Vaccinomics: A Scoping Review

Abstract

Background

This scoping review summarizes a key aspect of vaccinomics by collating known associations between heterogeneity in human genetics and vaccine immunogenicity and safety.

Methods

We searched PubMed for articles in English using terms covering vaccines routinely recommended to the general US population, their effects, and genetics/genomics. Included studies were controlled and demonstrated statistically significant associations with vaccine immunogenicity or safety. Studies of Pandemrix^®, an influenza vaccine previously used in Europe, were also included, due to its widely publicized genetically mediated association with narcolepsy.

Findings

Of the 2,300 articles manually screened, 214 were included for data extraction. Six included articles examined genetic influences on vaccine safety; the rest examined vaccine immunogenicity. Hepatitis B vaccine immunogenicity was reported in 92 articles and associated with 277 genetic determinants across 117 genes. Thirty-three articles identified 291 genetic determinants across 118 genes associated with measles vaccine immunogenicity, 22 articles identified 311 genetic determinants across 110 genes associated with rubella vaccine immunogenicity, and 25 articles identified 48 genetic determinants across 34 genes associated with influenza vaccine immunogenicity. Other vaccines had fewer than 10 studies each identifying genetic determinants of their immunogenicity. Genetic associations were reported with 4 adverse events following influenza vaccination (narcolepsy, GBS, GCA/PMR, high temperature) and 2 adverse events following measles vaccination (fever, febrile seizure).

Conclusion

This scoping review identified numerous genetic associations with vaccine immunogenicity and several genetic associations with vaccine safety. Most associations were only reported in one study. This illustrates both the potential of and need for investment in vaccinomics. Current research in this field is focused on systems and genetic-based studies designed to identify risk signatures for serious vaccine reactions or diminished vaccine immunogenicity. Such research could bolster our ability to develop safer and more effective vaccines.

Funding

This work was supported, in part, by the National Institutes of Health [grant number RM1HG009038-01]. The funder had no role in study design, collection, analysis or interpretation of data, report writing or the decision to submit to this publication.

Introduction

Immunization is one of the most effective and safe ways to prevent morbidity and mortality from infectious diseases [1, 2]. However, variations in human genetics are known to influence individual responses to vaccines [3]. Vaccinomics is a growing field that "aims to understand genomic and systems-level data to elucidate the basis of inter-individual variations in immune responses" [4]. The term "vaccinomics" was initially proposed by Poland et al. in 2007 [5, 6]. Adversomics is a subset of vaccinomics that focuses on adverse reactions to vaccination [7]. Advances in vaccinomics and adversomics have the potential to improve vaccine development and delivery, leading to safer and more effective vaccines [8-12].

This scoping review aims to summarize the known associations between human genetic and genomic variation and inter-individual heterogeneity in vaccine immunogenicity and safety, based on the peer-reviewed literature. Although this review focuses on vaccines routinely recommended for the general population in the United States (US), the findings are globally relevant.

Methods

We systematically reviewed the scientific literature to identify associations between human genetic and genomic variation and heterogeneity in vaccine immunogenicity (e.g., antibody response to vaccination). We searched PubMed on September 8, 2021. Initial search terms combined Medical Subject Headings (MeSH) indexing terms and text-word terms and covered four main concepts: general vaccine terms, US-specific vaccine terms, vaccine effects, and genetic and genomic terms (Appendix 9). Relevant synonyms for MeSH terms (listed as Entry Terms on the MeSH page) were included as text/word terms. Additional search terms were used to automatically exclude articles not published in English, articles exclusively studying animals (and not humans), and article types likely to be redundant (e.g., conference abstracts) or irrelevant due to lack of controls (e.g., comments, editorials, letters, newspaper articles, case reports, etc.). Search terms were developed and revised with input from subject matter experts and an informationist from the Welch Medical Library at Johns Hopkins University. Searches were performed iteratively to capture the number of results removed by search term limitations reflecting exclusion criteria (Appendix 1). Results were exported from PubMed to Covidence, EndNote (Clarivate Analytics) and Excel (Microsoft) for screening and data extraction.

Screening (for potential inclusion in data extraction) was based on a review of article titles and abstracts. Each potential screener began by performing at least two pilot rounds, in each of which they screened at least 20 articles that the first author of this paper had already screened, then reviewed their decisions with the first author for training and quality assurance purposes. Once deemed ready by the first author (based on their pilot round results consistently matching his own), screeners began reviewing remaining articles. Each article was screened by at least two reviewers, and conflicts between reviewers were adjudicated by the first author. To be considered for inclusion, articles were required to provide data from controlled studies showing at least one statistically significant genetic and/or genomic association with the safety and/or immunogenicity of at least one vaccine routinely recommended for use in the general US population. We also included articles addressing the widely publicized association between the 2009 pandemic H1N1 influenza vaccine, Pandemrix^®, and narcolepsy, since a genetic association with this adverse event was observed [13]. Articles were excluded from consideration for data extraction if they: did not have human data; reported only on vaccines not currently routinely recommended for the general US population per Advisory Committee on Immunization Practices (ACIP) recommendations (e.g., vaccines recommended only for travelers, such as yellow fever, or only for specific military populations, such as smallpox) [14]; did not have appropriate controls (e.g., studies that only included passive surveillance data, such as from the Vaccine Adverse Event Reporting System (VAERS), a US passive vaccine safety surveillance program run by CDC and FDA) [15]; were simulations, case reports, uncontrolled cases series, cross-sectional studies, or ecological studies; were letters, editorials, commentaries, or news articles; described the genetics or genomics of nonhumans (e.g., animals or a pathogen); contained no statistically significant associations of relevance; or were not available in English. Secondary data were not extracted from review articles to avoid duplication (with the exception of meta-analyses). However, articles referenced in reviews were screened, and relevant articles not already captured by our search that met our inclusion criteria were included and their primary data extracted. Once screening was completed, included articles were divided (based on vaccine) between two authors for data extraction. Both extracting authors again began by performing two pilot rounds, in which they extracted 17 total articles that the first author had already extracted, then reviewed their decisions with the first author for training and quality assurance purposes. Once deemed ready, extracting authors began extracting remaining articles. Extracting authors brought any questions to the first author, who obtained the input of a subject matter expert when needed. The first author also performed quality assurance checks by extracting 5 additional articles in common with each extracting author, in which no major differences between authors’ extracted data were found.

The initial stage of our search identified 3,271 articles (Appendix 1). Automatic exclusion via search terms removed non-English articles (n=140), non-human studies (n=604), and irrelevant article types (n=175), and our review software automatically removed 43 duplicate articles. During the screening of the remaining 2,309 articles, articles were manually excluded if they: focused on host genetics and the immune system but not response to vaccination (889); focused on variation by host genetics in response to infection, not vaccination (n=182); lacked a genetic or genomic component (n=316); lacked relevant data (n=110), lacked data on humans (n=276), lacked controls or active surveillance (n=61); or focused on vaccines other than Pandemrix^® not routinely recommended in the US (n=82). The full text of the remaining 393 articles were reviewed. Thirty-two additional articles identified from the reference lists of systematic reviews were screened, 15 of which were included for data extraction. Three articles did not have full-text available through the Welch Medical Library at Johns Hopkins and were thus excluded from data extraction. After full-text review was completed, 191 additional articles were excluded for not meeting inclusion criteria, leaving 214 articles eligible for data extraction. These articles were published between 1976 and 2021. Only statistically significant results were extracted; statistical significance was considered p<0.05, except genome-wide association studies (GWAS) using Bonferroni correction, for which the adjusted p-value was used (commonly P < 5 × 10⁻⁸) [16]. Included studies are described below, highlighting the genetic variants most commonly associated with vaccine immunogenicity (Table 1). Effect sizes calculated by meta-analyses for commonly found associations are noted when available (Figure 1).

Table 1.

Genetic Variants Most Commonly Associated with Vaccine Immunogenicity

Genetic Variant	Vaccine(s) (+/−)	Study Countries	Years	References
HLA- C4A*Q0	Hepatitis B (−)	Germany, Italy, Slovenia, US	1995-2020	[3, 32, 46, 65, 68, 86, 91, 104]
HLA-DPB1*0501	Hepatitis B (−)	China, Japan, Korea, Slovenia, Taiwan, Turkey, UK	2014-2021	[3, 36, 45, 56, 77, 78, 81, 85, 88, 91, 98, 102]#
HLA-DR3	Hepatitis B (−)	Belgium, France, Germany, Greece, Italy, Slovenia, Spain, UK, US	1986-2020	[42, 44, 46, 47, 64, 69, 85, 89, 91, 103]
HLA-DR7	Hepatitis B (−)	Belgium, Germany, Italy, Iran, Slovenia, Spain, Turkey, UK, US	1986-2020	[22, 32, 42, 46, 64, 69, 83, 91, 96, 103]*
HLA-DPB1 SNP: rs9277535	Hepatitis B (+)	China, Indonesia, Japan, Korea, Taiwan, Turkey	2011-2019	[37, 43, 55, 81, 88, 94]
HLA-DPB1*0402	Hepatitis B (+)	China, Japan, Korea, Taiwan	1996-2021 [102]	[3, 29, 36, 45, 78, 88, 99, 102]#
HLA-DRB1*0803	Hepatitis B (+)	Japan, Taiwan	1996-2018	[3, 29, 36, 45, 99]
HLA-DRB1*15	Hepatitis B (+)	China, France, Germany, Slovenia, UK, US	1993-2020	[3, 17, 82, 85, 86, 88, 91-93, 101]*
HLA-DQB1*0602	Hepatitis B, Influenza, Measles (+)	China, Iran, Japan, Sweden, UK	1997-2019	[3, 33, 34, 36, 83, 85, 88, 101, 105]*
HLA-DRB1*07	Hepatitis B, Influenza, Measles (−)	China, Germany, Iran, Slovenia, South Korea, UK, US	1997-2020	[3, 26, 28, 34, 49, 59, 82, 83, 85, 86, 88, 91, 101, 106, 120-124]*$
HLA-DQB1*0303	Influenza, Measles (−)	Italy, US	2002-2019	[3, 106, 124, 140, 147]$
HLA-W16	Influenza (−)	Australia, China, US	1978-2021	[110, 123, 127]
IFITM3 rs12252 C/C	Influenza (−)	China, US	2008-2020	[109, 113, 129]
HLA-DQB1*0603-9/14	Influenza (+)	China, Italy, UK, US	2002-2021	[3, 106, 118, 120, 122, 123]
HLA-DRB1*13	Influenza, Measles (+)	Italy, UK, US	1996-2019	[3, 106, 124, 157]$
CD46 rs2724384	Measles (−)	Australia, US	2011-2017	[3, 130, 132, 137, 139]
HLA-DQA1*0201	Measles (−)	Slovenia, US	2001-2020	[3, 91, 140, 148, 159]$
HLA-DQB1*0201	Measles (−)	US	2005-2014	[3, 147, 148]$
HLA-B*07	Measles (+)	US	1998-2014	[3, 140, 142-144]
HLA-B*3503	Measles (+)	Slovenia, US	2011-2020	[91, 140, 146, 159]
Blood Type 0, Se+	Rotavirus (+)	Pakistan, Sweden	2017-2021	[203, 206, 208]
Nonsecretor	Rotavirus (+)	Malawi, Sweden	2019-2021	[204, 206, 208]
Secretor	Rotavirus (+)	Ghana, Sweden	2019-2021	[206-208]
HLA-DPA1*0201	Rubella (−)	Slovenia, US	2009-2020	[3, 91, 159, 170, 171, 173]
HLA-DPB1*0301	Rubella (−)	Slovenia, US	2009-2020	[3, 91, 159, 171-173]
RARB rs6793694	Rubella (−)	US	2010-2014	[3, 162, 164, 165]
TRIM5 rs3740996	Rubella (−)	US	2010-2014	[3, 159, 162, 164, 165]
ADAR rs2229857	Rubella (+)	US	2010-2015	[162, 163, 167, 168]
HLA-DPB1*0401	Rubella (+)	US	2005-2014	[3, 159, 171-173]
HLA-DPB1*1501	Rubella (+)	US	2005-2014	[3, 171-173]

(+) = a positive association with vaccine immunogenicity; (−) = a negative association with vaccine immunogenicity

^*includes the meta-analysis Li et al. 2013 [101], which combined data from four other included articles [26, 34, 93, 95, 105].

^#includes the meta-analysis Ou et al. 2021 [102], which combined data from six other included articles [32, 36, 45, 50, 56, 103].

^$includes the meta-analysis Posteraro et al. 2014 [3], which combined data from seven other included articles [23, 26, 49, 95, 104, 106, 148].

Figure 1. Odds of Increased Immunogenicity by Commonly Studied Genetic Determinants.

Based on pooled odds ratios from three meta-analyses: Li et al. 2013 [101], Ou et al. 2021 [102], and Posteraro et al. 2014 [3]. All genetic determinants were associated with Hepatitis B vaccine. The following genetic determinants were also associated with influenza and MMR vaccines: HLA-DRB1*07, HLA-DQA1*0201, HLA-DQB1*0201, HLA-DQB1*0303, HLA-DRB1*13, and HLA-DRB1*13:01.

Results

Genetic associations with immunogenicity of hepatitis B vaccine

We identified 92 articles showing an association between host genetics and hepatitis B vaccine immunogenicity: 65 case-control studies [17-80], 11 reviews [81-91], 9 cohort studies [23, 92-99], 1 cross-sectional study [100], and 3 meta-analyses [3, 101, 102] (Appendix 2). Differences in immunogenicity were associated with 277 genetic determinants across 117 genes. HLA alleles accounted for 174 of these determinants.

The determinants found to have an impact on immunogenicity in three or more studies were all HLA alleles (Table 1). The presence of DRB1*07 [3, 26, 28, 34, 49, 59, 82, 83, 85, 86, 88, 91, 101], DPB1*0501 [3, 36, 45, 56, 77, 78, 81, 85, 88, 91, 98, 102], DR3 [42, 44, 46, 47, 64, 69, 86, 89, 91, 103], DR7 [22, 32, 42, 46, 64, 69, 83, 86, 96, 103], and C4A*Q0 [3, 32, 46, 65, 68, 86, 91, 104] alleles were negatively associated with immunogenicity. The presence of DRB1*15 [3, 17, 82, 85, 86, 88, 91-93, 101], DQB1*0602 [3, 33, 34, 36, 83, 85, 88, 101, 105], DRB1*0803 [3, 29, 36, 45, 99], DPB1*0402 [3, 36, 45, 78, 88, 99, 102], and DPB1 single nucleotide polymorphism (SNP) rs9277535 [37, 43, 55, 81, 88, 94] alleles were positively associated with immunogenicity.

Genetic associations with immunogenicity of influenza vaccine

We identified 25 articles demonstrating associations between host genetics and influenza vaccine immunogenicity: 12 case-control studies [106-117], 5 reviews [118-124], 5 cohort studies [125-129], and 1 meta-analysis [3] (Appendix 3). Differences in immunogenicity were associated with 48 genetic determinants across 34 genes. HLA alleles accounted for 20 of these: A11*01 [109, 123], A68*01 [109, 123], DPB1*0401 [3, 107, 123, 124], DQB1*0502 [109, 123], DQB1*0603-9/14 [3, 106, 118, 120, 122, 123], DR3, DR4, DR3/4 [116], DRB1*0401 [107, 121, 123], DRB1*1104 [109], DRB1*13 [3, 106, 124], DRB1*15/16 [108], DRB1*1601 [109, 123], and DRB3*0X [3, 106] were positively associated with immunogenicity; DQB1*0303 [3, 106,124], DRB1*07 [3, 106, 120-124], DRB1*1303 [109, 123], DR-DQB1*7/4-0302 [108, 123], DR-Y*7-11,13,14 [108, 123], and W16 [110, 123, 127] were negatively associated with immunogenicity. The determinants found to have an impact on immunogenicity in three or more studies were all HLA alleles except IFITM3 rs12252 C/C (Table 1).

Genetic associations with immunogenicity of measles, mumps, and rubella vaccines

We identified 33 articles showing associations between host genetics and measles vaccine immunogenicity: 26 cohort studies [130-138][139-156], 3 case-control studies [152, 157, 158], 3 reviews [90, 91, 159], and 1 meta-analysis [3] (Appendix 4). All but two of these studies were conducted in the US. Differences in vaccine immunogenicity were associated with 291 genetic determinants across 118 genes. HLA alleles accounted for 60 of these. The determinants found to have an impact on immunogenicity in three or more studies were all HLA alleles except CD46 rs2724384 (Table 1).

We identified 3 cohort studies showing an association between host genetics and mumps vaccine immunogenicity [151, 156, 160]. Differences in vaccine immunogenicity were associated with 29 genetic determinants across 10 genes.

We identified 22 cohort studies showing an association between host genetics and rubella vaccine immunogenicity [133, 151, 154, 156, 161-178]. Differences in vaccine immunogenicity were associated with 311 genetic determinants across 110 genes. HLA alleles accounted for 84 of these. The determinants found to have an impact on immunogenicity in three or more studies were all HLA alleles except ADAR-rs2229857, RARB-rs6793694, and TRIM5-rs3740996 (Table 1).

Genetic associations with immunogenicity of diphtheria, tetanus, and pertussis vaccines

We identified 11 cohort studies showing associations between host genetics and diphtheria [179-182], tetanus [181-185], and/or pertussis [186-189] vaccine immunogenicity (Appendix 5). These studies were conducted in the US [181, 183, 184], the Netherlands [186, 188], Australia [179, 180], Finland [187, 189], Italy [182], and Papua New Guinea [185]. Differences in vaccine immunogenicity were associated with 4 genetic determinants across 3 genes (diphtheria), 5 genetic determinants across 4 genes (tetanus), and 12 genetic determinants across 10 genes (pertussis). Most of these determinants were positively associated with immunogenicity. None of these determinants were associated with immunogenicity in multiple studies.

Genetic associations with immunogenicity of Hib, meningococcal, and pneumococcal vaccines

We identified 13 articles showing associations between host genetics and vaccine immunogenicity for Haemophilus influenzae type B (Hib) [190-192], meningococcal [193-197] and/or pneumococcal [181, 184, 191, 192, 198-200] (Appendix 6). Differences in vaccine immunogenicity were associated with 3 genetic determinants (Hib), 13 genetic determinants across 5 genes (meningococcal), and 25 genetic determinants across 14 genes (pneumococcal). All 3 associations with Hib were negative, and all but one association with pneumococcal was negative; about half of the associations with meningococcal were negative. None of these determinants were associated with immunogenicity in multiple studies.

A genetic association with Hib vaccine immunogenicity was identified in 2 cohort studies [191, 192] and 1 case-control study [190] published between 1979 and 2010 and conducted in the US and UK. The genetic association with meningococcal vaccine immunogenicity was identified in 3 cohort studies [194-196], 1 case control study [193], and 1 experimental study [197], published between 2004 and 2014 and conducted in the US and the UK. The genetic association with pneumococcal vaccine immunogenicity was identified in 6 cohort studies and 1 experimental study published between 2004 and 2014 and conducted in the US and the Netherlands [181, 184, 191, 192, 198-200].

Genetic associations with immunogenicity of poliomyelitis and rotavirus vaccines

We identified 8 articles showing associations between host genetics and vaccine immunogenicity for poliomyelitis [201] and/or rotavirus [202-207] (Appendix 7). Differences in vaccine immunogenicity were associated with 4 genetic determinants across 3 genes (poliomyelitis), and 2 genetic determinants across 2 genes along with Lewis Blood Group and Secretor Status (rotavirus). All but one of these determinants were positively associated with immunogenicity. None of these determinants were associated with immunogenicity in multiple studies with the exception of Blood Group and Secretor status (rotavirus) [203, 204, 206-208] (Table 1).

The genetic association with poliomyelitis vaccine immunogenicity was identified in a cohort study conducted in Finland in 2003 [201]. The genetic association with rotavirus vaccine immunogenicity was identified in 3 cohort studies [202-204], 1 case-control study [205], 1 randomized control trial [207], and 1 review paper [208], published between 2017 and 2021 and conducted in Sweden [208], Malawi [204], Nicaragua [202], Pakistan [203], Ghana [207], and China [205].

Odds of altered immunogenicity by genetic determinant, based on meta-analyses

Between 3 meta-analyses, pooled odds ratios (ORs) were compiled for 10 genetic determinants commonly associated with immunogenicity of vaccines such as hepatitis B, influenza, and MMR (Figure 1). In a 2013 meta-analysis [101], HLA-DRB1*15 had 2.29 (95% Confidence Interval: 1.61-3.27) and HLA-DQB1*0602 had 3.32 (95%CI: 1.80-6.15) times the odds of an increased response to hepatitis B vaccine, respectively. In a 2021 meta-analysis [102], the HLA-DPB1*0501 allele had 0.73 (95%CI: 0.69-0.78) and the HLA-DPB1*0402 allele had 4.20 (95%CI: 2.70-6.52) times the odds of a good antibody response to hepatitis B vaccine, respectively.

In a 2014 meta-analysis [3], HLA-DRB1*07 had 2.46 (95%CI: 1.60-3.77), HLA-DQA1*0201 had 2.21 (95%CI: 1.22-4.00), HLA-DQB1*0201 had 2.03 (95%CI: 1.35-3.07), HLA-DQB1*0303 had 3.31 (95%CI: 1.12-9.78), HLA-DRB1*13 had 0.52 (95%CI: 0.32-0.84), and HLA-DRB1*13:01 had 0.19 (95%CI: 0.06-0.58) times the odds of a decreased antibody response (to hepatitis B, influenza and MMR vaccines), respectively.

Genetic associations with vaccine safety

We identified 10 articles showing an association between host genetics and adverse events following immunization (AEFI): 7 with influenza vaccines [13, 112, 209-213] and 3 with measles vaccines [91, 214, 215] (Appendix 8). Seven of these articles were case control studies [13, 209, 211-215] and 3 were reviews [91, 112, 210]. Four articles identified associations with narcolepsy after influenza vaccine [13, 209-211]; two identified associations with febrile seizures after measles vaccine [91, 214], and one each identified associations with Guillain-Barré Syndrome (GBS) after influenza vaccine [212], giant cell arteritis/polymyalgia rheumatica (GCA/PMR) after influenza vaccine [213], high temperature after influenza vaccine [112], and fever after measles vaccine [215]. Differences in adverse reactions to influenza vaccines were associated with 95 genetic determinants across 46 genes.

The determinants found to have an association with narcolepsy after influenza vaccination in multiple studies were: TRA*rs12587781 [13, 210, 211], HLA-DQB1*0302 [13, 209], HLA-DQB1*0501 [13, 209], HLA-DQB1*0602 [13, 209-211], and HLA-DQB1*0603 [13, 209-211]. HLA-DQB1*0603, HLA-DQB1*0302, and HLA-DQB1*0501 had a protective effect, while the rest increased the likelihood of narcolepsy after vaccination. Narcolepsy studies were conducted in Sweden and the US, and their sample sizes ranged from 2,023 [13] to 7,032 [209]. The determinant most frequently and strongly associated with narcolepsy was HLA-DQB1*0602; in Swedish populations, the HLA-DQB1*0602 allele increased the odds of narcolepsy by a factor of up to 49 (95%CI: 12-202) [13, 209].

Discussion

This scoping review identified many genetic associations with heterogeneity in vaccine immunogenicity, and several genetic associations with heterogeneity in AEFI. Many of these associations included genes and genetic determinants involving the HLA genes. Most of the results summarized here were not reported in multiple studies, indicating the science of vaccinomics is nascent and evolving. This review can serve as a resource for other investigators interested in replicating and triangulating data obtained using various methods to identify, confirm, or contradict genetic associations described here.

Most data were extracted from studies conducted in Asia, Europe, and North America. The vaccine studied most commonly was hepatitis B; the next most studied vaccines were measles, rubella, and influenza. Other vaccines (e.g., mumps, diphtheria, tetanus, pertussis, poliovirus, Hib, pneumococcal, and meningococcal) each had fewer than 10 studies identifying genetic determinants of their immunogenicity. Many genetic associations with immunogenicity of hepatitis B and measles vaccines were validated in multiple studies; most associations with other vaccines were not. Six AEFI were identified, including 4 with influenza vaccination (narcolepsy, GBS, GCA/PMR, high temperature), and 2 with measles vaccination. The most cited association was between HLA-DQB*0602/0603, narcolepsy, and Pandemrix^® [13, 209-211].

The most common finding among the included studies was a genetic impact on the immunogenicity of hepatitis B vaccine, whether via a reduction or an improvement in immune response. Most of these findings involved individual genes, allotypes, extended haplotypes, or SNPs; however, advances in systems biology, immunology, and pharmacogenomics indicate that the interaction of multiple genetic factors is most likely to influence immune response [216, 217]. Vaccinomics has the potential to further elucidate the complexity of these relationships as the field grows.

The number and type of genetic associations with immunogenicity identified varied substantially between vaccines. This is likely mostly due to differences in which genes and vaccines have been studied more to this point. However, as further research in vaccinomics is conducted and synthesized, we may find that some such differences between vaccines in the number and type of genetic associations are truly indicative of differences in vaccines and not just of differences in research and/or funding priorities. Further study of these differences could help to illuminate specific causes of certain genetic associations, perhaps even inspiring ways to mitigate them.

This review has several limitations. Although much of the information presented has global relevance, this review focused on vaccines routinely recommended for the general US population, with the exception of Pandemrix^® [14]. This restriction was meant to limit the scope of the review to vaccines the authors were most familiar with and keep the number of articles manageable for a small team. Expanding the review to include vaccines not routinely used in the US will be an important next step to broaden our understanding of genetic heterogeneity in vaccine safety and immunogenicity. We focused exclusively on significant associations to limit the scope of the review thus we do not detail non-significant results, which increases the potential for publication bias. We did not evaluate study quality or how many independent populations were investigated due to resource constraints and the difficulty of determining these data. We limited our search to articles published in English. Our search was limited to the PubMed database, so we could have missed relevant articles only indexed in other databases. We did not perform a forensic analysis on the 15 additional articles included based on reviewing the reference lists of included systematic reviews, which could have elicited the reasons these articles were missed in the initial search. Genetics vary considerably by race/ethnicity and the studies we reviewed may underrepresent genetic traits associated with heterogeneity in vaccine immunogenicity among racial/ethnic minorities who are historically under-represented in research [218]. Few of the studies we reviewed underwent replication studies – a key component of understanding true genetic associations. However, this review comprehensively covers the most common vaccines protecting against thirteen infectious diseases and provides a solid foundation for those pursuing vaccinomics research especially those studying specific gene variants.

Despite the paucity of studies of genetic associations with vaccine safety, vaccinomics/adversomics could be a powerful tool for assessing causality of AEFI, both on the individual and population levels. A primary example of a potential genetic association with an AEFI is narcolepsy following the AS03 adjuvanted pandemic H1N1 2009-2010 vaccine (Pandemrix^®). Narcolepsy is strongly associated with genetics: 88–95% of narcolepsy patients with cataplexy are HLA-DQB1*06:02 positive although this association is weaker among those without cataplexy [219]. The prevalence of the HLA-DQB1*06:02 genotype is 25-30% in Northern Europeans and 25% in Chinese persons, much higher than in the rest of the world’s population [219]. A 2018 meta-analysis indicated Pandemrix^® increased the risk of narcolepsy in the year after vaccination by a factor of 5-14 in children and adolescents and 2-7 in adults [220] but significant country-level associations were limited to northern Europe – namely, Finland [219, 221], Sweden [210, 222], and England [223-225] – and, when data were combined with countries from other parts of Europe and the world, the significance of these associations disappeared [222]. Narcolepsy diagnoses also increased in China and Taiwan following the 2009-2010 influenza season yet an AS03 adjuvanted vaccine was not used there [226]. In Canada, the same AS03 adjuvant included in Pandemrix^® was used in a different H1N1 vaccine, Arepanrix^®, yet only two narcolepsy cases were reported [219]. This suggests that the risk of narcolepsy may increase after exposure to the H1N1 antigen and/or the Pandemrix^® vaccine and, though the AS03 adjuvant itself may not induce narcolepsy, it may enhance the Pandemrix^® antigen as a trigger [219, 226]. This also suggests that the HLA-DQB1*06:02 genotype may be a necessary but, by itself, insufficient cause of narcolepsy while also requiring an environmental trigger (like H1N1 infection or vaccination with Pandemrix^®) to induce the immune-mediated process [219]. Further exploration of this association could have important implications for vaccine development going forward, especially for use of the AS03 adjuvant which is currently included in several investigational products [227, 228].

Tailoring vaccine development and use to genetic subgroups of the population through vaccinomics is a natural extension of the current influenza vaccination paradigm, i.e., “individualized vaccinology”. Multiple influenza vaccines are licensed in the US for specific subgroups allowing for a personalized vaccinology approach [9]: high-titer inactivated influenza vaccines aim to increase immunogenicity in the elderly, live attenuated influenza nasal sprays are predominantly for children and needle-phobic adults, and recombinant and cell-based inactivated influenza vaccines are available for everyone else [229]. A relatively simple opportunity to further personalize vaccines is to tailor by sex, as females respond more strongly to vaccination than males on average both by developing higher antibody responses and by experiencing more AEFIs [230].

Genetic traits imperceptible to the naked eye could eventually influence many aspects of a provider's care for their patients, making such screening more likely as technology advances and costs continue to decrease – often driven by drug treatment. Genetic screening could even be incorporated into the newborn screening that is already conducted [231]. As genetic testing becomes more commonplace and knowledge of genetic predispositions to increased/decreased vaccine immunogenicity and safety evolve through advancements in vaccinomics, we may be able to further refine vaccine recommendations by subgroup to enhance vaccine immunogenicity and safety. Although such refinement carries the downside of complexity and thus could increase the mental and organizational burden on health care providers, advances in computer technology, artificial intelligence, and the proliferation of electronic records may be able to lighten this load. Vaccine confidence could increase if segments of the population who worry about genetic dispositions to rare side effects knew they could be screened to mitigate this risk [232, 233]. For example, specific haplotypes of the IL1, IL18, and IL4 genes are associated with a significantly increased risk of fever following smallpox vaccination and may be responsible for an increased risk of fever and febrile seizures following MMR or other vaccinations [7]. If these associations are determined to be causal, pre-vaccination screening could significantly decrease the occurrence of fever and febrile seizures in children [7] either by adjusting the dose for those testing positive for these genetic traits or by advising preemptive treatment such as immediate post-vaccination medication with NSAIDs or other fever reducers. In addition, knowledge of such associations, and the mechanisms behind such findings, allow the opportunity to reverse engineer new vaccine candidates to overcome or circumvent genetic barriers to protective immune responses or safety issues.

Admittedly, this vision of routine genetic screening and personalized vaccinology is highly speculative, and not necessarily indicated by the initial data presented in this review. There are also many barriers to overcome and counterarguments to consider before such a vision could or should be widely implemented. Further vaccinomics research may find that the prevalence of relevant mutations and/or the clinical relevance of their effects are so limited that personalizing vaccines based on them is not cost effective even given eventual reductions in associated costs. Much of the public may forego genetic testing due to privacy concerns, concerns which could be valid without adequate legal and data security measures in place. And the added complexity of personalized vaccinology could lead to increased errors in vaccine production or administration unless protected against with regular training and/or procedural checks, potentially decreasing vaccine confidence. Still, the nascent field of vaccinomics shows potential for improving vaccine efficacy and safety, and warrants further research and policy consideration.

Conclusion

This scoping review of vaccinomics identified numerous genetic associations with vaccine immunogenicity. Most of these associations were only identified in a single study and very few studies examined genetic associations with vaccine safety, illustrating both the potential of vaccinomics and the need for further investment in the field. Advancements in vaccinomics could bolster vaccine immunogenicity and safety, allow for reverse engineering of new vaccine candidates, and increase public confidence in vaccines.

Highlights.

• Many genetic associations with vaccine immunogenicity have been identified

• Most genetic associations identified with Hepatitis B, MMR, and influenza vaccines

• Genetic associations were identified with 6 adverse events following immunization

• Vaccinomics is a growing field with great potential but need for investment

• Progress in vaccinomics could improve vaccine safety and effectiveness