Evaluating the independent influence of sexual transmission on HBV infection in China: a modeling study

The data used in this paper were derived from three national serosurveys of hepatitis B in China (Table 1) [2, 14, 15]. The demographic data were collected from the National Bureau of Statistics of China. Based on the natural history of HBV infection and the main characteristics of HBV transmission in China, we developed a sex- and age-specific discrete model to describe the transmission dynamics of HBV. Figure 1 showed the structure of the compartment model. In the model, we divided the total population into three compartments: Susceptible individuals; Chronic HBV infections and Recovered cases who had recovered from HBV infection or had Hepatitis B vaccination. Particularly, we considered three different infection routes: one was transmission through sexual contacts, the second was perinatal infection and the third was transmission through contact with infected blood (a detailed model description was provided in the Supplementary Material). The initial conditions of the model were determined according to the age- and sex-specific national serosurvey data in 1992. Based on the age- and sex-specific national survey data of hepatitis B in 1992 and 2006, by using the MCMC method, we estimated the age- and sex-specific annual rates of HBsAg seroclearance and the sex-specific horizontal transmission rate, as well as their 95% confidence intervals (Table 2). The detailed description of other parameters and variables were summarized in Tables S1, S2 and Figure S1.

We calibrated and validated our model based on the national serosurvey data of hepatitis B in 2006 and 2014. More specifically, we substituted the estimated parameter values into the model, after 14 iterations, we obtained the estimated prevalence of HBsAg in males and females aged 1–59 years in 2006. By comparing with the national serosurvey data of hepatitis B in 2006, we can see that the serosurvey data were basically within the 95% confidence intervals and the largest errors in males and females was 1.04 and 0.91%, respectively, which indicated that the estimated prevalence of HBsAg fitted well with the serosurvey data of hepatitis B in 2006 (see Figure S2a,b). Then we substituted the estimated parameter values into the model, after 22 iterations, we obtained the estimated prevalence of HBsAg in males and females aged 1–29 years in 2014. We can see that the estimated prevalence of HBsAg in males and females also fitted well with the serosurvey data of hepatitis B in 2014 (see Figure S2c,d). Therefore, our model-based estimation were credible for evaluating the independent impact of sexual transmission on HBV infection and predicting the long-term effect of promotion of condom use in China in the future.

To evaluate the impact of sexual transmission, we estimated the sex-specific prevalence of HBsAg in the population aged 1–59 years from 1993 to 2014, the sex-specific number of acute HBV infections and HBV-related deaths with or without considering sexual transmission. Moreover, we estimated the increase in HBsAg prevalence due to sexual transmission for people aged 1–59 years in 2006 and 2014, as well as the proportion of sex-specific acute HBV infections and HBV-related deaths due to sexual transmission in 2006 and 2014.

We accessed two different condom use strategies: (i) the condom usage rate reached 100% starting in 2019; and (ii) the condom usage rate was increased by 10% annually starting in 2019. We compared the two strategies with the current practice by numerical simulation. Particularly, we predicted the total number of sex-specific chronic HBV infections, acute HBV infections and HBV-related deaths, as well as the corresponding number of sex-specific chronic HBV infections, acute HBV infections and HBV-related deaths in each age group in China from 2019 to 2035 under the two different condom use strategies.

The sensitivity analysis was performed from two aspects. First, we assessed which parameters would affect the estimated age- and sex-specific annual rates of HBsAg seroclearance. Second, we identified which parameters would affect the predictions of chronic HBV infections, acute HBV infections and HBV-related deaths of males and females in 2035. Specifically, the initial horizontal transmission rate \( {\beta}_a^g(1992) \), the vaccination protection rate p and the perinatal infection rate ε were assessed. Here, we assumed that each parameter changed in terms of their upper and lower limit values and that the other parameters remained unchanged.

The results showed that due to sexual transmission, the total number of chronic HBV infections in people aged 0–100 years in China increased 1,598,320 from 1993 to 2006 and 3,658,439 from 1993 to 2014. Particularly, the total number of chronic HBV infections in people aged 0–100 years due to sexual transmission increased 292,581 in 2014, of which males increased 189,200 and females increased 103,381 (see Fig. 2a,b). Furthermore, we found that sexual transmission had a more important impact on chronic HBV infections in males in China than in females.

The results showed that due to sexual transmission, the total number of acute HBV infections in people aged 0–100 years in China increased by 12.82% from 1993 to 2006 and 17.22% from 1993 to 2014. Particularly, among the total number of acute HBV infections in people aged 0–100 years in 2014, new infections due to sexual transmission accounted for 34.59% (=473,462/1368805, where the numerator 473,462 was the number of acute HBV infections due to sexual transmission in 2014, the denominator 1,368,805 was the number of acute HBV infections in people aged 0–100 years in 2014), of which males accounted for 42.93% (=302,737/705171) and females accounted for 25.73% (=170,725/663634) (see Fig. 2c,d). Furthermore, we found that sexual transmission had a more important impact on acute HBV infections in males in China than in females.

We estimated that due to sexual transmission, the number of HBV-related deaths of people aged 0–100 years in China increased by 0.10% from 1993 to 2006 and 0.19% from 1993 to 2014. Particularly, among the total number of HBV-related deaths of people aged 0–100 years in 2014, HBV-related deaths due to sexual transmission accounted for 0.38% (=1462/382845), of which males accounted for 0.42% (=1153/273670) and females accounted for 0.28% (=309/109175) (see Fig. 2e,f). Thus, the impact of sexual transmission on HBV-related deaths would gradually increase over time.

The results showed that a higher condom usage rate resulted in a significant reduction in the number of chronic HBV infections in both males and females. Under the current practice, the total number of chronic HBV infections in people aged 0–100 years from 2019 to 2035 is 863,550,362, of which infections in males is 514,551,487 and in females is 348,998,875. In comparison with the current practice, if the condom usage rate was increased by 10% annually starting in 2019, the total number of chronic HBV infections in people aged 0–100 years from 2019 to 2035 would be reduced by 0.07% (863,550,362 vs 862,955,659), of which infections in males would be reduced by 0.07% (514,551,487 vs 514,173,630) and in females by 0.06% (348,998,875 vs 348,782,029) (see Table 3). However, if the condom usage rate reached 100% starting in 2019, the number of chronic HBV infections in people aged 0–100 years from 2019 to 2035 would be reduced by 0.12% (863,550,362 vs 862,505,969), of which infections in males would be reduced by 0.13% (514,551,487 vs 513,889,292) and in females by 0.11% (348,998,875 vs 348,616,677) (see Fig. 3a,b,c). Particularly, in 2035, the number of chronic HBV infections in people aged 0–100 years would be reduced by 0.23% (38,820,497 vs 38,731,913), of which infections in males would be reduced by 0.25% (22,806,441 vs 22,750,481) and in females by 0.20% (16,014,056 vs 15,981,432). The HBsAg prevalence in people aged 1–59 years in 2035 would be reduced to 2.01%, of which the prevalence in males would be reduced to 2.40% and in females to 1.58%.

In addition, if the condom usage rate was increased by 10% annually starting in 2019, then from 2019 to 2035, in the 40–49 age group, the number of chronic HBV infections in males and females would reduce the most (males:146439, females:80273) (see Fig. 3d).

The results show that the promotion of condom use would play a significant role in reducing acute HBV infections in both males and females. Under the current practice, the total number of acute HBV infections in people aged 0–100 years from 2019 to 2035 is 14,989,270, of which new infections in males is 7,454,679 and in females is 7,534,591. If the condom usage rate was increased by 10% annually starting in 2019, the total number of acute HBV infections in people aged 0–100 years from 2019 to 2035 would be reduced by 16.68% (14,989,270 vs 12,488,681), of which new infections in males would be reduced by 21.49% (7,454,679 vs 5,852,870) and in females by 11.93% (7,534,591 vs 6,635,811). In contrast, if the condom usage rate reached 100% starting in 2019, the number of acute HBV infections in people aged 0–100 years from 2019 to 2035 would be reduced by 24.30% (14,989,270 vs 11,346,190), of which the number in males would be reduced by 31.25% (7,454,679 vs 5,125,174) and in females by 17.43% (7,534,591 vs 6,221,016) (see Fig. 4a,b,c).

In addition, if the condom usage rate was increased by 10% annually starting in 2019, then from 2019 to 2035, in the 40–59 age group, the number of acute HBV infections in males and females would reduce the most (males:559415, females:310901) (see Fig. 4d).

The results showed that a higher condom usage rate resulted in a significant reduction in the number of HBV-related deaths of both males and females. Under the current practice, the total number of HBV-related deaths in people aged 0–100 years from 2019 to 2035 is 7,534,126, of which the HBV-related deaths in males is 5,268,115 and in females is 2,266,011. However, if the condom usage rate was increased by 10% annually starting in 2019, the total number of HBV-related deaths of people aged 0–100 years from 2019 to 2035 would be reduced by 0.04% (7,534,126 vs 7,530,757), of which the number of deaths of males would be reduced by 0.05% (5,268,115 vs 5,265,513) and of females by 0.03% (2,266,011 vs 2,265,244). In contrast, if the condom usage rate reached 100% starting in 2019, the number of HBV-related deaths of people aged 0–100 years from 2019 to 2035 would be reduced by 0.08% (7,534,126 vs 7,528,113), of which the number of deaths of males would be reduced by 0.09% (5,268,115 vs 5,263,475) and of females by 0.06% (2,266,011 vs 2,264,638) (see Fig. 5a,b,c).

In addition, if the condom usage rate reached 100% starting in 2019, then from 2019 to 2035, in the 50–59 age group, the number of HBV-related deaths in males would reduce the most. In the 60–69 age group, the number of HBV-related deaths in females would reduce the most (see Fig. 5d).

As shown in Figures S3, S4, the sex-specific annual rate of HBsAg seroclearance would fluctuate only within a small range around the base-case result with the change of any one of the parameters between its upper and lower limit values. Among these parameters, the horizontal transmission rate had some impact on the annual rate of HBsAg seroclearance, the second parameter was the perinatal infection rate.

Moreover, from Table S3, we can see that the horizontal transmission rate had some impact on the number of chronic HBV infections in 2035. The vaccination protection rate had some impact on the number of acute HBV infections in males and females in 2035. However, the impact on the number of acute HBV infections in males and females in 2035 is small. If the vaccination protection rate took the upper limit value 0.95, then the number of acute HBV infections in males and females in 2035 would decrease by 31.31% (307571vs 211,263) and 31.02% (326,930 vs 225,501), respectively (see Table S3).