF grade were independent baseline factors associated with poor prognosis. Conclusions This retrospective study found that compared to SMT, the ALSS improved the short-term (28/90 days) survival rates and laboratory parameters in HBV-ACLF patients. The ALSS had a better therapeutic effect than SMT for patients with HBV-ACLF in China.


Introduction
Chronic hepatitis B (CHB) is a major public health challenge in China, with an estimated 78 million chronic carriers and 28 million patients with active hepatitis [1]. CHB is a significant risk factor that accounts for nearly 45% of cases of hepatocellular carcinoma (HCC) and 30% of cases of cirrhosis, causing nearly 1 million deaths each year worldwide [2,3]. e high prevalence of CHB causes that hepatitis B virus (HBV) infection absolutely predominated in the etiologies of ACLF, accounting for 96.5% cases in China, while alcoholism is the most common etiologies of ACLF in western developed countries [4,5].
In past decades, a series of artificial liver support systems (ALSSs) have been applied in liver failure which aim to detoxify blood and compensate liver function for patients with liver failure. Molecular adsorbent recirculating system (MARS) [8,9] and Prometheus [10,11] are the most widely used ALSS. Li's artificial liver system (Li-ALS, the low-volume plasma exchange-centered ALSS) was designed by Professor Li's team since 1986 [12], which is mainly used in China. Some studies demonstrated that ALSSs could detoxify and ameliorate hepatic encephalopathy during acute liver failure (ALF) [11,13,14]. However, several large randomized trials noted that patients with ALF [15] or ACLF [16,17], supported with ALSSs, did not show an increased short-term survival. A single-center study conducted in China reported that ALSS improved 90 days and 5 years outcomes of patients with HBV-ACLF [18].
In this study, we conducted a nationwide, multicenter, retrospective study to test whether ALSSs could improve the short-term (28/90 days) outcomes in patients with HBV-ALCF in China and identify predictive factors for the prognosis of these patients.

Patients.
is is a retrospective cohort study where all data were fully anonymized before access. We recruited hospitalized patients with HBV-ACLF from 11 liver centers of Chinese University hospitals between January 2014 and May 2017. e following clinical data were collected: demographic data (age, sex, and body mass index), cirrhosis, laboratory measurements (HBV-DNA level, ALT, total bilirubin (T-Bil), INR, and creatinine), hepatic encephalopathy (HE), nucleos(t)ide analogs (NA), and survival time.
e liver disease severity was assessed using the MELD scores and COSSH-ACLF grade (Chinese Group on the Study of Severe Hepatitis B-ACLF). e study was approved by the Ethics Committee of the First Affiliated Hospital, Zhejiang University School of Medicine.

Inclusion and Exclusion Criteria.
e inclusion criteria were as follows: Patients diagnosed with CHB [19] (HBV surface antigenpositive ≥6 months; serum HBV-DNA ≥20 000 IU/mL; a liver biopsy showing chronic hepatitis) and reached at least ACLF grade 1 diagnosed by the COSSH-ACLF criteria [6]: (1) ACLF grade 1: patients with kidney failure alone; patients with single liver failure (total bilirubin ≥12 mg/dL) with an international normalized ratio (INR) ≥1.5 and/or kidney dysfunction and/or HE grade I or II; patients with single type of organ failure of the coagulation, circulatory, or respiratory systems and/or kidney dysfunction and/or HE grade I or II; and patients with cerebral failure alone plus kidney dysfunction (2) ACLF grade 2: patients with failure of two organ systems (3) ACLF grade 3: patients with failure of 3 or more  organ systems   e exclusion criteria were as follows: (  [20], the treatment for severe liver disease mainly consists of restoring and preserving vital organ function and slowing down the progression of multiple organ failure. e standard medical therapy (SMT) included a high-calorie diet; enteral nutrition is recommended; correction hypoproteinemia; correction water-electrolyte and acid-base balance; nucleoside analogs for HBV-DNA-positive patients; anti-infective therapy for infection; restricted protein diet; lactulose, ammonia drugs, and L-ornithine aspartate for HE; diuretics and tolvaptan for ascites; tubular active drugs; maintenance of arterial blood pressure and water restriction for hepatorenal syndrome; and oxygen therapy for hepatopulmonary syndrome.
According to the Guideline (2012), patients with earlyor middle-stage liver failure were advised to receive ALSS treatment. For patients with early liver failure, PE was applied; for patients with metaphase hepatic failure, continuous blood purification (CBP) was applied; for patients with brain edema or renal failure or imbalance of water and electrolytes, CBP or plasma diafiltration (PDF) was applied; for patients with hyperbilirubinemia, plasma bilirubin absorption (PBA) was applied.
e ALSS sessions were scheduled as follows: the ALSS was usually performed in 48 hours after diagnosis. ALSS treatment was performed daily on the first 2 or 3 days; future treatments were offered according to the patients' condition. Overall, 841 ALSS treatment sessions were applied in 310 patients (average of 2.7 sessions per patient, ranging from 1 to 8 sessions). 9 (2.9%) patients received PE (2-3 h/per session), and 301 (97.1%) patients received continuous renal replacement therapy (CRRT, 8 h/per session).

Statistical Analysis.
Propensity score matching analysis was used to eliminate bias between the two groups. Propensity scores were computed using the following variables: age; sex; serum levels of HBV-DNA, ALT, T-Bil, and albumin; platelet count; white cell count; creatinine; INR; serum sodium; cirrhosis; HE; MELD score; and COSSH-ACLF grade. For propensity score matching, a nearestneighbor 1 : 1 matching scheme was used. Categorical data were compared using the χ 2 test or Fisher's exact test. Continuous variables were compared using the Mann-Whitney U test. e survival rate at 28 days and 90 days were calculated using a Kaplan-Meier graph. e difference in the survival rate was compared using a log-rank test. e relationship between baseline parameters and 28-day survival was studied using a multivariate logistic regression model. All statistical analyses were performed using IBM SPSS v. 24.0 for Windows (IBM Corp., Armonk, NY, USA). p values are two-tailed, and values less than 0.05 were considered statistically significant.

Patient Characteristics.
A total of 790 patients with HBV-ACLF were included in the study; 412 patients received standard treatment (SMT group) and 378 patients received ALSS treatment (ALSS group, Figure 1). Among them, 598 (85.3%) patients were male and 369 (52.5%) patients had no cirrhosis. Before matching, the baseline characteristics between two groups that differed significantly, such as ALT, T-Bil, and platelet count. Propensity score matching analysis generated 173 pairs, and the baseline characteristics of the pairs were balanced, with p > 0.05 for all baseline variables. After matching, the mean (SD) age of the patients in the SMT and ALSS groups was 45.4 (11.1) and 46.8 (11.3) for male, and the mean (SD) MELD scores were 25.1 (5.2) and 24.9 (5.6). In the SMT group, 118 (38.1%) patients had cirrhosis, and in the ALSS group, 133 (42.9%) patients had cirrhosis. All patients received NAs after diagnosis. In the SMT group, 155 patients received 100 mg of lamivudine (LAM) daily and 155 patients received 0.5 mg of entecavir (ETV) daily. In the ALSS group, 131 patients received 100 mg of LAM daily; 157 patients received 0.5 mg of ETV daily; 14 patients received 300 mg of tenofovir daily; and 8 patients received 600 mg of telbivudine daily. e characteristics of all patients are shown in Table 1.

Changes in Laboratory Parameters at Day 7 and Day 28.
e effect of treatment on laboratory parameters at day 7 and day 28 is shown in Tables 4 and 5, respectively. Compared with the SMT treated alone group, treatment with ALSS was associated with a significant reduction in serum T-Bil levels and MELD scores at both day 7 and day 28; however, the serum creatinine only decreased at day 7. e remaining analyzed parameters showed no significant difference between the two groups.

Risk Factors on 28-Day Survival.
Using multivariate logistic regression analysis, the independent baseline risk factors for 28-day survival were identified as age, T-Bil, low albumin, ALT, MELD score, and COSSH-ACLF grade (Table 6). Cirrhosis, INR, serum creatinine, and platelet count were not independent predictors of 28-day survival.

Discussion
HBV-ACLF is observed in populations with HBV-related chronic liver disease. Liver transplantation is the most effective therapy for patients with liver failure; however, less than 30% of patients have access to transplantation because of donor shortages and the extremely poor prognosis of HBV-ACLF [21]. Although NA treatment could effectively decrease the 3-month mortality for patients with HBV-ACLF, NAs are only valid in patients with MELD scores less than 30 [22,23]. In addition, mutations resistant to NAs are frequent precipitating events of HBV-ACLF, which is related to high mortality [24]. e development of ACLF includes the accumulation of various metabolites and toxins that vary in size, distribution volume, lipophilicity, and proteinbinding abilities [25,26]. Most ALSSs are capable of correcting the hemato-microenvironment, such as detoxification, synthesis, immune regulation, and reducing mortality in patients with ACLF, when compared with SMT [27]. erefore, ALSS has been recommended as an important method to treat ACLF. HBV-ACLF is a special type of ACLF and is the most frequent ACLF in China. is large, multicenter, nationwide, historical retrospective study showed that treatment with ALSS in patients with HBV-ACLF remarkably improved their short-term (28/90 days) survival rates compared with those receiving SMT only. e results indicated that ALSS is effective at removing the toxic substances from plasma that accumulate in patients with HBV-ACLF   Values are expressed as number and percentage or mean ± SD unless otherwise specified. SMT, standard medical therapy; ALSS, artificial liver support system; MELD, model for end-stage liver disease; COSSH-ACLF, Chinese Group on the Study of Severe Hepatitis B-ACLF.
(membrane pore size � 0.03 μm) of that of a normal plasma separator for direct PE, which could remove toxic substances effectively for patients with liver failure, retain important plasma components, and reduce the plasma dosage [28]. Although high-volume plasma (HVP, approximately 8 L) exchange has a beneficial therapeutic effect on patients with  2.9% (9/310) 11.1% (1/9) 0 (0/9) SMT, standard medical therapy; ALSS, artificial liver support system. ALF [26,29], it is usually performed over more than two sessions for patients with ACLF, which results in a shortage of fresh plasma. Moreover, the HVP removes almost all elements of plasma; part of them are beneficial substances (such as hepatic growth factor) for liver regeneration [30]. In this study, the ALSS using lower-volume plasma (LVP) exchange in which the total volume of exchanged fresh plasma is approximately 1500 mL [12]. e LVP is usually the first step of the ALSS, and subsequent adsorption and hemofiltration circulation are conducted through autologous plasma derived from waste plasma. e waste plasma is purified, which avoids wasting massive plasma and reduces the loss of essential substances. However, whether lowvolume PE provides better results compared with highvolume PE remains unclear. Selection of suitable inclusion criteria is crucial to evaluate the efficacy of ALSS for patients with HBV-ACLF. HBV-ACLF is an extremely special type of ACLF with some distinctive characteristics. Patients with HBV-ACLF have poorer prognosis than patients with non-HBV-ACLF (the 28/90 days mortality rates were 60.2% vs 52.1% and 73.9% vs 69.7%) [6]. Patients with HBV-ACLF have a higher incidence of liver and coagulation failure [31,32], whereas kidney failure and cerebral failure are the most common types of organ failure in patients with non-HBV-ACLF. Reactivation of HBV, an acute hepatic insult, is the leading cause of HBV-ACLF in the Asian region [33][34][35][36]. e European Association for the Study of the Liver (EASL) criteria and the American Association for the Study of Liver Diseases (AASLD) criteria are the major criteria for ACLF in patients from Europe and North America, where alcoholic liver disease is the major etiology [7,37]. ese criteria may not apply to China, where hepatitis B virus infection is the major etiology in China. e COSSH criteria were established based on a large Chinese HBV-ACLF group which is in accordance with our study group. erefore, we chose the COSSH criteria as the inclusion criteria.
A previous study showed that the first leading cause of HBV-ACLF was spontaneous severe acute exacerbation of CHB (62.5%) and the second leading cause was alcohol   [5]. e oral NA therapy effectively suppresses viral DNA and prevents the progression of liver inflammation; therefore, rapid initiation of oral NA treatment in patients with HBV-ACLF is recommended widely [38,39]. LAM and ETV are both widely used in China, though LAM has less potency and has higher resistance (LAM: 70%; ETV: 1.2%) than ETV [39]. is study revealed that compared with LAM, ETV did not improve the short-term survival rates in HBV-ACLF, neither in the SMT group nor ALSS group. ese results were similar to a largest meta-analysis, which demonstrated a comparable short-term mortality (within 4 months) of LAM and ETV; however, ETV revealed a more favorable long-term (beyond 4 months) outcome than LAM in patients with HBV-ACLF [40]. e results of multivariate logistic regression showed that cirrhosis is not a risk-independent predictor of 28-day survival. We also found a similarly short-term survival rate in patients with both cirrhotic HBV-ACLF (61.8%/46.2%) and noncirrhotic HBV-ACLF (63.6%/49.65%). However, this result was not in line with that of Wu's study, which indicated that cirrhosis patients had superior 28-day survival rate than noncirrhosis patients (47.9%: 39.8%, p < 0.05) [6]. e contradictory outcomes might result from small population samples of this study. In previous studies, T-Bil and platelet levels were found as independent risk factors of mortality among HBV-ACLF patients; however, in our study, the significant difference was only found in single-factor logistic regression analysis.
In conclusion, among patients with HBV-ACLF in China, ALSS has better therapeutic effect than SMT. ough the treatment for HBV-ACLF haas improved over the past three decades the short-term mortality of ACLF remains high. erefore, more effective therapeutic methods should be investigated.

Data Availability
All relevant data are within the paper and its supporting information file.
Ethical Approval e study was reviewed and approved by the ethics committee of the First Affiliated Hospital, College of Medicine, Zhejiang University.

Conflicts of Interest
e authors have declared that there are no conflicts of interest.

Authors' Contributions
Li LJ and Xiao LL designed the research. Xiao LL, Xu XW, Wu XX, Huang KZ, and Zhang LJ performed the research. Xiao LL and Zhao YL analyzed the data. All authors wrote the paper, had access to the study data, and have reviewed and approved the final manuscript.