Predictive value of HBeAg titer dynamics for HBsAg clearance in pediatric chronic hepatitis B

Abstract

Introduction: Achieving functional cure of chronic hepatitis B (CHB), characterized by the loss of HBV DNA and HBsAg, remains challenging in adults but demonstrates higher success rates in children. Elucidating the factors influencing HBsAg loss in pediatric patients is crucial for optimizing treatment strategies. This study aimed to evaluate the predictive value of HBeAg titer dynamics for HBsAg clearance in pediatric CHB and develop a predictive model incorporating these dynamics.

Material and methods: This retrospective cohort study analyzed 119 children aged 1-18 years with CHB treated with nucleos(t)ide analogues. Patient outcomes were evaluated using two independent classification approaches: HBsAg loss status and age stratification (≤6 vs. >6 years). Treatment response was assessed through longitudinal HBeAg titer measurements during the first 12 months. Based on identified predictors, a logistic regression model was developed incorporating age and HBeAg titer dynamics to predict HBsAg clearance probability.

Results: The study population exhibited a median age of 6.2 years. Factors associated with HBsAg loss encompassed younger age, female sex, and absence of breakthrough. In multivariate analysis, younger age was identified as the only significant factor. The cumulative HBsAg loss rate demonstrated markedly higher values in the ≤6 years group (Hazard ratio 7.69). HBeAg titer decline exhibited significantly steeper trajectories in the HBsAg loss group. The developed predictive model, "Log Odds = -1.182 + 0.308 × log_reduction-0.205 × age", demonstrated good performance with high accuracy.

Conclusions: Early HBeAg titer dynamics combined with age at treatment initiation may serve as useful predictors of HBsAg clearance in pediatric CHB. Our predictive model, utilizing readily available semi-quantitative HBeAg measurements, could potentially assist clinicians in therapeutic decision-making and individualized treatment strategies.

Keywords: antiviral agents; children; hepatitis B e antigens; hepatitis B surface antigens; seroconversion.

Figure 1

Flowchart outlining study design. CHB, chronic hepatitis B; UNL, upper normal limit; NUCs, nucleos(t)ide analogues.

Figure 2

Age-dependent analysis of cumulative HBsAg loss rates. Kaplan–Meier survival analysis demonstrates significantly enhanced clearance rates in the younger cohort. Cox Proportional Hazards Model analysis revealed a hazard ratio of 7.69 (95% CI: 3.03–20.0, p < 0.001) for the Age ≤6 group. The visualization employs Kaplan–Meier curves to illustrate age-group differences.

Figure 3

Comparative analysis of HBeAg titer reduction patterns. The illustration presents logarithmic HBeAg titer reduction trajectories for both outcome groups. The HBsAg loss group demonstrated an average slope of −5.628, in contrast to −3.740 in the non-HBsAg loss group (t-statistic = −3.982, p < 0.001).

Figure 4

Predictive model performance evaluation. The ROC curve illustrates the model's discriminative capability, with optimal prediction threshold denoted by an x-mark. ROC, Receiver Operating Characteristic; AUC, area under the curve; CI, confidence interval.