Relationship Between ABO Blood group Phenotypes and nCOVID-19 Susceptibility – A Retrospective Observational Study

Sudhir Bhandari SMS Medical College and Hospitals, Jaipur, Rajasthan, India Ajeet Singh Shaktawat SMS Medical College and Hospitals, Jaipur, Rajasthan, India Amit Tak (  dramittak@gmail.com ) SMS Medical College and Hospitals, Jaipur, Rajasthan, India https://orcid.org/0000-0003-2509-2311 Bhoopendra Patel Government Medical College, Barmer, Rajasthan, India Jyotsna Shukla SMS Medical College and Hospitals, Jaipur, Rajasthan, India Jitendra Gupta SMS Medical College and Hospitals, Jaipur, Rajasthan, India Shivankan Kakkar SMS Medical College and Hospitals, Jaipur, Rajasthan, India Amitabh Dube SMS Medical College and Hospitals, Jaipur, Rajasthan, India Sunita Dia Medstar Washington Hospital Center, Washington DC 20010, USA. Mahendra Dia North Carolina State University, Raleigh, NC 27695-7609, USA. Todd C Wehner North Carolina State University, Raleigh, NC 27695-7609, USA.


Introduction
According to the World Health Organisation report on 16 November 2020, there have been 54,075,995 con rmed cases of COVID-19, including 1,313,919 deaths globally. 1 The novel coronavirus 2019 (nCOVID- 19) has spread beyond the boundaries of Wuhan, China across the world. Researchers in many countries are engaged in discovering associations between nCOVID-19 mortality, severity and susceptibility and, other attributes such as demographic, clinical, and laboratory ndings. 2 A number of diseases have been found associated with various blood group phenotypes.
The ABO blood group antigens are widely distributed throughout the body along with the surface of red blood cells. 3 The association of ABO groups with malaria, pancreatic cancer, duodenal ulcer and other diseases has been elucidated. [3][4][5] Microorganisms interact with antibodies against blood group antigens, including ABO, T and Kell systems. 6 The objective of this study was to evaluate the relationships between the ABO, Rh, and Kell blood group systems and susceptibility for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.

Patients And Methods
A hospital-based, retrospective observational study was designed to evaluate ABO, Rh, and Kell blood group phenotypes and its relationship to susceptibility to SARS-CoV-2 infection. Laboratory con rmed, real-time reverse transcription polymerase chain reaction (RT-PCR) positive cases of COVID-19 numbering 132 were enrolled in the study from SMS Medical College and Hospitals in Jaipur (Rajasthan, India). The hospital provided medical care to patients of various states in North India, including Rajasthan, Harayana, Madhya Pradesh, Uttar Pradesh and Punjab. 7,8 The mean age of the participants was 36.68 years (SD = 17.87 years) and the men to women ratio was 8:3. The ABO, Rh, and Kell blood group phenotypes along with demographic data from the COVID-19 patients were collected from laboratory reports. The ABO, Rh (D antigen) and Kell phenotyping was carried out using fully automated solid phase red cell adherence (SPRCA) technology. 9 The observed frequencies of the blood groups -'A', 'AB', 'B' and 'O', 'Rh positive', 'Rh negative', 'Kell positive', and 'Kell negative' were calculated. The general (expected) proportions of the above blood groups were obtained from a study of the North Indian population. 10 The expected proportions of ABO blood groups 'A', 'AB', 'B' and 'O' were 22.3 %, 8.9 %, 39.2 % and 29.6 %, respectively ( Figure 1). Rh (D) phenotype were found positive in 93.8 % and negative in 6.2 %. For the Kell system, only 1.6 % were Kell positive, and the remaining were Kell negative. The expected frequencies of the blood group antigens were calculated by multiplying the total number of COVID-19 cases (n =132) with the general proportions of the respective blood groups (Table 1).

Statistical analysis
The observed frequency of blood group phenotypes of ABO, Rh and Kell system were compared against the expected population frequency using goodness of t chi-square test. The further hypotheses were tested using partition analysis in case of the ABO system. 11 In order to nd the effect of sample size on Type II error, post hoc power analysis was performed. The level of signi cance was considered at 5 %. Software used for the analysis was JASP (version 0.12.2) software and the Real Statistics package of MS Excel 2010. 12,13 Ethical aspects The study was conducted in accordance with the Declaration of Helsinki (as revised in 2013). The study was approved by Ethical Committee of the SMS Medical College, Jaipur (No.416 dated 26 Jun 2020) and individual consent for this retrospective analysis was waived.

Results
The observed frequencies of ABO, Rh and Kell blood group phenotypes with 95 % con dence intervals were calculated ( Figure 2). The comparison of the observed and expected frequencies of blood group phenotypes are shown in Figure 3.
Comparison of phenotypes of ABO Blood group system for COVID-19 susceptibility The goodness of t of observed and expected frequencies of the ABO blood groups were tested using Pearson's chi-squared test (Table 1 and Figure 3). There were signi cant differences between expected frequency and observed frequency in at least one of the blood group phenotypes (p = 0.013). The chi-squared test (χ2) does not reveal in which phenotype the difference actually exists. If the observed frequency is less than the expected frequency in one cell, it has to be more in one or the other cells because the total for both the observed and the expected frequencies is the same. The observed frequency was higher in blood group 'A' and lower in blood groups 'AB', 'B', and 'O'. In order to con rm the difference in observed frequencies across blood groups, a partition analysis was performed. ABO blood group phenotypes were further divided into two sub-groups and two hypotheses were tested using the chi-squared test (Table 2). First, the observed and expected frequencies of blood groups 'AB', 'B', and 'O' were tested and no signi cant difference was found (p = 0.89). Second, the observed and expected frequencies of blood group 'A' and 'non-A' were tested and found to differ signi cantly (p = 0.001). Table 1 was partitioned based on those comparisons and the χ 2 was calculated ( It can be inferred from the above analysis that persons with blood group 'A' were more susceptible to COVID-19 than 'non-A' groups. Further, there were no signi cant relationships between the 'AB', 'B', and 'O' blood group phenotypes for susceptibility to COVID-19.

Comparison of Rh (D antigen) System for COVID-19 susceptibility
The observed and expected frequencies of Rh blood group phenotypes showed no statistically signi cant differences (p = 0.51) ( Table 3).

Comparison of Kell antigen System for COVID-19 susceptibility
The observed and expected frequencies of Kell blood group phenotypes showed statistically signi cant differences (p < 0.001) ( Table 4). The Kell negatives were more susceptible to COVID-19 infection compared to Kell positives.

Discussion
The blood group antigens are an example of polymorphic traits inherited among individuals and populations. There are 34 recognised human blood groups and hundreds of individual blood group antigens and alleles. The variation in blood group system leads to variations in host susceptibility to many infections. Microorganisms interact with antibodies against blood group antigens, including ABO, T, and Kell systems. 6 The blood group antigens of the H, ABO, Lewis, and historical 'P' blood groups contains small carbohydrate epitopes expressed as post-translational modi cations on glycoproteins, mucins, and glycolipids. The ABO system (International Society of Blood Transfusion (ISBT) 001) contains two structurally related carbohydrate antigens, A and B. The O or H antigen is the biosynthetic precursor of A and B antigens and is classi ed as a separate blood group system (ISBT 018). All three antigens consist of 2 to 3 terminal oligosaccharides on glycoproteins and glycolipids. 14 The varying rates of evolution of COVID-19 in two populations may be due to different gene frequency distributions of the blood group and may lead to differences in susceptibility to  In order to evaluate the relationship between COVID-19 susceptibility and ABO blood group phenotypes, the null hypothesis H 0 that the blood group ratio The results showed that H 0 was rejected( Table 1). The observed frequency of blood group 'A' was found to be higher than the expected frequency. Also, observed frequencies in other blood groups were lower than the expected frequencies. If observed frequency is higher than expected frequency in one cell, it has to be lower in one or more of the other cells since the total for both the observed and expected frequencies is the same. Thus, rejecting H 0 does not indicate which observed frequency is different from the expected one. In order to check whether the observed frequency of blood group 'A' is higher than the expected frequency further partition analysis was performed, and two more hypotheses were tested. The corresponding hypotheses were: The results showed there was no signi cant difference in observed ratios of AB : B : O blood groups when compared with expected (p = 0.89). However, there was a signi cant difference in the observed ratio A : non-A blood groups when compared with expected (p = 0.001) (  17 In other recent ndings, Ziadi (2020) reported a decreased e ciency of adhesion of the Spike protein to the angiotensin converting enzyme 2 (ACE2) receptor by antibody A, as suggested in other studies. He further stated that a lower susceptibility of blood group 'B' and 'O' was the case, but it did not explain susceptibility of the 'AB' blood group without anti A and anti B in the blood serum. 18 In contrast to the above, a study of 397 patients reported an increase in COVID-19 infection associated with the 'AB' blood group. 19 However, the other ndings of Abdollahi (2020) regarding no signi cant relationship between Rh phenotypes with COVID-19 susceptibility were consistent with the results presented in this study. 19 A study using data from anti-A and anti-B antibodies rather than ABO blood group reported subjects with anti-A (ie, B and O blood groups) were signi cantly less susceptible to COVID-19 than those lacking anti-A regardless of group whereas there was no signi cant difference for anti-B in the serum. 20 Similarly, Guillon et al (2008) and Gustafsson et al (2005) argued that either a monoclonal anti-A antibody or natural plasma anti-A present in blood group 'O' speci cally inhibited the SARS-CoV S protein/ACE2-dependent adhesion to ACE2-expressing cell lines. Therefore, ABO polymorphism could contribute to substantially reducing SARS-CoV transmission. 21,22 The results from Gustafsson et al (2005) con rmed that the glycans have more potential to carry variations than proteins and nucleic acids. The well-known example is polymorphic terminal glycosylation of the ABO blood group family of antigens. The association between infectious diseases and ABO antigens at mucosal surfaces leads to differential adherence of pathogens. Another way the blood group types might affect susceptibility is interaction with the coagulation system. Some factors like von Willebrand factor and factor VIII affect in vivo half-life and clearance of blood type antigens. 23 Similarly, Dai et al (2020) studied the interaction among COVID-19 cases, hypertension, and ABO blood grouping. They reported that hypertensive patients the renin-angiotensin-aldosterone system (RAS) was overexpressed due to inhibitors of angiotensin converting enzyme (ACE2). The expression of ACE2 receptors was upregulated in hypertensives. The ACE2 receptor was the primary entry site for SARS-CoV-2. Dai (2020) mentioned that the ABO blood group was associated with ACE activity and ACE inhibitors induced cough.
The GATC haplotype of the ABO gene polymorphism is prevalent among the non-O blood type patients and is positively associated with an ACE activity. Therefore, O blood type carriers have lower ACE levels and are less susceptible to COVID-19 infection. [24] In a recent genome-wide association study between COVID-19 and disease severity (respiratory failure) 1,980 patients were included from Italy and Spain.
Ellinghaus et al (2020) analysed 8,582,968 single-nucleotide polymorphisms and conducted a metaanalysis of the two case-control panels. They found signi cant cross-replicating associations with rs11385942 at locus 3p21.31 and with rs657152 at locus 9q34.2. The association signal at locus 9q34.2 coincided with the ABO blood group locus. In a blood-group-speci c analysis, blood group A showed higher risk than in other blood groups (odds ratio, 1.45; 95% CI, 1.20 to 1.75) and blood group O showed a protective effect as compared to other blood groups (odds ratio, 0.65; 95% CI, 0.53 to 0.79). [25] There may be other mechanisms responsible for higher susceptibility of particular blood group phenotypes that require further studies.

Conclusion
The study shows a relationship between ABO blood grouping and COVID-19 susceptibility, with group A being more susceptible. The application of these relationship in clinical practice requires more exploratory studies.

Limitations of the study
The expected proportions used in this study lacked information regarding age and sex. Therefore, a multivariate analysis to adjust the effects of the two factors was not possible. However, this may not be necessary as since the distributions of blood group phenotypes were similar across age and gender. The sample size was not enough to provide powerful interpretation of Rh and Kell blood phenotypes and COVID-19 susceptibility.