Potential conflict of interest: Nothing to report.
Spontaneous hepatitis B surface antigen (HBsAg) seroclearance in chronic HBV infection has long been suggested as a rare event in high endemic areas. The prevalence of HBsAg in the general population of Taiwan, however, decreased remarkably from 15%-20% before age 40 to 5%-10% after age 60 or 70. This study aimed to reexamine the rates of HBsAg seroclearance by a long-term follow-up of 1965 hepatitis B e antibody–positive asymptomatic adult carriers. Of these, 1076 (55%) were males, the mean (± SD) age was 35.6 ± 9.2 years and the mean follow-up was 10.8 ± 5.4 years. Hepatitis relapsed in 314 patients, 0.5 to 18 (mean ± SD = 5.8 ± 4.4) years after the entry. The probability of hepatitis relapse correlated positively with male sex (P < 0.0001) and age at entry (P < 0.0001). Serum HBsAg cleared in 245 patients at the mean age of 47.8 ± 9.6 years. The cumulative probabilities of HBsAg seroclearance were 8.1% after 10 years, but increased disproportionally to 24.9% and 44.7%, respectively, after 20 and 25 years. In multivariate analysis, the probability of HBsAg seroclearance correlated positively with age at entry (P < 0.0001) and sustained remission of hepatitis (P < 0.0001) and marginally significantly with male sex (P = 0.053). Conclusion: Cumulative rate of HBsAg seroclearance in asymptomatic adult carriers from high endemic areas was approximately 40% after 25 years of follow-up. The low HBsAg seroclearance rates in previous studies might be due to the relative short period of follow-up. (HEPATOLOGY 2007;45:1187–1192.)
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
The epidemiological and clinical features of HBV infection varied considerably between high and low HBV endemic areas. In high endemic areas such as Southeast Asia and Taiwan, HBV infection occurred predominantly in the perinatal period or early childhood with high rate of persistent infection. The natural history of chronic HBV infection consists of 3 phases:1, 2 an initial immune tolerant phase (hepatitis B e antigen [HbeAg] –positive and normal ALT levels), followed by immune clearance phase (HBeAg–positive and abnormal ALT levels), and finally the low replication phase (hepatitis B e antibody [anti-HBe] –positive and normal ALT levels). Patients in the third phase are termed “inactive carriers”. Serum hepatitis B surface antigen (HBsAg) may become negative with time in inactive carriers. HBsAg seroclearance usually confers favorable outcome if there is no preexisted cirrhosis or viral superinfection,3–6 though adverse complications still can occur.7 HBsAg seroclearance in chronic HBV infection has long been recognized as a much rarer event in high endemic areas than in low endemic areas: the reported annular incidence of HBsAg seroclearance was 0.1%-0.8% in the former8, 9 and 1.0%-2.1% in the latter.10–14
Epidemiological studies on age-specific prevalence of HBsAg in the general population of Taiwan have shown that the prevalence of HBsAg decreased by 40% to 70% from age 40 to age 60 or 70 (15.4% to 6%, 17.1% to 9.7%, 22.7% to 6.9%).15–17 In addition to the possible birth cohort effect, this decrease may be due to spontaneous HBsAg seroclearance or a decrease in the number of HBsAg carriers through death of HBV-related liver disease. However, the annular incidence of cirrhosis and hepatocellular carcinoma in asymptomatic HBsAg carriers from Taiwan was estimated to be only 0.7% and 0.1%, respectively.18 Notably, in previous studies on the rates of HBsAg seroclearance in chronic HBV infection from Taiwan, the mean age at entry of the study patients was 32 to 33 years and the mean follow-up duration was only 2.5 to 4 years.8, 9 As the decrease in prevalence of HBsAg became remarkable after age 40 or 50,15–17 it seems likely that these studies may have underestimated the rates of spontaneous HBsAg seroclearance. Further long-term follow-up studies are indicated to delineate the incidence of spontaneous HBsAg seroclearance in chronic HBV infection in high endemic areas.
In this study, we report the incidence and determinants of spontaneous HBsAg seroclearance in 1965 anti-HBe-positive asymptomatic adult HBsAg carriers who were prospectively followed up for 3 to 26 (mean 10.8) years in Taiwan.
anti-HBe, antibody against hepatitis B e antigen; anti-HCV, antibodies against HCV; anti-HDV, antibody against HDV; HBeAg, hepatitis B e antigen; HBsAg, hepatitis B surface antigen.
Patients and Methods
Incidentally identified asymptomatic adult HBsAg carriers who visited the hepatitis B carrier clinic of Chang Gung Memorial Hospital at Taipei were registered from 1980. The clinical and laboratory data at baseline and during follow-up were recorded. Patients were recruited into this study if they fulfilled the following criteria: (1) HBsAg positive for at least 1 year; (2) HBeAg negative, anti-HBe–positive, normal ALT, no evidence of cirrhosis based on the clinical ground and liver ultrasonography19 and no concomitant hepatitis C or D virus infection at baseline; (3) no antiviral or immunomodulatory therapy before entry and during follow-up; (4) regular follow-up at least every year. Patients who had alcohol or drugs that might be possible etiological agents of hepatitis were excluded. No patient in this cohort admitted intravenous drug abuse or homosexuality.
Clinical Data at Baseline and During Follow-Up.
At entry, the patients were tested for liver biochemical tests, virological markers (HBsAg, HBeAg, anti-HBe, antibodies against HCV [anti-HCV], and antibody against HDV [anti-HDV]), α-fetoprotein, and real-time liver ultrasonography. Liver biochemical tests were was assayed every 6 month if ALT levels were normal, and every 1-3 months or more often if ALT levels were elevated. HBsAg, HBeAg, anti-HBe, and α-fetoprotein were assayed every 6 months or more often if clinically indicated. Assays for anti-HCV and anti-HDV were repeated if ALT levels were elevated to more than twice the normal upper limit. Serum HBV DNA was not routine assayed but was measured when ALT levels were elevated to more than twice the normal upper limit.
HBsAg, HBeAg, anti-HBe, and anti-HDV were assayed by radioimmunoassay (Abbott Diagnostics, North Chicago, IL). Anti-HCV was tested by a second- or third-generation enzyme immunoassay (Abbott Diagnostics). Hepatitis B virus DNA was tested by sandwich molecular hybridization assays using a Digene Hybrid Capture System (Digene Diagnostics Inc., Beltsville, MD). The detection sensitivity was 0.5 pg/ml.
Reversion of HBeAg was defined as reverse seroconversion from anti-HBe to HBeAg. Relapse of hepatitis was diagnosed when ALT levels elevated to more than twice the normal upper limit, accompanied by positive serum HBV DNA. HBsAg seroclearance was defined as persistent loss of serum HBsAg for at least 1 year and until last visit.
To compare characteristics between groups, either chi-squared test or Fisher exact test was used for analysis of categorical variables and Student t test or Wilcoxon nonparametric test was used for analysis of continuous variables. Estimates on the rate of hepatitis relapse or HBsAg seroclearance were calculated by the actuarial analysis method, and the difference was determined by the log-rank test. For each patient, there is a date of entry (the age [in year] when the study patients were first recognized as anti-HBe–positive HBsAg carriers with normal ALT) and a termination or censored date (the age [in year] when an event of interest or censoring occurred). Univariate and multivariate analyses were performed to identify factors associated with HBsAg seroclearance. Variables found to be significant in the univariate models were tested in a multivariate setting using the Cox proportional hazards regression models. Statistical procedures were performed with SPSS statistical software (version 13.0; SPSS, Chicago, IL). P values < 0.05 were considered significant.
A consecutive series of 1965 anti-HBe–positive HBsAg carrier with a minimal follow-up of 3 years were enrolled in this study. Their baseline and follow-up data are summarized in Table 1. The mean (± SD) age at entry was 35.6 ± 9.2 years (median, 34; range, 16 to 76), and 1076 (55%) were men.
Table 1. Baseline and Follow-Up Characteristics of Study Patients
During 20,298 person-years of follow-up, relapse of hepatitis occurred in 314 patients (236 males and 78 females), 0.5 to 18 (mean ± SD; 5.8 ± 4.4) years after the entry. Of them, only 5 patients had reversion of HBeAg. The calculated rate of hepatitis relapse was 1.55% per year. The interval between entry and hepatitis relapse showed no significant difference between males and females (mean ± SD; 5.7 ± 4.4 years versus 6.2 ± 4.6 years, P = 0.39), but the age at the onset of hepatitis relapse tended to be younger in males than in females (mean ± SD; 42.5 ± 9.5 years versus 45.0 ± 10.8 years, P = 0.046). The cumulative probabilities of hepatitis relapse were 9.7%, 17.1%, 19.6%, and 21.5%, respectively, after 5, 10, 15, and 20 years. Multivariate analyses using the Cox proportional hazards regression models showed that the probability of hepatitis relapse correlated significantly with male sex (hazard ratio [95%CI] = 2.55 [2.03-3.20], P < 0.0001) and age at entry (hazard ratio [95% CI] = 1.22 [1.09-1.37] per decade increase, P < 0.0001).
Spontaneous HBsAg seroclearance occurred in 245 patients during 21,267 person-years of follow-up. The calculated rate of HBsAg seroclearance was 1.15% per year. The person-years of follow-up as well as HBsAg seroclearance events in each age group is illustrated in Table 2.
Table 2. Calculated Rates of HBsAg Seroclearance in 1965 Anti-HBe-Positive Asymptomatic HBsAg Carriers
Age at entry
Person-years of follow-up
No. of patients with HBsAg seroclearance
Annual rate of HBsAg seroclearance
The mean (± SD) and median (interquartile range) ages of HBsAg seroclearance was 47.8 ± 9.6 and 48 (41-53) years, respectively. The peak age distribution was at 50 to 54 years and 79% (194/245) of HBsAg seroclearance occurred after age 40 years.
The cumulative probabilities of HBsAg seroclearance were 8.1%, 24.9%, and 44.7%, respectively, after 10, 20, and 25 years (Fig. 1). Assuming most patients in this cohort were infected at or soon after birth, then the probability of HBsAg seroclearance by age can be calculated, using 0 as the age at entry and the age when HBsAg seroclearance or censoring occurred as the termination or censoring date. The cumulative probability of HBsAg seroclearance increased substantially after age 50 (Fig. 2).
Multivariate analysis using the Cox proportional hazards regression models including relapse of hepatitis during follow-up as a time-dependent covariate showed that HBsAg seroclearance correlated significantly with age at entry (hazard ratio [95% CI] = 1.58 [1.39-1.79] per decade increase, P < 0.0001), sustained remission of hepatitis during follow-up (hazard ratio [95% CI] = 2.16 [1.42-3.29] versus those with relapse of hepatitis, P < 0.0001), and marginally significantly with male sex (hazard ratio [95% CI] = 1.30 [0.99-1.70], P = 0.053). The cumulative probability of HBsAg seroclearance in relation to the age at entry is shown in Fig. 3 and in relation to gender and sustained remission or relapse of hepatitis is shown in Fig. 4. HBsAg seroclearance was significantly more likely to occur in patients who were older at time of study entry and in those with sustained remission of hepatitis than in those who had relapse of hepatitis. In addition, HBsAg seroclearance tended to more likely occur in males than in females.
Among 314 patients with relapse of hepatitis, 24 patients underwent subsequent HBsAg seroclearance 3 to 20 (mean ± SD; 9.6 ± 4.6) years after the onset of hepatitis relapse. No patient had relapse of hepatitis after HBsAg seroclearance.
A total of 1241 patients were available for follow-up through 2005, and the remaining 724 patients were lost to follow-up before 2004, mostly before 2000. There was no significant difference in age at entry (mean ± SD; 35.6 ± 9.1 years versus 35.6 ± 9.4 years, P = 0.97) and gender [male ratio; 53% (661/1241) versus 57% (415/724), P = 0.16] between them. The follow-up duration was significantly longer (P < 0.0001) for the former (mean ± SD and median [interquartile range]; 12.3 ± 5.5 and 12 [8-17] years) than for the latter (8.3 ± 4.1 and 8 [5-11] years). During follow-up, 211 (17.0%) and 181 patients (14.6%), respectively, of the former had relapse of hepatitis and HBsAg seroclearance, as did 103 (14.2%) and 64 (8.8%) of the latter. The cumulative probability of hepatitis relapse after 5, 10, and 15 years were 10.2%, 17.4%, and 19.2%, respectively, for the former, and were 8.3%, 16.3%, and 23.1%, respectively, for the latter (P = 0.54). The corresponding figures of HBsAg seroclearance were 2.5%, 8.2%, and 15.6% for the former, and 1.8%, 8.1%, and 19.9% for the latter (P = 0.28). Among patients who were available for follow-up through 2005, 40 had liver cirrhosis based on the clinical ground and liver ultrasonography19, and 4 had HCC proved by image and histology/cytology until the last visit. Among patients who were lost to follow-up before 2004, data on ultrasonographic evidence of cirrhosis were incomplete as it was used mainly for screening of hepatocellular carcinoma in the early period, but none had ascites, hepatic encephalopathy or varices bleeding and one had suspected hepatocellular carcinoma by image study when last seen.
This study examined the rates of spontaneous HBsAg seroclearance in anti-HBe–positive asymptomatic HBsAg carriers who visited our hepatitis B carrier clinic from 1980. In keeping with previous observations in Taiwan where most HBV infections are acquired in the perinatal period or early childhood with high rate of persistent infection irrespective of sex,20 there is no sex dominance in this cohort of asymptomatic HBsAg carriers. The mean age at entry of this cohort was 35.6 years, whereas previous studies revealed that the mean age of HBeAg seroconversion for HBeAg-positive HBsAg carriers was 31-33 years.21–24 These data suggested that HBsAg carriers of this cohort were enrolled into this study 2-4 years after HBeAg seroconversion.
This study showed that cumulative probability of HBsAg seroclearance in anti-HBe–positive asymptomatic HBsAg carriers was approximately 40% after 25 years of follow-up (Fig. 1). These findings may suggest that HBsAg seroclearance is not as extremely rare as previously suggested in high HBV endemic areas.8, 9 Further analysis revealed the cumulative probability of HBsAg seroclearance was 8.1% after the first 10 years, a figure similar to those reported previously.8 However, HBsAg seroclearance rates increased disproportionally 10 years thereafter compared to the first 10 years: the cumulative probability of HBsAg seroclearance increased to 24.9% after 20 years (16.8% increase from 11th to 20th year of follow-up) and 44.7% after 25 years (19.8% increase from 21st to 25th year of follow-up).
The enhanced HBsAg seroclearance beyond 15 to 20 years of follow-up, though it might be attributed to the relatively few patients who were left in the study at this time, can be explained by the age of HBsAg seroclearance. The mean (± SD) age of spontaneous HBsAg seroclearance in this cohort was 47.8 ± 9.6 years. This figure is similar to that reported previously in Hong Kong (48.8 ± 13.8 years), Korea (mean, 50 [range, 27-72] years), and Japan (mean, 51 [range, 23-66] years).4–6 Detailed analysis of our data showed that the peak age distribution of HBsAg seroclearance was at 50-54 years and that 79% of HBsAg seroclearance occurred after age 40 years. These data suggest that, among HBsAg carriers with age at entry around 30-35, as reported in the previous series8, 9 and in the current series, HBsAg seroclearance should be relatively uncommon in the first 10 years of follow-up and then should tend to occur much more frequently 10 years thereafter. Previous studies from Taiwan should have much underestimated the rates of HBsAg seroclearance, because the follow-up period was inadequate (mean follow-up 2.5-4 years and follow-up duration <8 years in >90% of patients7, 8).
Interestingly, if we assumed that most patients in this cohort were infected at or soon after birth, the cumulative probability of HBsAg seroclearance increased substantially after age 50 (Fig. 2). Although this analysis is limited by the selection bias in the age at entry of study patients and by the relatively few patients who had long-term follow-up, these data might agree with the epidemiological features of chronic HBV infection in Taiwan that the prevalence rate of HBsAg in the general population decreased remarkably after age 5015–17 (Fig. 5).
Several factors independently associated with HBsAg seroclearance were identified in this study. In agreement with the results of previous studies from Taiwan and Alaska,9, 12 age at entry was the most significant factor associated with HBsAg seroclearance (Fig. 3). In the Alaska study, the cumulative probability of HBsAg seroclearance after 20 years was only around 6% for carriers with age at entry < 20 years but around 20% for those with age at entry ≥ 20 years.12 In high HBV endemic areas, age of entry presumably reflected duration of HBV infection. This study also showed that HBsAg carriers with sustained remission of liver disease were 2.2 times more likely to clear serum HBsAg than those who had relapse of hepatitis during follow-up. It is conceivable that levels of viral replication and serum titers of HBsAg tended to decrease with time in HBsAg carriers with sustained remission of hepatitis but less in those with relapse of hepatitis. Perhaps one limitation of this study is that HBV DNA was assayed by molecular hybridization assays, which may miss some patients with relapse of hepatitis. With regard to viral factors possibly associated with HBsAg seroclearance in chronic HBV infection, it has been suggested that HBsAg carriers with genotype B infection were more likely to clear HBsAg than those with genotype C infection, but the reason remained unclear.4 Unfortunately, data on HBV genotype were not available in this study, because patients in this cohort were entered into follow-up before the significance of HBV genotype had been explored. However, it has been shown that relapse of hepatitis was significantly more frequent in genotype C infection than in genotype B infection.24, 25 Our findings might suggest that the difference in HBsAg seroclearance with HBV genotype is possibly related to the genotype difference in the frequency of relapse of hepatitis. Further studies are indicated to clarify whether genotype is an independent factor for HBsAg seroclearance after correction of other confounding factors such as relapse of hepatitis. Perhaps another interesting factor is that male carriers were more likely to clear HBsAg than females. In keeping with the previous observations,21 male carriers were much more likely to have relapse of hepatitis than females. However, male carriers were 1.3 times more likely to clear HBsAg than female carriers after correction of the confounding factors such as age at entry and relapse of hepatitis during follow-up (P = 0.053). Whether this gender difference is related to genetic or hormonal factors remains unknown.
In conclusion, this study reveals that HBsAg seroclearance in chronic HBV infection is not as extremely rare as previously suggested in high HBV endemic areas. The cumulative probability of HBsAg seroclearance in anti-HBe–positive asymptomatic adult carriers is approximately 40% after 25 years of follow-up. Age at entry, sustained remission of liver disease and possibly male sex are factors independently associated with HBsAg seroclearance.