Efficacy and safety of hepatitis B vaccine: an umbrella review of meta-analyses

ABSTRACT

Background

There is a lack of synthesis of literature to determine hepatitis B vaccine (HepB) strategies for hepatitis B virus (HBV) supported by quality evidence. We aimed to explore the efficacy and safety of HepB strategies among people with different characteristics.

Research design and methods

PubMed, Cochrane Library, Embase, and Web of Science were searched for meta-analyses comparing the efficacy and safety of HepB up to July 2023.

Results

Twenty-one meta-analyses comparing 83 associations were included, with 16 high quality, 4 moderate, and 1 low quality assessed by AMSTAR 2. Highly suggestive evidence supports HepB booster and HepB with 1018 adjuvant (HBsAg-1018) for improved seroprotection, and targeted and universal HepB vaccination reduced HBV infection Suggestive evidence indicated that targeted vaccination decreased the rate of hepatitis B surface antibody positivity and booster doses increased seroprotection in people aged 10–20. Weak evidence suggests potential local/systemic reaction risk with nucleotide analogs or HBsAg-1018. Convincing evidence shows HLA-DPB1*04:01 and DPB1*04:02 increased, while DPB1*05:01 decreased, hepatitis B antibody response. Obesity may reduce HepB seroprotection, as highly suggested.

Conclusion

Targeted vaccination could effectively reduce HBV infection, and adjuvant and booster vaccinations enhance seroprotection without significant reaction. Factors such as obesity and genetic polymorphisms may affect the efficacy.

KEYWORDS:

• Hepatitis B vaccines
• efficiency
• safety
• systematic review
• meta-analysis

1. Introduction

Hepatitis B virus (HBV) infection is a major public health problem, which has caused a high direct and indirect disease burden worldwide [1]. The global burden of HBV infection is unevenly distributed, with particularly high prevalence in some countries in Africa and Southeast Asia [2]. However, persistent HBV infection can lead to varying degrees of liver damage, leading to various complications such as hepatitis, fibrosis, cirrhosis, and hepatocellular carcinoma (HCC) [3], causing nearly 900,000 deaths every year. The anti-HBV vaccine has been available to people since 1982 [4], study estimated that a timely scale-up of the birth dose vaccine to 90% by 2030 would immediately reduce the incidence of chronic HBV, and the number of deaths in the global population born between 2020 and 2030 will be reduced by 710,000 [5]. Studies have shown that hepatitis B vaccine (HepB) is an economically attractive option compared to other interventions [6,7], but it has not reached universal coverage worldwide, with global coverage remaining at only 42% and 17% in the WHO African Region by 2021. Between 1990 and 2020, 310 million people worldwide have benefited from hepatitis B vaccination and were protected from hepatitis B virus infect [2,5,6].

With medical progress, HepB and the corresponding vaccination program are constantly updated. As the specific situation varies from one country region to another, the World Health Organization (WHO) encouraged vaccination campaigns based on the specific prevalence of hepatitis B antigen carriers in the geographic area where vaccination is implemented [4]. Therefore, several vaccination strategies, such as vaccination protocols and different doses, have been proposed. For instance, the use of adjuvants and a booster dose has demonstrated the ability to enhance the immunological protection properties of the HepB vaccine according to studies [8,9]. The anti-HBs titers of two accelerated vaccination schedules for high-risk healthy adults [10], 0-7-21 days and 0-1-2-12 months, were confirmed to have the capacity of reaching seroprotective levels more rapidly than the standard group [11,12]. Also, individual compliance with a full course of vaccine should also be considered when assessing the efficacy of the vaccine program. However, the effectiveness and the safety of the hepatitis B vaccine varied among different populations.

In addition, it has been suggested that even if the same HepB strategy is used, the effectiveness of immunization is affected by different individual characteristics. Genome-wide association studies have shown that some DRB1 and DQB1 genes and associated single nucleotide polymorphisms (SNPs) play a key role in HepB response in different populations [13,14], studies shown that HLA–DPB1 alleles * 02:01, * 02:02, * 03:01, * 04:01, and * 14:01 were associated with vaccine response, while in contrast, * 05:01 was associated with non-response to HepB. It has also been suggested that HBV gene expression adopts a similar pattern to that of the major metabolic genes in the liver [15], so the impact that different metabolic states (e.g. diabetes, obesity, etc.) may have disorders of innate immunity that can negatively affect vaccine response [16–18]. In northern China, after the implementation of universal hepatitis B vaccination for infants starting in 1992, the prevalence of hepatitis B infection in children aged 1–14 years fell from 46% in 1992 to 4% in 2006 [19]. Age at acquisition of HBV infection is a major determinant of the clinical presentation of acute disease and the development of chronic infection, and perinatal and early postnatal transmission is the main cause of chronic hepatitis infection.

There is a growing number of meta-analyses summarizing the impact of current vaccination strategies and characteristics of the inoculated population for HepB. Due to the heterogeneity of study methods and findings, as well as other limitations, a single meta-analysis may not provide a comprehensive picture of its safety and efficacy. Just as systematic reviews synthesize data from existing individual studies [20], umbrella reviews summarized evidence from existing systematic reviews. Umbrella reviews can provide summaries of evidence from multiple systematic reviews covering different populations, research questions, or time periods [21]. It is important at this time to systematically assess the efficacy of HepB in vaccinated individuals, focusing on the overall picture of the higher level of evidence for all its effects.

Thus, the present umbrella review provides a comprehensive review of existing meta-analyses related to HepB, with the aim of summarizing the evidence on its efficacy and safety, including vaccine strategies, immune response, and vaccinated populations with different characteristics. In addition, it highlights areas of contradiction and consistency in the evidence base for HBV vaccination, as well as research gaps. This overview can be used as a guide for finding a high-quality systematic review of specific outcomes, patient populations, or vaccine types, with a view to bringing better available evidence to healthcare decision-makers [22].

2. Methodology

This review was registered and conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and the Grading of Recommendations, Assessment, Development and Evaluation (GRADE). Our study protocol is registered with PROSPERO (registration number: CRD42023413135).

2.1. Literature search strategy and eligibility criteria

An umbrella review was conducted on all available systematic reviews and meta-analyses centering on the topic of the efficacy and safety of HepB associated with different vaccine factors or populations. PubMed, Cochrane Library, Embase, and Web of Science were searched for papers published between the database inception and 14 July 2023, and no language restriction was applied. A manual search on Google Scholar was also conducted to include articles covering the topics of interest. The literature search strategy wasattached in the Supplement. Two independent reviewers (JQ and SZ) undertook screening, data extraction, and quality assessment according to the PRISMA guidelines and disagreements were judged by a third reviewer (FZ) to reach a consensus after discussions.

We included studies of HepB if they met the following inclusion criteria: (i) Systematic reviews and meta-analyses that focused on the safety or efficacy of HepB among various populations; (ii) Meta-analyses conducted based on original studies providing details about the efficacy (e.g. anti-HBs IgG titers; the response rate, etc.)  or safety (e.g. risk for cardiovascular-related mortality) with different HepB strategies or among populations with different characteristics; and (iii) Studies reported at least one of these outcomes as pooled odds ratios (ORs), relative risks (RRs), hazard ratios (HRs), or health-related diameters (e.g. standardized mean difference (SMD), weighted mean difference or mean difference [MD]) concerning health outcomes in association with HepB. When two or more systematic reviews existed for the same intervention and comparison, only the most recent systematic review with the largest number of individual studies providing study-level estimates would be included to avoid duplication of samples. No language or date restrictions were applied. Studies published in full peer-reviewed literature. Articles without study-level effect sizes and 95% confidence intervals (CIs) for systematic reviews, or those were conducted on animal models were excluded. We also exclude the conference abstracts, editorials, narrative reviews, or systematic review protocols.

We divided the included studies into three major categories, the effects of different vaccine strategies (e.g. vaccine composition, route of administration, vaccination procedure, dose, etc.), various characteristics of included populations (e.g. gender, race, genetics, disease status, etc.) on the effectiveness of HepB, and side effects (e.g. adverse effects, mortality, etc.) due to different vaccine strategies.

2.2. Data extraction and quality assessment

We retrieved the first author, the year of publication, population, reported HepB-related interventions and comparisons, outcomes, number of included studies, study design, and an aggregated meta-analysis estimate from each included systematic review. The following information was extracted from each individual study: publication year, study design (i.e. cohort design, case-control design, or clinical trial design), population (e.g. adolescents, the general population, chronic kidney disease population, etc.), sample size, reported HepB-related information (e.g. type, route, protocol, dosage of vaccination, etc.), and corresponding evaluation criteria for control information, and maximum adjusted study specific estimates [i.e. mean deviation (MD), normalized mean deviation (SMD); including Hedges’ g and Cohen’s d), OR, or RR], 95% confidence interval. We extracted any estimates reported by subgroup analysis as well. The protection threshold for anti-HBs concentration was set at 10 mIU/mL based on vaccine efficacy study [23].

AMSTAR (A Measurement Tool to Assessment Systems Reviews) 2 was used to evaluate the overall confidence in the results of all the included reviews, which was divided into four levels: high, moderate, low, and critically low to evaluate review design, literature screening, data extraction, and individual study quality assessment [24]. The Grading of Recommendations, Assessment, Development and Evaluation (GRADE) was also used to assess the certainty of the evidence. GRADE would be categorized as high quality, moderate quality, low quality, or very low quality, then evidence can be downgraded or upgraded by the following criteria [25,26].

2.3. Dataanalysis

Based on the re-analyses of the findings of the reviews, data were presented along with summary tables on the characteristics and findings of the reviews.

A unit of analysis was formed by a systematic review and meta-analysis that met the inclusion criteria. We used random-effects models (the DerSimonian and Laird) to re-estimate the pooled effect sizes, 95% CIs, and P-values for each meta-analysis by referring to the analysis of previous umbrella reviews [27–29].   P Value thresholds for statistical significance for the pooled effect estimate are set at <0.05, and P-value thresholds were set at <10⁻³ and <10⁻⁶ to assess the credibility of evidence [30,31]. For inter-group heterogeneity, we conducted Cochran’s Q (P < 0.10 indicated the existence of heterogeneity) testing and we calculated the I²statistic (I² ≥50% represented high inconsistency) [32,33]. In addition, to better compare the effects of different interventions on HepB safety and efficacy, we used Cohen’s method MDs were normalized to SMDs and further converted to ORs using Hasselblad and Hedges’ method. We also estimated 95% prediction intervals in order to investigate whether the effect on the unified study topic could persist in the future, and when the interval excluded null values (i.e. RRs or ORs equal to 1), the effect was inferred to occur in the new study [34,35]. To identify potential publication bias, we used Egger’s regression asymmetry test and presented it as a funnel plot [36]. Small-study effect was established with Egger’s P < 0.10, with the estimation for the largest component study (the study on the lowest SE) being more conservative than the summary estimate based on the random-effects models [27–29].

We assessed the excess significance for any observed number of studies (O) with nominally meaningful results from every association (P < 0.05) in order to see whether or not it is greater than their expected number (E) [37]. In each association, an expected amount of significant studies was estimated by summing up the estimated statistical power of each individual study using an algorithm from the non-central t distribution, in which the largest effect in each association is the plausible power of the tested association [38]. Each association was compared with the results of the studies, and the excess significance thresholds for the association were established at P < 0.10. At P < 0.10, excess significance was established for a single association. All analyses were performed by Stata version 17.7. The P-values of the test results were two-tailed.

2.4. Determining the credibility of evidence

Based on previous umbrella reviews [27–29,39], we used the following criteria to determine the level of evidence: (1) P < 10⁻⁶based on random-effects meta-analysis; (2)  >1000 participants; (3) P < 0.05 of the largest study; (4) between-study heterogeneity with I² <50%; (5) no evidence of small-study effects; (6) 95% prediction interval that excluded the null value; and (7) no excess significance bias.

According to the results of statistical analysis, we sorted out each evidence’scredibility as Table 1.

Table 1. Rating overall evidence’s credibility of the review.

Classes	Grading of evidences	Eligibility
I	Convincing evidence	Met all criteria
II	Highly suggestive evidence	Met 1 to 3 of the criteria
III	Suggestive evidence	Required only a P < 0.001 by random-effects and > 1000 participants
IV	Weak evidence	Required only a P < 0.05 under random-effects
V	No significant evidence	P ≥ 0.05under random-effects

3. Results

By systematic database searching, we identified 840 records. Forty full-text articles were screened for eligibility after being cleared of duplicates and examined for titles and abstracts. Ultimately, 21 meta analyses comprising 89 associations were included for re-analysis [6,24,40–58] (Figure 1). Supplemental results provide a detailed explanation for the exclusion reviews and the corresponding reasons. Of the 21 included metaanalyses, 15 were rated as high quality according to the AMSTAR 2 scoring system [41–49,52,53,57], 4 were rated as moderate quality (40, 49, 51, 54), and 2 were rated as low quality [50,58] (Figures 2–4, sTables 2–4). Supplemental file 7 shows the poolGRADE evidence of association of different factors and HepB.

Figure 1. Flowchart of the study selection process.

Figure 2. Pool GRADE of evidence of association of different factors and HepB.

^aProtect free from prevalence of anti-HBc (+++) and HBAsg (++); ^bBooster dose (+++), Booster in 10–20-year-old (++), Booster in 20–30-year-old (+); ^cSeroprotection in 28 wks (++), in 50 wks (+); ^dHepB: Hepatitis B virus vaccine; ^eGM-CSF: granulocyte macrophage colony-stimulating factor; ^hHBIG: Hepatitis B immunoglobulin; ⁱAnti-TNFs: Anti-tumor necrosis factors.

Figure 3. Forest plot of vaccination strategies and the efficacy of HepB.

Figure 4. Forest plot of vaccination strategies and the safety of HepB.

3.1. Effect of different vaccine strategies on the efficacy of HepB

Intotal, 45 associations in 11 articles evaluated the effectiveness of various HepB strategies, including different vaccination protocols, doses, routes, and origins of HepB, and whether universal vaccinated. Of these, only 24 re-analyses reported nominally statistically significant pooled effects by the random-effects estimate (P < 0.05), and only 6 had 95% of the prediction intervals excluding null values (sTable 5). In these comparisons, we observed 13 associations with significant heterogeneity (I² >50%). Twenty-eight analyses consisted of fewer than five separate studies, in which case the power of the test was diminished.

A total of 45 GRADE evidence quality classifications were conducted to evaluate the impact of different vaccine strategies on hepatitis B efficacy (Supplementary file 7). Among these classifications, 15 were rated as very low quality (15/45, 33.3%), 7 were rated as low quality (7/45, 15.6%), 11 were rated as moderate quality (11/45, 24.4%), and 12 were rated as high quality (12/45, 26.7%) (Figure 5).

Figure 5. Forest plot of different characteristics and the efficacy of HepB.

*AIR: adequate immune response, AIR.

Out of 45 associations, 4 showed highly suggestive evidence, 2 suggestive evidence, and 20 weak evidence, according to the quantitative umbrella review criteria (Figure 2, sTable 5). Highly suggestive evidence showed that vaccination of HepB booster (pooled RR: 2.023; P-random effects: <0.001) and HepB with 1018 as adjuvant (HBsAg-1018, pooled RR: 1.235; P-random effects: <0.001) could improve the seroprotection rate of vaccine, targeted vaccination could effectively reduce HBV infection (with the positive expression of hepatitis B surface antigen, HBsAg as an indicator) compared to non-vaccinated populations (pooled OR: 0.194; P-random effects: <0.001). Similarly, universal vaccination could also effectively have the same effect on HBV infection (with the positive of hepatitis B surface) (pooled OR: 0.217; P-random effects: <0.001), just as the suggestive evidence suggested that targeted vaccination decreased HBsAg positivity in the population (pooled OR: 0.261; P-random effects: <0.001). Suggestive evidence showed that HepB booster could improve the seroprotection of 10–20 years old people (pooled RR: 1.729; P-random effects: <0.001). Limited evidence suggests that HepB vaccination effectively boosts immune response, especially when combined with hepatitis B immune globulin (HBIG) in newborns of HBsAg-positive mothers, reducing HBsAg and HBV-DNA positivity rates. After a year, this combo continues to increase anti-HBs rates. Using granulocyte macrophage colony-stimulating factor (GM-CSF) as adjuvant therapy, both first and last doses enhance anti-HBs titers. Intradermal vaccination improves seroprotection more than intramuscular, but the trend reverses with reduced doses. 0-7-21 days and 0-1-2-12 months schedules enhance seroprotection. The 20 μg vaccine dose is effective, while 5 μg has the opposite effect. When 1018 was used as an adjuvant (HBsAg-1018), seroprotection could also be improved, extending to boosters in 21–30-year-olds.

3.2. Effect of different vaccine strategies on the safety of HepB

A total of 19 associations assessing the vaccine strategies on vaccine safety were reassessed, and 11 associations reported statistically significant results (p < 0.05) using random-effects model analysis, with null values excluded from the 95% prediction interval in 2 of these analyses. Excess of significance bias was detected in three comparisons and five comparisons consisted of less than 5 individual studies (sTable 6).

In the 17 GRADE evidence quality classifications of effect of different vaccine strategies on the safety of HepB, the quality of evidence was very low in 9 items (9/17, 52.9%), moderatein 3 items (3/17, 42.9%), and high in 5 items (5/17, 29.4%), with no items graded as low (Supplemental file 7).

None of the 19 associations had a convincing, highly suggestive or suggestive strength of evidence according to the quantitative omnibus evaluation criteria. In addition, nine associations had weak strength of evidence (Figure 3, sTable 6). For example, weak evidence suggested that vaccine combined with nucleotide analogues compared to nucleotide analogues (NAs) increased the risk of injection site pain, fatigue, myalgia, headache, and nausea.

3.3. Effect of populations with different characteristics on the efficacy of HepB

Twenty-two associations assessing the impact of populations with different characteristics (e.g. gender, disease or health status, genetics, etc.) on vaccine efficacy were reassessed, and 14 associations reported statistically significant results by using random-effects model analysis (p < 0.05), 6 of these analyses showed null values excluded from the 95% prediction interval. Heterogeneity (I² >50%) was observed in people who were obese, with HLA-DPB1*02:02 allelepolymorphisms, using immunosuppression (anti-TNFs, immunomodulators) (sTable7). We found a risk of small-study effect bias in one pooled analysis, excess of significance bias was detected in three comparisons and 5 comparisons consisted of less than 5 individual studies (Figure 5).

There were a total of 21 results of GRADE evidence quality classification on populations with different characteristics and the efficacy of HepB outcomes. Of these outcomes, the quality of evidence was very low in 9 (9/21, 12.3%), low in 19 (10/21, 47.6%), moderate in 2 (2/21, 9.5%), and there were no items with a high grade (Supplemental file 7). Of the 23 associations, we found three convincing evidence, one highly suggestive evidence, and nine evidence of weak strength (Figure 4, sTable 7). Convincing evidence suggested that alleles HLA-DPB1 *04:01 (pooled OR: 3.224; P-random effects: <0.001), DPB1*04:02 (pooled OR: 3.601; P-random effects: <0.001) were found to be associated with a significant increase in the antibody response to HepB, whereas DPB1*05:01 (pooled OR: 0.736; P-random effects: <0.001) showed the opposite association. Highly suggestive evidence indicated obesity decreased HepB seroprotection (pooled OR: 0.468; P-random effects: <0.001). Moreover, weak evidences suggested that Asian with IL4 genetic polymorphisms and population with the T allele of rs2070874 or HLA-DPB1* 02:02 allele polymorphism  increased HepB protection. Diabetes, preterm infants, males end-stage renal disease patients were associated with the reduction of the seroprotection, and those using  immunosuppression, including anti-TNFs andimmunomodulators  also have the same effect.

4. Discussion

This umbrella review includes 83 associations on the residual safety of the effectiveness of HepB, and it presents compelling evidence suggesting HepB boosters, HBsAg-1018, targeted and universal vaccination were effective in increasing the immune response and thus protection against HBVinfection without major side effects. The finding also shows that people with different alleles have different immune responses to hepatitis B vaccines, while obesity may reduce the immune protection.

The findings of this umbrella review show the effectiveness of booster doses in augmenting seroprotection rates across diverse age groups, with a notable impact observed in individuals aged 10–20 years. Previously, various attempts have also been made to improve the protective effect of the HepB, including higher antigen concentration, vaccine boosters, the intradermal vaccine route, and new adjuvant systems. The concentration of protective antibody decreases over time, with less than 70% of people having more than 10mIU/ml of seroprotection present 10 years after vaccination. The term ’booster’ in the context of immunization refers to a subsequent vaccination administered after a primary vaccination series [59]. Based on current scientific evidence, primary vaccination is required for all people [60,61]. However, in immunocompromised patients, or for major breakthrough infections, serologic monitoring of this population is necessary, and if their anti-HBs level is below 10 mIU/mL, they need to receive a booster vaccination to provide rapid protective immunity, while booster vaccination of immunocompetent children and adults is not recommended [62].

Additionally, we demonstrated the efficacy of hepatitis B vaccine in combination with 1018 adjuvant to significantly increase hepatitis B vaccine response rates. HBsAg-1018 is a HepB that uses cytidine-phosphate-guanosine oligodeoxynucleotide (CpGODN), 1018, as an adjuvant. This adjuvant stimulates hepatitis B surface antigen (HBsAg) directly, producing a targeted immune response, compared to a multi-pathway, broad immune stimulation response [63]. In two previous phase 3 trials, HBsAg-1018 has been shown to induce higher and earlier seroprotection rate in healthy adults [64,65]. It has been noted that the seroprotection rate in the HBsAg-1018 group was 95% or higher in all pre-specified populations except for diabetics or those aged 60–70 years [65]. Accordingly, this greater immunogenicity may not be due to the induction of higher overall antibody levels in the vaccinated population, but to the induction of higher antibody levels in a broad of the recipient population, and the induction of more uniform and effective levels of protective antibodies.

Similarly, universal and targeted vaccination strategies have proven effective in reducing HBV infection in our study. Universal infant and birth dose immunization against hepatitis B is key in global efforts to eliminate HBV infection. WHO estimated a significant reduction in HBsAg prevalence among children under 5, from 4.7% to 1.3% following the implementation of universal vaccination. This highlights the vital role of vaccination programs in curbing HBV transmission and moving closer to global eradication goals [66]. Certain countries have implemented targeted vaccination programs, including catch-up vaccination initiatives, for children born shortly before the introduction of universal infant immunization [57]. These targeted efforts have successfully narrowed national coverage disparities and achieved HepB coverage of ≥80% in all districts [67].

The assessment related to safety to HepB in this study was generally limited, and safety assessments indicate evidence supporting a slight increase in local and systemic reactions such as injection-site pain, fatigue, and myalgia when HepB is administered concurrently with nucleotide analogues or HBsAg-1018. According to the Centers for Disease Control and Prevention (CDC) [68], administration of the HepB is just like taking any other medications, and there may be side effects, but the HepB is so safe that most people do not experience side effects, and if they do occur, they are relatively mild [69]. Our study concluded that all-cause mortality and cardiovascular disease mortality were significantly lower among dialysis patients who were HBV vaccine responders than among non-responders [56], suggesting that immunization against HBV in dialysis patients is essential for the protection of dialysis patients.

Some of the factors associated with vaccine non-responsiveness include obesity [17], age at first vaccination, gender, immune status, and genetic factors [70].   In the present study, the DPGly84-related alleles DPB1*04:01 and DPB1*04:02 were associated with high HepB response, whereas DPB1*05:01 was related to lower HepB response [71]. The reason for this may be due to the fact that DPGly84-related alleles can use class I and class II antigen processing pathways to present endogenous and exogenous peptides, while DP84Asp (including DPB1*05:01) does not have this function of endogenous antigen presentation.

Obesity status also affects the immune response to HepB. Hepatitis B virus is considered to be a metabolic virus and obesity is known to be a risk factor for many metabolic diseases [15], with one study finding diabetes to be an independent risk factor for liver cirrhosis [72]. There may be an impaired immune response in obese individuals due to leptin-induced systemic and B-cell intrinsic inflammation [73], and there is also a hypothesis pointed that obesity or obesity-related disease influences HepB escape mutations [74], and even more researches have concluded that individuals with a BMI of 25–30 and greater than 30 were less likely to be seroprotected after hepatitis B vaccination than those with a BMI less than 25 [17]. The short-term response of HepB to obesity can be enhanced by improved immunisation methods, but the impaired vaccine response due to obesity still needs to be addressed by more robust mechanistic studies.

The main limitation of this study is that, similar to other umbrella reviews [75], our study only reported published systematic reviews or meta-analyses, and even if a factor has a strong effect but is not systematically reviewed or meta-analyzed, it may be excluded from the umbrella evaluation, and if it is little studied, it may also be classified as weak evidence only because it involves <1000 patients. For example, the assessment related to adverse reactions to HepB in this study was generally limited, with only weak evidence showing an increased risk of fatigue, nausea, and pain with vaccine plus nucleotide analogues (NAs) compared to NAs, and for those HepB responders of dialysis patients, they would have a higher all-cause mortality rate as well as cardiovascular-related mortality rate. Possible explanations for the above results can be attributed to the selection criteria of the original study design, which are often not appropriate for very rare outcomes, such as rarely reported adverse reactions. Therefore, we may have overlooked other influences that have not been studied by meta-analysis. Secondly, different meta-analyses may differ in their selection criteria and analysis methods, which may bias the assessment of the evidence. For review included preliminary studies differing significantly in basic characteristics, the heterogeneity should be considered. Therefore, this study used a criteria of I² <50% between studies for type I evidence for studies. However, clinical heterogeneity could exist even in the absence of statistical heterogeneity [76]. Moreover, some of the biases contained in the primary studies could not be fully excluded by the comprehensive review criteria of this study, and it is beyond the scope of this study to assess the quality of primary studies, which should be conducted by the original meta-analysis [29]. The entire analysis was conducted using published data and further analysis to estimate dose–response relationships was not possible due to the variability of data reported across studies. Thirdly, as many of the systematic reviews had similar objectives and were conducted in a relatively limited period of time, the results are likely to have been derived from overlapping initial studies.

5. Limitation

Our study has several limitations. Firstly, as with other umbrella reviews [75], our study is based solely on published systematic reviews or meta-analyses. This approach may exclude factors with strong effects that have not been systematically reviewed or meta-analyzed. For instance, factors involving fewer than 1000 patients are often classified as weak evidence, not necessarily reflecting their actual impact. For instance, the higher mortality rates in HepB responders among dialysis patients could not be thoroughly explored due to the selection criteria of the original study designs not being suited for rare outcomes. Secondly, different meta-analyses may have varying selection criteria and analysis methods, potentially biasing the evidence assessment. While we used a criterion of I² < 50% to identify type I evidence, clinical heterogeneity can still exist without statistical heterogeneity [76], which could affect the reliability of our findings. Thirdly, the biases in primary studies included in the meta-analyses are not fully addressed in our review, which should be conducted by the original meta-analysis [29]. Fourthly, our analysis was limited to published data, preventing further estimation of dose–response relationships due to data variability. Fifthly, it is crucial to mention that our included meta-analyses did not limit anti-HBs measurements, potentially leading to some discrepancies. While the predominant method was radioimmunoassay (Ausab, Abbott Laboratories, Chicago, IL, U.S.A.), certain studies still employed alternative measurement methods. Finally, many of the systematic reviews we included had similar objectives and were conducted in alimited timeframe, likely leading to overlapping initial studies. This overlap might have influenced our results.

6. Conclusions

In conclusion, our study demonstrates that the efficacy of the hepatitis B vaccine is enhanced with booster shots and the addition of adjuvant 1018. We also found that both targeted and universal vaccination strategies significantly reduce hepatitis B incidence, thereby improving population health. Notably, individuals with HLA-DPB1*04:01 and DPB1*04:02 genotype exhibit a higher immune response to the vaccine, in contrast to those with HLA-DPB1*05:01 and obese individuals. Our analysis on the safety and efficacy of HBV vaccine provides valuable insights for future research, highlighting the need for varied study designs and outcome measures. This comprehensive overview contributes to refining hepatitis B vaccination strategies and supports the development of effective booster programs. These findings are crucial for public health authorities in optimizing hepatitis B prevention and control practices.

Article highlights

• Targeted hepatitis B vaccination is an effective way in reducing hepatitis B infection

• Booster doses increase hepatitis B immune response

• Population with specific factors such as obesity may reduce the efficacy of the vaccine

• Hepatitis B vaccine strategies have no significant side effects

Declaration of interests

The authors have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or material discussed in the manuscript. This includes employment, consultancies, honoraria, stock ownership or options, expert testimony, grants or patents received or mending, or royalties.

Reviewer disclosures

A peer reviewer on this manuscript has received an honorarium for their review work. Peer reviewers on this manuscript have no other relevant financial or other relationships to disclose.

Author contributions statement

Jiamin Qiu did the literature search and screening and extracted the data; conducted the data analyses; made the figures and tables; and drafted manuscript.

Shiwen Zhang did the literature search and screening and extracted the data; conducted the data analyses.

Yonghui Feng conducted the data analyses, participated in the interpretation of results, and critically revised the manuscript.

Xin Su and Jun Cai conducted the data analyses.

Shiyun Chen, Jiazi Liu, Sui Zhu, Shiqi Huang, Haokun Huang, Huiyan Wen, Jiaxin Li, Haoyu Yan, and Zhiquan Diao participated in the interpretation of results and critically revised the manuscript.

Xiaofeng Liang was the co-corresponding author who took responsibility for the integrity of the data and the accuracy of the data analysis.

Fangfang Zeng designed the study and also was the co-corresponding author who took responsibility for the integrity of the data and the accuracy of the data analysis.

All authors participated in the interpretation of results, critically revised the manuscript, and approved the final version for submission.

Supplementary material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/14760584.2023.2289566

Additional information

Funding

This manuscript funded by Guangdong Provincial Basic and Applied Basic Research Fund Natural Science Foundation (2023), the National Natural Science Foundation of China (No.81602853) and Major Talent Program of Guangdong Provincial (No.2021JC02Y144). The funding sponsors had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript and in the decision to publish the results.