The antigens found on the surface of red blood cells determine the blood type of an individual. Blood grouping is essential for donor-patient blood transfusion match.¹ Red cell antigen mismatch between donor and patient can result in alloimmunisation, which is the formation of red cell antibodies in the patient following transfusion with red cells that the patient lacks. Clinically, alloimmunisation can result in mild to fatal haemolytic transfusion reactions.² In neonates, alloimmunisation may cause mild to fatal haemolytic disease of the foetus or newborn.^3,4

The current serological phenotyping method of red cell antibody detection used in South Africa is simple and inexpensive. However, inconclusive serology results may result due to the interference from donor red blood cells in chronically transfused patients, a rare blood type causing false-positive results or autoantibodies in patients with autoimmune diseases.⁵ Also, sourcing blood for patients with rare blood types is challenging due to the scarcity of commercial rare antisera for serological testing.⁶ The term rare blood type describes the absence of an antigen that is found normally in 99% of the country’s population and is described as: ‘negative for high-frequency-antigens (HFA)’ or the presence of antigens found in only 1% of the population termed ‘positive for low-frequency-antigens’ or the presence of an unusual or rare Rh subtype in a very small percentage of the population (1:1000, 1:10 000).⁷ Hence, rare commercial antisera are often not readily available and are expensive. Therefore, they are not cost-effective for large-scale blood type screening.⁶

Molecular genotyping overcomes the limitations of serology. The mapping of the human genome provided knowledge on the molecular backgrounds and polymorphisms of the blood groups that enabled the development of DNA-based assays for red cell genotyping.⁸ Genotyping utilises patients’ DNA information to infer the red cell phenotype. Genotyping assays have been widely used in transfusion medicine for over 20 years.⁸ This technology only became an affordable service in South Africa in 2015 and has to date not translated into the genetic identification of rare and unique red cell genotypes in South Africa or the wider African context. Smaller studies in some African countries have been completed; however, these were limited to group-specific antigen coverage. Moreover, the sample numbers were too low across the regions to be representative of the African diaspora and its multiple ethnic groups.⁹ Therefore, this study aimed to provide the first comprehensive red cell genotyping of rare blood donors referred to the South African National Blood Service, Immunohaematology Reference Laboratory (IRL).

Most of the study participants were men (63%, n = 95/150) with 80% of the race group distribution comprising 41% (n = 61) White donors and 39% (n = 59) Black South African donors. The 2015–2016 red cell genotyping data was skewed by a selection bias for Group O+ (71%, n = 107) donors which accounted for 33 rare blood types found in 32 donors. In comparison, Group O– donors that made up 13% (n = 20) of the study population showed the presence of 15 rare blood types in 14 donors. The remaining 16% (n = 23) of the Group A+/− and Group B+/− had four rare blood types in four donors, cumulatively resulting in a total of 50 rare donors with 52 rare blood types at a molecular level. Of the 58 serological known rare donors, 44 donors had the serologically defined rare blood types HrB–, HrS–, k–, U–, U-variant+, Jsb, Kpb–, r̎ r̍, RzR1, Lub– and Yta– that are covered by the ID CORE XT. However, molecular genotype correlated with serotype in only 34 of the 44 donors (Figure 1). Ten of the 44 serology-defined rare blood types were absent upon molecular testing giving a 23% discrepancy rate between serotyping and genotyping. More so, rare blood types previously not detected by serology were found in two of these 10 donors (Figure 1). Fourteen of the 58 serologically known rare donors had one of the eight rare blood types: Bombay Oh, Vel–, Kna–, In(Lu), LAN–, Dantu–, Milton (Mi II) and Henshaw that were not covered in the ID CORE XT assay. Therefore, these could not be confirmed genotypically (Figure 1).

From among the 92 random donors, genetic and molecular screening identified an additional 14 donors with a total of 16 rare blood types. This brings the total number of rare donors confirmed using genotyping to 50, which comprises 33.3% of the study population (Figure 1).

Studying the red cell genetic variation among the four major South African race groups showed that the Rh subtype R_o is found in 42 out of 59 Black donors, R₁R₁ in 10 out of 19 Indian donors while six different Rh subtypes are spread among the White donors. Also, rare blood types such as the HrB– and HrS– were found more in Black donors while other rare types such as k–, Kpb– were found among White donors suggesting genetical association between Rh subtype, race group and blood type. This knowledge allows for more effective rare donor screening strategies to be implemented to enhance rare donor-patient blood matching in South Africa. In this study, screening of random Group O+ donors identified 14 new rare donors that would have been missed by routine testing. Hence, identifying markers suggesting a rare blood type in donors such as Rh subtype, race and patterns of genetic variations will assist the development of algorithms for rare blood type testing and identification in donors.

On comparing our results to those from historical serological prevalence studies,¹¹ we observed that White and Indian South Africans share similar alleles for the RHCE*ce/ce, *Ce/cE, KEL*K_KPB_JSB (k–), FY*A, FY*A and YT*A, YT*B genotypes.¹¹ Similarly, common red cell antigens among the Black and mixed race South Africans were the RHCE*cde/Cde, KEL*k_KPB_JSB/JSA and FY*A/B_GATA.¹¹

According to the Department of Statistics South Africa,¹⁰ the four major race groups are defined as Black, White, Indian and mixed race people while internationally and outside Africa, the broadly defined ethnic groups are termed Caucasian, African American, Asian, other or mixed race. It has been proposed that global emigration patterns account for the red cell antigens of White South Africans being related to Caucasians, whereas Black and mixed race South Africans are similar to African Americans and Indian South Africans similar to Asians, and in some cases, Caucasians.^11,12,13,14 In this study, the presence of the RHCE*Cde/cde (r’r), *Cde/cdE (r’r”), *cDe/cDe(Ro), KEL*K (k–), KEL*KPA (Kpb–) and LU*A (Lub–) in White South Africans matched the findings in Caucasian populations of Italy, Netherlands, America, Spain, Germany and Austria.^{15,16,17,18,19} Further, the rare E-Hrb- and E-HrS- found among Black South Africans matched findings in the African American population.^20,21,22 The JS*A,JS*A genotype (Jsb–) was found in three Black South Africans and one Indian donor which is not common in South Africa. However, this phenotype is present in America and Japan and is reportedly associated with Asian populations.^12,23 The detection of the RHCE*cDe/cDe (Ro) in all the race groups represented in the study contradicts the assumption of a race-specific blood type. Interestingly, the RHCE*cDe/cDe (Ro) was most detected in the Black South Africans but only found in the Netherlands.¹⁶ This finding suggests that this is a very rare blood type associated with African ancestry. The absence of the Yta(Yta–) Cartwright antigen shown by the YT*B,YT*B genotype, was found in one Indian South African but was not detected in a study by Kahar and Patel,²⁴ which surveyed the prevalent red cell phenotypes in central India. The n of the Yta(Yta–) can be attributed to the small study size or area of India sampled. The Yta- antigen was also identified in Italy,¹⁵ America,¹⁷ Spain,¹⁸ Japan²³ and Canada²⁵ prompting a decision to expand screening for this rare antigen in South Africa to all race groups.

Similar to the European study by Finning et al.,²⁶ three of the negative for HFA Js(b–), Kp(b–) and Jk(b–) were also found in the current study. Although there have been suggested similarities between South African White donors and Caucasians or Europeans, it is evident that some uniqueness exists between the two geographic regions. The Dib–, RHCE*CeRN, Dia–, Mia+, Cob+, Lua+, Dob+, which were unique to the European study by Finning,²⁶ were absent in the present study possibly due to the small sample size. More studies will be required to interrogate and investigate the underlying reasons for these potential differences at a genetic level or whether this pattern changes and becomes more similar when more donors are tested.

The RHCE*CE/Ce genotype is a rare Rh subtype (R_zR₁) that was found in one mixed race South African. A study completed by Sharma et al.²⁷ reported R_zR₁ to be found in 6% of Native Americans and 2.2% of Indians in Central India. The low prevalence antigen RHCE*Cw was found in two White South Africans only.

The ID CORE XT assay does not differentiate the HrB– (Rh:-34) from the hrB– (Rh:-31). The formation of anti-HrB is due to the absence of the high-frequency antigen HrB (HrB–) and anti-hrB (Rh:-31) is an ‘anti-e-like’ antibody. The absence of the ‘E’ antigen together with the HrB– (E-HrB)– indicates the true rare blood type. Besides, the ID CORE XT assay does not differentiate the HrS– (Rh:-18) from hrS– (Rh:-19) subtype. Similar to the HrB– subtype, the absence of the ‘E’ antigen combined with the HrS– subtype indicates the true rare antigen type. This study had revealed a possible differentiation of the two HrB–/hrB– and HrS–/hrS– types due to the presence of specific single nucleotide polymorphisms in the genotype and in comparison to genotype patterns produced when historically known HrB– (Rh:-34) samples are genotyped. This is reflected in Table 2 where the r’s haplotype with the presence of the 733G and 1006T pointed to the rare HrB– type while the presence of homozygous ceAR or with the single nucleotide polymorphism 712G was an indication of the HrS– rare blood type. A total of 12 HrB– and six HrS– donors were reported in this study as these two rare types originated in South Africa, hence found in more of our South African donor population.^20,21

The rare negative high-frequency antigens k–, Kp(b–), Jk(b–), Jo(a–), Hy– and Lu(b–) were found only in White donors and this is consistent with reports in Europeans and Caucasians from America and Europe.²⁸

The Fy(a-b–) rare phenotype was not identified in the current study and may be attributed to the purposive sampling techniques used in the study. The GATA mutation identified by the ID CORE XT assay predicts Fy(a–,b–); however, this is not the true rare phenotype, but rather an assay limitation.

The genotype GYPB*S, GYPB*S or s– phenotype is currently not listed as a rare type in South Africa as it is not difficult to locate; however, s– is listed on the rare donor files of Spain¹⁸ and Japan.²³ Although the S-s-U– phenotype predicted by the GYPB*deletion is rare in Caucasians and found mainly in African Americans,¹⁷ it was found in one White and one Black donor in the present study. This suggests the presence of this rare blood type among White South Africans. Interestingly, this study revealed many null alleles among the Ss alleles of the MNS blood group system, which was missed by serology. These are important as the null alleles were mainly associated with the presence of the U-variants. Patients with U-variant phenotypes can only be given blood matched from U-variant donors or risk the formation of anti-U antibodies that will result in transfusion reactions.

Three red cell antigens, the Jk(a-b–), Co(a–) and Ko, were not identified in the present study which could be explained by the small sample size. However, donors for these three antigens were previously active rare donors on the South African Rare Donor file but became lapsed donors having not donated blood in over 3 years.²⁹ This highlights the need to increase stores of rare blood units in frozen storage so that patient-donor blood matching is not challenged by the lack of active blood donors. Mass-scale cost-effective red cell genotyping can be used to screen donors routinely for this purpose in line with international practice as in the United States.³⁰

The most important finding of this study was that among the mixed race group, 73% (8/11) of the donors showed the presence of rare antigens. Thus, mass-scale rare donor screening among mixed race donors can increase the pool of donors with rare blood types.