Advanced searchCitation search
Skip main navigation
Close Drawer MenuOpen Drawer Menu
Original ResearchMarch 2021

Association Between ABO and Rh Blood Groups and SARS-CoV-2 Infection or Severe COVID-19 Illness

FREE
A Population-Based Cohort Study
    Author, Article and Disclosure Information
    https://doi.org/10.7326/M20-4511

    Visual Abstract. Association Between ABO and Rh Blood and SARS-CoV-2 or Severe COVID-19

    This cohort study used data from Ontario, Canada, to examine whether ABO and Rh blood groups are associated with risk for SARS-CoV-2 infection and severe COVID-19 illness.

    Abstract

    Background:

    The ABO and rhesus (Rh) blood groups may influence risk for severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection.

    Objective:

    To determine whether ABO and Rh blood groups are associated with risk for SARS-CoV-2 infection and severe coronavirus disease 2019 (COVID-19) illness.

    Design:

    Population-based cohort study.

    Setting:

    Ontario, Canada.

    Patients:

    All adults and children who had ABO blood group assessed between January 2007 and December 2019 and who subsequently had SARS-CoV-2 testing between 15 January and 30 June 2020.

    Measurements:

    The main study outcome was SARS-CoV-2 infection, determined by viral RNA polymerase chain reaction testing. A second outcome was severe COVID-19 illness or death. Adjusted relative risks (aRRs) and absolute risk differences (ARDs) were adjusted for demographic characteristics and comorbidities.

    Results:

    A total of 225 556 persons were included, with a mean age of 54 years. The aRR of SARS-CoV-2 infection for O blood group versus A, AB, and B blood groups together was 0.88 (95% CI, 0.84 to 0.92; ARD, −3.9 per 1000 [CI, −5.4 to −2.5]). Rhesus-negative (Rh−) blood type was protective against SARS-CoV-2 infection (aRR, 0.79 [CI, 0.73 to 0.85]; ARD, −6.8 per 1000 [CI, −8.9 to −4.7]), especially for those who were O-negative (O−) (aRR, 0.74 [CI, 0.66 to 0.83]; ARD, −8.2 per 1000 [CI, −10.8 to −5.3]). There was also a lower risk for severe COVID-19 illness or death associated with type O blood group versus all others (aRR, 0.87 [CI, 0.78 to 0.97]; ARD, −0.8 per 1000 [CI, −1.4 to −0.2]) and with Rh− versus Rh-positive (aRR, 0.82 [CI, 0.68 to 0.96]; ARD, −1.1 per 1000 [CI, −2.0 to −0.2]).

    Limitation:

    Persons who rapidly died of severe COVID-19 illness may not have had SARS-CoV-2 testing.

    Conclusion:

    The O and Rh− blood groups may be associated with a slightly lower risk for SARS-CoV-2 infection and severe COVID-19 illness.

    Primary Funding Source:

    Ontario Academic Health Sciences Centre AFP Innovation Fund and the Ontario Ministry of Health and Long-Term Care.

    Type O blood may protect against severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (1–3). Non–peer-reviewed data from New York City suggest that rhesus-negative (Rh−) status may also be protective (4). Prior studies did not distinguish between SARS-CoV-2 infection and severe coronavirus disease 2019 (COVID-19)–related illness, were not done at a population level within a universal health care environment, comprised few patients typically recruited within a single hospital setting (5, 6), and were prone to ascertainment bias. These studies did not use appropriate control participants, such as those who had also undergone SARS-CoV-2 testing, and typically did not consider important confounders, especially preexisting comorbidities.

    The uncertainty around ABO or Rh blood groups and SARS-CoV-2 infection persists (7). Accordingly, this population-based study was done to evaluate SARS-CoV-2 infection and severe COVID-19 illness in relation to ABO and Rh status.

    Methods

    This population-based, retrospective cohort study was done across Ontario, Canada, which has universal health care. Existing patient-level data sets for all of Ontario capture all hospitalizations, all emergency department visits, and most laboratory tests for SARS-CoV-2 (Supplement Table 1). These data sets were linked using unique encoded identifiers and analyzed at ICES. All Ontario hospitals submit demographic and clinical information about all inpatient admissions and discharges, including transfers and deaths, using standard diagnostic and procedure codes from the International Classification of Diseases, 10th Revision with Canadian Enhancements. The Canadian Institute for Health Information's Discharge Abstract Database comprises all hospital admissions, and the National Ambulatory Care Reporting System database captures all emergency department visits (Supplement Table 1). The Ontario Health Insurance Plan claims database identifies preexisting health conditions using an International Classification of Diseases, 9th Revision, diagnostic code for every outpatient visit. The Ministry of Health's Registered Persons Database contains vital status and demographic information for all persons ever eligible for the Ontario Health Insurance Plan. Residential income quintile and rural residence were identified using Statistics Canada's Census data (Supplement Table 1).

    Table 1. Characteristics of 225 386 Persons in Ontario, Canada, With Known ABO Blood Group Who Subsequently Had SARS-CoV-2 Viral RNA Polymerase Chain Reaction Testing Between 15 January and 30 June 2020

    Table 1.

    Study entry required that a person had ABO blood group assessed between January 2007 and December 2019—before any known international cases of COVID-19—and then subsequently had SARS-CoV-2 viral RNA polymerase chain reaction (PCR) testing between 15 January and 30 June 2020. Persons were excluded if they were not tested for SARS-CoV-2, if sex or birth date were missing, and if they were not Ontario residents.

    The main study outcome was SARS-CoV-2 infection (a positive result for SARS-CoV-2 on viral RNA PCR testing). The test date was the date on which the SARS-CoV-2 specimen was collected. For any given person, if more than 1 positive specimen date was available, then the earliest specimen with a positive result was considered; otherwise, the earliest specimen with a negative result was considered.

    A second outcome was a composite of severe COVID-19 illness or death. Severe illness was defined as an admission to an intensive care unit, hospitalization with a length of stay of 7 days or more, or diagnosed myocardial infarction or viral pneumonia, each within 14 days before or after the SARS-CoV-2 test date. Death was captured from 1 day before to up to 14 days after the SARS-CoV-2 test date. We permitted death to precede testing by 1 day to allow for specimen labeling that may occur on the next calendar day.

    Statistical Analysis

    Baseline characteristics, such as demographics, prior pregnancy, and preexisting health conditions, were assessed relative to the SARS-CoV-2 specimen collection date and presented by ABO blood type.

    For each study outcome, analyses were done by each ABO blood group (with A as the reference group), O blood group versus all others (reference group), Rh− versus Rh-positive (Rh+) (reference group) status, and O-negative (O−) versus all other ABO and Rh+ blood groups (reference group).

    Unadjusted probabilities (percentages and 95% CIs) of SARS-CoV-2 infection, as well as severe COVID-19 illness or death, were each estimated in relation to ABO and Rh blood groups. Next, relative risks (RRs) and absolute risk differences (ARDs) based on marginal probabilities of the outcome of interest (also called population-average probabilities of success for exposed and unexposed participants) were calculated using a modified approach to logistic regression analysis developed by Austin (8). The 95% CIs were estimated by bootstrapping with resampling 1000 times (8). The RRs and ARDs were adjusted for age, sex, area-level income quintile, rurality, and local health integration network, each at the time of the SARS-CoV-2 test; for a history of cardiac ischemia or arrhythmia, cancer, or chronic kidney disease diagnosed within 5 years before the SARS-CoV-2 test; and for congestive heart failure or diabetes mellitus diagnosed anytime before the SARS-CoV-2 test (9, 10).

    For the main outcome of SARS-CoV-2 infection, analyses were further stratified by age younger than 70 years versus 70 years or older (9, 10), and age was excluded from those related multivariable models.

    In 1 additional analysis, the association between ABO or Rh blood group and risk for severe illness or death was reanalyzed. This analysis was limited to those who tested positive for SARS-CoV-2 and used the modified logistic regression approach described earlier (8). It is known that some persons with SARS-CoV-2 may initially have a false-negative viral RNA PCR test result (11, 12), which may mean that some persons with severe COVID-19 illness could have a negative swab. Accordingly, another additional analysis used multinomial logistic regression to assess the relation between ABO or Rh blood groups and the adjusted odds ratio (aOR) of SARS-CoV-2 negativity with severe illness or death, SARS-CoV-2 positivity without severe illness or death, and SARS-CoV-2 positivity with severe illness or death versus SARS-CoV-2 negativity without severe illness or death (the baseline category). Adjusted models were run with the same covariates as in the aforementioned modified logistic regression models.

    Persons who have ABO testing may differ from those who do not. Accordingly, we assessed all persons who had SARS-CoV-2 testing in Ontario during the study period and contrasted the characteristics and outcomes among those whose ABO blood group was known versus not known before SARS-CoV-2 testing, expressed as standardized differences.

    Statistical significance was set at a 2-sided P value of less than 0.05, and analyses were planned a priori. Statistical analyses were done using SAS, version 9.4 for UNIX (SAS Institute). PROC LOGISTIC was used to compute adjusted RRs (aRRs) and ARDs from logistic regression models using a marginal probabilities approach (https://support.sas.com/resources/papers/proceedings11/345-2011.pdf).

    Role of the Funding Source

    The funding sources played no role in the design, conduct, or reporting of this study.

    Results

    Among 2 659 328 persons who had an ABO blood group test from January 2007 to December 2019, a total of 2 432 155 did not have a SARS-CoV-2 laboratory test in the subsequent period of observation (Appendix Figure 1). In total, 225 556 persons were included in the final cohort. Of these, 36.3% had blood type A, 4.5% had type AB, 14.9% had type B, and 44.3% had type O (Table 1). The proportion with Rh− status was 13.1%. Mean age was 53.8 years, and about 29% were men.

    Appendix Figure 1. Study flow diagram.

    Rh(D) = rhesus D; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

    Within 5 years before SARS-CoV-2 specimen collection, about 13% to 15% of persons had preexisting cardiac disease, 11% had chronic kidney disease, 21% had anemia, and 27% to 29% had cancer (Table 1). Asthma (18% to 21%), chronic obstructive pulmonary disease (13% to 17%), and heart failure (10% to 11%) were prevalent, in addition to dementia or frailty (33% to 38%), diabetes mellitus (21%), and chronic hypertension (39% to 41%) at any preceding time.

    SARS-CoV-2 Infection

    The lowest unadjusted probability of SARS-CoV-2 infection was among the O− blood group (2.1% [95% CI, 1.8% to 2.3%]), and the highest was in the B-positive blood group (4.2% [CI, 4.0% to 4.5%]) (Table 2).

    Table 2. Unadjusted Probabilities of SARS-CoV-2 Infection or Severe COVID-19 Illness or Death, Each in Relation to ABO and Rh Blood Groups

    Table 2.

    The aRR for SARS-CoV-2 infection was higher with blood type AB than with type A (1.15 [CI, 1.03 to 1.28]) and B (1.21 [CI, 1.13 to 1.29]) and slightly lower with type O (0.95 [CI, 0.91 to 1.01]) (Table 3). The aRR was 0.88 (CI, 0.84 to 0.92; ARD, −3.9 per 1000 [CI, −5.4 to −2.5]) when comparing O versus all other blood groups. An Rh− status seemed protective against SARS-CoV-2 infection (aRR, 0.79 [CI, 0.73 to 0.85]; ARD, −6.8 per 1000 [CI, −8.9 to −4.7]), especially for those who were O− (aRR, 0.74 [CI, 0.66 to 0.83]; ARD, −8.2 per 1000 [CI, −10.8 to −5.3]). In analyses stratified by age, the relative protective effects of O, Rh−, and O− blood groups were more pronounced among those younger than 70 years than those aged 70 years or older (Appendix Figure 2).

    Table 3. Association Between ABO, Rh, and ABO-Rh Blood Groups and Risk for SARS-CoV-2 Infection

    Table 3.
    Appendix Figure 2. Risk for SARS-CoV-2 infection associated with ABO, Rh, and ABO-Rh blood groups, stratified by age <70 or ≥70 y.

    Relative risks are adjusted for sex, area-level income quintile, rurality, and local health integration network, each at the time of SARS-CoV-2 testing, as well as for any history of congestive heart failure, cardiac ischemia or arrhythmia, cancer, diabetes mellitus, or chronic kidney disease diagnosed before the SARS-CoV-2 specimen collection date. Rh = rhesus; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2.

    Severe COVID-19 Illness

    There were 1328 cases of COVID-19 with severe illness or death, with higher probabilities among AB and B blood groups as well as those who were Rh+ (Table 2). A breakdown of the components of the composite outcome of severe COVID-19 illness or death is shown in Supplement Table 2.

    Those with blood type B were at significantly higher risk for severe illness than those with type A (aRR, 1.21 [CI, 1.04 to 1.40]; ARD, 1.2 per 1000 [CI, 0.2 to 2.2]) (Table 4). The aRR was 0.87 (CI, 0.78 to 0.97; ARD, −0.8 per 1000 [CI, −1.4 to −0.2]) when comparing O blood group versus all others (Table 4). Compared with Rh+ blood type, Rh− had a lower aRR of severe COVID-19 illness or death (0.82 [CI, 0.68 to 0.96]). Type O− blood did not seem protective against severe COVID-19 illness or death (aRR, 0.84 [CI, 0.64 to 1.07]) (Table 4).

    Table 4. Association Between ABO, Rh, and ABO-Rh Blood Groups and Risk for Severe COVID-19 Illness or Death

    Table 4.

    In the first additional analysis, restricted to 7071 persons who tested positive for SARS-CoV-2, there was no observed association between blood group and the risk for severe illness or death (Table 5). However, in a more thorough analysis of all 225 556 patients, including those with a negative SARS-CoV-2 test result, type O blood versus others was protective against SARS-CoV-2 positivity without severe illness or death (aOR, 0.89 [CI, 0.84 to 0.94]) and also SARS-CoV-2 positivity with severe illness or death (aOR, 0.87 [CI, 0.78 to 0.97]) (Table 6). A similar pattern was seen for Rh− status, with respective aORs of 0.80 (CI, 0.73 to 0.87) and 0.82 (CI, 0.68 to 0.98), as well as for O− blood group, with respective aORs of 0.72 (CI, 0.63 to 0.83) and 0.84 (CI, 0.65 to 1.08) (Table 6).

    Table 5. Association Between ABO, Rh, and ABO-Rh Blood Groups and Risk for Severe COVID-19 Illness or Death, Restricted to 7071 Persons Who Tested Positive for SARS-CoV-2 Infection

    Table 5.

    Table 6. Association Between ABO, Rh, and ABO-Rh Blood Groups and Risk for Being SARS-CoV-2–Negative With Severe Illness or Death, SARS-CoV-2–Positive Without Severe Illness or Death, or SARS-CoV-2–Positive With Severe Illness or Death, Each in Contrast to Having Neither SARS-CoV-2 Infection nor Severe Illness or Death

    Table 6.

    Among those who had SARS-CoV-2 testing in the study period, there were some differences in those whose antecedent ABO blood group was known or unknown (Supplement Table 3). For example, those with a known blood type were slightly older, were less likely to be male, and had more comorbidities than those whose ABO status was unknown. However, subsequent rates of SARS-CoV-2 infection or related severe illness did not differ appreciably by ABO status (Supplement Table 3).

    Discussion

    In this study, which was done within a universal health care system with widespread SARS-CoV-2 testing, O and Rh− blood groups were associated with a slightly lower risk for SARS-CoV-2 infection as well as severe COVID-19 illness or death.

    Our study has several strengths. Testing policies for SARS-CoV-2 have evolved during the Canadian pandemic, from mostly symptomatic persons to broader population screening (https://bit.ly/3e2ar0U); yet, no directive was based on a person's blood group, and the observed subsequent rates of SARS-CoV-2 infection were similar by antecedent ABO test status (Supplement Table 3, lower rows). Persons with early SARS-CoV-2 infection may have a false-negative viral RNA PCR test result (11, 12) and become severely ill days later. This possibility was handled by our second additional analysis considering the outcome of SARS-CoV-2 negativity with severe illness or death, in which there was no important variation in that outcome by blood group (Table 6).

    Our study also has limitations. Selection bias was reduced by the requirement that ABO status precede SARS-CoV-2 testing and by further covariate adjustment. Even so, it is possible that those most susceptible to severe COVID-19 illness, such as an elderly resident living in a long-term care facility, died before arriving at the hospital for SARS-CoV-2 testing or died without antecedent symptoms of COVID-19 illness (13). This could be true given that a protective effect from O and Rh− blood type was less pronounced in those older than 70 years (Appendix Figure 2). If O or Rh− blood type is truly protective against SARS-CoV-2 infection, then it is possible that a person who was type O or Rh− would remain asymptomatic and thus not even have viral testing. Accordingly, a study from a setting in which universal screening was done may optimally handle some of these potential issues related to selection bias.

    Our study findings align with prior work. In 1 study from China of 2173 patients with COVID-19 and 27 080 unmatched control participants, the unadjusted OR for COVID-19 was 0.67 (CI, 0.60 to 0.75) when comparing O versus non-O blood groups (2). In a recent study of 1610 Italian and Spanish patients with COVID-19 and 2205 unmatched control participants, the age- and sex-adjusted OR for mechanical ventilation was 0.65 (CI, 0.53 to 0.79) when comparing O versus other blood groups (3). A non–peer-reviewed study from NewYork–Presbyterian Hospital comprised 14 112 patients who were tested for SARS-CoV-2 and whose age was about 57 years; 62% were women (4). After adjustment for ethnicity, the risk for SARS-CoV-2 infection was no different between ABO blood types but was lower for Rh− (RR, 0.85 [CI, 0.73 to 0.96]). The risk for intubation did not differ by ABO or Rh status, but the risk for death was lower among Rh− patients (RR, 0.44 [CI, 0.21 to 0.74]) (4).

    A retrospective study from Turkey comprised 227 patients who tested positive for SARS-CoV-2 on PCR and another 165 possible cases on the basis of computed tomography scans, who were compared with historical population controls (6). The unadjusted OR of SARS-CoV-2 infection was 0.28 (CI, 0.17 to 0.48) in the presence of Rh− blood type. A cross-sectional study at a single Iranian hospital compared 397 admitted patients with PCR-diagnosed COVID-19 to 500 patients with negative COVID-19 blood samples obtained from outpatient and inpatient services (5). The age- and sex-adjusted ORs were 0.68 (CI, 0.50 to 0.92) when comparing O blood group versus others and 0.91 (CI, 0.58 to 1.43) when comparing Rh− versus Rh+. Moreover, among the 397 hospitalized patients with COVID-19, neither O (aOR, 1.17 [CI, 0.73 to 1.89]) nor Rh− (aOR, 0.70 [CI, 0.33 to 1.49]) blood type was associated with a lower risk for admission to the intensive care unit versus a general ward (5). Taken together, the current body of evidence (some still lacking peer review) suggests that O and Rh− blood types may protect against SARS-CoV-2 infection and, possibly, severe COVID-19 illness.

    Our findings may have implications for clinicians and policymakers. If O or Rh− blood type is associated with SARS-CoV-2 infection, then that effect is likely small, which should not undermine the importance of other public health and therapeutic measures aimed at reducing viral transmission or progression to severe COVID-19 illness.

    This study was done within a multiethnic Canadian province, but participant ethnicity was not known. Among 3.1 million American blood donors, O− was seen in 8.0% of White non-Hispanic donors, 3.9% of Hispanic donors, 3.6% of Black non-Hispanic donors, and 0.7% of Asian donors (14). Given that Black and Asian persons may be at increased risk for SARS-CoV-2 infection and, possibly, worse clinical outcomes than White persons (15), future large-scale epidemiologic research should consider contrasting the risk for SARS-CoV-2 infection between different subpopulations on an international and regional level. For example, the Basque people of Spain are more likely to be O and Rh− (16) and thus would be expected to be at lower risk for SARS-CoV-2 infection. Other epidemiologic research could expand our current knowledge by measuring not only ABO and Rh status but also the presence of erythrocyte alloimmunization—namely, the formation of antibodies against non–self-antigens on erythrocytes (17).

    Although we saw a statistically significant association between blood group and severe disease or death, it was also assumed that we had correctly identified severe illness associated with COVID-19. For example, our composite outcome did not include venous thromboembolism, which is a well-described complication of COVID-19 (18). It is also of interest that the O blood group phenotype and genotype is associated with a decreased risk for venous thromboembolism, possibly because O group members have lower plasma levels of procoagulant factor VIII and von Willebrand factor (19). Thus, the moderating effect of ABO status on venous thromboembolism risk among patients with COVID-19 should be tested.

    Studies of the accuracy of serologic tests for anti–SARS-CoV-2 immunoglobulins (20) may assess whether there is variation in antibody titers by ABO and Rh status. Furthermore, among ongoing clinical trials of immunotherapy using convalescent plasma or of SARS-CoV-2 vaccines (21), the interaction between participant blood groups and therapeutic efficacy could be measured.

    In conclusion, type O blood may be associated with a lower risk for SARS-CoV-2 infection and severe COVID-19 illness or death. At most, a small proportion of SARS-CoV-2 infection or related illness in the entire population could be prevented by some undetermined property conferred by O blood type and, perhaps, further enhanced by Rh− status. Whether this information can influence COVID-19 prevention or treatment strategies remains to be determined.

    References

    • 1. Cheng Y , Cheng Y , Cheng G , et al. ABO blood group and susceptibility to severe acute respiratory syndrome [Letter]. JAMA. 2005;293:1450-1. [PMID:15784866] MedlineGoogle Scholar
    • 2. Zhao J, Yang Y, Huang HP, et al. Relationship between the ABO blood group and the COVID-19 susceptibility. medRxiv. Preprint posted online 27 March 2020. doi:10.1101/2020.03.11.20031096 Google Scholar
    • 3. Ellinghaus D , Degenhardt F , Bujanda L , et al; Severe Covid-19 GWAS Group. Genomewide association study of severe Covid-19 with respiratory failure. N Engl J Med. 2020;383:1522-1534. [PMID: 32558485] doi:10.1056/NEJMoa2020283 CrossrefMedlineGoogle Scholar
    • 4. Zietz M, Zucker JE, Tatonetti NP. Testing the association between blood type and COVID-19 infection, intubation, and death. medRxiv. Preprint posted online 10 September 2020. doi:10.1101/2020.04.08.20058073 Google Scholar
    • 5. Abdollahi A , Mahmoudi-Aliabadi M , Mehrtash V , et al. The novel coronavirus SARS-CoV-2 vulnerability association with ABO/Rh blood types. Iran J Pathol. 2020;15:156-160. [PMID: 32754209] doi:10.30699/ijp.2020.125135.2367 CrossrefMedlineGoogle Scholar
    • 6. Arac E, Solmaz I, Akkoc H, et al. Association between the Rh blood group and the Covid-19 susceptibility. Int J Hematol Oncol. 2020;30:81-86. doi:10.4999/uhod.204247 Google Scholar
    • 7. Rubin R . Investigating whether blood type is linked to COVID-19 risk. JAMA. 2020;324:1273. [PMID: 32936219] doi:10.1001/jama.2020.16516 CrossrefMedlineGoogle Scholar
    • 8. Austin PC . Absolute risk reductions, relative risks, relative risk reductions, and numbers needed to treat can be obtained from a logistic regression model. J Clin Epidemiol. 2010;63:2-6. [PMID: 19230611] doi:10.1016/j.jclinepi.2008.11.004 CrossrefMedlineGoogle Scholar
    • 9. Richardson S , Hirsch JS , Narasimhan M , et al; the Northwell COVID-19 Research Consortium. Presenting characteristics, comorbidities, and outcomes among 5700 patients hospitalized with COVID-19 in the New York City area. JAMA. 2020;323:2052-2059. [PMID:32320003] doi:10.1001/jama.2020.6775 CrossrefMedlineGoogle Scholar
    • 10. Grasselli G , Zangrillo A , Zanella A , et al; COVID-19 Lombardy ICU Network. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy region, Italy. JAMA. 2020;323:1574-1581. [PMID: 32250385] doi:10.1001/jama.2020.5394 CrossrefMedlineGoogle Scholar
    • 11. Woloshin S , Patel N , Kesselheim AS . False negative tests for SARS-CoV-2 infection—challenges and implications. N Engl J Med. 2020;383:e38. [PMID: 32502334] doi:10.1056/NEJMp2015897 CrossrefMedlineGoogle Scholar
    • 12. Kucirka LM , Lauer SA , Laeyendecker O , et al. Variation in false-negative rate of reverse transcriptase polymerase chain reaction-based SARS-CoV-2 tests by time since exposure. Ann Intern Med. 2020;173:262-267. doi:10.7326/M20-1495 LinkGoogle Scholar
    • 13. Graham NSN , Junghans C , Downes R , et al. SARS-CoV-2 infection, clinical features and outcome of COVID-19 in United Kingdom nursing homes. J Infect. 2020;81:411-419 [PMID: 32504743] doi:10.1016/j.jinf.2020.05.073 CrossrefMedlineGoogle Scholar
    • 14. Garratty G , Glynn SA , McEntire R ; Retrovirus Epidemiology Donor Study. ABO and Rh(D) phenotype frequencies of different racial/ethnic groups in the United States. Transfusion. 2004;44:703-6. [PMID: 15104651] CrossrefMedlineGoogle Scholar
    • 15. Pan D , Sze S , Minhas JS , et al. The impact of ethnicity on clinical outcomes in COVID-19: a systematic review. EClinicalMedicine. 2020;23:100404. [PMID: 32632416] doi:10.1016/j.eclinm.2020.100404 CrossrefMedlineGoogle Scholar
    • 16. Flores-Bello A , Mas-Ponte D , Rosu ME , et al. Sequence diversity of the Rh blood group system in Basques. Eur J Hum Genet. 2018;26:1859-1866. [PMID: 30089826] doi:10.1038/s41431-018-0232-1 CrossrefMedlineGoogle Scholar
    • 17. Hendrickson JE , Tormey CA . Understanding red blood cell alloimmunization triggers. Hematology Am Soc Hematol Educ Program. 2016;2016:446-451. [PMID: 27913514] CrossrefMedlineGoogle Scholar
    • 18. Connors JM , Levy JH . COVID-19 and its implications for thrombosis and anticoagulation. Blood. 2020;135:2033-2040. [PMID: 32339221] doi:10.1182/blood.2020006000 CrossrefMedlineGoogle Scholar
    • 19. Heit JA , Armasu SM , Asmann YW , et al. A genome-wide association study of venous thromboembolism identifies risk variants in chromosomes 1q24.2 and 9q. J Thromb Haemost. 2012;10:1521-31. [PMID:22672568] doi:10.1111/j.1538-7836.2012.04810.x CrossrefMedlineGoogle Scholar
    • 20. Lisboa Bastos M, Tavaziva G, Abidi SK, et al. Diagnostic accuracy of serological tests for Covid-19: systematic review and meta-analysis. BMJ. 2020;370:m2516. [PMID: 32611558] doi:10.1136/bmj.m2516 CrossrefMedlineGoogle Scholar
    • 21. Corey L , Mascola JR , Fauci AS , et al. A strategic approachto COVID-19 vaccine R&D. Science. 2020;368:948-950. [PMID: 32393526] doi:10.1126/science.abc5312 CrossrefMedlineGoogle Scholar

    Comments

    0 Comments
    Sign In to Submit A Comment
    Mostafa Saadat18 December 2020
    Importance of Rh blood groups in the current COVID-19 pandemic

    TO THE EDITOR: I read the article by Ray and colleagues (1) with great interest. A total of 7071 confirmed COVID-19 cases and 218485 non-infected participants were included in their retrospective cohort study to investigate the association of ABO and Rh blood groups with susceptibility/mortality of COVID-19. There are several case-control studies investigating the association between ABO/Rh blood groups and susceptibility/outcome of COVID-19. However, there is no cohort study. The authors should be complimented for having now addressed this point. Here, I would like to make two points. Today, Canada is recognized as an immigrant country. Although we know that the Canadian/Ontarians population composition should include a majority of European descent, there are residents of other descents whose abundance cannot be ignored. I also know that the relationship between polymorphisms and multifactorial diseases can vary from ethnicity to ethnicity (2,3). Ray and colleagues did not mention the ethnicity of their participants and did not adjust the associations for ethnicity. Providing this information can make a significant contribution to understanding the subject. Second, I reported an ecologic study investigating the association between ABO/Rh blood groups and epidemiologic parameters of COVID-19 in 86 countries. There were significant associations between COVID-19 epidemiologic parameters and both polymorphisms of Rh and ABO. Countries with higher Rh-negative blood group, showed higher prevalence/mortality of COVID-19 (4). At that time, the correlation between mortality due to COVID-19 and Rh-negative blood group was a novel finding. Although I and my colleague tried to find some evidence in favor of the independent roles of Rh and ABO blood groups as explanatory factors, due to inherent limitation of the ecologic studies, this point remained as an open question. Our findings are very similar to those in Ray and colleagues’ study. Given that Rh-negative phenotype has an interesting geographical distribution, the results reported by the Ray and colleagues are of particular importance. The Rh-negative phenotype is rare (about 1%) in East Asian populations (such as Japan, China, and Korea) and is frequent (mainly higher than 15%) in European populations. This is reminiscent of the geographical distribution of prevalence/mortality of covid-19 worldwide. Taken together, I suggest that in the future, a multicenter study be designed to identify the effects of the ABO and especially Rh polymorphisms on the following subjects: a) susceptibility/mortality of COVID-19; b) remarkable difference between East Asians and Europeans for susceptibility/mortality of COVID-19.

    Competing interests: Author state no conflict of interest.

    Research funding: None declared.

    References

    1. Ray JG, Schull MJ, Vermeulen MJ, Park AL. Association between ABO and Rh blood groups and SARS-CoV-2 infection or severe COVID-19 illness: A population-based cohort study. Ann Intern Med. 2020:M20-4511. doi: 10.7326/M20-4511

    2. Yu C, Hequn C, Longfei L, et al. GSTM1 and GSTT1 polymorphisms are associated with increased bladder cancer risk: Evidence from updated meta-analysis. Oncotarget. 2017;8:3246-3258. doi: 10.18632/oncotarget.13702

    3. Saadat M. Genetic polymorphisms of glutathione S-transferase T1 (GSTT1) and susceptibility to gastric cancer: a meta-analysis. Cancer Sci. 2006;97:505-509. doi: 10.1111/j.1349-7006.2006.00207.x 4. Ansari-Lari M, Saadat M. The morbidity and mortality of COVID-19 are associated with ABO and Rh blood groups. Eur J Prev Cardiol. 2020 Jul 7:2047487320939216. doi: 10.1177/2047487320939216

    Joel G. Ray, MD MSc, Alison L. Park, MSc9 December 2020
    Authors' Response to Alifano

    Pietro Alifano and colleagues make a noteworthy observation about the global prevalence of SARS-CoV-2 and its relation to ABO blood group dominance within each of the 109 countries they considered. However, in doing so, they have likely generated a classic example of ecological fallacy (also known as ecological inference fallacy). With ecological fallacy, correlations (r) for the same two variables can be different at the individual- and ecologic- (in this case, country) level, leading to the invalid application of aggregate results to individuals.1 At times, when such a fallacy is present, a negative r value at the aggregate level can actually be contrasted by a positive r value at the individual level.2

     

    It is unclear if Alifano and colleagues were capturing country-reported SARS-CoV-2 infections, or actual COVID-19 severe disease “cases”. It is also unknown which of the 109 included countries (there are 193 in the world) actually had available to them widespread viral PCR testing for SARS-CoV-2. At a statistical level, the type of data available to Alifano and colleagues did not permit them to generate a multilevel model, which could consider individual-level variables (e.g., age,3 obesity and diabetes) as well as country-level factors (e.g., SARS-CoV-2 test availability, or systematic reporting of infected cases).1 Accordingly, drawing ecological correlations between the number of “cases” per 1 million inhabitants of each country and the blood groups that are found in each country may generate spurious results.

     

    Joel G. Ray, MD MSc

    Alison L. Park, MSc

     

     

    References

    1. Subramanian SV, Jones K, Kaddour A, Krieger N. Revisiting Robinson: the perils of individualistic and ecologic fallacy. Int J Epidemiol 2009;38:342-60. doi: 10.1093/ije/dyn359.
    2. Robinson WS. Ecological correlations and the behaviour of individuals. Am Soc Re. 1950;15:351-57.
    3. Fisman DN, Greer AL, Tuite AR. Age Is Just a Number: A Critically Important Number for COVID-19 Case Fatality. Ann Intern Med 2020;173:762-763. doi: 10.7326/M20-4048.

     

     

     

    Pietro Alifano, Matteo Calcagnile, Marco Alifano7 December 2020
    Blood types and heterogeneity in diffusion and mortality of COVID-19.

    Dear Editor,

    The paper by Ray et al.1 provides interesting insight in understanding basis of genetic susceptibility to COVID-19 infection and its severe forms. The authors concluded that O and Rh− blood groups may be associated with a slightly lower risk for SARS-CoV-2 infection and severe illness. Clinical series on ABO groups in Covid-19 infection provided so far somewhat conflicting results, and the methodology of the study of Ray et al   (a population-based cohort study on 225,556 persons) provides a different, robust,  angle of attack to study the question. While the authors should be commended for the excellent work, generalization of results deserves some commentaries.

    ABO and Rh Blood Groups are distributed worldwide in an extremely heterogeneous manner and one would have the temptation to see if distribution would mirror COVID-19 spread and lethality. We retrieved from Wordometer2, for each country, data on number of confirmed cases and number of deaths/Million inhabitants and built an Excel file adding for each country (when information was available3, n=109, accounting for 7,081,438,437 inhabitants) the prevalence of each ABO-Rh phenotype and correlated these variables by Spearman rank tests. Worldwide, number of cases/1M was positively correlated with prevalence of A group (r=0.530,p<0.0000001) and negatively with prevalence of B group (r=-0.396,p=0.000041). Significance was not reached for O group (r=-0.185,p=0.052) which showed a trend toward protective effect. Interestingly, regardless from ABO groups, prevalence of Rh+ phenotype was negatively associated with number of cases/1M inhabitants (r=-0.641,p<0.000001). Thus, prevalence of A Rh+ (r=0.410,p=0.000023)) and, at higher extent, A Rh- (r=0.651,p<0.0000001), but also of B Rh- (r=0.489,p=0.0000004), AB Rh- (r=0.534,p<0.0000001), and O Rh- (r=0.606,p<0.0000001) was associated with increased number of cases/1M inhabitants, whereas prevalence of B Rh+ and O Rh+ were associated with lower number of cases (r=-0.454,p=0.0000028 and r=-0.348,p=0.00031, respectively). Correlation of number of deaths/1M inhabitants and ABO Rh prevalence mirrored these data: it was positive for A Rh+ (r=0.342,p=0.00038) and, at higher extent, A Rh- (r=0.617,p<0.0000001), as well as for B Rh- (r=0.439,p=0.0000056), AB Rh- (r=0.508,p=0.0000014), and O Rh-(r=0.592,p<0.0000001), but not for B Rh+ (r=-0.533,p<0.0000001) or O Rh+ (r=-0.244,p=0.011), both showing a “protective” effect.

    These simple, easily reproducible, observations would confirm that individual-based risk factors other than age, sex and co-morbidities are determinants of susceptibility to the infection and related mortality: however the impact of specific  ABO Rh blood groups probably applies differently in different populations, because of accumulation and/or interference of a multitude of factors in a single scenario.

     

    References

    1. Ray JG, Schull MJ, Vermeulen MJ, Park AL. Association Between ABO and Rh Blood Groups and SARS-CoV-2 Infection or Severe COVID-19 Illness : A Population-Based Cohort Study. Ann Intern Med 2020 Nov 24;M20-4511. doi: 10.7326/M20-4511.
    2. Wordometer. COVID-19 CORONAVIRUS PANDEMIC. https://www.worldometers.info/coronavirus/. Accessed November 21, 2020.
    3. Wikipedia. Blood type distribution by country. https://en.wikipedia.org/wiki/Blood_type_distribution_by_country. Accessed November 21, 2020.
    Previous
    Next

    Submit a Comment

    Contributors must reveal any conflict of interest. Comments are moderated. Please see our information for authorsregarding comments on an Annals publication.
    *All comments submitted after October 1, 2021 and selected for publication will be published online only.
    Cancel