Universal infant immunization and occult hepatitis B virus infection in children and adolescents: A population-based study
This work was supported by Grant NSC 99-2314-B-002-010-MY3 and Grant NSC 102-2628-B002-026-MY3 from the National Science Council, Taiwan, and by the Liver Disease Prevention and Treatment Research Foundation, Taiwan.
Potential conflict of interest: Nothing to report.
Abstract
To determine whether universal infant immunization affects occult hepatitis B virus (HBV) infection (OBI), serum samples from hepatitis B surface antigen (HBsAg)-negative subjects <18 years enrolled during six sequential seroepidemiological surveys conducted between 1984 (just before universal infant immunization) and 2009 were analyzed. Study subjects were divided into unvaccinated cohorts (born before 1984) and vaccinated cohorts (born after 1984). HBV-DNA positivity was determined by positivity of nested polymerase chain reaction in at least two of three regions (pre-S, S, and pre-core/core genes). OBI frequency was lower in vaccinated than unvaccinated antibody to hepatitis B core antigen (anti-HBc)-negative subjects (0 of 392 [0%] vs. 4 of 218 [1.8%]; P = 0.007), tended to be higher in vaccinated than unvaccinated anti-HBc-positive subjects (16 of 334 [4.8%] vs. 3 of 181 [1.7%]; P = 0.072), and was higher in vaccinated than unvaccinated subjects seropositive for both antibody to hepatitis B surface antigen (anti-HBs) and anti-HBc (13 of 233 [5.6%] vs. 3 of 170 [1.8%]; P = 0.025). By using known anti-HBc seropositivity rate in children in our serosurveys, the estimated OBI frequency per 104 HBsAg-negative subjects declined from 160.7 in unvaccinated cohorts to 11.5 in vaccinated cohorts. In vaccinated cohorts, OBI frequency was higher in anti-HBc-positive subjects than in anti-HBc-negative subjects (16 of 334 [4.8%] vs. 0 of 392 [0%]; P < 0.001). Subjects with OBI had much lower viral load (P < 0.001) and a trend of higher mutation rates in “a” determinant of HBsAg than age-comparable, HBsAg-positive subjects. Conclusions: Reduction of OBI in immunized subjects complements the well-documented universal infant immunization-related benefit of markedly reduced overt HBV infection. Breakthrough infections in immunized subjects seem to associate with more occurrence of OBI than natural infections in unvaccinated subjects. In the postvaccination era, anti-HBc seropositivity is a useful marker for OBI screening in HBsAg-negative subjects, and a very-low-level viral replication and HBsAg expression is the major mechanism underlying OBI. (Hepatology 2015;61:1183–1191)
Abbreviations
-
- aa
-
- amino acids
-
- cpm
-
- counts per minute
-
- ALT
-
- alanine aminotransferase
-
- BCP
-
- basal core promoter
-
- EIA
-
- enzyme immunoassay
-
- anti-HBc
-
- antibody to hepatitis B core antigen
-
- anti-HBs
-
- antibody to hepatitis B surface antigen
-
- HBeAg
-
- hepatitis B e antigen
-
- HBIG
-
- hepatitis B immune globulin
-
- HBsAg
-
- hepatitis B surface antigen
-
- HBV
-
- hepatitis B virus
-
- HCC
-
- hepatocellular carcinoma
-
- OBI
-
- occult HBV infection
-
- PCR
-
- polymerase chain reaction
-
- RIA
-
- radioimmunoassay
-
- UII
-
- universal infant immunization
In areas highly endemic for hepatitis B virus (HBV) infection, primary HBV infection occurs mainly during infancy or early childhood.1, 2 Perinatal transmission from highly infectious mothers to their infants is the most important route of transmission, which leads to high rate of chronicity.3 Universal infant hepatitis B immunization worldwide has been proven to be a powerful strategy to combat HBV-associated diseases, and in Taiwan, has led to marked decrease in the incidence of acute, fulminant, and chronic hepatitis B (reduced hepatitis B surface antigen [HBsAg]-positive chronic carrier rate from 9.8% in 1984 to 0.5% in 2009) and of hepatocellular carcinoma (HCC).4, 5 However, there are still problems that hamper the success of global control and the ultimate eradication of HBV infection. For example, by using highly sensitive polymerase chain reaction (PCR) methods, some earlier studies have demonstrated the presence of HBV DNA in vaccinated children without detectable HBsAg.6, 7 This is so-called occult HBV infection (OBI), which has been defined as the presence of HBV DNA in livers or sera in subjects who are serologically negative for HBsAg.8
Although its clinical significance remains incompletely understood, subjects with OBI may retain the HBV genome indefinitely in previously infected liver cells, are susceptible to HBV reactivation (if the person becomes immunodeficient), and may have an increased risk for cirrhosis and HCC.9 Subjects with OBI may also create problems for the safety of blood or organ donation.10
The world's first universal hepatitis B vaccination program was launched in Taiwan in July 1984.11 From 1984 (just before universal immunization) to 2009, we have conducted six quinquennial seroepidemiological surveys in children and adolescents in Taipei.2, 4 By analyzing serum samples from these surveyed subjects seronegative for HBsAg, we studied the prevalence of OBI and genetic characteristics of the HBV isolates in these apparently healthy children and adolescent population with or without receiving infant hepatitis B immunization and aimed to elucidate the real influence of universal infant immunization on the emergence of OBI. We also analyzed the mutation patterns in children with OBI and HBsAg-positive carrier children to clarify the major mechanisms related to OBI.
Subjects and Methods
Universal Infant Immunization Program, HBV Seroepidemiological Studies, and Serological Testing for HBV Infection
A nation-wide vaccination program against HBV was launched in Taiwan in July 1984 to immunize the infants of HBsAg-carrier mothers.11 For neonates of mothers with positive hepatitis B e antigen (HBeAg) or high titers of HBsAg, hepatitis B immune globulin (HBIG) was also given within 24 hours of life. The program was extended to all newborns since July 1986. Six sequential HBV seroepidemiological studies were performed in Taipei City in 1984 (just before the start of the vaccination program), 1989, 1994, 1999, 2004, and 2009.2, 4
We recruited participants for the baseline and follow-up seroepidemiological studies in Cheng-Chung District, Taipei City. Participants included 1,200 children (<15 years of age) in 1984, 1,134 children in 1989, 1,515 children in 1994, 1,916 subjects (≤20 years) in 1999, 18,779 subjects (<30 years) in 2004, and 3,332 subjects (<30 years) in the 2009 survey. All of the above participants were voluntarily recruited through poster advertisements or health staff invitations. They consisted of children <3 years enrolled from the well-baby clinic and day care centers, children ages 3-6 years enrolled from preschools and kindergartens, children ages 7-12 years enrolled from public elementary school, children ages 12-15 years enrolled from public junior high school, adolescents ages 15-18 years enrolled from high school, college students ages 18-22 years enrolled from universities, and subjects ages 22-29 years enrolled from graduate schools, new employees of general hospitals, and local commercial companies.
In 1984, 1989 and 1994 surveys, serum HBsAg, antibody to hepatitis B surface antigen (anti-HBs), and antibody to hepatitis B core antigen (anti-HBc) were tested by radioimmunoassay (RIA) using Ausria II, Ausab, and Corab (Abbott Laboratories, North Chicago, IL). HBsAg and anti-HBs were considered positive when the value ≥2.1 S/N (the ratio of the counts in the sample to the count for a negative control). Sensitivity and specificity of the HBsAg assay were 0.6 ng/mL and >99%, respectively. Anti-HBc levels were measured as percent inhibition calculated using the following formula: 100 − (Test counts per minute [cpm]/Negative control cpm). Serum samples demonstrating inhibition ≥50% were considered positive.
In 1999, 2004, and 2009 surveys, serum HBsAg, anti-HBs, and anti-HBc were tested by using a microparticle enzyme immunoassay (AxSYM Systems; Abbott Laboratories). HBsAg was considered positive when the value was >2.0 S/N. Sensitivity of the HBsAg assay was 0.1-0.6 ng/mL. The value of anti-HBc was considered positive if S/CO (sample rate/cut-off rate) was <1.0 and negative if the value was >1.0. Sensitivity and specificity of the anti-HBc assay was 1 PEI U/mL and >99%, respectively. For anti-HBs, serum titers >10 mIU/mL were considered positive.
After assaying for serum HBV markers, serum samples were stored in −70oC until analyzed. Written informed consent to collect serum samples was obtained from the subjects themselves or their guardians. The study protocol was approved by the institutional review board of the National Taiwan University Hospital (Taipei, Taiwan).
Study Subjects
Subjects enrolled in the above six serosurveys who were HBsAg negative and <18 years of age were considered for entry into this study.
For recruiting anti-HBc-positive subjects (group 1), all subjects who were anti-HBs (−), but anti-HBc (+), and subjects who were anti-HBs (+) and anti-HBc (+), if they had sufficient serum samples for HBV-DNA testing, were included.
For recruiting anti-HBc-negative subjects (group 2), those who were both anti-HBs (−) and anti-HBc (−), and those who were anti-HBs (+), but anti-HBc (−), were included. Because such serum samples were available in five surveys (1989-2009) at the time of this study, and that numerous subjects in each survey possessed these marker profiles, appropriate selection of study subjects was performed: (1) For recruiting unvaccinated cohorts (subjects born before 1984), only 5-7 subjects in each age group were randomly selected from those ages 5-14 years enrolled in the 1989 survey, those ages 10-14 years enrolled in the 1994 survey, and those ages 15-17 years enrolled in the 1999 survey; (2) for recruiting vaccinated cohorts (subjects born after 1984), only 4-7 subjects in each age groups (<18 years) were randomly selected from the 2004 and 2009 surveys.
-
1. Subjects with positive anti-HBc (group 1)
-
a. Unvaccinated cohorts (group 1a)
-
i. Subjects who were anti-HBc (+) and anti-HBs (−): 11 (37.9%) of 29
- ii.Subjects who were anti-HBc (+) and anti-HBs (+): 170 (51.5%) of 330
-
-
b. Vaccinated cohorts (group 1b)
-
i. Subjects who were anti-HBc (+) and anti-HBs (−): 101(93.5%) of 108
- ii. Subjects who were anti-HBc (+) and anti-HBs (+): 233(54.1%) of 431
-
-
-
2. Subjects with negative anti-HBc (group 2)
-
a. Unvaccinated cohorts (group 2a)
-
i. Subjects who were anti-HBc (−) and anti-HBs (−): 112 (19.4%) of 576
- ii. Subjects who were anti-HBc (−) and anti-HBs (+): 106 (49.8%) of 213
-
-
b. Vaccinated cohorts (group 2b)
-
i. Subjects who were anti-HBc (−) and anti-HBs (−): 165 (2.0%) of 8,442
- ii. Subjects who were anti-HBc (−) and anti-HBs (+): 227(3.4%) of 6,709
-
-
There was no statistical difference in age, sex, or district of residence between group 2a and 2b subjects (see the Supporting Appendix), suggesting that these children selected for PCR represented unbiased samples that could reflect the status of those with negative anti-HBc.
Control Group for Subjects With OBI
For comparison with 19 anti-HBc-positive subjects with OBI, serum samples from another 28 HBsAg-positive, anti-HBc-positive children recruited from each of the corresponding seroepidemiological surveys served as control groups. Their age, sex ratio, HBV genotypes, and the proportion distributed in each survey were comparable to those subjects with OBI.
Amplification and Direct Sequencing of the Pre-S, Surface, and Pre-Core/Core Genes
HBV DNA was extracted from 200 uL of serum using the QIAamp DNA Blood mini kit (Qiagen, Hilden, Germany). Pre-S, S, and pre-core/core genes were amplified by using nested PCR. Extracted HBV DNA was subjected to PCR using outer and inner primer pairs shown in Supporting Table 1. PCR details are described in the Supporting Appendix. PCR results were read after gel electrophoresis. Strict precautions were taken to avoid possible contamination.12 Only the data that revealed no false-positive results in the negative controls and that were reproducible were used. The PCR products were then directly sequenced by use of an automatic ABI DNA Sequencer (Model 377A; Applied Biosystems, Foster City, CA) to determine the nucleotide sequences.
| 1984 | 1989 | 1994 | 1999 | 2004 | 2009 | 1984∼2009 | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| HBsAg | – | – | – | – | – | – | – | – | – | – | – | – | – | – | |
| Anti-HBs | – | + | – | + | – | + | – | + | – | + | – | + | – | + | |
| Anti-HBc | + | + | + | + | + | + | + | + | + | + | + | + | + | + | Total |
| Unvaccinated cohorts (birth before 1984) | |||||||||||||||
| No. of children found during the survey | 14 | 210 | 14 | 76 | 0 | 18 | 1 | 26 | 29 | 330 | 359 | ||||
| No. of children analyzed for serum HBV DNA | 8 | 92 | 2 | 35 | 0 | 17 | 1 | 26 | – | – | – | – | 11 | 170 | 181 |
| No. of occult HBV infection (%) | 0 | 2 | 0 | 1 | 0 | 0 | 0 | 0 | – | – | – | – | 0 | 3 | 3 |
| (0) | (2.2) | (0) | (2.9) | (0) | (0) | (0) | (0) | (0)a | (1.8)b | (1.7)c | |||||
| Vaccinated cohorts (birth after 1984) | |||||||||||||||
| No. of children found during the survey | 1 | 25 | 2 | 30 | 20 | 54 | 59 | 203 | 26 | 119 | 108 | 431 | 539 | ||
| No. of children analyzed for serum HBV DNA | – | – | 1 | 14 | 2 | 28 | 17 | 31 | 58 | 102 | 23 | 58 | 101 | 233 | 334 |
| No. of occult HBV infection (%) | – | – | 0 | 1 | 0 | 2 | 1 | 2 | 0 | 3 | 2 | 5 | 3 | 13 | 16 |
| (0) | (7.1) | (0) | (7.1) | (5.9) | (6.5) | (0) | (3) | (8.7) | (8.6) | (3)a | (5.6)b | (4.8)c | |||
- a P = 1.00, by Fisher's exact test.
- b P = 0.025, by Fisher's exact test.
- c P = 0.072, by Fisher's exact test.
In this study, HBsAg-negative children who were HBV-DNA positive in at least two of three (pre-S, S and pre-core/core) regions by nested PCR were considered as OBI.
Quantification and Genotyping of HBV DNA
Serum HBV DNA was extracted and quantified by a real-time PCR amplification assay using a LightCycler (Roche Diagnostics, Basel, Switzerland) with a detection sensitivity of 5 × 102 copies/mL of HBV in serum.13 HBV genotype in patients was determined at the same time.13
Reanalysis of HBsAg in OBI Children With Architect Assay
The Architect i1000 analyzer (Abbott Diagnostics) uses a chemiluminescent microparticle immunoassay. Positive cut-off value for HBsAg and anti-HBs are ≥0.05 and ≥10 mIU/mL, respectively. HBeAg, anti-HBe, and anti-HBc are interpreted using the ratio of the sample's rate to the cut-off rate (S/CO), where S/CO ≥1.00 indicates HBeAg or anti-HBc positive, whereas S/CO ≤1.00 indicates anti-HBe positive.
Statistical Analysis
The chi-squared (χ2) test with or without Yate's correction, Fisher's exact test, and contingency tables with Yate's correction were used for statistical comparison. Mann-Whitney's U test was used to compare continuous variables. A P value <0.05 was considered statistically significant and a P value between 0.1 and 0.05 as showing a trend.
Results
OBI in HBsAg-Negative, Anti-HBc-Positive Subjects (Group 1)
In all 112 tested children seropositive for anti-HBc alone, 3 (2.7%; 1 in 1999 and 2 in 2009) had OBI. In 403 tested children seropositive for “both anti-HBs and anti-HBc only,” 16 (4.0%; 2 in 1984, 2 in 1989, 2 in 1994, 2 in 1999, 3 in 2004, and 5 in 2009) had OBI (Table 1). Prevalence of OBI was not different between the above two groups (P = 0.521).
Influence of Universal Immunization on the Prevalence of OBI in HBsAg-Negative, Anti-HBc-Positive Subjects
In subjects seropositive for both anti-HBs and anti-HBc, OBI frequency was higher in those born after 1984 than in those born before 1984 (13 of 233 [5.6%] vs. 3 of 170 [1.8%]; P = 0.025). Such difference was not observed in subjects seropositive for anti-HBc alone. Taken together, in HBsAg-negative subjects seropositive for anti-HBc, OBI frequency was 1.7% (3 of 181) in those born before 1984 and was 4.8% (16 of 334) in those born after 1984 (P = 0.072; Table 1).
OBI in HBsAg-Negative, Anti-HBc-Negative Subjects (Group 2)
In subjects born before 1984, 1 of 112 who were anti-HBs (−) and anti-HBc (−) and 3 of 106 who were anti-HBs (+) and anti-HBc (−) had OBI.
In subjects born after 1984, 0 of 227 subjects who were anti-HBs (+), but anti-HBc (−), and 0 of 165 subjects who were both anti-HBs (−) and anti-HBc (−) had OBI.
Taken together, in HBsAg-negative, anti-HBc-negative subjects, OBI frequency was 1.8% (4 of 218) in those born before 1984 and was 0% (0 of 392) in those born after 1984 (P = 0.007; Table 2)
| HBV Marker Status | ||
|---|---|---|
| HBsAg (–), Anti-HBs (+/–), Anti-HBc (–) | HBsAg (–), Anti-HBs (+/–), Anti-HBc (+) | |
| Unvaccinated (birth before 1984) | ||
| No. of children analyzed for serum HBV DNA | 218ac | 181bc |
| No. of occult HBV infection (%) | 4 (1.8)ac | 3 (1.7)bc |
| Vaccinated (birth after 1984) | ||
| No. of children analyzed for serum HBV DNA | 392ad | 334bd |
| No. of occult HBV infection (%) | 0 (0)ad | 16 (4.8)bd |
- a P = 0.007, by Fisher's exact test.
- b P = 0.072, by Fisher's exact test.
- c P = 0.89, by Fisher's exact test.
- d P < 0.001, by Fisher exact test.
Role of Anti-HBc Seropositivity in Screening OBI in Postvaccination Era
The prevalence of OBI in HBsAg-negative subjects born after 1984 was significantly higher in those seropositive for anti-HBc than in those seronegative for anti-HBc (16 of 334 [4.8%] vs. 0 of 392 [0%]; P < 0.001; Table 2).
Estimation of the Prevalence of OBI in HBsAg-Negative Subjects With or Without Receiving Infant Hepatitis B Immunization
Our previous studies showed that the HBsAg-positive rate was markedly declined from 9.8% in the 1984 survey to 0.5% in the 2009 survey and that the anti-HBc seropositivity rate decreased from 26.2% in the 1984 survey to 2.9% in the 2009 survey.2, 4 These data were adopted because the latter two are the best representative data for the rates of anti-HBc seropositivity in a child population without infant HBV immunization and that fully covered by infant HBV immunization at birth, respectively. Namely, the rate of HBsAg-negative, but anti-HBc-positive, subjects decreased from 16.4% in 1984 to 2.4% in 2009; and the rate of both HBsAg-negative and anti-HBc-negative subjects increased from 73.8% in 1984 to 97.1% in 2009. Thus, the estimated frequency of OBI per 10,000 HBsAg-negative children declined from 160.7 in unvaccinated cohorts to 11.5 in vaccinated cohorts (Table 3).
| Unvaccinated Cohorts | Vaccinated Cohorts | |
|---|---|---|
| Prevalence rate of specific HBV marker profile in total enrolled children population in previous serosurveys, % | ||
| HBsAg (+) rate* | 9.8 | 0.5 |
| Anti-HBc (+) rate† | 26.2 | 2.9 |
| HBsAg (–) but anti-HBc (+) rate‡ | 16.4 | 2.4 |
| HBsAg (–) but anti-HBc (–) rate§ | 73.8 | 97.1 |
| Prevalence rate of OBI in children with or without anti-HBc positivity who were sampled for analysis in the present study | ||
| Rate of OBI in HBsAg (–) but anti-HBc (+) subjects‖, % | 1.7 | 4.8 |
| Rate of OBI in HBsAg (–) but anti-HBc (–) subjects¶, % | 1.8 | 0 |
| Estimated frequency of OBI per 104 HBsAg-negative children# | 160.7 | 11.5 |
- *†From references 2 and 4.
- ‡Deduced from † − *.
- §Deduced from 100% − †.
- ‖¶Data from Table 2.
- #Calculated by (‡×‖ + §×¶).
Demographic, Serological, and Virological Characteristics of 23 Children With OBI
Serum HBV DNA by real-time PCR was detected only in cases 4 and 14. Among 16 OBI children seropositive for anti-HBs and anti-HBc, documented anti-HBs titer was available in 11. Whereas cases 5 and 6 had a significant level of anti-HBs by RIA of 117 and 189 S/N, respectively, cases 9, 10, 11, 12, 15, 16, 17, 18, and 19 had a protective level of anti-HBs by enzyme immunoassay (EIA; 18, 375, 158, 449, >1,000, 10.6, 99.2, 193, and 861 mIU/mL, respectively). All HBV isolates from 23 subjects with OBI belonged to genotype B (Table 4).
| Case No.a | Age (years) | Sex | Geno-type | HBsAg/Anti-HBs/Anti-HBc | Viral Load (log copies/mL) | Maternal HBsAg | Amino Acid Substitutionb | Nucleotide Mutation | ||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Pre-S1 | Pre-S2 | S | BCP | Pre-C | Core | |||||||
| 1 | 9.8 | M | B | –/+/+ | <3 | ND | S78N/L108I (B/B) | Wild | Wild | wild | wild | wild |
| 2 | 9.8 | F | B | –/+/+ | <3 | ND | S78N(B) | Wild | Wild | A1752G | wild | C1913A |
| 3 | 10.5 | M | B | –/+/+ | <3 | ND | Wild | Wild | Wild | A1752G | wild | C1969T |
| 4 | 4.4 | M | B | –/+/+ | 3.34 | ND | A90V (Hsc, CAD) | Wild | Wild | A1752G | wild | wild |
| 5 | 0.5 | F | B | –/+/+ | <3 | + | Wild | Wild | I110L | wild | wild | C2015A |
| 6 | 11.3 | F | B | –/+/+ | <3 | – | Wild | Wild | Wild | A1752G | wild | wild |
| 7 | 13.9 | M | B | –/–/+ | <3 | ND | Wild | Wild | Q129R | wild | A1846T/G1896A | C1913A |
| 8 | 14.7 | F | B | –/+/+ | <3 | ND | Q118L (Hsc, NB) | Wild | T123A | A1752G | wild | wild |
| 9 | 7.6 | M | B | –/+/+ | <3 | ND | Wild | Wild | Wild | wild | wild | wild |
| 10 | 6.04 | F | B | –/+/+ | <3 | – | Wild | Wild | Wild | A1752G | wild | wild |
| 11 | 5.9 | F | B | –/+/+ | <3 | + | T87S/A90T (Hsc/Hsc, CAD) | Wild | Wild | wild | wild | wild |
| 12 | 13.88 | M | B | –/+/+ | <3 | + | Wild | Wild | S117N | wild | wild | wild |
| 13 | 13.1 | M | B | –/–/+ | <3 | + | Wild | Wild | Wild | C1773T | wild | wild |
| 14 | 0.5 | M | B | –/–/+ | 3.84 | + | L45F/P47T/N56T/G73E/ P94S (B/B/Ud/B/B) | Wild | G145R | wild | wild | wild |
| 15 | 1.0 | M | B | –/+/+ | <3 | + | Wild | Wild | I110L | wild | wild | wild |
| 16 | 8.5 | F | B | –/+/+ | <3 | – | Wild | Wild | Wild | wild | wild | wild |
| 17 | 5.1 | M | B | –/+/+ | <3 | + | L108I (B) | Wild | Wild | wild | wild | wild |
| 18 | 15.7 | M | B | –/+/+ | <3 | + | Wild | Wild | T115A | C1773T | wild | wild |
| 19 | 11.46 | M | B | –/+/+ | <3 | + | Wild | Wild | Wild | wild | wild | wild |
| 20 | 6.9 | F | B | –/+/– | <3 | ND | Wild | Wild | Wild | wild | wild | wild |
| 21 | 6.4 | M | B | –/+/– | <3 | ND | Wild | Wild | T126A | C1773T | wild | wild |
| 22 | 7.0 | M | B | –/+/– | <3 | ND | Wild | Wild | wild | A1752G | wild | wild |
| A1808G | ||||||||||||
| 23 | 10.6 | M | B | –/–/– | <3 | – | Wild | Wild | P111Q | C1773T | wild | wild |
- a Cases 1-19 were anti-HBc (+). Cases 20-23 were anti-HBc (–). Cases 1-3 and 20-23 were born before 1984 and unvaccinated. Cases 4-19 were born after 1984 and vaccinated. Protective anti-HBs levels was noted in cases 5 and 6 were 117 S/N and 189 S/N, respectively (by RIA), and in cases 8-12 and 15-19 (ranging from 10.6 to >1,000 mIU/mL by EIA).
- b Data denote the amino acid variation and its location at the region with immune epitopes and functional domains (from reference 17). In parentheses, B indicates B-cell epitopes, both Hsc (heat shock protein 70 binding site) and CAD (cytosolic anchorage determinant) are important for dual topology of L protein. NB (nucleocapsid binding site) is important for viron morphogenesis. Ud indicates undetermined function.
- Abbreviation: ND, no data available.
Genomic Variability of HBV in 23 Children With OBI and 28 HBsAg-Positive, Anti-HBc-Positive Control Subjects
Table 4 shows that HBV surface “a” determinant (amino acids [aa] 110-160) variants were found in 7 anti-HBc (+) children with OBI (cases 5, 7, 8, 12, 14, 15, and 18) and in 2 anti-HBc (−) children with OBI (cases 21 and 23). These S mutants were with or without mutations in other regions of the HBV genome.
Pre-S1 variants with wild-type S region were found in 5 anti-HBc (+) children with OBI (cases 1, 2, 4, 11, and 17). In our immunized subjects with OBI, pre-S1 mutations were located at the region with known B-cell epitopes or other important functions. Nucleotide sequence analysis of basal core promoter (BCP) and pre-core/core gene showed that one had G1896A in the pre-core gene.
Of 27 HBsAg (+) and anti-HBc (+) controls with successful sequencing of S gene, 4 (13.3%) had “a” determinant mutant, including Q129H, M133T, T140I, and G145R (see Supporting Table 2). The aa substitutions of the pre-S (pre-S1+pre-S2) region were more diverse than those of the S region.
Table 5 shows that the rate of aa substitutions/100 aa within the “a” determinant (aa 110-160) in anti-HBc (+) subjects with OBI tended to be higher than in anti-HBc (+), HBsAg (+) control subjects (P = 0.088). Viral load was much higher in the latter (P < 0.001). aa variation in B-cell epitopes of the pre-S region was not different between OBI and control subjects (see Supporting Table 3).
| Occult (n = 19) | Carrier (n = 28) | P Value | ||
|---|---|---|---|---|
| Age (years) | 9.8 (0.5-15.7) | 8.2 (0.35-17.64) | 0.871 | |
| Sex (male/female) | 12/7 | 15/13 | 0.514 | |
| Genotype (B/C) | 19/0 | 28/0 | 1 | |
| Viral load (log copies/mL) | <3 (<3.0-3.84) | 6.83 (<3.0-8.32) | <0.001 | |
| 3.06 ± 0.2 | 6.21 ± 1.66 | |||
| Regions | ||||
| Pre-core/core | N = 19 | N = 28 | ||
| BCP (%) | 8/18 (44.44) | 17/28 (60.71) | 0.210 | |
| Pre-C (%) | 0/19 (0) | 1/28 (3.57) | 1 | |
| Core (%) | 4/19 (21.05) | 7/28 (25) | 1 | |
| Mutation rates per hundred of nucleotides | 0.14 (0-0.28) | 0.14 (0.0-0.42) | 0.275 | |
| 0.08 ± 0.1 | 0.14 ± 0.14 | |||
| Pre-S1/pre-S2 | N = 19 | N = 28 | ||
| Pre-S1 (%) | 7/19 (36.84) | 12/28 (41.93) | 0.680 | |
| Pre-S2 (%) | 0/19 (0) | 4/28 (12.9) | 0.137 | |
| Mutation rates per hundred of amino acids | 0 (1.0-1.69) | 0 (0.0-0.68) | 0.592 | |
| 0.23 ± 0.42 | 0.15 ± 0.20 | |||
| B-cell epitope mutants (%) | 4/19 (21.1) | 5/58 (17.9) | 1 | |
| T-cell epitope mutants (%) | 0/19 (0) | 4/28(12.9) | 0.137 | |
| S | N = 19 | N = 27* | ||
| “a” determinant mutants (%) | 7/19 (36.84) | 4/27 (13.33) | 0.159 | |
| Mutation rates per hundred of amino acids (median and range) | 0 (0.0-1.96) | 0 (0.0-1.96) | 0.088 | |
| 0.76 ± 0.98 | 0.29 ± 0.71 | |||
- Nucleotide sequence of surface gene was unsuccessfully sequenced in 1 case (case 22 in Supporting Table 2).
Use of HBIG in Relation to the Rates of OBI and Surface “a” Determinant Mutants
Immunization history was known in 14 OBI children and 17 control HBsAg-positive children. Among them, 8 (57%) of the 14 and 13 (76%) of the 17 were given HBIG plus vaccination, implying that HBIG usage was not associated with the occurrence of OBI or overt HBV infection (P = 0.44). HBIG usage appears to select more S-escape mutants in OBI than in overt HBV infection because 4 (50%) of 8 OBI subjects and 2 (15.4%) of 13 HBsAg (+) control subjects harbored such mutants (P = 0.16, possibly owing to small subject sample).
Reanalysis of HBsAg With High Sensitive Architect Assay
Of 14 OBI children with sufficient serum volume for reexamination with Architect assay, 13 still showed HBsAg-negative (<0.05 IU/mL) and one (initially HBsAg negative by RIA) showed a marginally positive value (0.06 IU/mL).
Follow-up of OBI Children Seropositive for Anti-HBc
Although no serially collected serum samples from OBI children was available for testing, a second follow-up for serum alanine aminotransferase (ALT) and HBV-DNA analyses could be performed for 5 OBI children (2 were anti-HBc positive alone and 3 were seropositive for both anti-HBs and anti-HBc) at 5 years after the first examination. Supporting Table 4 shows that case 13 was anti-HBc (+) alone and HBV DNA (+) by nested PCR at first examination. He became HBsAg (+) anti-HBc (+), but HBV DNA (−), by nested PCR at second examination. Cases 16, 18, and 19 were HBsAg (−) anti-HBc (+), anti-HBs (+), and HBV DNA (+) by nested PCR at the first examination, but turned out to be HBsAg(−), anti-HBc (−), and HBV DNA(-) by nested PCR at the second examination.
Discussion
Our results show that in HBsAg (−), but anti-HBc (+) subjects, OBI frequency was lower in those unvaccinated (1.7%) than in those vaccinated (4.8%), whereas in HBsAg (−) and anti-HBc (−) subjects, it was higher in those unvaccinated (1.8%) than in those vaccinated (0%). However, the anti-HBc-positive rate was 26.2% in the 1984 survey (a population without universal infant immunization [UII]) and 2.9% in the 2009 surveys (a population fully covered by UII), respectively. Thus, the estimated OBI frequency per 104 HBsAg-negative subjects declined from 160.7 in unvaccinated cohorts to 11.5 in vaccinated cohorts, greater than 10-fold reduction in OBI. Obviously, a markedly decreased rate of anti-HBc seropositivity within the 25-year period after implementation of UII has largely contributed to the reduction of OBI in immunized children. This reduction complements the well-documented UII-related benefit of markedly reduced overt HBV infections.
Prevalence of OBI was significantly higher in HBsAg-negative, immunized children seropositive for anti-HBc than in those seronegative for anti-HBc (4.8% vs. 0%), indicating that serum anti-HBc can be considered as a very useful marker for screening OBI in HBsAg-negative, immunized children population. Once the majority of the donor population is born after UII, we can exclude subjects with positive anti-HBc as donors to minimize the risk of HBV transmission in recipients.
Anti-HBc testing by RIA or EIA was associated with a high false-positive rate, especially in samples with low-level anti-HBc (near cut-off value or 50%-70% inhibition).14 Because HBV DNA still could be detected in some subjects with low-level anti-HBc,15 we included every HBsAg-negative, anti-HBc-positive subject with sufficient serum samples for HBV-DNA analysis, regardless of their anti-HBc levels, thereby reduced the possibility of underestimation of OBI prevalence. On the other hand, the “true prevalence” of OBI among those with “true positive anti-HBc” is likely higher than identified in this study. This is supported by a previous study showing a positive correlation between anti-HBc titers and HBV-DNA prevalence in HBsAg-negative donor blood16 and might also explain the variability in OBI prevalence among anti-HBc-positive cohorts in different countries.
OBI was previously detected in 19.4% (6 of 31) of vaccinated, HBsAg-negative Gambia children,6 10.9% (5 of 46) of vaccinated, HBsAg-negative Taiwanese children visiting the hospital with other medical illnesses,17 and 28% (21/75) of immunized HBsAg-negative Iranian children born to carrier mothers.18 In Indian infants born to carrier mothers, OBI frequency was 64% at 18 weeks, which decreased to 42% by 2 years of age.19 These studies were limited by either small subject sample or a specific group of children with high risk of HBV infection, with apparently higher OBI frequencies than our results. Subjects in different studies differed in age at enrollment, immunization schedule, and presence of infection source in close contact. The apparently healthy individuals enrolled in this study might be a better representative of the general population of children and adolescents.
In HBsAg-negative subjects seropositive for both anti-HBs and anti-HBc, a significantly higher prevalence of OBI was observed in those born after 1984 than in those born before 1984, suggesting that breakthrough HBV infection in immunized subjects might associate with more OBI than natural HBV infection in unimmunized subjects. In the prevaccination era, the serological profile (HBsAg−, anti-HBs+, and anti-HBc+) in children (born before 1984) represents those who have cleared their infection and usually do not have viremia by PCR. However, in immunized children (born after 1984), such a profile is mostly owing to breakthrough HBV infections in subjects previously responsive to HBV immunization. In immunized children with circulating anti-HBs, very small quantities of HBV might still survive and retain in the liver after perinatal exposure to high viral load of carrier mothers or after frequent exposure to HBV through close contact with highly infective mothers later in life, leading to OBI. This is supported by the fact that most OBI children born after 1984 had a protective level of anti-HBs, implying that an anti-HBs response had been adequately induced and maintained by vaccination. In addition, 73% of OBI children with known maternal HBV status had carrier mothers.
We found that 2 (29%) of 7 unimmunized children (born before 1984) with OBI, but 7 (44%) of 16 immunized children (born after 1984) with OBI, had surface antigen variants in or around the “a” determinant region. Epitope mapping with monoclonal antibodies shows the I110L change results in epitope loss.20 T115A and S117N lie within the major hydrophilic region of HBsAg that specifies HBsAg epitopes. T123A is responsible for both diagnostic and HBIG therapeutic failure.21 Q129R and G145R have previously been confirmed to be vaccine escape mutants.22, 23 However, functional analysis of these variants is needed to determine whether they lead to immune evasion. Additionally, direct population sequencing (vs. clonal sequencing or next-generation assays) could actually underestimate the presence of minor surface variants.
Taken together, it seems that current immunization does not induce full protection to HBV infection in some vaccinated individuals and result in OBI; and selection pressure by this immunization further led to the generation of diverse mutation patterns capable of causing immune evasion.23 However, the majority (61%) of 23 OBI cases were without detectable “a” mutants, and all 23 OBI cases had very low viral load. Thus, undetectable or extremely low HBV-DNA level, which was enough for viral assembly, but might lead to low expression levels of HBsAg below the sensitivity threshold of the detection tests, is the major mechanism underlying our OBI cases.
Another important issue is the impact of antiviral therapy on the rates of OBI derived from vertical and horizontal transmission. Among the 10 OBI children identified in later (2004 and 2009) surveys, none of their mothers received specific antiviral treatment before and during pregnancy. Further large-scale study is needed.
The detection of HBV DNA in blood may be an insensitive means of defining OBI because alternative sources of virus (peripheral blood mononuclear cells or liver) have been shown to be more sensitive in other studies.24, 25 Repeat measures of HBV DNA in serum using highly sensitive PCR/hybridization techniques are also important because the presence of both covalently closed circular DNA and viral transcripts in the liver with intermittent viremia may occur in HBsAg-negative, anti-HBc-positive patients.8, 15
It is quite difficult for us to reanalyze all HBsAg-negative serum samples in previous serosurveys because of the large number of such samples and insufficient serum volume in many subjects. Reexamination of HBsAg in 14 OBI children with the Architect assay suggested that HBsAg-negative subjects screened by RIA or EIA may be HBsAg positive when tested by more-sensitive HBsAg assays, indicating the importance of using the most sensitive HBsAg assay in screening OBI. The detection of HBsAg can also be enhanced by ultracentrifugation of a larger volume or pooled serum to concentrate viral paticles.24
Because of the remoteness of study the population and that many parents did not consent to further examinations, follow-up examinations could be performed for only 5 OBI children (4 had carrier mothers and 1 had a carrier father) with a history of HBV immunization. Changes of their HBV serological profile suggest that OBI is not a stable status of HBV infection. Case 13 was HBsAg negative initially and became weakly HBsAg positive on follow-up. Persistent presence of anti-HBc and traces of HBV DNA with negative HBeAg/hepatitis B e antibody suggested a case of chronic HBV carriage with non- or low productive infection. Cases 16, 18, and 19 were initially anti-HBc (+) and anti-HBs (+) with detectable serum HBV DNA. Although the manufacturer of the AxSYM CORE kit reports a specificity of 99.9%-100%, false-positive anti-HBc results cannot be ruled out. It is also possible that under the protection of vaccine-induced anti-HBs, breakthrough HBV infection may become a low-dose virus infection, which develops short-duration viremia, absent expression of HBsAg, and low-level anti-HBc, but is insufficient to allow full maturation of protective memory, resulting in negative anti-HBc and HBV DNA on follow-up. However, intrahepatic latency in these subjects remains possible.
This study has other limitations. Because of the cross-sectional nature of previous serosurveys, a single-serum sample was used to determine OBI, which might underestimate OBI prevalence. Some serum samples have been stored frozen for prolonged periods and might have impact on HBV DNA testing. Taq polymerase used for HBV sequence is potentially for greater error/nucleotide misincoporation rate. Given that different HBV genotypes replicate and undergo spontaneous mutations at different rates, and that the HBV genotypes in the study population are predominantly genotype B, the findings may not be generalizable to other genotypes.26 No measurement of ALT levels for OBI subjects precluded us to understand whether HBV DNA can be detected easily in OBI patients with hepatocyte damage.
In conclusion, reduction of OBI in immunized children and adolescents complements the well-documented UII-related benefit of markedly reduced overt HBV infection. Breakthrough infection in vaccinated subjects may associate with more occurrence of OBI than natural infection in unvaccinated subjects. In the postvaccination era, the presence of anti-HBc is a very useful marker for OBI screening in HBsAg-negative subjects. A very-low-level viral replication and HBsAg expression, rather than surface gene mutations that may escape detection by HBsAg screening assays, is the major mechanism related to OBI.
