Assessment of common red blood cell pretreatments to yield an accurate serologic antigen phenotype compared with genotype-predicted phenotype

For patients requiring multiple transfusions and patients with positive direct antiglobulin tests (DATs), an extended red blood cell (RBC) phenotype can provide valuable information and help to determine the risk of forming alloantibodies. In some instances, the phenotype may be used for prophylactic matching. Phenotyping in this patient population is often hindered by the presence of circulating donor cells and/or by a positive DAT. Several methods, such as EDTA glycine acid (EGA) treatment to remove IgG, hypotonic saline wash to separate autologous RBCs, or reticulocyte separation, are often used in these situations to isolate patient RBCs for serologic phenotyping. This study aimed to determine the accuracy of each RBC pretreatment method by comparing serologically determined antigen types with those predicted by RBC genotyping. Forty-eight peripheral blood samples from recently transfused patients were phenotyped for selected antigens in the Rh, Kell, MNS, Duffy, and Kidd systems. Treatment methods for the sample sets were reticulocyte separation (N = 12), EGA (N = 16), and hypotonic saline wash (N = 20). DNA was extracted using standard methods, and genotyping was performed using the HEA BeadChip panel. In addition, 21 samples positive for RBC-bound IgG were EGA-treated up to two times. These samples were analyzed pre- and post-EGA treatment for RBC-bound IgG by tube DAT and by flow cytometry with fluorescein isothiocyanate–labeled anti-human IgG. After reticulocyte separation, 3 of the 12 samples had discordant types with one antigen each: Fy^b, N, and K; serologic results were negative compared with genotype-predicted positive phenotype results. The EGA-treated sample set showed one discordant type: Fy^b; serologic results were negative compared with genotype-predicted positive phenotype results. Four of the 20 samples had discordant types involving the following antigens: Fy^b, N, e, and M; serologic results were negative compared with genotype-predicted positive phenotype results. After EGA treatment of 21 samples, 14 (67%) were negative for RBC-bound IgG by tube DAT, and 7 remained positive. Using flow cytometry, EGA treatment rendered only 4 samples negative, and 17 remained positive. In the antigen testing sample set of 48 samples, 10 of 511 total antigen types tested were discordant. Discordant types were most frequent in the hypotonic saline wash sample set (N = 6). In the flow cytometry sample set, 48 percent of the samples negative by tube DAT after EGA elution had detectable RBC-bound IgG by flow cytometry. These findings suggest that caution should be taken when using phenotype results from all pretreated RBCs and support the use of RBC genotyping to predict RBC antigen expression in samples from recently transfused patients. Immunohematology2017;33:147–151.