Red Cell Alloimmunization in Sickle Cell Disease: Benefit of Extended Crossmatching in Adults

Patients with sickle cell disease (SCD) receive frequent transfusions and have been observed to have a high rate of transfusion reactions and alloantibody formation after transfusion. Alloimmunization to antigens of the Rh system, Kell, Duffy, MNSs, and Kidd are most common. It has become accepted practice at most sickle cell centers to match for C, E, and Kell antigens in order to prevent alloimmunization. Most studies of the effectiveness of red cell (RBC) phenotype matching have been in pediatric patients and have shown a decrease in alloimmunization beyond those specific antigens that are pre-emptively matched. Leukoreduction (LR) of blood products is known to decrease HLA alloimmunization and may also reduce the risk of RBC antibody formation. It is also unclear whether there is value in continuing phenotype matching beyond the pediatric age and whether new RBC antibodies may be prevented by LR, independent of phenotype matching. Patients with SCD at our institution are routinely matched for C, E, and Kell antigens and given leukoreduced blood. Here, we study our past experience in reducing the frequency of new alloantibody formation.

We reviewed charts of patients with SCD, including sickle cell anemia, sickle/thalassemia, and SC disease, who were transfused at Thomas Jefferson University (TJU) and had documentation of presence or absence of alloantibodies on presentation to our center. Data extracted from chart included patient's age, disease type, known antibodies, and time of first documentation of antibody. Antibodies were considered to be definitely unrelated to transfusion at TJU if they were present before first transfusion at TJU or if a previous screen not showing the antibody in question had been performed and no transfusions given between that screening and the development of the relevant antibody. LR was used routinely at TJU starting in 2003 and was used in selected cases, including many sickle patients, earlier. In 1996, TJU began crossmatching for C, E, and Kell antigens.

68 charts of patients with SCD transfused at least once at TJU were reviewed, including 8 patients undergoing routine exchange transfusion for stroke prevention. Of these, 33 (48.5%) had no alloantibodies detected at any time. Of the remaining 35, 16 had 1 antibody, 8 had 2 antibodies, and 11 had 3 or more antibodies, including 3 patients with 8–10 each. A total of 92 antibodies were identified among these 35 patients, 60 of which were not related to transfusion at TJU. Of the remaining 32 antibodies, the most common antibodies identified were C (5 instances) and E (6 instances). Antibodies to Kell, Fya, and Jkb developed in 3 patients apiece. Two instances each of C, E, and Kell antibodies developed after the initiation of crossmatching for these antigens, but in all but one case, the records suggested that the antibody was most likely developed by transfusion at an outside institution not practicing phenotype matching, but this could not be confirmed (i.e. the patient was not transfused at TJU for >1 year prior to the positive screen appearing, but there were no screens in between to identify whether the antibody became evident ). One instance of C developed after emergency transfusion without extended crossmatch and thus the antibody likely developed from transfusion at TJU.

Patients who developed antibodies had a higher number of transfusions than non-immunized patients (median 117 vs 74 units, respectively) and were older (38 vs 35 years, respectively), but neither was statistically significant (p>0.1 in each). Of 8 patients undergoing chronic RBC exchange transfusion, 3 have no alloantibodies, 3 had antibodies that developed from transfusion elsewhere, 1 had a new alloantibody but had prior alloimmunization and 1 developed new alloantibodies at TJU.

Our results confirm that C, E, and Kell continue to be the most common antigens leading to alloimmunization in transfused adult SCD patients. A combination of LR and extended phenotype matching appears to be effective in reducing antibody formation compared to historical reports. Our retrospective study has limitations, notably the possibility that patients may have been transfused at other institutions, which may not use phenotype matching or leukoreduced products. Further prospective studies are indicated to aid in determining whether phenotype matching is justified in adult sickle cell patients in the face of growing global LR.

No relevant conflicts of interest to declare.