2.1 Study Design, Patient Recruitment and Clinical Evaluation

In this observational cohort study, we identified 68 HBV-ALF patients from the US-ALFSG prospective registry between 1998 and 2015 (> 3400 patients). All participating sites within the US-ALFSG were tertiary academic liver transplant referral centers. This study was approved by the institutional review boards or health research ethics boards of all US-ALFSG enrolling sites. All participants or next of kin provided informed consent following the principles of the Declaration of Helsinki. The NIH guidelines for the inclusion of women and minorities as participants in clinical research were also observed. All authors had access to the study data and reviewed and approved the final manuscript.

The inclusion criteria included age of > 18 years and diagnosis of ALF, defined as presence of 1) hepatic encephalopathy of any degree, 2) evidence of coagulopathy with INR ≥ 1.5, 3) acute illness onset of < 26 weeks and 4) no evidence of cirrhosis and hepatocellular carcinoma (HCC). Transplant hepatology expert clinical decision makers verified absence of cirrhosis in enrolled patients in this study. In total 68 individuals tested HBsAg positive and were diagnosed with HBV-ALF, and further sub-classified as de novo acute hepatitis B associated ALF (HBsAg and anti-HBc IgM positive, n = 48) vs. history of hepatitis B and reactivation (HBsAg positive, anti-HBc IgG positive n = 20). HBsAg positive HBV-ALF patients with acetaminophen overdose, alcohol misuse disorder, antibodies to human immunodeficiency virus (anti-HIV), hepatitis C virus (HCV) RNA, anti-HCV, antibodies to hepatitis D virus (anti-HDV) and anti-HEV positivity were excluded (n = 4) (Figure S1).

Demographical, clinical, virological and outcome data were collected for all enrolled patients from admission (baseline) up to day 21 follow-up, including age, sex, ethnicity (race), biochemistry, viral serology (HBV, HCV, HDV, HEV and HIV), hepatic encephalopathy grade (West Haven criteria, on admission), MELD score, KCC on admission, therapy including previous use of immunosuppressive treatments, day 21 transplant-free survival (TFS) and liver transplantation. Serum samples were collected on admission to hospital (baseline) and days 3–5 follow-up and stored long-term at 80°C and shipped by overnight courier on dry ice to the investigator for further analysis.

2.2 Novel HBV Biomarkers and Next-Generation Sequencing Analysis

Quantification of novel NRAg and qAHBc biomarkers was performed using an alternative enzyme-linked immunosorbent assay-based method (Beijing Wantai Biological, Beijing, China) [11]. HBV sequencing of the HBV pre-core (C)/C and pre-surface (S)/S regions was done as previously described [18-20]. In brief, total DNA was isolated using a commercial kit (DNeasy Qiagen, Hilden, Germany) in parallel with negative mock and positive controls. HBV DNA was PCR-amplified (sensitivity of 10 IU/mL or 50 virus copies/mL), and gel purified (QIAquick gel extraction kit, Qiagen, Hilden, Germany) amplicons were analysed by Sanger sequencing (University of Calgary Sequencing Services, Calgary, AB, Canada). HBV genotype was determined using the National Center for Biotechnology Information Basic Local Alignment Search Tool (NCBI BLAST) HBV genotyping tool. Next-generation sequencing adaptors were added to HBV preS/S and preC/C amplicons using high-fidelity polymerase (Phusion, New England BioLabs, Whitby, Canada), and then gel amplicons were purified and sequenced using the Illumina MiSeq (San Diego, USA), along with preS/S and preC/C PCR clones as an internal control to determine the error rate (mean error rate calculated as < 1%). Sequence alignment and mutational analysis were performed using MEGA 10.0 with Clustal W alignment.

2.3 Serum Cytokine Quantification

Multiplex Luminex assay was used to measure a panel of 42 serum cytokines, chemokines and growth factors (Luminex Eve Technologies, Calgary, AB, Canada). The panel analytes include soluble CD40 ligand (sCD40L), epidermal growth factor (EGF), Eotaxin, fibroblast growth factor (FGF)-2, FMS-like tyrosine kinase 3 ligand (FLT-3 L), Fractalkine, granulocyte colony-stimulating factor (G-CSF), granulocyte-macrophage colony-stimulating factor (GM-CSF), growth-related oncogene (GRO)-α, IFN-α2, IFN-γ, interleukin (IL)-1α, IL-1β, IL-1 receptor antagonist (RA), IL-2, IL-3, IL-4, IL-5, IL-6, IL-7, IL-8, IL-9, IL-10, IL-12p40, IL-12p70, IL-13, IL-15, IL-17A, IL-18, interferon gamma-induced protein (IP)-10, monocyte chemoattractant protein (MCP)-1, MCP-3, macrophage-derived chemokine (MDC), macrophage inflammatory protein (MIP)-1α, MIP-1β, platelet-derived growth factor (PDGF)-AA, PDGF-BB, regulated on activation, normal T expressed and secreted (RANTES), transforming growth factor (TGF)-α, TNF-α, TNF-β and vascular endothelial growth factor (VEGF)-A.

2.4 Statistical Analysis

TFS at day 21 was used as the primary outcome. Mann Whitney U-test was performed for continuous variables, a chi-square test or Fisher's exact test for cells < 5 was performed for categorical variables. Wilcoxon signed-ranked test was performed to compare cytokine and chemokine levels of HBV-ALF patients from baseline to follow-up. Comparisons between unpaired groups were performed using Mann Whitney U-test. Logistic regression was used for univariate analysis. Given the exploratory nature of this study, adjustment for multiple comparisons was not performed. GraphPad Prism 9.0.0 and IBM SPSS Statistics 27.0.1.0 were used to perform all the statistical tests.

3.1 Demographic, Clinical and Viral Characteristics of Patient Cohorts

From the ALFSG registry, we identified 48 de novo acute HBV-ALF patients and 20 patients that were classified with viral reactivation and/or history of hepatitis B (Figure S1). In 48 acute HBV-ALF individuals (median age 40.0 years, 50% Female), 26/48 (54%) were transplant-free survivors (median age 42.0 years, 50% female and 54% White) compared to 22/48 (46%) that received liver transplantation and/or died (non-TFS) at day 21 post-admission (median age 39.0 years, 50% female and 68% White) (Table 1). Data on 20 individuals with ALF due to history of HBV and reactivation are provided in Table 2 [(5 were TFS with median age 51.0 years, all male and 40% White) and (15 received liver transplantation and/or died at day 21 post-admission with median age 53.0 years, 53% female and 46.7% Asian)]. As expected, all de novo acute HBV-ALF patients that survived spontaneously required less organ support including mechanical ventilation (11.5% vs. 59.1%, p = 0.0007), vasopressors (7.7% vs. 36.4%, p = 0.029), renal replacement therapy (3.9% vs. 31.8%, p = 0.017) and sedatives (15.4% vs. 54.6%, p = 0.006) as reflected by established prognostic scores (i.e., KCC and MELD), Table 1. No significant changes in organ support, intracranial pressure (ICP) direct therapies and blood products (except for fresh frozen plasma, 20% TFS vs. 80% LT/died, p = 0.015) were observed for ALF patients due to reactivation of known hepatitis B infection (Table 2). We performed univariate logistic regression analysis to analyse possible predictors of transplant-free survival in HBV-ALF patients (Tables 3 and S1). As expected, poor KCC, 3/4 admit coma grade, high MELD score and increased INR were predictors for non-TFS in acute HBV-ALF patients.

3.2 HBV Viral Load and qAHBc Was Associated With TFS in Acute HBV-ALF Patients

Assessment of hepatitis B virus in 48 patients with acute HBV-ALF showed all were HBsAg positive, 22 (46%) were HBeAg positive, with a median HBV DNA of 3.71 log₁₀ IU/mL and predominant HBV genotype A (40%). Overall, detectable DNA levels and higher levels of qAHBc were predictors for TFS (Tables 1 and 3). A higher number of TFS HBV-ALF patients had detectable HBV DNA levels compared to non-TFS HBV-ALF patients (n = 23/26, 88.5% vs. 13/22, 59.1%, p = 0.042). Additionally, higher qAHBc levels were observed for TFS HBV-ALF patients (median 4.5 log₁₀ IU/mL vs. 4.2 log₁₀ IU/mL, p = 0.044). No significant changes in other HBV biomarkers tested (i.e., HBeAg status, anti-HBc total and/or IgM, anti-HBs and HBV-NRAg) and HBV genotypes A–E in TFS vs. non-TFS were observed. Univariate logistic regression analysis showed that detectable HBV DNA levels (adjusted odds ratio of 5.308 with 95% Cl of 1.217–23.155, p = 0.026) and increased qAHBc (adjusted odds ratio of 4.466 with 95% Cl of 0.968–20.608, p = 0.050) were predictors of transplant-free survival acute in HBV-ALF patients. ALF patients due to reactivation and/or history of hepatitis B did not show any significant changes in HBV biomarkers and genotype per primary outcomes of TFS (p > 0.05). (Tables 2 and S1).

3.3 Elevation of Growth Factors and Decline in Pro-Inflammatory Cytokines Observed in TFS

A significant increase in growth factors was observed in acute HBV-ALF patients who were transplant-free survivors from baseline to days 3–5 follow-up (Figure 1). These angiogenic factors included PDGF-AA (p = 0.008), PDGF-BB (p = 0.0006) and VEGF-A (p = 0.014), Figure 1A. In comparison, non-TFS acute HBV-ALF patients that received liver transplantation and/or died showed little to no increase/no change or declining levels of angiogenic factors during follow-up. Additionally, non-TFS acute HBV-ALF patients showed an increase in serum pro-inflammatory cytokines and chemokines levels from baseline to days 3–5 follow-up, including interleukin IL-12p70 (p < 0.0001), TNF-α (p = 0.0004), IFN-α2 (p = 0.0006), G-CSF (p = 0.002), RANTES (p = 0.018), IL-6 (p = 0.034), IL-8 (p = 0.035) and fractalkine (p = 0.039), Figures 1B and S2. Whereas TFS acute HBV-ALF patients had decreasing levels of pro-inflammatory IL-8 (p = 0.0003), IL-1α (p = 0.031), IL-2 (p = 0.014) and IL-6 (p = 0.039) compared to non-TFS acute HBV-ALF patients. In addition, an increase in anti-inflammatory IL-10 was observed in both TFS (p = 0.036) and non-TFS (p = 0.0009) acute HBV-ALF patient cohorts, Figure 1C. Analysis of other cytokines showed no difference between groups (Figure S2). Collectively, the multiplex Luminex data indicate that despite an increase in pro-inflammatory cytokine response, acute TFS HBV-ALF patients exhibited heightened angiogenic response and a lower pro-inflammatory cytokine response compared to non-TFS acute HBV-ALF patients.

Serum cytokine and chemokine levels in HBV-ALF patients with history and/or reactivation of HBV showed elevation of pro-inflammatory cytokines such as IL-3 (p = 0.0005), MIP-1α (p = 0.004), TNF-α (p = 0.002), TGF-α (p = 0.016), IL-12p70 (p = 0.016), fractalkine (p = 0.029) and IL-13 (p = 0.031) in non-TFS compared to TFS patients from baseline to days 3–5 follow-up (Figure 1D,E). Only the angiogenic factor PDGF-BB (p = 0.027) was higher in TFS in this group. No significant difference in analysis of other angiogenic factors, cytokines and chemokines were observed (Figure S3).

3.4 Higher Frequency of HBV Immune Escape Variants in TFS Acute HBV-ALF Patients

Next-generation sequencing analysis of HBV surface and core genes showed that TFS acute HBV-ALF patients had higher levels of HBV variants associated with immune escape phenotype compared to non-TFS acute HBV-ALF patients at baseline and days 3–5 follow-up (Figure 2A). These variants include P127L/T (p = 0.030 baseline, p = 0.007 d 3–5), T140I (p = 0.035 baseline) and S143L or T143S (p = 0.041 baseline) and E164D/G (p = 0.046 d 3–5) (Table S2). In TFS compared to non-TFS acute HBV-ALF patients, the overall frequency of HBV immune escape variants (D144A/E) declined from baseline to days 3–5 follow-up p = 0.047. HBV surface gene variants associated with the development of anti-viral therapy resistance were not statistically different between groups (Figure 2B). As well, HBV core gene variant at baseline only, F24Y levels (associated with increased risk of cirrhosis and HCC development) were greater for non-TFS than TFS acute HBV-ALF patients (p = 0.037, Figure 2C).

In ALF patients with history and/or reactivation of hepatitis B, there weren't significant differences in immune escape variants frequencies between TFS and non-TFS groups at baseline (Figure 3A). Increased frequency of variants associated with anti-viral therapy resistance (F161H/L p = 0.034, S193F/L p = 0.011 and V194F/S p = 0.019, Figure 3B), and variants associated with increased risk of cirrhosis and HCC (F24Y p = 0.036 and L116 I p = 0.030, Figure 3C) were observed at baseline. No significant differences in HBV surface and core gene variants were observed at days 3–5 follow-up.