Does hepatitis B seroconversion affect survival outcome in patients with hepatitis B related hepatocellular carcinoma?

BACKGROUND

Little is known about whether hepatitis B surface antigen (HBsAg) seroconversion (SC) contributes to any survival benefits for patients with hepatocellular carcinoma (HCC).

METHODS

All patients with hepatitis B-related HCC and HBsAg seroclearance between 1989 and 2013 were identified. Case- and control-groups were matched according to their stage of disease and mode of treatment. Baseline demographics, liver function, and overall survivals (OS) were compared between these two groups.

RESULTS

Thirty-nine HCC cases with HBsAg SC were identified, and 312 non-seroconversion (NSC) HCC cases were matched. Forty-eight percent of patients had curative resections, 14% were treated with ablation and 38% were for palliation. Age of patients in SC group was older than those in NSC group (P=0.026). Although there was significantly better liver function in SC vs. NSC groups in terms of bilirubin (P=0.027), albumin (P=0.003), AST (P=0.001) and ALT (P<0.001), there was no overall difference in Child-Pugh grade among the two groups. In regarding tumour pathology, SC commonly presented with solitary tumour nodule as compared to multiple nodules in NSC (P=0.027), and was also frequently associated with a normal background liver parenchyma (P<0.001). Although no survival benefit was confirmed in log-rank analysis between SC and NSC, the absolute 5-year survival of SC group was better in resection (72.2% vs. 55.3%), ablation (83.3% vs. 57.4%) and palliation (24.4% vs. 14.4%).

CONCLUSIONS

HCC patients with HBsAg SC are associated with a better background liver parenchyma and function, and might contribute to an improved long-term survival.

Longitudinal profiles of highly sensitive hepatitis B surface antigen levels: re-evaluation of HBsAg seroclearance

BACKGROUND & AIMS

Serologic profiles after hepatitis B surface antigen (HBsAg) seroclearance in chronic hepatitis B (CHB) have not been well-studied.

METHODS

We employed a highly sensitive HBsAg (hs-HBsAg) assay (lower detection limit 0.5 mIU/ml), 100 times more sensitive than conventional HBsAg measurements. CHB patients achieving HBsAg seroclearance defined by conventional assays were followed up for serum hs-HBsAg, HBV DNA and antibody to HBsAg (anti-HBs) levels at 0 months, 6-12 months and 3-5 years after HBsAg seroclearance. Factors associated with hs-HBsAg detectability were determined.

RESULTS

One hundred and nine patients were recruited; 94 (86.2%) were followed up to years 3-5; and 25 patients (22.9%) were on nucleoside analogue therapy for a median duration of 6.0 (range 1.5-12.7) years before HBsAg seroclearance. Detectable hs-HBsAg was noted in 88 (80.7%), 60 (55.0%) and 20 (21.3%) patients at 0 months, 6-12 months and 3-5 years respectively. At years 3-5, genotype B patients, when compared to genotype C patients, had a higher anti-HBs positive rate (63.2% and 41.1% respectively, P = 0.036). Serum anti-HBs positivity, when compared to persistent anti-HBs negativity, was associated with a lower rate of hs-HBsAg detection (7.4% and 40% respectively, P < 0.001). Multivariate analysis showed anti-HBs negativity at years 3-5 to be independently associated with persistently positive hs-HBsAg (P = 0.007, odds ratio 7.1, 95% confidence interval 1.7-29.3).

CONCLUSION

Serum hs-HBsAg could detect HBsAg presence in a substantial proportion of CHB after HBsAg seroclearance defined by conventional assays, especially among anti-HBs negative individuals. Serum hs-HBsAg could potentially assist differentiating HBsAg-negative CHB from individuals with only past HBV exposure without carrier state.

Comparing Normothermic Machine Perfusion Preservation With Different Perfusates on Porcine Livers From Donors After Circulatory Death

The utilization of normothermic machine perfusion (NMP) may be an effective strategy to resuscitate livers from donation after circulatory death (DCD). There is no consensus regarding the efficacy of different perfusates on graft and bile duct viability. The aim of this study was to compare, in an NMP porcine DCD model, the preservation potential of three different perfusates. Twenty porcine livers with 60 min of warm ischemia were separated into four preservation groups: cold storage (CS), NMP with Steen solution (Steen; XVIVO Perfusion Inc., Denver, CO), Steen plus red blood cells (RBCs), or whole blood (WB). All livers were preserved for 10 h and reperfused to simulate transplantation for 24 h. During preservation, the NMP with Steen group presented the highest hepatocellular injury. At reperfusion, the CS group had the lowest bile production and the worst hepatocellular injury compared with all other groups, followed by NMP with Steen; the Steen plus RBC and WB groups presented the best functional and hepatocellular injury outcomes, with WB livers showing lower aspartate aminotransferase release and a trend toward better results for most parameters. Based on our results, a perfusate that contains an oxygen carrier is most effective in a model of NMP porcine DCD livers compared with Steen solution. Specifically, WB-perfused livers showed a trend toward better outcomes compared with Steen plus RBCs.

Relation of Tricuspid Regurgitation to Liver Stiffness Measured by Transient Elastography in Patients With Left-Sided Cardiac Valve Disease

The aim of the study was to evaluate the relation between tricuspid regurgitation (TR) severity and liver stiffness (LS) in patients with TR. A total of 131 patients with various degrees of TR secondary to left-sided heart valve disease were enrolled. Severity of TR was quantitatively assessed by proximal isovelocity surface area-derived effective regurgitant orifice (ERO). Patients were divided into 2 groups: 48 with mild-moderate TR (ERO <0.4 cm(2)) and 83 with severe TR (ERO ≥0.4 cm(2)). Transient elastography was used to measure the level of LS, an established marker of liver fibrosis, with the threshold of significant LS set at ≥12.5 kPa. Patients with severe TR had a higher LS and prevalence of significant LS than those with mild-moderate TR. Furthermore, LS and significant LS independently correlated with TR-ERO, right atrial pressure and inferior vena cava (IVC) diameter. The presence of a large TR-ERO (≥0.4 cm(2)) and IVC diameter (>2.15 cm(2)) provided a high specificity of 78% for significant LS. In conclusion, the present study demonstrates that TR-ERO, right atrial pressure, and IVC diameter are important parameters associated with LS in patients with TR.

Outcomes including liver histology after liver transplantation for chronic hepatitis B using oral antiviral therapy alone

The outcomes of hepatitis B virus (HBV)-related hepatitis after liver transplantation (LT) without hepatitis B immune globulin (HBIG) is not well documented. This study aims to determine the outcomes of chronic hepatitis B (CHB) patients using an HBIG-free regimen. All biopsies performed 3 months or more after LT in consecutive CHB patients transplanted from 2003 to 2012 were reviewed. None of the patients received HBIG. Results of all liver histologies were reviewed to determine the cause of graft dysfunction. Of the 435 patients transplanted during this period, 263 liver biopsies were performed in 144 patients. Thirty-six patients were positive for hepatitis B surface antigen (HBsAg) with undetectable HBV DNA at the time of biopsy, and none had histological evidence of HBV infection. Of the 263 biopsies, 44 (17%) had evidence of fibrosis. There was a significantly higher rate of fibrosis in those with large duct obstruction compared to those without (51% versus 9%, respectively; P < 0.001). Of the 291 patients without a liver biopsy during the same period, 43 were HBsAg+. Seven patients had evidence of virological rebound, of whom 6 had evidence of rtM204V/I mutation and 1 had recurrence of hepatocellular carcinoma with low-level rebound and wild-type virus. In conclusion, for patients without virological rebound, positive serum HBsAg was not associated with histological evidence of HBV-related hepatitis after LT. To prevent virological rebound, nucleos(t)ide analogues with higher barriers to resistance should be used.

Intrahepatic hepatitis B virus replication and liver histology in subjects with occult hepatitis B infection

We studied the intrahepatic hepatitis B virus (HBV) replicative status in 40 people with occult hepatitis B infection (OBI) and 40 patients with chronic hepatitis B (CHB). Intrahepatic HBV DNA, covalently closed circular DNA (cccDNA), and pre-genomic RNA (pgRNA) were quantified. Patients with OBI had median necroinflammation and fibrosis scores of 1 and 0, respectively. Intrahepatic total HBV DNA, cccDNA and pgRNA were detectable in 30 (77%), one (3%) and five (13%) of the participants with OBI, respectively. People with OBI had lower median intrahepatic total HBV DNA than the patients with CHB (p < 0.0001). They had nearly normal liver histology and low intrahepatic HBV replication.

Era of direct acting antivirals in chronic hepatitis C: Who will benefit?

In the era of highly effective direct acting antiviral (DAA) drugs for the treatment of chronic hepatitis C (CHC) infection, where eradication is almost ensured with minimal side effects, all hepatitis C carriers should benefit theoretically. In the real world setting however, only a small proportion will benefit at this time point due to the multiple barriers to accessing therapy. Given that universal treatment is unlikely, treatment with DAAs will likely be restricted to those with the highest health benefits, and for those who can afford the high expense of a treatment course. Those with the highest unmet needs include those who have failed previous interferon-based therapy or who are interferon-ineligible with evidence of active disease, those with advance liver disease, and those with recurrence of hepatitis C after liver transplantation. In the future, the focus should be on increasing access to treatment for those infected with CHC.

Evolutionary Changes of Hepatitis B Virus Pre-S Mutations Prior to Development of Hepatocellular Carcinoma

BACKGROUND AND AIMS

Deletions/mutations in the hepatitis B virus (HBV) pre-S region have been associated with hepatocellular carcinoma (HCC). We aimed to study the evolutionary changes of pre-S mutations prior to HCC development.

METHODS

We studied the HBV pre-S sequences at 1 to 10 years preceding diagnosis of HCC in 74 patients with HBV-related HCC (HCC group). 148 chronic hepatitis B patients matched for sex and age in 2:1 ratio, who had been followed up for at least 3 years without HCC (HCC-free group) were recruited as controls. 56 and 47 patients of HCC and HCC-free groups respectively had serially stored sera for longitudinally examination at 1-3 years, 4-6 years, 7-9 years and ≥10 years prior to the recruitment of the study.

RESULTS

Compared to the HCC-free group, higher frequencies of pre-S deletions and point mutations (at 11 codons) were observed in the HCC group (p<0.05). Multiple logistic regression analysis showed that pre-S deletions, point mutations at codon 51 and 167 were independent factors associated with HCC. Longitudinal observation showed that pre-S deletions and most of the 11 HCC-associated pre-S point mutations existed at least 10 years before HCC development, and were more prevalent preceding HCC development in patients from HCC groups than HCC-free group. The number of HCC-associated pre-S point mutations increased over time preceding HCC development, and correlated positively with the time to HCC diagnosis (r = 0.220, p = 0.005).

CONCLUSIONS

High prevalence and cumulative evolution of pre-S mutations preceding HCC development suggested a possible carcinogenic role of pre-S mutations and their potential application in HCC risk prediction.

A detailed evaluation of cardiac function in cirrhotic patients and its alteration with or without liver transplantation

BACKGROUND

Cirrhosis has been shown to be associated with left ventricular (LV) myocardial dysfunction, but studies of right ventricular (RV) function in cirrhotic patients compared with controls are scarce. Limited studies have prospectively evaluated the progression of myocardial function in patients with cirrhosis and assessed changes in cardiac function following liver transplantation (LTx). So the aim of the study was to evaluate biventricular myocardial function in cirrhotic patients and its alteration with or without liver transplantation.

METHODS

A total of 103 patients with cirrhosis (age 55±7 years, male 75%) were recruited. Conventional and 2-dimensional speckle tracking echocardiography was performed to determine the presence of LV and RV (biventricular) dysfunction. For comparison, 48 matched control subjects were included. Follow-up echocardiography was performed in 41 patients following LTx and in 26 patients who did not undergo LTx.

RESULTS

Patients with cirrhosis had biventricular dilatation, increased LV mass, impaired LV diastolic function, and biventricular systolic strain compared with controls. Following LTx, cirrhotic patients had reduced biventricular dilatation, a smaller LV mass, and improved biventricular systolic strain after a mean duration of 18.2±6.6 months. Patients who did not undergo LTx had a further increase in LV mass but no significant change in biventricular dimensions or systolic strain (mean duration of 20.4±8.3 months).

CONCLUSIONS

The present study demonstrates that patients with cirrhosis had biventricular dilatation and impaired biventricular systolic strain compared with controls. Following LTx, biventricular dilatation reduced and biventricular systolic strain improved. In contrast, patients who did not undergo LTx experienced a further increase in LV mass.

Management of chronic hepatitis B before and after liver transplantation

Liver transplantation remains the only curative option for eligible patients with complications of chronic hepatitis B (CHB) infection, including severe acute hepatitis flares, decompensated cirrhosis, and hepatocellular carcinoma. In general, all patients with CHB awaiting liver transplantation should be treated with oral nucleos(t)ide analogs (NAs) with high barriers to resistance to prevent potential flares of hepatitis and reduce disease progression. After liver transplantation, lifelong antiviral therapy is also required to prevent graft hepatitis, which may lead to subsequent graft loss. Although combination therapy using NA and hepatitis B immune globulin (HBIG) has been the regimen most widely adopted for over a decade, recent studies have demonstrated that newer NAs with low rates of resistance are effective in preventing graft hepatitis even without the use of HBIG, achieving excellent long term outcome. For patients without pre-existing resistant mutations, monotherapy with a single NA has been shown to be effective. For those with resistant strains, a combination of nucleoside analog and nucleotide analog should be used. To date, clinical trials using therapeutic vaccination have shown suboptimal response, as CHB patients likely have an immune deficit against HBV epitopes. Future strategies include targeting different sites of the hepatitis B replication cycle and restoring the host immunity response to facilitate complete viral eradication.

Hepatitis B surface antigen levels after hepatitis B e-antigen seroclearance: a longitudinal follow-up study

BACKGROUND & AIMS

The role of quantitative hepatitis B surface antigen (HBsAg) after hepatitis B e-antigen (HBeAg) seroclearance is not well defined. To determine the role of HBsAg levels in predicting significant viremia and hepatitis flares after HBeAg seroclearance.

METHODS

A total of 228 chronic hepatitis B patients with spontaneous HBeAg seroclearance were included. Patients were followed up regularly at 3-6 monthly intervals with routine liver biochemistry and hepatitis B serology. Levels of HBV DNA and HBsAg were measured at yearly intervals for up to 5 years after HBeAg seroclearance.

RESULTS

The median log HBsAg and HBV DNA level after HBeAg seroclearance was 3.52 IU/ml and 4.13 IU/ ml respectively, with no significant correlation observed between them (P = 0.572). The HBV DNA at HBeAg seroclearance was 4.13 log IU/ml, compared with 3.12 log IU/ml after 5 years (P < 0.001). No significant change was observed for HBsAg levels (P = 0.991). Hepatitis B flares occurred in 76 (33.3%) patients. Patients who developed hepatitic flares compared with those without hepatitic flares were older (40 vs. 36 years, P = 0.001), had a higher HBV DNA at the time of HBeAg seroclearance (4.70 vs. 3.77 log IU/ml, P =< 0.001), and more likely to be males (42.7% vs. 23.4%, P = 0.002) respectively. There was no difference in HBsAg levels between those with and without hepatitis flare (3.54 vs. 3.52 log IU/ml respectively, P = 0.555).

CONCLUSION

HBV DNA levels, but not HBsAg levels, after HBeAg seroclearance were associated with subsequent significant viremia and hepatitic flares. Male gender and older age was associated with significant viremia.

High prevalence of non-alcoholic fatty liver disease in the Chinese - results from the Hong Kong liver health census

BACKGROUND & AIMS

The prevalence of non-alcoholic fatty liver disease (NAFLD) in the general Chinese population is not clear, as there have been few large population studies focussing on healthy individuals. We aim to determine the prevalence and risk factors of NAFLD in a healthy Westernized Chinese population.

METHODS

This is a prospective cross-sectional study of 2493 volunteers recruited from the general population and the Red Cross Transfusion Center in Hong Kong. All underwent comprehensive questionnaire survey, blood test, abdominal ultrasound and transient elastography.

RESULTS

The prevalence of NAFLD was 42%. Gender, age, waist circumference, systolic blood pressure, fasting cholesterol and glucose levels were significant factors. The relative risk of NAFLD in those with high waist circumference, diabetes, hypertension, and hypercholesterolaemia was 2.99, 2.01, 1.79 and 1.54 respectively. The degree of steatosis increased with levels of fasting glucose, cholesterol, systolic blood pressure and waist circumference (all P < 0.001). The optimal waist circumference was found to be 84 and 74 cm for male and females respectively, with a relative risk of 5.16 for those above this limit. In total, 1.2% and 0.002% with NAFLD had advanced liver fibrosis and cirrhosis respectively.

CONCLUSION

Non-alcoholic fatty liver disease was found to be highly prevalent in the Chinese population. Increasing levels of risk factors were associated with increasing severity of NAFLD. Lower cut-off levels of waist circumference to predict NAFLD should be adopted for Chinese population.

Management of chronic hepatitis B in severe liver disease

In the past few decades, chronic hepatitis B (CHB) has evolved from a disease that was untreatable and progressive, to one that can be easily controlled with antiviral therapy. However, patients with severe liver disease still remain difficult to treat despite the availability of highly potent nucleos(t)ide analogs. These include those with underlying cirrhosis, severe flares of CHB, hepatocellular carcinoma (HCC), and for those undergoing liver transplantation. For those with established cirrhosis, antiviral therapy should be considered for all, as unpredictable flares can still occur, which can be fatal for those with advanced chronic liver disease. However, even with effective viral suppression, the development of HCC can still occur. For patients with severe flares of CHB, although the use of antiviral can improve long term outcomes, a significant proportion may still die without liver transplantation. The short term prognosis of these patients is dependent on both the severity of flare and underlying pre-existing liver disease. In patients with decompensated cirrhosis, liver failure secondary to severe flares, or those with HCC, liver transplantation may be curative. After liver transplantation, long term antiviral therapy is required to prevent graft loss from recurrent hepatitis B infection. The use of hepatitis B immune globulin (HBIG) in combination with an oral antiviral agent has been the mainstay of post-transplant antiviral regimen for over a decade. With newer and more potent antiviral agents such as tenofovir and entecavir, use of these agents along with HBIG have demonstrated to be effective in preventing significant recurrence in the long term.

Hepatitis B reactivation in patients with previous hepatitis B virus exposure undergoing rituximab-containing chemotherapy for lymphoma: a prospective study

PURPOSE

Patterns of hepatitis B virus (HBV) reactivation in hepatitis B surface antigen (HBsAg) -negative, antihepatitis B core antigen antibody (anti-HBc) -positive patients with lymphoma receiving rituximab-containing chemotherapy have not been well described.

PATIENTS AND METHODS

HBsAg-negative, anti-HBc-positive Chinese patients with undetectable serum HBV DNA (< 10 IU/mL), diagnosed with hematologic malignancies and receiving rituximab-containing chemotherapy, were prospectively monitored every 4 weeks for up to 2 years. Entecavir was started when HBV reactivation (defined as detectable HBV DNA) was encountered.

RESULTS

Among 260 patients receiving rituximab-containing chemotherapy, 63 patients (24.2%) who were HBsAg negative and anti-HBc positive underwent follow-up for a median of 70 weeks (range, 6 to 104 weeks). The 2-year cumulative rate of HBV reactivation was 41.5%, occurring at a median of 23 weeks (range, 4 to 100 weeks) after rituximab treatment. The median HBV DNA level at reactivation was 43 IU/mL (range, 14 to 920 IU/mL). A baseline undetectable antibody to HBsAg (anti-HBs; < 10 mIU/mL) was the only significant risk factor that was positively associated with HBV reactivation (hazard ratio, 3.51; 95% CI, 1.37 to 8.98; P = .009). Patients with negative baseline anti-HBs, compared with those with positive anti-HBs, had a significantly higher 2-year cumulative rate of HBV reactivation (68.3% v 34.4%; P = .012). At HBV reactivation, all patients had normal ALT, and all patients but one were HBsAg negative. Entecavir successfully controlled HBV reactivation in all patients.

CONCLUSION

A high rate of HBV reactivation was observed in HBsAg-negative, anti-HBc-positive patients undergoing rituximab-containing chemotherapy, with the risk of reactivation significantly higher in anti-HBs-negative patients. Periodic HBV DNA monitoring was an effective strategy in preventing HBV-related complications.

Effect of hepatitis B virus reverse transcriptase variations on entecavir treatment response

BACKGROUND

Entecavir therapy often reduces hepatitis B virus (HBV) DNA to an undetectable level, but HBV DNA remain detectable in some patients. We investigated whether baseline HBV reverse transcriptase (rt) polymorphism and quasispecies complexity and diversity were associated with treatment response.

METHODS

Pretreatment HBV DNA levels, HBV rt sequence, serology, and quasispecies complexity and diversity from 305 entecavir-treated patients were determined. These data were tested for their association with year 1 virological outcome, defined by optimal response (undetectable HBV DNA; lower limit of detection, ≤12 IU/mL) or partial response (detectable HBV DNA).

RESULTS

Four rt variants were more frequently detected in the 64 partial responders than in the 241 optimal responders (all P < .05). Multivariate analysis revealed that high baseline HBV DNA level (P < .0001; odds ratio [OR], 2.32), HBV e antigen (HBeAg) positivity (P < .001; OR, 3.70), and rt124N (P = .002; OR, 3.06) were associated with a partial entecavir response. Compared with the optimal responders, the partial responders had a lower quasispecies complexity and diversity.

CONCLUSIONS

Apart from the known factors (high baseline HBV DNA level and HBeAg positivity), a novel single nucleotide polymorphism (rt124N) and lower quasispecies complexity and diversity were associated with partial entecavir response at year 1.

Association between liver transplant center performance evaluations and transplant volume

There has been increased oversight of transplant centers and stagnation in liver transplantation nationally in recent years. We hypothesized that centers that received low performance (LP) evaluations were more likely to alter protocols, resulting in reduced rates of transplants and patients placed on the waiting list. We evaluated the association of LP evaluations and transplant activity among liver transplant centers in the United States using national Scientific Registry of Transplant Recipients data (January 2007 to July 2012). We compared the average change in recipient and candidate volume and donor and patient characteristics based on whether the centers received LP evaluations. Of 92 eligible centers, 27 (29%) received at least one LP evaluation. Centers without an LP evaluation (n = 65) had an average increase of 9.3 transplants and 14.9 candidates while LP centers had an average decrease of 39.9 transplants (p < 0.01) and 67.3 candidates (p < 0.01). LP centers reduced the use of older donors, donations with longer cold ischemia, and donations after cardiac death (p-values < 0.01). There was no association between the change in transplant volume and measured performance (R(2)  = 0.002, p = 0.91). Findings indicate a strong association between performance evaluations and changes in candidate listings and transplants among liver transplant centers, with no measurable improvement in outcomes associated with reduction in transplant volume.

Linearized hepatitis B surface antigen and hepatitis B core-related antigen in the natural history of chronic hepatitis B

Changes in two novel HBV serological markers, linearized hepatitis B surface antigen (HQ-HBsAg) and hepatitis B core-related antigen (HBcrAg), in the natural history of chronic hepatitis B (CHB) have not been well characterized. Serum HQ-HBsAg and HBcrAg levels of 404 Asian treatment-naïve CHB patients were analysed in a cross-sectional manner. Patients were categorized into five groups: immune tolerant (IT group, n=52), immune clearance (IC group, n=105), hepatitis B e antigen (HBeAg)-negative hepatitis (ENH group, n=97), HBeAg-negative quiescent group (ENQ group, n=95) and CHB with hepatitis B surface antigen (HBsAg) seroclearance (SC group, n=55). HQ-HBsAg and HBcrAg were measured and correlated with HBV DNA, HBsAg, HBV genotype and clinical parameters. HQ-HBsAg showed good correlation with HBsAg, especially in the ENQ group (r=0.874, p<0.001). Correlation of HQ-HBsAg with HBV DNA was less prominent and weakest in the ENH group (r=0.268, p 0.008). HBcrAg correlated best with HBV DNA in the ENQ group (r=0.537, p<0.001). In the ENQ group, 42.1% of patients had undetectable HBcrAg; this subgroup of patients, when compared with those with detectable HBcrAg, had significantly lower median HBV DNA (3.17/4.48 log IU/mL, p<0.001) and HBsAg (5.05/5.96 log mIU/mL, p<0.001) levels. Forty per cent of the SC group patients had detectable HQ-HBsAg and/or HBcrAg up to 42 months after HBsAg seroclearance. When comparing anti-HBs positivity and median time after HBsAg seroclearance in the SC group with and without detectable HQ-HBsAg/HBcrAg, there was no significant difference (22.7% and 36.4%, respectively, p 0.284, and 76.5 and 93.2 months, respectively, p 0.245). HQ-HBsAg and HBcrAg showed unique patterns of distribution throughout the five disease phases of CHB, including high detectability rates after HBsAg seroclearance, opening up different possibilities for their applicability.

Artificial neural network accurately predicts hepatitis B surface antigen seroclearance

BACKGROUND & AIMS

Hepatitis B surface antigen (HBsAg) seroclearance and seroconversion are regarded as favorable outcomes of chronic hepatitis B (CHB). This study aimed to develop artificial neural networks (ANNs) that could accurately predict HBsAg seroclearance or seroconversion on the basis of available serum variables.

METHODS

Data from 203 untreated, HBeAg-negative CHB patients with spontaneous HBsAg seroclearance (63 with HBsAg seroconversion), and 203 age- and sex-matched HBeAg-negative controls were analyzed. ANNs and logistic regression models (LRMs) were built and tested according to HBsAg seroclearance and seroconversion. Predictive accuracy was assessed with area under the receiver operating characteristic curve (AUROC).

RESULTS

Serum quantitative HBsAg (qHBsAg) and HBV DNA levels, qHBsAg and HBV DNA reduction were related to HBsAg seroclearance (P<0.001) and were used for ANN/LRM-HBsAg seroclearance building, whereas, qHBsAg reduction was not associated with ANN-HBsAg seroconversion (P = 0.197) and LRM-HBsAg seroconversion was solely based on qHBsAg (P = 0.01). For HBsAg seroclearance, AUROCs of ANN were 0.96, 0.93 and 0.95 for the training, testing and genotype B subgroups respectively. They were significantly higher than those of LRM, qHBsAg and HBV DNA (all P<0.05). Although the performance of ANN-HBsAg seroconversion (AUROC 0.757) was inferior to that for HBsAg seroclearance, it tended to be better than those of LRM, qHBsAg and HBV DNA.

CONCLUSIONS

ANN identifies spontaneous HBsAg seroclearance in HBeAg-negative CHB patients with better accuracy, on the basis of easily available serum data. More useful predictors for HBsAg seroconversion are still needed to be explored in the future.