Reduced susceptibility to COVID-19 associated with ABO blood group and pre-existing anti-A and anti-B antibodies

Abstract

Background

Susceptibility to severe acute respiratory syndrome coronavirus 2 shows individual variability in un-vaccinated and previously un-exposed individuals. We investigated the impact of ABO blood group, titers of anti-A and anti-B, other blood group antigens, and the extracellular deposition of ABH antigens as controlled by secretor fucosyltransferase 2 (FUT2) status.

Study design and methods

We studied incidents in three different hospitals between April to September 2020, where un-diagnosed coronavirus disease 2019 (COVID-19) patients were cared for by health care workers without use of personal protection and with close contact while delivering therapy. We recruited 108 exposed staff, of whom 34 were diagnosed with COVID-19. ABO blood type, titer of anti-A and -B, blood group specific alleles, and secretor status were determined.

Results

Blood group O was associated with lower risk of COVID-19 (OR 0.39, 95 %CI (0.16–0.92), p = 0.03) compared to non-O, i.e., blood groups A, B and AB. High titer anti-A immunoglobulin G (IgG) compared to low titer was associated with lower risk of COVID-19 (OR 0.24 95 %CI (0.07–0.78), p = 0.017). High titer of anti-B immunoglobulin M (IgM) compared to no anti-B (IgM) was associated with lower risk of COVID-19 (OR 0.16, 95 %CI (0.039–0.608), p = 0.006) and the same applies to low titer anti-B (IgM) compared to no titer (OR 0.23, 95 %CI (0.07–0.72), p = 0.012).

The 33Pro variant in Integrin beta-3, that is part of human platelet antigen 1b (HPA-1b), was associated with lower risk of COVID-19 (OR 0.23, 95 %CI (0.034–0.86), p = 0.028).

Conclusion

Our data showed that blood group O, anti-A (IgG) titer, anti-B (IgM) titer as well as HPA-1b are associated with lower risk for COVID-19.

1. Introduction

Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) causing coronavirus disease 2019 (COVID-19) was discovered in late 2019 and has since evolved into a global pandemic. Elucidation of the mechanisms underlying individual variability in susceptibility to the virus is especially of great interest to optimize treatment and prevention.

ABO blood group has previously been reported to be associated with susceptibility to several viruses including SARS-CoV-2 (Severe Covid et al., 2020, Barnkob et al., 2020, Zhao et al., 20212021, Zietz et al., 2020, Cheng et al., 2005). However, knowledge of the mechanism behind the variability in susceptibility is still limited. It has been demonstrated that viral particles carry A and B blood group antigens corresponding to the producing host cell’s ABO type (Deleers et al., 2021, Gerard et al., 2020, König-Beihammer et al., 2021). Thus, pre-existing anti-A and/or anti-B in a virus recipient might reduce the infectivity, and the number of viral particles, and susceptibility to infection might be inversely associated with antibody level. Protection against COVID-19 infection would thus be linked to ABO blood group incompatibility between the virus-producing host and the recipient. A derived consequence would be that in populations with very different frequencies of ABO, the protection may be quantitatively different (Guillon et al., 2008, Marionneau et al., 2001). From the perspective of the individual, a minority of blood group O individuals would be well protected in a population with a high fraction of blood groups A or B as they encounter incompatible viral particles. And, in contrast, the same blood group O individuals mixed in a population of predominantly blood group O, would encounter viral particles carrying neither A nor B antigens, and thus have no targets for anti-A and anti-B.

Another cause of variable susceptibility to infection by noro- and rotaviruses has been shown to be the recipient secretor status encoded by the fucosyltransferase 2 (FUT2) gene. The FUT2 gene facilitates high levels of ABH antigens in secretions, in airway and gastrointestinal tract fluids and on mucosal surfaces (Marionneau et al., 2001). It has been demonstrated by genome-wide association study (GWAS) that blood group O, non-secretors are less susceptible to SARS-CoV2 (OR 0.69) compared to blood group non-O, secretors (Nishida et al., 2022).

It has been suggested that SARS-CoV-2 may also use integrins as a cellular receptor and may bind through the conserved viral tripeptide arginine-glycine-aspartic acid (Arg-Gly-Asp/RGD) amino acid motif that is present in the receptor-binding domain of the spike proteins of all published sequences of isolates of SARS-CoV-2 (Sigrist et al., 2020). Integrins are heterodimeric transmembrane surface molecules that function in cell-to-cell interaction, cell migration, and intercellular signaling processes. Viral proteins with RGD motifs promote infection by binding integrin heterodimers such as αVβ1, αVβ3, αVβ5, αVβ6, αVβ8, α5β1, α8β1 and αIIbβ3. The integrin subunit beta 3 (ITGB3) gene codes for the protein product integrin beta-3 chain which is found in platelets. The single nucleotide polymorphism (SNP) rs5918 (Leu33Pro: substitution of leucine by proline) in the ITGB3 is responsible for the platelet antigen polymorphism designated human platelet antigen 1a and 1b (HPA-1a and -b) (Hussein et al., 2015).

Interaction with integrins by the viral RGD tripeptide motif is assumed to be central for the infective mechanism of several other viruses. West Nile virus (Schmidt et al., 2013), human cytomegalovirus (Feire et al., 2004) and Kaposi's sarcoma-associated virus (Hussein et al., 2015) have been implicated as binding specifically to integrins including ITGB3.

The individual susceptibility may be influenced by recipient integrin genotype. Currently, there are no studies on integrins and susceptibility to SARS-CoV-2.

1.1. Objectives

This study aimed to examine the association between susceptibility to infection with SARS-CoV-2 and the magnitude of anti-A and anti-B titers. Also, we examined the association between susceptibility to SARS-CoV-2 and secretor status and selected blood group antigens.

2. Methods and materials

We recruited staff from three SARS-CoV-2 outbreaks that occurred between April and September 2020 at three hospitals designated NY, HIL and HVI. For an extended period, undiagnosed COVID-19 patients were cared for by health care workers without strictly using personal protective measures. Only later a positive Polymerase Chain Reaction (PCR) test diagnosed the patients with COVID-19, and strict use of protective measures was implemented. In one hospital (HVI), the participants were exposed to unconfirmed as well as to confirmed COVID-19 patients with the use of protective measures while delivering therapy within a distance less than 1 m. None of the subjects were purposefully exposed to contract COVID-19, as the status of the COVID-19 of the patients was either unknown/not initially suspected or the regulations for protective measures were limited at the time as the pandemic was in the very early stages.

We collected our samples before the availability of a vaccine and therefore this study exclusively investigates naturally occurring factors of susceptibility.

The study was carried out with approval from the ethics committee of the Capital Region of Denmark (H-20031631) and data management approval number P-2020–639. Recruitment of participants was carried out independently of the leadership of clinical departments to avoid a pressure. Recruitment and consent procedures among the centers were identical. Written informed consent was obtained from all participants.

We recruited exposed health care workers to the study. Exposure was defined as close contact under one meter distance or contact over 15 min with a COVID-positive patient. Participants were recruited at least 30 days after exposure to allow for testing and recovery. Blood was examined for ABO and RhD blood grouping, for genotyping of blood group specific alleles, for quantification of titers of anti-A and anti-B, and for COVID-19 serology by enzyme-linked immunoassay (ELISA).

From each participant a total of 3 vials of each 6 mL was drawn: 2 EDTA vials, 1 serum tube. EDTA blood was used for automated (NEO, Immucor Inc., Norcross, GA, USA) standard forward and reverse serological ABO blood grouping, and RhD blood grouping utilizing Immucor and BAG Diagnostics GmbH (D-35423 Lich, Germany) monoclonal anti-A, anti-B, and anti-D antibodies according to routine patient blood grouping.

Titers of anti-A and anti-B antibodies, Immunoglobulin G (IgG), and Immunoglobulin M (IgM), were determined by testing with the NEO analyzer (Immucor Inc., Norcross, GA, USA) of fixed dilutions of plasma as previously validated (Sprogoe et al., 2021, Pendu et al., 2021). Two different assays were used, a Solid Phase Red Cell Adherence assay (SPRCA) specifically detecting IgG, and a direct agglutination assay preferentially detecting IgM on the Neo analyzer according to the manufacturer’s instructions (Ching, 2012). The IgG assay was done in Capture-R Select microplates (catalog number 0006446, Immucor Inc.). Briefly, the wells were coated with a monolayer of either A1 or B type red blood cells (RBCs). Reagent RBCs were in-house single donation glycerol frozen-thawed heterozygous A1 and heterozygous B RBCs suspended in Cellstab (catalog number B005652, DiaMed, Cressier, Switzerland). The entire process was fully automated with an intra-laboratory variation of +/- 1 titer step and a CV% of 5.8. Serial dilutions of the participants’ plasma samples were prepared in phosphate buffered saline (PBS) to the following dilutions: 1:1, 1:2, 1:4, 1:8, 1:16, 1:32, 1:64, 1:128, 1:256, 1:512, 1:1024, 1:2048 and 1:4096. From these dilutions, 50 µL was tested in 100 µL low ionic strength solution (catalog number 0006420, Immucor Inc.). Visualization of the reaction was done by anti-IgG coated indicator RBCs (catalog number 0006428, Immucor Inc.).

The IgM assay was a direct agglutination assay using either A1 or B RBCs and 50 µL from the plasma dilutions. Absence or presence of agglutination was determined by the NEO analyzer.

For the statistical analysis we divided participants into, a no antibodies-group, and antibody-positive participants were divided into 2 equally large groups, designated low and high titer. This equals titers, none (0), low (1–8) and high (>= 16), except for anti-A IgG which included titers 1–32 in the low group and >= 64 in the high group, due to the ranges in anti-A IgG. Antibody titer 0 is not recorded for participants with blood groups where an antibody would not be expected, e.g., anti-B titer 0 is not recorded in blood group B.

Additionally, 200 µL of EDTA blood was used for genotyping for blood group and secretor status according to a previously described procedure (Krog et al., 2019). Briefly, genotypes encompassing 59 alleles enables prediction of 50 blood group antigens. 44 RBC antigens and 6 Human platelet antigens (HPAs) were determined for each participant. The following blood group antigens were amenable to analysis for association with COVID-19, Le^b, RhD, RhC, Rhc, Do^a, Fy^a, Fy^b, HPA-1b, HPA-15a, HPA-15b, HPA-5b, Jk^a, Jk^b, N, S, secretor. The rare alleles (e.g., Yta, Kpb, Lub, Vel etc.) pose an inherent methodological challenge for a study of this size and were therefore not included in the analysis. Plasma and serum were stored for supplemental use.

Furthermore, we collected limited information on past medical history and smoking status.

2.1. Statistical analyses

Descriptive statistics were used to summarize the study population. Due to the low number of subjects, ABO blood groups were grouped in blood group O versus non-O, the latter including A, B, AB, for the analysis of ABO blood group and association to susceptibility. Odds Ratios (OR) were estimated by logistic regression with 95% confidence intervals based on profile likelihoods. Anti-A/B titres were categorized and analyzed a priori. The p-values are based on likelihood ratio tests from logistic regression models. For the statistical analyses, we used R version 4.1.2. software (R Core Team 2021, Vienna, Austria, https://www.R-project.org/).

3. Results

A total of 108 SARS-CoV-2 exposed healthcare workers were included, with 51, 39 and 18 subjects from each hospital, respectively ( Table 1 ). Among the participants, 34 had contracted COVID-19. We compared risk of being infected with COVID-19 with staff from NY as a reference. Staff from HVI had lower risk of being infected OR 0.27 (95% CI(0.09–0.69), p-value 0.006, and the same trend was observed for HIL OR 0.35 (95% CI(0.09.1.12), p-value 0.079. No difference was observed for male or female participants and risk of COVID-19 (28.6 vs 32.0%). Smokers had a lower risk of COVID-19 (16.7% vs 33.7%) (Table 1).

Table 1.

Participant Data and COVID status.

	COVID-19, n = 34	Non-COVID-19, n = 74	Total
Male, n (%)	2 (29%)	5 (71%)	7
Female, n (%)	32 (32%)	68 (68%)	100
Missing		1	1
Smoker, n (%)	2 (17%)	10 (83%)	12
Non-smoker, n (%)	32 (34%)	63 (66%)	95
Missing		1	1
Comorbidity, n (%)
Pulmonary	0 (0%)	1 (100%)	1
No pulmonary diseases	34 (32%)	73 (68%)	107
Cardiovascular	1 (50%)	1 (50%)	2
No Cardiovascular diseases	33 (31%)	73 (69%)	106
Diabetes Mellitus	1 (100%)	0 (0%)	1
No Diabetes Mellitus	33 (31%)	74 (69%)	107
Hypertension	4 (100%)	0 (0%)	4
No hypertension	30 (29%)	74 (70%)	104
Centers, n (%)
Nykobing (n = 51)	23 (45%)	28 (55%)	51
Hvidovre (n = 39)	7 (18%)	32 (82%)	39
Hillerod (n = 18)	4 (22%)	14 (78%)	18

3.1. ABO blood group and susceptibility to COVID-19

Blood group O compared to non-O was associated with a lower risk of contracting COVID-19 (21% vs 40%) (OR 0.39, 95 %CI (0.16–0.92) p = 0.031). That is, blood group O is estimated to have 61% lower odds of being infected compared to non-O ( Table 2 ).

Table 2.

Analysis of association between blood groups and susceptibility to COVID-19.

Blood groups	COVID-19 (n = 34)	Non-COVID-19 (n = 74)	OR (95% CI)	p-value
A	14 (31.8%)	30 (68.2%)
B	9 (64.3%)	5 (35.7%)
AB	1 (50.0%)	1 (50.0%)
Non-O*	24 (40.0%)	36 (60.0%)	Reference
O	10 (20.8%)	38 (79.2%)	0.39 (0.16–0.92)	0.031
Non-secretor	11 (42%)	15 (58%)	Reference	-
Secretor	20 (26%)	56 (74%)	0.49 (0.19–1.25)	0.13
RhD negative	9 (50%)	9 (50%)	Reference	–
RhD positive	25 (28%)	65 (72%)	0.39 (0.14–1.09)	0.07
HPA-1b negative	31 (36%)	56 (64%)	Reference	–
HPA-1b positive	2 (11%)	16 (89%)	0.23 (0.03–0.86)	0.03

*Non-O combines ABO blood groups A, B and AB.

There are missing data on secretor status for 3 participants in the COVID-19 group and 3 in the Non-COVID-19 group.

There are missing data on HPA-1b status in 1 participant in the COVID-19 group and 2 in the Non-COVID-19 group.

Analysis of further blood group antigens are presented in supplementary.

3.2. Levels of pre-existing anti-A and anti-B and susceptibility to COVID-19

High levels of anti-A and anti-B antibodies were associated with lower odds of contracting COVID-19, for IgG as well as for IgM ( Table 3 ). This was statistically significant for anti-A (IgG) when comparing high levels versus low levels (OR 0.24 95 %CI (0.07–0.78) p = 0.017). Furthermore, anti-B (IgM) was significantly associated with a lower risk of contracting COVID-19, both when comparing low levels with none (OR 0.23, 95 %CI (0.07–0.72), p = 0.012), as well as when comparing high levels to none (OR 0.16, 95 %CI (0.04–0.61), p = 0.006). For anti-A (IgM) and anti-B (IgG) the same trend was observed, with high level of antibodies being associated with lower odds of COVID infection. Statistical significance was obtained only for the former 2 comparisons (Table 3).

Table 3.

Analysis of pre-existing natural antibody titers and isotypes and the association with risk of COVID-19 calculated as OR with susceptibility of the no antibody group as the reference.

	COVID-19	Non-COVID-19	OR (95% CI)	p-value
Anti-A, IgG				0.047*
None	17	31	Reference	–
Low levels	12	15	1.46 (0.55, 3.84)	0.442
High levels	5	26	0.35 (0.10, 1.02)	0.055
High vs. low		–	0.24 (0.07, 0.78)	0.017
Missing data		2
Anti-A, IgM				0.467*
None	15	31	Reference	–
Low levels	12	20	1.24 (0.48, 3.20)	0.656
High levels	7	23	0.63 (0.21, 1.75)	0.379
High vs. low	–	–	0.51 (0.16, 1.51)	0.224
Anti-B, IgG				0.180*
None	15	19	Reference	–
Low levels	8	22	0.46 (0.15, 1.30)	0.144
High levels	11	32	0.44 (0.16, 1.13)	0.088
High vs. low	–	–	0.95 (0.33, 2.80)	0.917
Missing data		1
Anti-B, IgM				0.016*
None	10	6	Reference
Low levels	18	46	0.23 (0.07, 0.72)	0.012
High levels	6	22	0.16 (0.04, 0.61)	0.006
High vs. low	–	–	0.70 (0.22, 1.93)	0.496

Antibody titer levels are divided into three groups: none (0), low (1–8 (32)) and high (≥16 (64)). Antibody titers are not noted for blood groups where they would not be expected (e.g., anti-B in blood group B etc.).

*Likelihood ratio test of the null hypothesis that the risk of COVID infection is the same in all three groups (none/low/high).

3.3. Secretor status and susceptibility to COVID-19

There is no evidence that the risk for COVID infection is different for secretors and non-secretors (p = 0.13) (Table 2).

3.4. Blood group antigens and susceptibility to COVID-19

Selected blood group antigens of the cohort are presented in Table 2, and the risk for COVID-19 is compared to the blood group antigen-negative individuals as a reference. The 33Pro variant in ITGB3, that is part of platelet antigen HPA-1b, was associated with lower susceptibility (11.1% vs 35.6%), OR 0.23 95 %CI (0.03–0.86), p = 0.028. However, only 2 participants were HPA-1b positive and infected with COVID-19.

The RhD positive blood group was associated with a lower susceptibility for COVID-19 in comparison with RhD negative blood group (27.8% vs 50%), OR = 0.39, 95 %CI (0.14–1.09), however, not significantly so (p = 0.07). (Table 2).

We also determined additional blood group antigens (Le^b, RhC, Rhc, Do^a, Fy^a, Fy^b, HPA-15a, HPA-15b, HPA-5b, Jk^a, Jk^b, N, S) and examined association with susceptibility for COVID-19. However, no other blood group antigen - besides ABO and the antigen HPA-1b presented in Table 2 - was found to be associated with susceptibility for COVID-19. The results of the analysis of additional blood group antigens and association with risk of COVID-19 are presented in the Supplementary Table.

4. Discussion

In this study we used incidents of unprotected exposure to examine the possible protective role of pre-existing natural anti-A and anti-B antibodies, secretor status, and blood group antigens to infection with SARS-CoV-2. We tested the frequency of infection in the cohort of hospital health care workers exposed to SARS-CoV-2 while caring for COVID-19 patients. Majority of the subjects in this study were female, as are the gender ratio in our healthcare system. We sorted participants according to level of pre-existing anti-A and anti-B antibody, carriers and non-carriers of several erythrocyte and platelet blood group antigens and compared infection rates. The inspiration to this study was provided by a similar setup used in the early SARS epidemic in 2003 where exposure to virus from a single patient caused a similar accidental exposure and infection scenario that demonstrated a lower frequency of infection in participants with blood group O (Cheng et al., 2005).

4.1. ABO blood group antigen, pre-existing anti-A and anti-B antibodies and COVID-19

Previous studies have found blood groups to be associated with specific outcomes of disease processes, most extensively studied in infectious diseases (Cheng et al., 2005, Lindesmith et al., 2003, Wang et al., 2012, Rowe et al., 2007, Tiongco et al., 2018, Foster and Labrum, 1976). In accordance with previous studies (Cheng et al., 2005, Barnkob et al., 2020, Zhao et al., 20212021), we found blood group O to be significantly associated with lower susceptibility to infection with SARS-CoV-2 compared to non-O.

Viral particles carry the ABH glycosylation corresponding to the ABO blood group antigens of the virus-producing human host-cell (Deleers et al., 2021, Arendrup et al., 1991). This is a potential target for protective interaction with recipient pre-existing anti-A and anti-B, in case of ABO major incompatibility (Deleers et al., 2021, Gerard et al., 2020, Guillon et al., 2008). There is a variation in titers between individuals of the same blood group as shown in Table 3, and individuals with lower antibody levels have a higher rate of COVID-19. Accordingly, a recent study found significantly lower anti-A/anti-B antibodies in COVID-19 patients compared to healthy controls (Deleers et al., 2021).

Here, we found anti-A of the IgG isotype to be a protective factor and we found an enhanced protective effect when comparing high levels versus low levels of antibody (Table 3). Pre-existing anti-A antibodies of the IgG isotype are found in 89% of blood group O individuals (Stussi et al., 2005) and to a lesser extent in blood group B. Efficacy of resistance might vary depending on the isotype of the anti-A and anti-B thus accounting for the observed difference between IgG and IgM antibodies with specificity for the prevalent blood group A. We also found anti-B (IgM) to be associated with lower susceptibility to SARS-CoV-2. This antibody is found in individuals of blood group O and A. Blood group A plus O is represented by 85% of individuals, which might explain the significant finding for anti-B (IgM).

We speculate that the protective interaction with pre-existing antibodies takes place anatomically in secretions of airways as well as in the airway tissues, and in the cytoplasm of epithelial and other recipient cells hosting the virus with antibodies. The cytoplasmic Fc receptor TRIM21 is a likely cytoplasmic candidate of an effector molecule responsible for directing virus-antibody immune complexes to degradation in the proteasome, and for subsequent presentation for cytotoxic T cells of peptides derived from the antigen (Caddy et al., 2021). Pre-existing anti-A and anti-B could thus act as adjuvants eliciting T cell mediated immunity to SARS-CoV-2 derived peptides. Thus, the immediate effect of reducing susceptibility could be accompanied by a potential long-term anti-viral immunity.

From the population perspective of viral susceptibility, the observed resistance in the individuals with blood group O has the effect to slow down the dissemination of ABO incompatible non-O virus particles, and thereby to decrease the spreading of COVID-19 in all individuals (Guillon et al., 2008, Ellis, 2021). For this protective mechanism to be effective group O recipients should encounter ABO non-compatible virus particles. Similarly, a recent study by Boukhari et al., found ABO incompatibility to be significantly associated with decreased risk of contracting SARS-CoV-2 (Boukhari et al., 2021).

From a methodological point of view, it is interesting that GWAS and epidemiological studies have identified the ABO locus as correlated to susceptibility. Our results are in accordance with the interpretation that the mechanistically important prerequisite for resistance is the possession, and more precisely the levels, of pre-existing natural antibodies, anti-A, and anti-B, in association with ABO major incompatibility between virus-producing host and recipient.

4.2. Additional blood group antigens

We observed that the blood group RhD was associated with reduced susceptibility for COVID-19, (OR 0.39, 95 %CI (0.14–1.09), p = 0.07), however, not significantly so. A significant correlation has previously been reported by others (Rahim et al., 2021).

The single nucleotide polymorphism (SNP) rs5918 in ITGB3 is responsible for the platelet polymorphism designated HPA-1a and HPA-1b. Also, due to the combination of integrin subunit beta-3 (b₃) with the integrin subunit alpha-V (a_V), the SNP rs5918 potentially could influence several distinct integrins located on separate cell types and tissues. We found that the presence of “C” at position 176 of the ITGB3 gene in SNP rs5918, corresponding to variant 33Pro in ITGB3, was associated with a significantly lower risk of contracting COVID-19. Previous studies have demonstrated an association between rs5918 genotyping and susceptibility to Hanta virus. Specifically, the rs5918 CC genotype was found to be associated with protection and the rs5918 TT genotype was associated with increased susceptibility (Matthys et al., 2010). This SNP has not been demonstrated in the published GWASs of COVID-19.

4.3. Strengths and limitations of this study

The strengths of our study lie in the use of incidents of exposure to SARS-CoV-2 in 2020 before vaccination was introduced. The cohort consists of a relatively homogeneous population of essentially healthy hospital staff, and with the majority being female. It may have resulted in lower risk of confounding by gender.

A major limitation of our study is the small sample size. This potentially causes the wide 95% CI of all ORs, including those of anti-A (IgM) og anti-B (IgG), and a possible lack of statistical significance for some blood groups, such as RhD. Furthermore, we were unable to include all exposed staff at each of the centers, participation in our study was voluntary, and not all exposed healthcare providers may have reached out to be included in the study. Like previous studies, this study was not designed to explore the mechanisms behind the susceptibility, but to explore blood type groups and antibodies responsible for susceptibility to COVID-19. Overall, we aimed to create a hypothesis-generating study for future investigations.

Another limitation of our study is the examination of the effect of many blood group antigens (Le^b, RhC, Rhc, Do^a, Fy^a, Fy^b, HPA-15a, HPA-15b, HPA-5b, Jk^a, Jk^b, N, S) besides ABO, HPA-1b, secretor, and RhD. This increases the risk of finding false associations between blood groups and risk of COVID-19. We decided not to correct for multiple comparisons.

Our findings on anti-A and anti-B, and the 33Pro in ITGB3 variant should therefore be reproduced in other cohorts.

5. Conclusions

We reproduced the finding that blood group O is associated with lower susceptibility to COVID-19.

A dose–response relationship was observed between anti-B (IgM) and of anti-A (IgG) and susceptibility to SARS-CoV-2. Furthermore, we found that the 33Pro variant of ITGB3 was associated with a lower risk of contracting COVID-19.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data availability

Data will be made available on request.