Estimating age-related incidence of HBsAg seroclearance in chronic hepatitis B virus infections of China by using a dynamic compartmental model

Abstract

The age-specific seroclearance pattern of hepatitis B surface antigen (HBsAg) in chronic hepatitis B virus (HBV) infections of China remains unclear. In this study, based on three national serosurvey data of hepatitis B in China, we propose an age- and time-dependent discrete model and use the method of non-linear least squares to estimate the age-specific annual rate of HBsAg seroclearance. We found that the HBsAg seroclearance in chronic HBV infections of China aged 1-59 years occurred at an average annual rate of 1.80% (95% CI, 1.54-2.06%) from 1993 to 2006. The HBsAg seroclearance occurred predominantly in the early childhood, 20-24 and 35-39 year age groups. Moreover, our model estimated that HBsAg seroclearance resulted in 23.38% of the decrease of total HBsAg prevalence for population aged 1-59 years in 2006. It also prevented 9.30% of new HBV infections (about 7.43 million people) and 9.95% of HBV-related deaths (about 0.25 million people) from 1993 to 2006. This study develops a new and efficient method to estimate the age-specific incidence of HBsAg seroclearance at a population-level and evaluate its effect.

Figure 1

Flow chart of HBV transmission. Here “a” denoted the age and “t” denoted the time. β _a(t) denoted the average transmission rate in age group a in year t in the sense that an age a susceptible had contacted with an infectious individual and was successfully infected; q _a denoted the age-specific proportion of acute HBV infections that became chronic per year; r _a denoted the annual rate of HBsAg seroclearance in chronic HBV infections; m _a denoted the age-specific mortality rate of HBV-related diseases per year; θ _a(t) denoted catch-up vaccination coverage rate for population aged 8–15 years in 2009–2011; μ _a(t) denoted age- and time-dependent death rate of non-HBV related diseases in year t; b(t) denoted birth rate in year t; v(t) denoted vaccination coverage rate of newborns in year t; p denoted vaccination protection rate per year; ε denoted perinatal infection rate per year; w denoted the proportion of HBsAg positive mothers in the total prevalence of HBsAg aged 15–49 years; q _p denoted the proportion of acute HBV infections that became chronic during delivery; N(t) denoted the total number of population in year t; C(t) denoted the total population number of chronic HBV infections in year t; N _a(t) denoted the number of people in age category a in year t.

Figure 2

Model input parameters. (a) Birth rate from 1991 to 2014. (b) Vaccination coverage rate of newborns from 1993 to 2014. (c) Age-specific mortality rate of HBV-related diseases. (d) The age-specific total death rate in 1990, 2000, 2005 and 2010. (e) The age-specific transmission rate of HBV in 1992 and its 95% confidence intervals.

Figure 3

Comparison of estimated HBsAg prevalence with national survey data. (a) In 2006. (b) In 2014.

Figure 4

The influence of parameters on estimation of annual rate of HBsAg seroclearance. (a) The influence of vaccination coverage rate for newborns. (b) The influence of vaccination protection rate. (c) The influence of perinatal infection rate.

Figure 5

The effect of HBsAg seroclearance. (a) On total prevalence of HBsAg for population aged 1–59 years in China. (b) On the age-specific prevalence of HBsAg in 2006. (c) On the age-specific prevalence of HBsAg in 2014. (d) On the total number of new HBV infections. (e) On the total number of HBV-related deaths.