How I treat challenging transfusion cases in sickle cell disease Free

Red blood cell (RBC) transfusions are recommended to treat or prevent multiple complications of sickle cell disease (SCD).^1-4 Although often lifesaving, transfusions are not risk-free.⁵ One common serious hazard of transfusion for individuals living with SCD is the formation of RBC alloantibodies to non-self antigens. Although some patients may receive hundreds of RBC transfusions without forming antibodies, others form multiple antibodies after just a few transfusions. Approximately 30% to 50% of transfused patients with SCD will develop RBC alloantibodies over their lifetimes,^6,7 potentially leading to difficulties in identifying compatible RBCs, increasing the risk of hemolytic transfusion reactions, and putting women of childbearing age at risk of hemolytic disease of the fetus and newborn. This How I Treat article reviews possible strategies for difficult-to-transfuse scenarios affecting patients with SCD, providers caring for these patients, and transfusion services/blood suppliers across the world.

Our awareness of the dangers of RBC antibodies in SCD, including a higher overall mortality in patients who are alloimmunized,^8,9 has increased over time. Some of this mortality can be attributed to delayed hemolytic transfusion reactions (DHTRs).^10,11 It is estimated that 4% of transfusions administered for acute indications in adults with SCD result in a DHTR.¹² Most patients who experience DHTRs are alloimmunized, though new antibodies cannot always be identified at the time of hemolysis using traditional blood banking methodologies.¹³ Strategies for treating or preventing DHTRs are reviewed in detail elsewhere,^2,14 with case 1 highlighting additional considerations, including antibody evanescence and the importance of sharing alloantibody data between hospitals.

Sensitization to Rh and K antigens underlies most of the RBC antibodies formed by patients with SCD; therefore, provision of RBCs phenotype-matched for Rh (C, E or C/c, E/e) and K antigens is recommended as one strategy to mitigate alloantibody formation.^2-4 With ABO/RhD matching alone, combined data in patients with SCD show an RBC alloantibody incidence rate of 1.94 antibodies per 100 RBC units transfused. In contrast, combined data show lower RBC alloantibody incidence rates of 0.25 to 0.4 antibodies per 100 RBC units transfused when RBCs are phenotype-matched for at least Rh (C, E, or C/c, E/e) and K antigens.² Most large SCD programs in high-income countries have been providing Rh- and K-matched RBCs for >20 years; consensus guidelines aim to broaden this practice, including the 2018 International Collaboration for Transfusion Medicine Guidelines and the American Society of Hematology (ASH) 2020 guidelines for SCD transfusion support.^2,3 Transfusion with RBC units from donors of the same racial background as the patient has also been suggested as a strategy to decrease exposure to non-self antigens because they are more likely to have similar blood group antigen profiles as the recipients. In the United States, most individuals with SCD are Black, but SCD also affects individuals of other races.

Despite providing serologic Rh-matched blood, Rh alloimmunization persists because of the high frequency of RH variants among patients with SCD and Black blood donors.^6,15,16 Nearly 85% of individuals of African descent carry variant RHD or RHCE alleles that result in the loss or alteration of Rh antigenic epitopes, compared with <3% of individuals from other racial populations.¹⁷ Variant Rh antigens are not identified by traditional blood bank phenotyping tests, which detect only the common Rh antigens (D, C, c, E, e); thus, serologic-matched RBCs are not truly Rh-matched. Although inheritance of variant RH alleles explains approximately one-third of Rh antibodies formed by patients with SCD, the remainder are likely stimulated by altered Rh proteins on Black donor RBCs.^18,19 

The rate of Rh alloimmunization in chronically transfused patients with SCD is comparable at hospitals in Philadelphia and Atlanta, although one provides CEK-matched RBCs via a blood center program to recruit Black donors, whereas the other requests CEK-matched RBCs only.¹⁶ This is likely due to the widespread recruitment of Black donors to support the provision of C, E, and K negative units for patients with SCD; thus, the avoidance of Rh immunization will require improved matching using molecular approaches. Rh antibodies remain the most common specificities identified in patients with SCD, leading to complex transfusion management, including donor selection. More precise matching of blood donor and recipient using DNA-based methods that match according to RH variants is under investigation. Cases 2, 3, and 4 highlight the clinical implications of variant Rh antigens and how we approach the challenges of identifying compatible donor RBCs for patients who are Rh-alloimmunized.

A 31-year-old woman with hemoglobin (Hb) SS from Atlanta developed acute chest syndrome (ACS) while visiting New York City. Her Hb was 6 g/dL, her RBC antibody test was negative, and she was transfused with 2 units of RBCs that lacked the C, E, and K antigens per the New York hospital’s transfusion protocol for patients with SCD. After transfusion, her Hb increased to 9 g/dL, and her respiratory status improved. Ten days later, she developed whole-body pain, her Hb fell to 5 g/dL, and her laboratory evaluation showed signs of hemolysis. A repeat antibody test showed Jk^b and S alloantibodies, along with a direct antiglobulin test positive for immunoglobulin G and C3. She was diagnosed with a DHTR with bystander hemolysis (also referred to as hyperhemolysis), involving the destruction of self RBCs in addition to transfused RBCs. Despite supportive care, her Hb continued to drop to 4 g/dL. Her hematology team contemplated whether they should transfuse additional RBCs, this time lacking the C, E, K, Jk^b, and S antigens. The blood bank ordered additional RBCs from their blood supplier that lacked those antigens; all units were crossmatch compatible through the anti-human globulin phase (note: patients with a history of RBC alloimmunization require crossmatching through the phase where anti-human globulin is added to increase the sensitivity of detecting incompatibility). Despite this in vitro compatibility, the transfusion medicine team warned the hematology team that even fully crossmatch compatible RBCs may exacerbate ongoing DHTRs with bystander hemolysis.

This case illustrates a few dangers: (1) the potential risk of additional transfusions to a patient with SCD who is actively having a DHTR and (2) the danger of evanescent antibodies.

Although this patient would not be classified as nontransfusable (crossmatch compatible RBCs are available), transfusing RBCs during a DHTR with bystander hemolysis may exacerbate ongoing hemolysis and result in a lower posttransfusion Hb and other adverse consequences. One author typically recommends supportive care and avoiding transfusions during a DHTR with bystander hemolysis, except with critical illness. The other author considers the risk/benefit ratio of treatment with intravenous immunoglobulin and steroids followed by RBC transfusion, with a low reticulocyte count and/or hemodynamic instability with an Hb <4 g/dL swaying her toward additional transfusion. Erythropoietin and intravenous iron may mitigate the need for further RBC transfusions. Guidelines published by ASH^2,11,20 address potential approaches in this acute setting, including supportive care to increase erythropoiesis as well as treatment with drugs such as steroids, intravenous immunoglobulin, eculizumab, tocilizumab, or rituximab. Table 1 shows possible doses for these therapies, with these drugs being used off-label for treatment of DHTRs and the risk/benefit ratio needing to be carefully considered.

Beyond the acute setting, the question may arise about future transfusions. For example, could this patient safely receive RBC transfusions before hip replacement surgery 5 years later? The answer varies by patient. In those without a prior DHTR, guidelines typically recommend preoperative transfusions to lower the HbS percentage for surgeries requiring anesthesia of any significant duration.² However, DHTRs with bystander hemolysis are known to recur in some patients, even when future transfusion of RBCs is closely antigen-matched between donor and recipient.^10,22 The ASH guidelines² address this scenario, though the evidence to guide recommendations is weak. Patients with prior DHTRs have successfully been transfused, with adverse outcomes upon re-exposure to RBCs potentially less likely in the setting of immunomodulatory treatments.²³ In any case, the risk/benefit ratio of the surgery, the anesthesia, and future transfusions should be carefully considered in patients with a history of DHTRs with bystander hemolysis.

Most RBC alloantibodies evanesce or fall below the level of detection by traditional blood banking methods over time.^24,25 During this patient’s DHTR, communication with 5 hospital blood banks in the Atlanta area where the patient had previously been treated uncovered that her anti-S had been identified a decade before at one hospital. This antibody was not detectable at the time of testing in New York, was not known by the 4 other hospitals in the Atlanta area, the patient, or her family, and was not present in the Care Everywhere section of the electronic medical record system.

Although some countries (such as the Netherlands, who use the Transfusion Register of Irregular Antibodies and Cross-match Problems²⁶) have systems to share antibody data between hospitals, the United States continues to struggle to develop antibody-sharing strategies that are scalable beyond a hospital system or a region. This is an area of interest to many,²⁷ given the safety implications of not knowing a patient’s antibody history.²⁸ In the absence of widespread antibody data sharing in the electronic medical record, it is critical that the clinical team obtains a list of all hospitals previously involved in the patient’s transfusion care and that the treating hospital contacts the blood banks of those prior hospitals to obtain an accurate RBC antibody history. If this patient’s anti-S had been known, the New York hospital would have selected S antigen–negative RBCs for transfusion. Further, depending on their transfusion service’s policy for a patient who is alloimmunized, they may have extended the phenotype match to include Fy and Jk antigens, thus preventing this patient from making the anti-Jk^b. Extended phenotype matching refers to additional antigen matching beyond Rh and K, typically including Fy^a, Fy^b, Jk^a, Jk^b, S, and s, and can provide further protection from RBC alloimmunization.² 

A 12-year-old girl with HbSS disease and autoimmune hepatitis, on hydroxyurea, was admitted with splenic sequestration, an Hb of 5 g/dL (baseline, 7-7.5 g/dL), and a reticulocyte count of 5%. She had 3 lifetime transfusions with 1 unit of packed red blood cells (PRBCs) per episode, and ∼1 year after the last transfusion, an anti-C, -hr^B, and -Lu^a were identified by the reference immunohematology laboratory. In addition, the serologic evaluation suggested an additional antibody directed toward another Rh antigen, and an anti-Hr^B was reported. Her RH genotype obtained in 2012 revealed the variant alleles DIIIa-ce^S and DAU0-ce^S (International Society of Blood Transfusion nomenclature: RHD∗03.01-RHCE∗01.20.03 and RHD∗10.00-RHCE∗01.20.03), with a predicted phenotype of D⁺, C^–, partial c⁺, E^–, partial e⁺, hr^B–; Hr^B status was not reported because it has not been well defined. Her Hb declined to 4.2 g/dL, and the hematology service requested RBCs for transfusion. Her current antibody test demonstrated no antibodies. The blood center queried the American Rare Donor Program (ARDP) for C^–, E^–, K^–, hr^B–, Hr^B–, and Lu^a– units, which identified 8 potential donors nationally, but no donor units were available in the inventory. She was treated with erythropoietin, iron, and supportive care and discharged with an Hb of 6 g/dL and a reticulocyte count of 15%. Two months later, she returned with recurrent splenic sequestration and an Hb of 4 g/dL. With 2 life-threatening events, surgery was consulted for a splenectomy, for which the patient would require preoperative transfusion. From an RH perspective, given her history of Rh antibodies including anti-C, -hr^B, and possible anti-Hr^B and her need for prophylactic E-RBCs, a compatible donor match would be a homozygous DIIIa-ce^S donor who is also lacking K and Lu^a antigens. One frozen homozygous DIIIa-ce^S genotype-matched unit was identified. She was treated with rituximab during admission to prevent further alloimmunization, and the splenectomy was scheduled after a B lymphocyte panel 2 weeks later demonstrated B-cell depletion. She received the transfusion without complication or additional antibody development.

This case illustrates the clinical implications of Rh variants in patients with SCD requiring transfusion, how RH genotyping can inform the antibody identification and support donor selection, and judicious use of rare RBCs.

The Rh proteins are encoded by RHD and RHCE that are inherited as haplotypes and encode the D and CE antigens in various combinations (ce, cE, Ce, or CE), respectively (Figure 1). Standard serologic Rh typing of RBCs tests for 5 principal antigens: D, C, c, E, and e. These 5 antigens are responsible for most Rh incompatibilities, but the Rh system is much more complex and includes >50 antigens that encompass polymorphic epitopes. RH variants result from the large number of single-nucleotide polymorphisms and gene rearrangements that are facilitated by the inverted orientation and proximity of the duplicated RHD and RHCE genes (Figure 1A). Patients with variant RH are at risk of antibody production if exposed to non–self Rh epitopes via transfusion or pregnancy. Variant RHD and RHCE alleles encoding partial D, C, c, or e antigens occur frequently in Blacks, and are missing epitopes expressed on the corresponding common Rh antigen; those RBCs can also lack high prevalence Rh antigenic epitopes (present on >99% of RBCs of most populations and including hr^B and hr^S) or, in contrast, express new epitopes (eg, V, VS) (Figure 2). These lead to Rh mismatches between patient and donor that are not detected by standard testing.^29-31 

In this case, the patient was homozygous for the variant RHCE∗ce^S (Figure 2), which results in partial c and e antigens as well as loss of the high prevalence Rh antigen hr^B; thus, her immunization to hr^B was not surprising because exposure to this antigen was inevitable with Rh-matching strategies in which she would be matched C-c⁺E-e⁺ RBCs. The anti-C she formed despite prophylactic-matched RBCs lacking C antigen was likely due to exposure of a non-self RhD or RhCE antigen on donor RBCs. The reported anti-Hr^B was unexpected as one of her haplotypes was DAU0-ceS, but a RH genotype-matched donor homozygous for DIIIa-ceS would meet the requirements for C^–, E^–, hr^B–, and Hr^B– RBCs. The donor should also lack K and Lu^a antigens.

Because transfusion was desired by the clinical team for preoperative preparation in this patient with multiple alloantibodies after a low number of RBC exposures, immunosuppression with rituximab was provided to prevent further alloimmunization. Extended matched RBCs to include Fy, Jk, and S antigens may be recommended for patients with multiple alloantibodies, but in this patient’s case, additional antigen-negative requirements would likely have precluded the identification of units to transfuse. Given the rarity of compatible RBCs for this individual, we would recommend supportive care with erythropoietin and iron for the future if her Hb downtrends at the start of admissions with complications of SCD. In addition to the hydroxyurea, we would also recommend preventive maintenance therapy with voxelotor, an HbS polymerization inhibitor, to increase her baseline Hb level²⁰ and decrease the need for future transfusions.

A 3-year-old boy with HbSS, a history of recurrent splenic sequestration requiring a 6-month period of regular transfusion therapy, and now splenectomized, was admitted for ACS with an Hb of 6.6 g/dL and hematocrit of 19.1%. On hydroxyurea at the maximal tolerated dose, his baseline Hb was 7.7 to 8.6 g/dL. Given his history of ACS requiring pediatric intensive care unit admission and RBC transfusion, an RBC unit was ordered. The antibody test demonstrated a new anti-e with a panreactive eluate (an eluate refers to an enzymatic technique used to elute antibodies bound to RBCs, which is subsequently used for an antibody test). His blood bank records were significant for an RH genotype of RHD-ce733G/DIVa-ceTI (RHD∗01.01-RHCE∗01.20.01 and RHD∗04.01-RHCE∗01.02.01) that predicts he is D⁺, C^–, E^–, partial c⁺, partial e⁺, hr^B+, and hr^S+ (Figure 3). He had no history of RBC alloantibodies. He was admitted 2 months prior for elective tonsillectomy and adenoidectomy, at which time he received an RBC transfusion preoperatively and had an uneventful postoperative course. An Hb quantification revealed 10.6% HbA remaining, suggesting the transfused RBCs had not all been cleared.

This case highlights a relatively frequent conundrum where a patient phenotypes as E-e⁺, whose RH genotype predicts partial e expression, and who, not unexpectedly, forms an alloantibody with e specificity after serologic Rh-matched RBC transfusion (Figure 3). Donor selection for this scenario can be challenging.

For prophylactic Rh- and K-matched transfusions, he would require RBCs lacking C, E, and K antigens because he lacked all 3. E and e are antithetical antigens, meaning an individual’s RBCs must express either one by homozygosity or both by heterozygosity. Therefore, providing e- RBCs to this patient with anti-e would necessitate exposure to E⁺ RBCs and the risk of anti-E formation. In the authors’ experience, patients who are E^– frequently form anti-E after E⁺ RBC exposure because the E antigen is highly immunogenic. Although the anti-e is identifiable in the patient’s plasma, we would recommend RH genotype-matched RBCs that are E^– and e⁺ (Figure 3). The RHCE∗ce733G allele is much more common than RHCE∗ceTI; identifying a donor homozygous or compound heterozygous for RHCE∗ce733G or RHCE∗ce48C,733G would be most feasible. The allele frequency of RHCE∗ce733G and RHCE∗ce48C,733G in Black donors is 14.3% and 2.7%,¹⁷ respectively; ∼3% to 4% of Black donors would be homozygous or compound heterozygous for these alleles and be an appropriate RH genotype match for this patient. Some blood centers screen Black donors for the RHCE∗ce733G allele because individuals homozygous for this variant lack the high prevalence of Rh antigen hr^B and are valuable donors. Thus, a request to the blood supplier or ARDP should identify RH genotype-matched RBCs, unless the patient has many additional antigen-negative requirements.

If RH genotype-matched RBCs are not available, whereas the anti-e is detectable, and if only E⁺ RBCs are offered, the risk-benefit ratio of transfusion must be considered with the patient. We typically do not rely on crossmatching alone to find a compatible E^–e⁺ unit because the anti-e may have a broader reactivity and even RH genotype–matched donor red cells can appear crossmatch incompatible; moreover, identifying a compatible unit by crossmatching is highly inefficient in comparison to donor RH genotyping. If the anti-e is not currently identified in the patient’s plasma, it has been our experience in a small number of patients that re-exposure to non-RH genotype–matched E^–e⁺ RBCs did not result in an overt DHTR. For this patient, an RBC unit that was homozygous for RHCE∗ce733G was identified in the case transfusion was needed.

If the patient expressed a conventional E antigen, transfusion with E⁺ RBCs would be a simple solution. For example, if the patient’s genotype was RHD-ce733G/RHD-cE, the straightforward match would be a donor who lacks the C and e antigens (DcE/DcE, known to transfusion specialists as R2R2). However, patients with partial e antigens who lack E are challenging for transfusion management.

Can we prevent Rh antibodies in an era when RH genotype-matched RBCs are not yet widely available? The feasibility and efficacy of RH genotype-matched RBCs to prevent alloimmunization are under investigation (ClinicalTrials.gov identifier: NCT04156893). The ASH guidelines suggest an extended RBC antigen profile by genotype if possible at the earliest opportunity.² RBC antigen genotype tests include DNA arrays that test for up to 37 antigens in 12 blood group systems (excluding ABO and RhD) and are licensed by the US Food and Drug Administration.^32,33 These arrays can identify the presence of C, c, E, and e antigens, as well as some variant RhCE antigens, but do not detect most RH variants.

To identify RH variants found in Blacks, we recommend RHD and RHCE genotyping by blood group genomics laboratories that use a combination of polymerase chain reaction–based laboratory-developed tests and DNA arrays.^6,33 The RHD and RHCE Beadchip arrays target 35 RHD and 25 RHCE single-nucleotide variations or insertions, respectively. Laboratory-developed tests for clinically relevant single-nucleotide polymorphisms not covered by the RH Beadchips should include RHD exon 8 c.1136C>T (RHD∗DAU alleles), RHCE exon 2 c.254C>G (RHCE∗ceAG), and RHCE exon 4 c.577A.G (RHCE∗CeRN). RHD zygosity should also be examined for comprehensive RH genotype determination. We suggest obtaining this comprehensive RH genotype for all patients with SCD to identify those with partial Rh antigens. For patients with partial C or D antigens who do not express the conventional antigen, we provide prophylactic antigen-negative RBCs. The hybrid RHD∗DIIIa-CE(4-7)-D allele (case 4), where RHCE exons 4 to 7 have replaced the corresponding exons of RHD, is the most common allele resulting in partial C antigen expression and is found in 2% to 3% of Blacks. Among over 1400 patients with SCD at 1 author’s institution, RH genotyping has revealed 6% and 5.7% of individuals express partial D or C antigen only, respectively, so this strategy should not place a significant strain on the antigen-negative inventory.

A 25-year-old man with HbSS presented to a transplant physician at a tertiary care hospital for a second opinion. He was interested in curative therapy but had been told by a hematologist at another institution that he was untransfusable and therefore likely ineligible for an unrelated stem cell transplantation or gene therapy. His past or currently detected RBC antibodies included anti-D, anti-E, and anti-hr^B, and his RH genotype DIIIa-ce^s/DIIIa-CE(4-7)-D-ce^S (RHD∗03.01-RHCE∗01.20.03 and RHD∗03N.01-RHCE∗01.20.03) predicted partial antigens for D, C, c, and e. The patient was also negative for the Fy^a, Jk^b, and S antigens. The transplant physician reached out to her transfusion medicine colleagues to determine whether sufficient compatible RBC units could be located for exchange transfusions, leading up to curative therapy.

This case illustrates an issue that arises with some frequency, given the increasing number of emerging cell-based therapies for individuals living with SCD. Allogeneic stem cell transplantation is one such curative option, with potential indications described in the ASH guidelines for stem cell transplantation.³⁴ Before the allogeneic transplantation, hematologists often request that the percent HbS of the recipient be lowered to <30%.³⁵ Assuming a baseline HbS level of 90%, an exchange transfusion with an ending hematocrit of 30% and goal HbS <30% would typically require 6 to 10 RBC units, depending on the size of the patient and the baseline hematocrit. In the case described, it may be possible to locate RBC units over time, taking an approach involving the ARDP and recruiting donors to freeze sufficient units for an exchange transfusion and to sustain the patient through engraftment. A recent report describes such an approach for a difficult-to-transfuse, patient who is highly alloimmunized, followed by a successful transplantation.³⁶ 

One consideration is whether special preparations are needed for the hematopoietic stem cell product based on recipient alloantibodies. At some institutions, RBC reduction of stem cell products to lower the RBC volume to under 30 to 50 mL is completed for donors expressing an RBC antigen that the recipient has alloantibodies against (regardless of whether that antibody is currently detected or is evanescent). At other institutions, RBC reduction of the hematopoietic stem cell products is completed only if the recipient’s RBC alloantibody is detectable at the time of the transplantation. Another consideration is whether the donor’s RBC antigen and recipient’s RBC alloantibody status should affect recipient conditioning intensity; 1 case describes a patient with SCD and multiple RBC alloantibodies who continued producing one of the RBC antibodies after a reduced intensity transplantation, leading to a mixed chimeric state with a donor that expressed that RBC antigen and autoimmune hemolytic anemia.³⁷ 

Patients who were RBC alloimmunized may also have HLA alloantibodies.^38,39 Donor-specific antibodies may potentially lead to future graft rejection^40-43 and/or acute graft-versus-host disease.⁴⁴ Further, HLA antibodies may increase the risk of platelet transfusion refractoriness and the need for more platelet transfusions in the peritransplantation period.^44-46 Adequate pretransplantation planning and close communication between the transplantation, hematology, and transfusion services are recommended in all transplantation cases involving patients with alloantibodies (RBC, HLA, or both).^36,47,48 

As of December 2023, there are 2 gene therapies approved by the US Food and Drug Administration for patients with SCD: exagamglogene autotemcel⁴⁹ and lovotibeglogene autotemcel.^50,51 These therapies are clustered regularly interspaced short palindromic repeats (CRISPR)-based and lentiviral-based, with the goal of increasing fetal Hb or producing a modified antisickling Hb (HbA^T87Q), respectively. Recommendations for gene therapy typically include maintaining the HbS <30% for 8 to 12 weeks before stem cell collection to suppress erythropoiesis and improve mobilization and stem cell collection. Autologous RBCs with HbS, in addition to reticulocytosis, presumably contribute to inefficient stem cell collection.⁵² In addition to precollection RBC transfusions, transfusions are recommended until product infusion to minimize SCD-related complications. In patients who are highly alloimmunized and have few compatible RBC units available, a goal HbS <30% may not be feasible. For case 4, it is unlikely that sufficient RBC units will be identified to support multiple months of transfusion therapy. However, it may be feasible to locate sufficient units over time to freeze and use for a single exchange transfusion or to support a period of simple transfusions before stem cell collection.

RBC transfusions for patients with SCD remain a critically important therapeutic option. When alloantibodies prevent antigen-matched RBCs from being identified for future transfusions, an individualized transfusion management plan must be developed. Cases involving Rh alloimmunization despite serologic Rh-matched RBCs may be particularly difficult.¹⁶ Immunization events involving DHTRs with bystander hemolysis are potentially deadly. In these challenging transfusion cases, multidisciplinary involvement and care coordination between the primary medical team, the transfusion medicine service, and the blood supplier are required for optimal patient outcomes. Further, complex cases may benefit from evaluation by an immunohematology reference laboratory and discussions with experts in transfusion complications for patients with SCD. At times, patients are deemed nontransfusable for safety concerns, and at other times, prophylactic antigen matching must be waived to allow lifesaving transfusions to occur. As patients seek care in different geographic regions, sharing RBC alloantibody information among health care systems nationally is needed to optimize transfusion safety; high RBC antibody evanescence rates^53,54 make this a high-priority issue for patients with SCD. Importantly, as curative options for patients with SCD become more widespread, creative solutions are needed to safely transfuse challenging patients who are highly alloimmunized who would otherwise be ineligible for these therapies.

The authors thank the clinical teams caring for the patients included in this manuscript, the blood bank staff at the Children’s Hospital of Philadelphia and Emory University, the Immunogenetics Laboratories at the American Red Cross and the New York Blood Center, and the American Rare Donor Program. The authors also thank Kaoru Takasaki and Sunitha Vege for their valuable discussion and insights.

This work was supported in part by the National Institutes of Health (NIH)/National Heart, Lung, and Blood Institute (NHLBI) (grants U01 HL134696 and R01 HL147879), and a Distinguished Chair in Pediatrics to S.T.C. J.E.H. receives funding from NIH/NHLBI (grants R21HL165306 and HHSN26819HB00003R/75N92019D000036).

Contribution: S.T.C. and J.E.H. wrote the manuscript.

Conflict-of-interest disclosure: S.T.C. is a consultant for Pfizer for a real-world evidence study of voxelotor, has received honoraria from American Society of Hematology for speaker engagements and from Wolters Kluwer for UpToDate chapters, and has provided paid expert testimony within the past 2 years for pediatric hematology cases unrelated to sickle cell disease and alloimmunization. J.E.H. declares no competing financial interests.

Correspondence: Stella T. Chou, Department of Pediatrics, University of Pennsylvania, The Children’s Hospital of Philadelphia, 3615 Civic Center Blvd, Philadelphia, PA 19104; email: chous@chop.edu.