Molecular genotyping for RHD: What (not) to do?

Genomics and the development of DNA-based clinical assays are having a major impact on testing methods in the clinical laboratory, as well as in the blood bank and transfusion service. Although unaccustomed to riding the first wave of new technology, transfusion medicine is poised to embrace DNA methods because, similar to many disease gene markers, blood group antigens result from single-nucleotide gene polymorphisms (SNPs) inherited in a straightforward Mendelian manner. SNPs account for much of the diversity of the human genome, and the ability to link SNPs to disease risk, and even to interindividual variations in drug responses, makes SNP profiling the promise of personalized medicine for the future. This potential has fueled the development of high-throughput SNP genotyping platforms that utilize microarray or beadchip technologies, and these are being applied for testing blood group antigens.1,2 When fully developed and validated, this approach will enable fast, reliable, cost-effective antigen screening for all major blood group antigens in a single assay, paving the way for major improvements in the selection of blood for transfusion therapy. We can now consider, for the first time, individualized transfusion treatment, that is, matching donor units with the patient at multiple blood group loci, which may reduce, or potentially eliminate, alloimmunization. The validity of DNA typing for blood groups has been demonstrated in the past decade by studies correlating the RBC serologic phenotype with the genotype in a large number of samples. For some blood group systems, the detection of silencing mutations in addition to the specific allelic polymorphism is required for accurate interpretation. The ABO and RH genes are more diverse, with more than 100 alleles in each system and additional alleles still being discovered. Which of these numerous alleles will be important for clinical transfusion practice has not yet been definitively determined. Numerous reports investigating molecular testing for blood groups in TRANSFUSION have focused on RHD. This emphasis is not surprising, because RBC serologic typing for D can be challenging in the approximately 2 percent of samples with a weak D or partial D antigens. Individuals with partial D, as well as some weak D types, lack or have altered D epitopes and are at risk for production of anti-D when exposed to conventional D. If present in women under the age of 50, it is preferable to consider them D– for transfusion and RhIG prophylaxis to avoid the possibility of hemolytic disease of the fetus and newborn (HDFN). Serologic reagents, however, cannot distinguish the majority of these RBCs from those with conventional or normal D. As an additional challenge, when typing donors some weak D RBCs are “missed,” and RBCs with very low levels of D are only detected by adsorption-elution of anti-D (designated Del). This observation has resulted in concerns about “contamination” of the D– donor pool with RBCs carrying low levels of D. RHD genotyping would eliminate the ambiguity associated with detection of partial D (or weak D) RBCs based on the strength of serologic reactivity and would ensure that all donor units labeled as D– lacked expression of D antigen. In this issue of TRANSFUSION, Polin and colleagues3 report their approach to molecular testing for RHD. They characterized RHD alleles in 201 samples in which the RBCs were weakly reactive with anti-D and also screened samples from 2427 D– donors for the presence of the RHD gene to identify any with low levels of D not detected by serologic testing. Among the weak D+ samples, 120 of 201 (60%) were categorized as weak D Types 1, 2, and 3 with three molecular assays. The high prevalence of these three weak D types in donors in Upper Austria was consistent with other studies from Europe.4,5 Eighty-one (40%) of the weak D+ samples required further testing to determine the specific RHD allele present. Ten exon-specific PCR assays were performed to determine the presence/absence of RHD exons, and in 52 samples (26%) sequencing of the products of all 10 RHD exons was required to ultimately classify the samples into 16 different weak or partial D alleles. Forty-one (20%) samples actually had a normal RHD or a RHD deletion. Screening of the 2427 D– donor samples was performed in pools of 20 with three assays targeting RHD exons 4, 7, and 10. For samples that tested positive for the presence of RHD, sequencing of all 10 exons was required to determine the specific allele present. One weak D and two Del were found, which translates to an estimated incidence of 0.12 percent (3/2427) of D– blood donors expressing a low level of D antigen, similar to the incidence of 0.20 percent (20/8442) in a report from Germany.6 The study is of interest from a genetic and population perspective, because it gives insight into the diversity of the RHD locus in blood donors in Upper Austria. The authors recommend these methods for molecular RHD characterization “as a complement to serologic D typingand suggest they may reduce the occurrence of transfusion incidents.” Does this study lend support for implementation of RHD genotyping of donors? Is there clinical relevance or merit in classifying donor samples demonstrating weak reactivity with anti-D (which would be labeled as D+) into numerous specific allele categories? It would be difficult to envision clinical benefit from knowing the specific weak D allele in a donor unit, which would be appropriately transfused as D+. Rarely, knowledge of a partial D in a donor unit might be helpful when transfusing a patient with an antibody to a low-incidence Rh antigen encoded by a specific partial D, for example, avoiding RHD*IVa, which are Go(a+), for a patient with that antibody. If all patients and donors were genotyped for RHD, one could transfuse a patient carrying a partial D with a donor unit carrying the same partial D allele, thereby eliminating use of a D– unit. Nevertheless, these applications would not be frequent. Of note, comprehensive characterization of a donor population for partial D could not use weak reactivity with anti-D as a criterion for molecular investigation, as was used in the study reported here, because many partial D RBCs react strongly with anti-D reagents. All D+ donors, regardless of strength of reactivity, would need to be investigated. Most importantly, although the approach of Polin and coworkers3 uses real-time PCR with automated allele detection, the large number of amplification reactions, combined with the large number of samples requiring extensive gene sequencing, would be too costly and time-consuming for practical use. Does the discovery of three donors of 2427 with serologically undetectable levels of D antigen support screening first-time D– donors for RHD, as has been implemented in a few European centers? 7,8 The primary consideration is the risk of immunization (or not) by RBCs with weak D antigen. This topic has recently been debated in the pages of TRANSFUSION7,9-11 following reports of five cases of anti-D (the majority involving secondary responses) in D– recipients inadvertently transfused with Del and weak D Type 1 and Type 2 RBCs.12-15 The only clinical trial to assess the potential of RBCs with weak D to stimulate anti-D in D– recipients found no antibody production in 49 patients transfused with 68 units of weak D RBCs.16 Evidence-based medicine requires “application of formal rules of evidence to evaluate the available data and tries to avoid anecdotal clinical experience as well as expert intuition.”17 The case reports would be considered Level 5, low-level evidence, defined as data from small case series, uncontrolled observations, or anecdotal observations. Because additional clinical trials are unlikely, to gather additional data it is important that blood products labeled as D– that may have stimulated anti-D in a recipient be investigated by RHD genotyping to confirm the D status. Evaluation of risk for anti-D includes factors other than the antigen dose, and not all D– patients make anti-D when stimulated. An immune response involves the complex interaction of many gene products in a manner that is not yet fully understood. Although it is often indicated that 75 to 80 percent of D– patients will make anti-D, the incidence may not be that high. In D– hospitalized patients switched to D+ blood products, the incidence was much lower, approximately 32 percent.18 Is zero risk for anti-D an appropriate goal? The precautionary principle indicates that “measures to be employed should not be disproportionate to the desired level of protection, and must not aim at zero risk.”19,20 In actual clinical practice, the liberal use of RhIG in many US centers to prevent anti-D even in male patients reflects the aversion to anti-D. It is important to put the risk of anti-D into perspective with other clinically significant blood group antigens. We accept that 10 percent of recipients are potentially exposed to K and approximately 18 percent to the c antigen. Severe anemia and HDFN have been reported due to maternal anti-K or anti-c stimulated by transfusion. Anti-K is present in approximately 1 per 1000 pregnant women, and 40 percent of K+ babies of women with anti-K suffer from severe anemia.21 Anti-c was associated with 32 deaths from HDFN in England and Wales from 1977 to 1990,22 and in the United States, 8 of 55 pregnancies complicated by anti-c required fetal transfusion.23 Even pregnancies with unaffected newborns require costly monitoring. In my opinion, avoiding exposure to c and Kell in female recipients under the age of 50, as practiced in Europe, should be a priority for our limited resources before focusing on detecting rare donors who may carry low levels of D and whose units might be transfused to women of child-bearing age who are responders. Importantly, the current lack of consensus surrounding the value of RHD genotyping should not overshadow the real power of this approach to improve the practice of transfusion medicine. Genotyping for blood groups offers potential far beyond cataloging the dozens of RHD genetic variants. The availability of automated, high-throughput, cost-effective donor screening, with the development of RHD genotyping strategies, will change the cost–benefit equation and will alter, and perhaps eliminate, the debate. Although RH genotyping will remove the ambiguity associated with determining the D status, which of the numerous alleles will be important for clinical decision making in transfusion medicine remains to be determined. The clinical relevance of each must be established by the scientific method, and that requires serology (yes, agglutination and all that).