Introduction

Hepatitis B as an acute and chronic communicable disease, has been a worldwide health problem estimated to lead to between 500,000 to 1.2 million deaths every year through causing chronic hepatitis, cirrhosis and hepatocellular carcinoma1. The prevalence of HBV infection varies significantly in different areas: prevalence of chronic infection with HBV estimates range between 0.1–0.7% in Western, Northern, and Central Europe, while those considerably higher in Eastern and Southern European countries, such as Italy (0.2–4.3%), Turkey (2.5–9%), and Romania (5.6%)2,3. In Alaska, 41% had anti-HBs levels of >10 mIU/ml 7 to 9 years after booster vaccination at birth4, even 51% had this protective levels 30 years after receiving the primary series without subsequent doses in Alaska native persons5. In China, the HBsAg carrier rate was 8.75% in 1979, 9.75% in 1992, and 7.18% in 20066; in Taiwan, the values are as high as 15–20% in adults7; and in the Middle East and North Africa region, the HBV infection estimates are various such as 9.8% in Egypt, 7.4% in Iran, 2.4% in Lebanon and 6.9% Libya from the prisoners; 50.7% in Iran, 8.6% in Israel, 2.8% in Lebanon, 4.5% in Libya, 2.6% in Palestine, 6.1% in Saudi Arabia from the injecting drug users8. In Gambia, 13.2% were found to carry HbsAg9 and national infant HBV vaccination controlling chronic infection had 94% vaccine efficacy10. HBV can be transmitted in many ways, with sexual intercourse and mother-to-child transmission being the most common. Between 15% and 40% of those infected develop acute or chronic liver disease and liver failure, cirrhosis or hepatocellular carcinoma may result.

Many countries have gradually adopted the HBV vaccine in national immunization programs since the World Health Organization (WHO) recommended vaccination for children in 1990s. Most individuals with chronic hepatitis B are asymptomatic and therefore ignorant of their infection status but HBV vaccination, if used for primary prevention, can significantly lower the risk of infection. HBV vaccination triggers antibody response and antibody to hepatitis B surface antigen (anti-HBs) levels ≥10 IU/L are usually regarded as seroprotection for most vaccinees.

Vaccination efficacy among children has been widely studied, but there remains a large proportion of adult populations who are as yet unvaccinated. A previous meta-analysis in 2002 observed many factors influencing response to hepatitis B vaccine, especially a decrease response to recombinant HBV vaccine at higher ages11, which suggested that earlier vaccination should be prioritized for prevention at the population level. However, in the last decade, numerous emerging reports, which focused on the seroprotection rate of hepatitis B vaccine in adults12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48, are still inconclusive to immunize what adults are the most appropriate in order to increase the seroprotection rate. Factors influencing immunologic response to hepatitis B vaccine in adults have been inconsistently examined in existing studies. In this study, we conducted a systematic review and meta-analysis to update and assess a more precise estimation of factors that influence the response to HBV vaccine.

Material and Methods

Search strategy

To find all relevant publications that investigated the association between adult and hepatitis B vaccine and seroprotection, a systematic literature search was independently conducted by two individual investigators with the same method in PubMed, Embase and Cochrane Library using the keywords “hepatitis B vaccine”, “HBV”, “adult”, “anti-HBs” were used. Data were collected from the full-published paper and no language or race restriction was used. Bibliographies of relevant review articles were also screened to supplement the electronic searches.

Inclusion criteria

Included studies met the following criteria (1) original research papers; (2) prospective or retrospective studies, including cohorts and trials; (3) sample size ≥10; (4) healthy subjects, pregnant women, participants with diabetes, chronic renal failure or other diseases but without congestive hepatopathy or infectious diseases; (5) mean sample population age ≥18 years; (6) populations are largely vaccine naive; (7) seroprotection (generally defined as antibody-HBs at a titer of >10 mlU/mL) is assessed at least one month after last recombinant vaccine dose in the majority of participants.

Article screening

Citations were electronically downloaded into reference management software and duplicate citations were electronically/manually excluded. Where studies had multiple reports, the most recent or most complete article was retained. The remaining citations were screened independently by two reviewers using pre-defined criteria. Full-text versions of potentially relevant citations were obtained and again screened independently by two reviewers according to pre-defined criteria. Disagreement was resolved by the opinion of a third reviewer.

Data extraction and quality assessment

The data was independently extracted and then cross-checked by two investigators according to a standard format as follows: author, publication year, age, country, male/female participants, body mass index (BMI), vaccination schedule, time of immunological assessment after last vaccine dose, vaccine characteristics and injection pathwayIf necessary data were unavailable in articles, a request was sent to the author for relevant data. The definition of age strata varied among studies. We considered age ≥16 years as adult and age ≥40 years as a default definition of older age for study participants. We considered individuals with anti-HBs titers ≥10 IU/L to be seroprotective after completion of vaccination against HBV. In a few articles, those data are also available if the seroconversion was defined as anti-HBs titers ≥10 IU/L. The articles were divided into four quality levels such as high, moderate, low, and very low by GRADE evidence profile, which allocates original ranks of low score to observational studies and high score to RCTs49.

Statistical analysis

In this meta-analysis, we calculated the relative risks (RRs) and 95% confidence intervals (CIs) by comparing the valid and invalid participators in the experimental group and control group of recruited articles. Statistical heterogeneity in the studies was examined by the Q statistic. We evaluated the heterogeneity in these studies by this method, I2 = 100 %*(Q-df)/Q. A fixed-effect model was used to analyze the data if there was no statistical difference of heterogeneity (p ≥ 0.05). Otherwise, a random-effect model would be selected.

Subgroups analyses were defined in advance/defined according to the reported data, and studies or results were grouped according to age (older or younger than 40), sex, smoking status, alcoholism, vaccine administration (0-1-12/0-1-6 vaccination schedule), geographical location (Asians/Non-Asian). Sensitivity analysis was performed to estimate the stability of the model by removing each study in turn. Additionally, publication bias was assessed through the funnel plot and Egger’s linear regression test50. All statistical analyses were conducted by Stata 12.0 software.

Results

Characteristics of eligible studies

We finally identified 21053 adults from 37 articles up to June 30, 2015 through electronic and manual searches (Fig. 1). Nine hundred and forty-six studies were excluded according to the mentioned criteria. The characteristics of included studies for this meta-analysis are presented in Table 1 and the majority of the studies were assessed as being of good quality (Table 2).

Figure 1
figure 1

Flow diagram of the study selection process.

Full size image
Table 1 Summary of studies investigating the response to hepatitis B vaccine in adults.
Full size table
Table 2 The absolute and relative risk of non-response to HBV vaccine by subgroup and evidence quality grading*.
Full size table

The studies were largely prospective cohort (n = 5), retrospective cohorts (n = 23), or randomized trials (n = 10). Studies varied considerably in size and were conducted among many countries. The three vaccine doses tended to be administered either at months 0, 1, 6 or 0, 1, 12 and recombinant vaccine doses ranged from 10 μg to 40 μg.

Meta-analysis results

Heterogeneity test result and subgroup analysis

The Q-tests of heterogeneity were marked in partial groups and then the pooled RRs were calculated by the random-effect models and fixed-effect models. Meta-analysis revealed that vaccine non-response rates were significantly greater in older participants (age ≥40 vs. <40 years, RR:1.86, 95% CI:1.55 to 2.23, I2 = 56%, P = 0.001; age ≥30 vs. <30 years, RR:1.77, 95% CI:1.48 to 2.10, I2 = 37%, P = 0.074; age ≥60 vs. age<60 years RR:1.30, 95% CI:1.01 to 1.68 I2 = 33.4%, P = 0.199). Non-response was also more likely among males (male adults vs. female adults, RR:1.40, 95% CI:1.22 to 1.61, I2 = 44.3%, P = 0.005); overweight participants (BMI ≥ 25 adults vs. <25, RR:1.56, 95% CI:1.12 to 2.17, I2 = 77.3%, P < 0.001); smokers (smoker vs. nonsmoker, RR:1. 53, 95% CI:1.21 to 1.93, I2 = 52.1%, P = 0.01) and those with concomitant disease compared to healthy participants (RR:1.39, 95% CI:1.04 to 1.86, I2 = 63.4%, P = 0.002) (Figs 2a, 3a, 4a,b,d and 5a,d). However, there were no differences in response to HBV by alcoholic status (alcoholic vs. nonalcoholic, RR:0.90, 95% CI:0.64 to 1.26, I2 = 0, P = 0.941) or vaccination schedule (vaccine delivered at months 0-1-12 vs. 0-1-6, RR:1.39, 95% CI:0.41 to 4.67, I2 = 77.4%, P = 0.004).

Figure 2
figure 2

(a) The relative risks of response to HBV vaccine between adults age ≥40 and adults age <40. (b) The relative risks of response to HBV vaccine between adults age ≥40 and adults age <40. Comparing with adults age <40, the RRs show decreased response to HBV vaccine among adults age ≥40 in cohort and overall studies.

Full size image
Figure 3
figure 3

(a) The relative risks of response to HBV vaccine between adults age ≥30 and adults age <30. (b) The relative risks of response to HBV vaccine between adults age ≥30 and adults age <30 grouped by study design. Comparing with adults age <30, the RRs indicate reduced response to HBV vaccine among adults age ≥30 both in cohort and RCT studies.

Full size image
Figure 4
figure 4

(a) The relative risks of response to HBV vaccine between adults age ≥60 and adults age <60. (b) The relative risks of response to HBV vaccine between male adults and female adults. (c) The relative risks of response to HBV vaccine between male adults and female adults grouped by study design. Comparing with female adults, The RRs suggest declined response to HBV vaccine among male adults both in cohort and RCT studies. (d) The relative risks of response to HBV vaccine between BMI ≥25 adults and BMI <25 adults.

Full size image
Figure 5
figure 5

(a) The relative risks of response to HBV vaccine between smoker and nonsmoker. (b) The relative risks of response to HBV vaccine between alcoholic and nonalcoholic. (c) The relative risk of response to HBV vaccine between adults with concomitant disease and healthy adults. (d) The relative risk of response to HBV vaccine between 0-1-12 and 0-1-6 vaccination schedule.

Full size image

Subgroup analysis by study location and age indicates that older adults (≥40 years) from non-Asian countries revealed that in contrast with Asians, specially non-Asians in older adults (age ≥ 40) may be slightly less response to hepatitis B vaccine than younger adults (age < 40) (RR:2.02, 95% CI:1.59–2.58; RR:1.60, 95% CI:1.24–2.08), consistent with the region result of older adults (age ≥ 30) (RR:2.16, 95% CI:1.66–2.80; RR:1.46, 95% CI:1.15–1.86). Particularly comparing with Asians, the male in non-Asians has a similar nonresponse to females (RR:1.42, 95% CI:1.18–1.71; RR:1.40, 95% CI:1.11–1.77).

When studies were subdivided by study design results were consistent and lower response was seen among studies with older participants and male participants and again, no difference was observed by alcoholic status (Figs 2b, 3b, 4c,d and 5a,b).

Sensitivity analysis and publication bias

Sensitivity analysis was conducted to estimate the stability of the results and indicated no significant change if any one study was excluded. Funnel plot asymmetry was assessed by means of Egger’s linear regression test and showed that there was significant publication bias in the following groups: age of 40 years, gender and BMI (age40: t = 2.54, P = 0.019; sex: t = 2.99, P = 0.006; BMI: t = 2.70, P = 0.025).

Discussion

When stratified by demographic features, our study showed a lower response in older adults (especially age ≥ 40), male adults and overweight adults (BMI ≥ 25), smoker and adults with concomitant disease after completion of vaccination against hepatitis B.

Our study indicated that young adults have a higher seroprotection rate to hepatitis B vaccine than other age groups (age30: RR = 1.77; age40: RR = 1.86; age60: RR = 1.30). It means that the earlier adult vaccination was inoculated at an age, the better efficiency is. The lower responsiveness to hepatitis B vaccines in older adults might result from the waning immunity with age. In previous studies, it did not find a significant association between age and the immune response43,44. The reason may be that most adults in the study were under the age of 40 years. However, in an observational prospective study of 666 participants, the percentage of nonresponders elevated gradually with age51. Another study aligned with our findings also found that younger age and female gender were predictive of better response52. It indicated in our study that population in non-Asians were both better in age of 40 or 30 years (age30: RR = 2.16; age40: RR = 2.06). Surprisingly our result also showed that the response rate in the younger adults (age < 60) was better than those older (age ≥ 60), different from the previous study44. Some studies reported seroprotection rates of hepatitis B vaccine in older adults (aged ≥ 60 years) range from 30% to 80% and rely on these factors such as study population, vaccination plan, vaccination history and type of vaccine40.

Besides age in our study, male gender both in Asians and non-Asians may be associated with nonresponse to hepatitis B vaccine. It may be owing to the opposite effects of sex hormone androgen and estrogen. This difference is experimentally repeated in animal models, which indicated to be activated by sex hormones in genetic regulation. Moreover, there are numerous immunological genes appearing on the X chromosome while few ones are mapped on the Y chromosome. Estrogen activates monocytes to secrete IL-10, which induces IgG and IgM secretion through B-cells in turn53, while testosterone damages the production of IgG and IgM from B-lymphocytes, as well as restrains producing IL-6 from monocytes54. The hormones’ joint effects on the epigenetic adjustment of genetic expression, and gene structure on the X chromosome differing between XX females and XY males, will partly account for vaccine response heterogeneity in gender55. Based on our results, future programme should be emphasized on males both Asians and non-Asians, who tend to have less response to hepatitis B vaccine.

BMI might influence the level of vaccine response25. The low response to vaccination of overweight on vaccine could be due to the main distribution of the vaccine in fat not in muscle. This could hinder absorption and enable denaturation of the vaccine antigen by enzymatic action25. Another possible interpretation is damaged proliferation and function of the antibody-secreting plasma cells.

The lower immunogenicity of hepatitis B vaccine was linked with smoking and male gender. In smokers, smoking can affect cells and humoral mediated immune responses in humans and animals. Nicotine restrains the antibody-forming cell response by damaging antigen-mediated pathway in T cells and intracellular calcium response. In addition, a high prevalence of HBV markers has been reported in alcoholics. Persistent alcohol intake could restrain immune responses especially in female56. But some studies also reported that difference was undetected between alcohol consumption and seroprotection of hepatitis B vaccination31,38. In this study, the inapparent association within alcoholic subgroup may result from the small sample size and drinking is not common in females.

Aside from those factors, patients with concomitant disease usually have a complicated and inconstant status due to diverse pathogenesis. Although some researches found no association between comorbidity and seroprotection36,37, comorbidity may be a significant element decreasing the efficacy of hepatitis B vaccine from this analysis and others23,27, which could bring immunity disturbance. However, the detailed mechanisms between the poor response to hepatitis B vaccine and adults suffering from concomitant disease are still incompletely understood.

What’s more, the four reports regarding different vaccination schedules in adults such as 0-1-6 and 0-1-12 schedules are still controversial13,39,47,48. Three studies among them13,39,47 observed no difference in seroconversion rates between these two schedules (as in these cases they were all nearly 100%), while another study48 showed a higher seroconversion rate in individuals with the 0-1-12 schedule. Our meta analysis found no difference for seroconversion rate one month after the third injection both in 0-1-6 and 0-1-12 vaccination schedules. Thus in consideration of timing and vaccination compliance, the conventional 0-1-6 vaccination schedule could be still worth to be recommended.

Recently, emerging studies tended to suggest the genetic determinants of heterogeneity in response to the vaccines against hepatitis B. In twins study, 60% of the phenotypic variance was interpreted for the anti-HBs immune response by additive genetic while 40% by non-shared environmental effects57. Asians and non-Asians as study location may also play an important role in seroprotection efficiency of hepatitis B vaccine in adults. The percentage of nonresponders after hepatitis B vaccine remarkably varied in ethnic groups, which may result from the difference of environmental surroundings, the mutation rate and genetic variability, especially at the human leucocyte antigen (HLA) genetic region. However, it is also hard to accurately locate the variation affecting the HBV response in the HLA locus as a result of the long-range linkage disequilibrium in this area58. It needs further studies to explore.

In a word, the factors mentioned above suggested these factors consisting of elder adults, male, BMI ≥ 25, smoking and concomitant disease would be the significant variables reducing the immune response to hepatitis B vaccination. Those who are more likely to have non-response should be checked for seroprotection level and offered additional booster vaccinations. Thereby finding those without immunization and improving overall immunization rates across the population should be emphasized.

Our results should be interpreted in view of the following limitations. First, publication bias was identified among studies reporting rates by age groups, gender and BMI groups. Second, for the various subgroup analyses, sample size is diminished and therefore CIs are wide leading to less accurate estimates of response. Third, due to differences in lifestyle characteristics in different studies’ population, significant heterogeneity was present in study, even among subgroup estimates. In addition, there were poor reporting in some included studies and limited inclusion in subgroup analyses such as BMI, smoking status, alcohol status and concomitant disease. What’s more, several different vaccines were used in the different studies, which had different immunogenicity. Engerix B (with 20 μg HBsAg per dose) is more immunogenic than Recombivax (with 10 μg per dose), the difference being seen especially in older individuals. Twinrix is similar immunogenicity to Engerix-B. A multi-center study found that a S-PreS1/PreS2-vaccine (Hepacare) is also more immunogenic than Engerix B2. Despite limitations, in this work, we systematically sought out all published literature relevant to our research question and then carefully screened studies and extracted data in duplicate using protocols to ensure high quality and consistency in the extracted data. Missing data were sought from authors and studies results were statistically combined to provide robust estimates of the factors associated with poorer immunological response. To our knowledge, this is the first study to examine multiple factors associated with vaccine response and important differences have been found.

Conclusions

Taken together, this meta-analysis indicated that there were lower seroprotection rates to hepatitis B vaccine in the subgroups of increasing aged adults, male, BMI ≥ 25, smoking and concomitant disease, and more vaccine doses, supplemental/additional strengthening immunity should be focused on this specific population. No difference in seroconversion rates between 0-1-6 and 0-1-12 vaccination schedule was observed, but in consideration of timing and vaccination compliance, the vaccination 0-1-6 schedule could be still worth to be recommended. In order to obtain accurate effectiveness of hepatitis B vaccine in adults, more large-scale studies should be conducted in the future.

Additional Information

How to cite this article: Yang, S. et al. Factors influencing immunologic response to hepatitis B vaccine in adults. Sci. Rep. 6, 27251; doi: 10.1038/srep27251 (2016).