Emergence of red blood cell alloantibodies and transfusion management in patients with warm autoantibodies at a tertiary care centre in British Columbia, Canada

Abstract

Background and Objectives

The presence of warm autoantibodies (WAAs) complicates pre‐transfusion and compatibility testing. Despite attempts to provide antigen‐matched red blood cells (RBCs), the risk of alloimmunization remains. Rates of alloimmunization and indications for transfusion were reviewed to streamline testing and RBC provision algorithms at a large tertiary care centre serving patients with lymphoid cancers and complex surgical needs.

Materials and Methods

This retrospective observational study investigated the development of new RBC alloantibodies in patients with WAAs. This included 295,109 antibody screenings and 3129 antibody investigations (AIs) performed on 2493 patients between 1 September 2019 and 30 June 2024. AI results for patients with a history of WAAs were reviewed, along with diagnoses, transfusion data, and where applicable, phenotyping and genotyping results.

Results

Ninety‐four patients had WAAs. Twenty‐three of them (24%) had lymphoproliferative disorders (LPDs) and 21 (22%) required urgent antibody tests for surgical procedures. Fifty‐one patients (54%) received RBC transfusions, and 30 of them (59%) had anaemia with haemoglobin below 70 g/dL. Thirteen patients (14%) required RBC genotyping because of recent transfusions or indeterminate results. The alloimmunization rate was 10%, including anti‐Jk^a, anti‐Kp^a, anti‐Jk^b, anti‐C^w, anti‐Js^a and anti‐Le^a, after RHDCE/K or more extended‐matched RBC transfusions.

Conclusion

RBC alloantibodies develop in patients with WAAs, as the urgency of transfusions often limits the complete identification of antibodies and extended phenotype matching. With prompt investigation and RBC preparation, the risk of alloimmunization to major antibodies can be minimized.

Highlights.

• The presence of warm autoantibodies (WAAs) interferes with antibody identification and compatibility testing.

• Patients with WAAs remain at risk of developing alloantibodies, even with efforts to provide phenotype‐matched transfusions, because of limited time to find compatible blood or the absence of specific antigen‐negative units in the local inventory.

• For patients with WAAs and anaemia, clinicians should ensure prompt preparation of phenotype‐matched red blood cells.

INTRODUCTION

Warm autoantibodies (WAAs) cause panreactive red blood cell (RBC) agglutinations, masking underlying alloantibodies. They also interfere with compatibility testing, which complicates RBC provisions. WAAs can mimic alloantibodies, most commonly against the Rh system antigens [1, 2]. Alloantibodies of clinical significance coexisting with WAAs are not uncommon and have been found in up to 12%–40% of patients [3]. Additional serological investigations are required to identify alloantibodies and inform of RBC selection to prevent haemolysis. Various methods for identifying alloantibodies, such as RBC adsorption and dilution techniques, have demonstrated varying sensitivities.

Panreactivity is a challenge faced by most transfusion medicine laboratories. It can be due to a variety of aetiologies, including WAAs, antibodies against high‐incidence antigens or multiple antibodies, drug interference or false reactivity from non‐specific agglutination related to the assay. This can result in heightened manual testing workload and delayed RBC transfusions. At our institution, we provide extended phenotype‐matched RBC antigens, including D, C, E, c, e, K, Fy^a, Fy^b, Jk^a, Jk^b, S and s, when available, to patients with WAAs to prevent alloimmunization. This approach also reduces the need to repeat extensive antibody investigations (AIs) and alloadsorptions.

To minimize non‐specific panreactivity or antibody of undetermined significance (AUS), the saline‐indirect antiglobulin test (SIAT) was incorporated into the new AI algorithm at our institution in September 2019. Following this change, the detection rate of WAAs decreased from 11% to 6%, reducing the workload of further AIs, limiting unnecessary use of phenotypically matched RBCs and preventing delays in providing RBCs [4]. Despite this improvement, ongoing monitoring of alloimmunization rates in patients previously identified with WAAs and who have been transfused is necessary. We sought to evaluate the effectiveness of the current investigation and RBC provision strategy (Figure 1) in preventing seroconversion and to assess the feasibility of integrating transfusion indications and anaemia status to guide antibody testing and provision of phenotypically matched RBCs for patients with WAAs.

FIGURE 1.

Antibody investigation and red blood cell (RBC) provision algorithms in patients with warm autoantibody (WAA). DAT, direct‐antiglobulin test; Hb, haemoglobin; IAT, indirect antiglobulin test; PEG, polyethylene glycol; SIAT, saline‐indirect‐antiglobulin test; SPRCA, solid‐phase red cell adherence assay; TM, transfusion medicine. *Exclude cases with a known history of anti‐CD38 monoclonal antibody therapy in the past 6 months. **Genotyping in patients with recent transfusions or in indeterminate phenotypes.

MATERIALS AND METHODS

A retrospective observational assessment was conducted to examine the frequency of new alloantibody formation in patients with WAAs, using pre‐transfusion AIs from 1 September 2019 to 30 June 2024, at the largest tertiary care centre in British Columbia, Canada. Our institution provides specialized treatment for patients with complex medical conditions, including trauma, haematopoietic stem cells and solid organ transplants, haematolymphoid cancers and cardiovascular surgeries. Our centre serves referrals for complex surgical procedures in British Columbia. RBC antibody results are not always accessible for patients referred from different health authorities, and not all patients are assessed by the pre‐admission clinic. Consequently, we treat all group and screen samples for patients listed for same‐day surgery as urgent when there are no up‐to‐date antibody screen results within 72 h. RBC allocation for these surgical cases is also performed urgently, with a typical turnaround time of 2 h. Our institution performs over 60,000 antibody screens annually. Our primary screening and confirmation method was the solid‐phase red cell adherence assay (SPRCA), complemented by reflex saline‐ and polyethylene glycol (PEG)‐based tube testing.

For all patients with WAAs, the following variables were collected: patient's diagnosis, transfusion indications, results of previous and subsequent AIs, direct antiglobulin test (DAT) results and IgG strength, haemoglobin levels, RBC alloantibody specificity, times to alloantibody detection, the patient's and donor's RBC phenotypes and the number of transfused RBC units. In patients with RBC alloimmunization, the alloantibody types, donor RBC antigens in transfused units and anaemia status were reviewed. The study assessed seroconversion after transfusion; however, the lack of a control group of patients who did not receive phenotypically matched RBCs prior to the 2019 algorithm implementation prevented comparison of alloimmunization rates.

The University of British Columbia Institutional Research Ethics Board waived the requirement for informed consent, considering the retrospective nature of this study.

Antibody screening and identification methods

RBC antibody screening was performed using SPRCA employing automated instruments (Echo and Neo; Immucor, Norcross, GA), which utilized two‐ or three‐cell antibody detection test plates (Capture‐R Ready‐Screen). The automated instruments interpreted and graded the antibody results, which were confirmed by medical laboratory technologists as negative, weak+, 1+, 2+, 3+ or 4+. We considered 3+ and 4+ reactions strong. Samples newly positive on the SPRCA antibody screen (weak+ to 4+) were re‐tested using PEG (Dominion Biological Limited) to confirm the positive reaction.

All samples positive at screening by SPRCA and confirmed by PEG were further investigated with extended RBC panels (Capture‐R Ready‐ID by SPRCA) for AI. We used homozygous antigen‐positive RBCs to confirm or exclude the presence of clinically significant antibodies.

When extended RBC panels were panreactive, further manual methods using SIAT were performed with panel cells (Immucor Panocell), manual DAT using polyspecific antibodies, followed by anti‐IgG and anti‐C3d testing if indicated. When DAT was positive by anti‐IgG in patients who had received RBC transfusions in the past 90 days, an eluate from the patient's RBCs was tested using an antiglobulin elution kit (Gamma ELU‐KIT II, Rapid Acid Elution, Immucor).

Reinvestigation was also performed every 3 months if WAAs persisted, when the extended phenotype‐matched RBC unit(s) were transfused, or monthly when partially extended phenotype‐matched RBC unit(s) were transfused.

Serological diagnosis of WAA

WAA was defined when panreactivity was observed in the antibody screening and panel investigation at SIAT, accompanied by a positive DAT with anti‐IgG.

Panreactivity in SPRCA/PEG antibody screening and panel with negative SIAT, where clinically significant antibodies could be ruled out, was considered an AUS. Reinvestigation with a complete AI was performed if there was an increase in reactivity strength greater than 1+ compared to the previous results.

RBC phenotyping

Phenotyping of donor RBC units was performed using the tube technique with licensed antisera, following the manufacturer's directions (Immucor and Ortho Clinical Diagnostics).

RBC allocation and provision

In patients with WAA, extended phenotype‐matched RBC units (D, C, c, E, e, K, Jka, Jkb, Fya, Fyb, S, s) were preferred for transfusion. However, Rh and K only (D, C, c, E, e, K) were considered sufficient if the patient's RBC extended phenotypes were unavailable. Crossmatch‐compatible RBC units by any of the following methods, namely SPRCA, PEG–indirect antiglobulin test (PEG‐IAT) or SIAT, were provided.

Once the most appropriate RBC units were identified, the allocation within the transfusion medicine laboratory information system for the patient with WAAs was performed, ensuring they were readily available for transfusion orders. Patients with AUS did not require phenotypically matched RBCs, unlike WAA patients. We previously reported that the new AI algorithm implemented in 2019, which integrated the negative SIAT result to exclude AUS, did not increase alloimmunization. It primarily reduced the WAA detection rate and referred out tests to the immunohaematology reference laboratory to exclude antibodies against highly prevalent antigens [4].

In Canada, Canadian Blood Services (CBS) is the centralized blood supplier, except in Quebec. Approximately 40% of RBC units are phenotyped beyond ABO, RhD and K to aid RBC selection in selected patient populations.

Assessing indications for RBC transfusion

When extended phenotype‐matched RBC units were unavailable, the transfusion medicine physician would be notified to review the patient's electronic medical record and/or discuss the clinical situation with the most responsible physician. Transfusions were generally indicated when haemoglobin was <70 g/L or when symptomatic anaemia was present. The transfusion medicine physician would then decide which RBC antigen(s) to omit. For same‐day surgical cases, RBC allocation was considered urgent regardless of haemoglobin levels, with a minimum of two antigen‐matched RBC units readily available. Our policy requires that two crossmatch‐compatible RBC units be prepared and allocated for any patient with a history of antibodies. As noted previously, our institution manages complex surgical and medical cases. A large portion of RBC transfusions are issued on an urgent (STAT) basis. Two RBC units are designated as the minimum allocation in the laboratory information system. After one unit is transfused, additional cross‐matched units are allocated to replace the issued unit.

AIs in recently transfused patients

When a recent history of RBC transfusion within 3 months was noted, the sample was referred to CBS for allogeneic RBC adsorption. Genotyping was also considered for patients anticipated to require ongoing RBC transfusions, such as those with haemoglobinopathies, severe autoimmune haemolytic anaemia or aggressive lymphomas undergoing high‐dose chemotherapy or stem cell transplantation.

RBC alloadsorptions

PEG alloadsorption was performed using group O red cells that were phenotypically similar to the patient or a combination of R1R1, R2R2 and rr where one of the cells was K− and both Jk^a−Jk^b+ and Jk^a+Jk^b− phenotypes were represented.

RBC genotyping

RBC genotyping was performed on the ID CORE XT platform (Progenika Biopharma‐Grifols, Bizkaia, Spain).

Data analysis

Frequencies and percentages for categorical variables, as well as medians and ranges for continuous variables, were calculated using SPSS 30.0 software.

RESULTS

During the study period, a total of 295,109 antibody screenings were conducted. Of these, 3129 AIs were performed on 2493 patients. Ninety‐four patients were identified with WAAs (87 with IgG‐only DAT positivity and 7 with both IgG and C3d positivity). Eighty‐eight patients had IgG DAT strengths of weak+ to 1+ and 2+, while six had strengths of 3+ to 4+. Nineteen patients underwent an elution study as part of their subsequent AIs following recent transfusions, and all exhibited panreactivity. Among these patients, 23 (24%) had lymphoproliferative disorders (LPDs), and 15 of them received RBC transfusions. Twenty‐one patients (22%) required urgent group and screen testing for surgical procedures. The median haemoglobin level at the time of WAA identification was 77 g/L (interquartile range [IQR] 65–105 g/L). Fifty‐one patients (54%) received RBC transfusions, with a median of two units (IQR 1–46), of which 30 (59%) had a pretransfusion haemoglobin level below 70 g/dL.

Because recent transfusions prevented reliable RBC phenotyping due to donor RBC interference or the strong WAAs, limiting indirect antiglobulin‐phase phenotyping of Fy^a, 13 patients (14%) underwent RBC genotyping to guide future selection of antigen‐negative units.

Three patients were identified as having pre‐existing WAA and alloantibodies, including anti‐E and anti‐K, and one patient had anti‐C due to previously unmatched RBC transfusions. Follow‐up AIs carried out at a median interval of 12 months in 51 patients who received RBC transfusions (range, 0–49) showed that 5 patients (10%) subsequently developed alloantibodies: anti‐Jk^a, anti‐Jk^b, anti‐C^w, anti‐Kp^a, anti‐Js^a and anti‐Le^a after RHDCE/K‐ or more extended‐matched RBC transfusions.

For the patient who developed anti‐Jk^a, unmatched RBCs were transfused 11 years ago following severe traumatic injuries. After receiving a total of 11 RBC units, WAAs were detected before discharge. The patient later returned to the hospital because of osteomyelitis requiring urgent hip debridement surgery. During this admission, only one of the screen cells reacted with the pre‐transfusion sample, and other clinically significant alloantibodies were ruled out. The patient then received two units of crossmatch‐compatible RBCs. Subsequently, at 20 days post transfusion, anti‐Jk^a and weak WAAs were identified. Retrospective phenotyping revealed that two of the three transfused RBC units were Jk^a‐positive. This case indicated the potential presence of either a weak WAA or anti‐Jk^a in the pre‐transfusion sample from the second admission. Following discharge, the patient remained well without requiring any RBC transfusions or antibody testing during the interim; anti‐Kp^a was detected approximately 2 years later.

The other patient who developed anti‐Jk^b presented with anaemia of chronic kidney insufficiency. Because of acute limb ischemia and the urgent need for an arteriovenous fistula to facilitate haemodialysis, the patient received two units of C‐, E‐, K‐negative RBCs that were not matched for Jk^b. Nine days after transfusion, anti‐Jk^b was detected, and retrospective testing of the donor RBCs confirmed they were Jk^b‐positive, implicating these units in the alloimmunization. The patient demographics, RBC phenotypes and transfusion history for both patient groups are shown in Tables 1 and 2, respectively.

TABLE 1.

Demographics, phenotypes and transfusion history of patients with warm autoantibodies with pre‐existing red blood cell alloantibodies.

Age/gender/diagnosis	Pre‐transfused Hb level	RBC transfusion requirements within 24 h of arrival	Patient's RBC phenotype	Types of pre‐existing alloantibodies	Time interval from previous RBC transfusions
29/male/severe aplastic anaemia	20 g/L	6 units	E‐, c‐, K‐, Fyb‐, Jkb‐, S‐	Anti‐E, ‐K	4 years
63/female/SLE, α‐thalassaemia	68 g/L	3 units	E‐, K‐	Anti‐K	7 months
23/female/Evans syndrome	56 g/L	3 units	C‐, E‐	Anti‐E	Unknown
49/male/mantle cell lymphoma	67 g/L	2 units	C‐, N‐	Anti‐C	6 months

Abbreviations: Hb, haemoglobin; RBC, red blood cell; SLE, systemic lupus erythematosus.

TABLE 2.

Demographics, phenotypes and transfusion history of patients with warm autoantibodies with newly developed red blood cell alloantibodies.

Age/gender/diagnosis	Pre‐transfused Hb level	Patient's RBC phenotype	New alloantibodies	Time post transfusion to detect new alloantibodies	RBC transfusion requirements within 24 h of request
70/female/indolent B‐cell lymphoma	69 g/L	C‐, E‐, S‐, Cw‐	Anti‐Cw	4 years	2 units
64/male/hip surgery/osteomyelitis	98 g/L	K‐, Jka‐	Anti‐Jka, ‐Kpa	20 days (anti‐Jka) 2 years (anti‐Kpa)	2 units
43/male/renal failure, ischemic limb requiring urgent arterio‐venous fistula	81 g/L	c‐, E‐, K‐, M‐, Jkb‐	Anti‐Jkb	9 days	2 units
79/female/Evans syndrome	52 g/L	E‐, K‐, S‐, Fyb‐	Anti‐Jsa	3 years	26 units
68/female/emergency cholecystectomy and anaemia	65 g/L	C‐, K‐, S‐, M‐, Fyb‐	Anti‐Lea	1 year	2 units

Abbreviations: Hb, haemoglobin; RBC, red blood cell.

DISCUSSION

WAAs are commonly seen in patients with LPDs, with a prevalence of 5%–10% for warm autoimmune haemolytic anaemia in individuals with chronic lymphocytic leukaemia alone [5, 6]. In our retrospective observational study, we observed a 10% seroconversion rate in patients with WAAs who received RBC transfusions. The alloantibody types identified ranged from major clinically significant antibodies—such as anti‐Jk^a and anti‐Jk^b—in two patients who required urgent surgeries and did not receive full extended phenotype matching to low‐prevalence or clinically insignificant RBC alloantibodies, including anti‐C^w, anti‐Kp^a, anti‐Js^a and anti‐Le^a. Owing to limited published evidence on alloimmunization rates, logistical challenges in excluding alloantibodies in patients with WAAs and other operational difficulties in providing phenotypically matched RBCs, our study provides insights for large hospitals serving diverse and complex patient populations.

Our observed seroconversion rate is lower than that reported in other studies assessing alloimmunization in patients with WAAs who received RBC transfusions. Delaney et al.'s study from the Biomedical Excellence for Safer Transfusion (BEST) Collaborative is one of the largest, with 372 patients with WAAs and serological testing post transfusion. They reported a 15% incidence of new RBC alloimmunization in WAAs, where the use of a prophylactic antigen‐matching strategy was not found to be protective [7]. However, the extent of prophylactic antigen matching in that study varied in application from partial matching (Rh and Kell antigens) to full extended matching (Rh, Kell, Kidd, Duffy and MNS antigens), as well as in compliance in urgent situations. This most likely explains why antibodies against C, c, E and K antigens accounted for 19 out of 41 (46%) of new alloimmunization events in the prophylactic matching arm. Interestingly, in their study, antibodies against low‐incidence antigens accounted for 9 of 41 (22%) new alloimmunization events in the prophylactic arm, compared with only 1 of 41 (3%) at sites without prophylactic matching. This enrichment of antibodies against low‐incidence antigens in the prophylactic matching arm was also observed in our study, accounting for the majority of new alloimmunization events. Potential explanations for this finding include that certain full extended phenotype requests may be selected for donor populations that have a higher expression of low‐incidence antigens (i.e., V antigen in the black population) or that demonstration of antibodies against low‐incidence antigens in the non‐prophylactically matched arm was obscured by the presence of more common antibody specificities [8, 9]. The ability to provide any level of prophylactic antigen matching relies on the presence of a phenotype.

Phenotyping patients' RBCs is sometimes not feasible because of recent transfusions within the past 3 months. Additionally, when RBCs are heavily coated with IgG—whether from WAAs or alloimmunization—the RBC phenotype testing can yield false‐positive results, and the accurate antigen profile may not be reliably determined. Fy^a blood group phenotyping requires an indirect antiglobulin test, making it difficult to type RBCs from patients with a positive DAT. These observations were noted at our institution and have also been described by Nathalang et al. [10]. Without confirmation of patients' phenotypes, it is not possible to ensure phenotypically matched RBCs, which are intended to prevent haemolysis. Genotyping is frequently unavailable in hospital transfusion medicine laboratories, and its long turnaround time limits its utility for timely transfusion management.

The limitations of our study include the following: (1) the study lacked a comparator cohort of patients who did not receive phenotypically matched RBCs before 2019, limiting evaluation of seroconversion rate differences; (2) the patient cohort was relatively small, consisting of 94 patients with WAA, of whom only 51 (54%) received RBC transfusions; and (3) the study also did not evaluate the presence of haemolysis or clinically evident autoimmune haemolytic anaemia, whereas Delaney et al. found that patients with these conditions were more likely to develop alloantibodies.

In conclusion, efforts to diagnose alloantibodies and prevent alloimmunization in patients with WAA appear effective, as evidenced by low seroconversion rates. Nonetheless, providing extended antigen‐matched RBCs remains challenging owing to the urgent nature of transfusion needs. Implementing a clear AI algorithm to assist in serving WAA patients and promptly allocating the most appropriate RBC units would enhance transfusion support efficiency. While genotyping is presently guided by future transfusion requirements, it has the potential to become a standard of care, given the significant time and workload involved in WAA investigations and alloantibody exclusion.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Hutspardol S, Mi J, Denesiuk C, Kalar D, Sham L, Roche M, et al. Emergence of red blood cell alloantibodies and transfusion management in patients with warm autoantibodies at a tertiary care centre in British Columbia, Canada. Vox Sang. 2026;121:339–344.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.