Predicting the effect of transfusing only phenotype‐matched RBCs to patients with sickle cell disease: theoretical and practical implications

BACKGROUND: Transfusing only phenotype-matched RBCs has been recommended to reduce the incidence of alloimmunization to blood group antigens in patients with sickle cell disease (SCD). STUDY DESIGN AND METHODS: The expected benefit of phenotype matching was determined by identifying which of the existing blood group alloantibodies in patients with SCD who received conventional transfusions would not have been formed if they had received only phenotype-matched RBCs. By use of each patient's alloantibodies as a baseline, it was possible to identify specific alloantibodies that would not have been formed if each of five different phenotype-matching protocols had been used. RESULTS: During a 12-year period, 351 patients received transfusions with 8939 units of ABO- and D-matched RBCs. Of these, 102 patients (29.1%) formed at least one blood group alloantibody. An additional 35 patients with SCD with alloantibodies were identified by reviewing clinical records, yielding a total of 137 alloimmunized patients for inclusion in this study. If all transfusions had been selected by limited phenotype matching (C, c, E, e, and K, as well as for ABO and D), all alloantibodies would have been prevented for more than half (53.3%) of the 137 alloimmunized patients. If all transfusions had been matched for C, c, E, e, K, S, Fya, and Jkb, all antibodies would have been prevented for 70.8 percent of the 137 alloimmunized patients. Approximately 13.6 percent of random white blood donors would be expected to match a limited phenotype-matching protocol, whereas only 0.6 percent would match an extended phenotype-matching protocol. CONCLUSION: Limited phenotype matching would have prevented all alloantibodies in 53.3 percent of the patients who formed alloantibodies. This protocol requires RBCs that are readily available. Extended phenotype matching would have prevented alloimmunization in 70.8 percent of patients who formed alloantibodies. However, this would require phenotypes that are 22.7 times less prevalent among random blood donors and is therefore impractical for a long-term strategy.