The impact of an antibody investigation algorithm emphasizing specificity on reducing potential false‐positive warm autoantibody detection at a Canadian tertiary care centre

To reduce potential false‐positive warm autoantibody (WAA) by solid‐phase red cell adherence assay (SPRCA), our centre implemented a new antibody investigation algorithm (AIA) by classifying cases with panreactive SPRCA but negative saline‐indirect antiglobulin test as ‘antibody of undetermined significance’ (AUS) after excluding clinically significant antibodies. We assessed the effects of the new AIA and subsequent alloantibody formation in patients with AUS.


INTRODUCTION
The solid-phase red cell adherence assay (SPRCA) was developed to improve sensitivity in red blood cell (RBC) antibody detection and identification using standardized and automated technology.With the higher sensitivity, non-specific reactions such as panreactivity are frequently observed in SPRCA testing [1,2].This higher sensitivity is also seen in other automated methods, such as gel agglutination column methods [3].Reactivity with all RBCs tested or panreactivity by SPRCA can be due to warm autoantibody (WAA), alloantibody to high prevalence antigen, multiple alloantibodies to common antigens, therapeutic substances-anti-CD38 monoclonal antibody, and other non-specific reactions.Panreactivity from WAA commonly interferes with RBC antibody detection, causing incompatible crossmatch and obscuring underlying alloantibodies, possibly increasing the risk of haemolytic transfusion reaction due to incompatible RBC transfusion.Further specialized testing, such as RBC adsorption and recipient RBC antigen phenotyping or genotyping, is often required to exclude alloantibodies and provide antigen-matched RBCs for patients.As this specialized testing may only be available at either external reference immunohematology laboratories or large transfusion medicine services, RBC provision in patients with possible WAA is often delayed.Overdiagnosis of WAA may also lead to patients being inappropriately treated with immunomodulatory therapy if associated with anaemia, given a false diagnosis of warm autoimmune haemolytic anaemia.
While accurate detection of WAA is important, the differential diagnosis of panreactive antibody screen includes an antibody of undetermined significance (AUS).AUS is commonly used to report unexplained reactions when antibodies against specified RBC antigens have been ruled out.AUS interferes with pre-transfusion testing by reacting similar to WAA, but unlike WAA detection, it is unlikely to be clinically significant to the patient and does not warrant further investigation.Chemicals in testing reagents, enhancement media, drugs and other non-RBC antibodies producing anomalous agglutination are responsible for AUS.Liu et al. described a high frequency of AUS at 18% using gel technology, but the data on AUS frequency by the SPRCA platform are still lacking [4].
Vancouver General Hospital (VGH) is a quaternary care centre in Vancouver, British Columbia, Canada, which performs approximately 46,000 pre-transfusion tests and transfuses 25,000 RBC units per annum.VGH provides multiple subspecialty medical services, including haematology-oncology and bone marrow transplant, subspecialty surgical services, including cardiac surgery and organ transplantation, and an adult Level 1 trauma service.In modern transfusion medicine services that serve a large patient population, a high-throughput automated RBC antibody screening method is required to serve patients promptly.After SPRCA was implemented as a primary RBC antibody screening and identification platform in 2014, panreactive antibody screens were frequently observed, leading to a high volume of further antibody investigations to exclude RBC alloantibodies.
To distinguish AUS from potential WAA, saline-indirect antiglobulin test (SIAT) via manual tube testing was introduced into the antibody investigation algorithm (AIA) to increase the specificity of findings.SIAT determines the reactivity of the antibodies at 37 C in saline without enhancement media, whereas a low ionic strength solution (LISS) is used in SPRCA.We hypothesize that panreactivity present in SPRCA alone, but not in SIAT, should be considered false-positive reactivity or AUS.After implementing the new AIA integrating the SIAT results, the frequencies of potential WAA, AUS and other RBC antibodies at our centre were evaluated, comparing the old and new AIAs.Historical and subsequent antibody results and transfusion history of patients with AUS were also reviewed, focusing on RBC alloantibody development.

SAMPLES AND METHODS
A retrospective observational assessment examining frequencies of RBC antibody results before and after implementation of the old and new AIAs between 1 September 2017 and 31 August 2021 was completed, with the implementation occurring on 1 September 2019.
In patients with AUS, the following variables were collected: patient demographics, previous and subsequent antibody investigation results, auto-and alloantibody formation, transfusion history and direct antiglobulin test (DAT) results.A review by the University of British Columbia Research Ethics Board determined that this was a quality improvement initiative involving antibody investigation findings that may not be necessarily generalizable to other centres or patient populations.Only demographic and laboratory findings were utilized without patient outcomes.

Institutional antibody screening and identification methods
RBC antibody screening was performed using one of two SPRCA automated instruments (Echo and Neo; Immucor, Norcross, GA).Both instruments utilized the two-or three-cell antibody detection test plates (Capture-R Ready-Screen, Immucor).The automated instruments interpreted and graded the antibody results, and were confirmed by medical laboratory technologists as negative, weak+, 1+, 2+, 3+ or 4+.Samples positive at the screening by SPRCA (weak+ to 4+) were typically re-tested using a trio panel PEG-IAT additive tube method (Gamma PEG, Immucor).
After a confirmed positive antibody screen by PEG-IAT, further antibody identifications were performed using the SPRCA 14-cell panels (Capture-R Ready-ID, Capture-R Ready-ID Extend I and Extend II; Immucor).If needed, the testing was supplemented by a PEG additive tube method using the previously described grading strategy.We used double-dose antigen-positive RBCs for the exclusion of clinically significant antibodies.
All samples positive at screening by SPRCA and confirmed by PEG-IAT were investigated further by SIAT screening, DAT using polyspecific antibodies followed by anti-IgG and anti-C3 if indicated.
When DAT was positive in patients who received RBC transfusion in the past 90 days, an eluate from patient RBCs was tested with an antiglobulin (Gamma ELU-KIT II, Rapid Acid Elution, Immucor).
For patients with a known history of anti-CD38 monoclonal antibody therapy within the past 6 months and the antibody screen by SPRCA or PEG-IAT showed panreactivity, no further investigation would be required if the phenotypically matched RBC units had been transfused.However, if the patient was transfused with nonphenotypically matched RBC units, the antibody screen would be repeated monthly, and the specimen would be submitted to the reference laboratory for further investigation if the screen was positive.
Samples with panreactive SPRCA and PEG-IAT screens were assessed for cold agglutinin by step (1), cold screen, where the patient's plasma and three-screen cells were incubated at room temperature for 30 min.In cases of the positive cold screen, step (2), a pre-warmed PEG screen, would be performed for confirmation [5].

Development of our new institutional AIA
The general principle of developing our new AIA was to add the SIAT to assess the reactivity of antibodies at 37 C without adding enhancement media [6].SIAT is performed to eliminate non-specific reactions that may have led to false-positive determinations of potential WAA previously and to rule out clinically significant alloantibodies on the antibody panel, starting by using three screening cells in SIAT after a positive PEG-IAT antibody screen.Before 1 September 2019, a panreactive SPRCA on the antibody panel and confirmed positive PEG-IAT screen was considered potential WAA, regardless of DAT, elution study and SIAT findings.The definition of panreactive screen by SPRCA, PEG-IAT and SIAT shows no difference in the reactivity strength on each test cell greater than 1+ to exclude potential specificity.This algorithm resulted in frequent cases with possible WAA.From 1 September 2019, SIAT results were integrated into the new RBC AIA to distinguish AUS from potential WAA.In the new algorithm, panreactivity in SPRCA antibody panels where clinically significant antibodies could be ruled out with negative reactions on SIAT is considered AUS rather than potential WAA.Reinvestigation with a complete antibody identification will be performed if there is an increase in reactivity strength greater than 1+ compared to the previous results.
If there is a positive reaction that was not panreactive, the antibody is considered an unidentified antibody to denote a potentially clinically significant antibody may exist, but specific clinically significant RBC alloantibodies on the panels and autoantibodies have been ruled out.Reinvestigation will be performed on the next admission or if serological reactions change.If the reaction is panreactive with a positive SIAT, regardless of the DAT results, it is considered a potential WAA.We do not routinely refer all the specimens with potential WAA to the immunohematology reference laboratory to exclude the alloantibody of high-prevalence antigen but opt to perform RBC phenotyping locally and provide extended phenotypically matched RBC units to the patient, considering the referral process and the turnaround time generally takes approximately 24-48 h, which is impractical for our service.
A standardized approach for providing RBCs and reinvestigating potential WAA was also adopted.As fewer cases were predicted to have potentially false-positive WAA, a policy to provide patients with potential WAA prophylactic extended phenotype-matched RBC units (D, C, c, E, e, K, Jk a , Jk b , Fy a , Fy b , S, s) rather than units that were phenotypically matched for Rh and Kell only (D, C, c, E, e, K) was discussed and agreed upon with the jurisdictional Canadian Blood Services' blood supplier medical team (Canadian Blood Services acts as the national blood supplier in Canada except for Quebec).Reinvestigation will then be performed for potential WAA every 3 months if the patient has been transfused with extended phenotype-matched RBC unit(s) or monthly if the patient has been transfused without or partially extended phenotype-matched RBC unit(s).Given that our centre does not perform autoadsorption or alloadsorption, further investigations would be referred to the immunohaematology reference laboratory if the patient received non-phenotypically matched RBCs.For AUS, reinvestigation will be performed every 3 months if the patient has been transfused.Crossmatch-compatible RBC units by any of the following methods, SPRCA, PEG-indirect antiglobulin test (PEG-IAT) and SIAT, will be provided for patients with unidentified antibodies or AUS; if no crossmatch-compatible units can be found, least-incompatible units will be provided with consultation between the transfusion medicine and clinical service.Figure 1 illustrates the key comparison between old and new AIA.

Statistical analysis
R-software (version 4.1.1)was used for statistical analysis [7].Count data were analysed using Fisher's exact probability method, and continuous variables were analysed using the Mann-Whitney U test.A p-value of less than 0.05 was considered statistically significant.

RESULTS
A total of 2485 antibody investigation tests were performed among 1926 patients in the 4-year period.There were 464 repeat tests performed on the same patients using the same algorithm, which was considered duplications.After removing duplicate test results, 2021 antibody results remained in the analysis, with 1167 and 854 obtained using the old and new AIA, respectively.
The remaining alloantibodies, including anti-C, anti-C w , anti-Jk b , anti-Fy a , anti-Fy b , anti-S, anti-M, anti-Kp a and anti-Wr a , were less common with a prevalence of under 5%.
Frequencies of the categorized antibodies affected by the new AIA compared before and after implementing the new AIA with the odds ratios and p-values are included in Table 1.A significant reduction of potential WAA frequencies from 127 (11%) to 53 tests (6%) was observed (p-value <0.001).Frequencies of passive anti-D were also significantly reduced from 68 (6%) to 25 (3%) after implementing the new AIA ( p = 0.002).There was no significant difference in fre-  While no patients with AUS later transitioned to potential WAA using the new AIA, four patients developed specific RBC alloantibodies, anti-E, anti-C, both anti-C and anti-E and anti-Wr a , at a median time of 3 months post-AUS (IQR 0.9-4.8)with detailed transfusion histories described in Table 3.

DISCUSSION
SPRCA detects RBC antibodies by capturing their adherence to lysed RBC membranes immobilized on the polystyrene micro-wells [8].
Despite its high sensitivity, excellent throughput and less manual work for the initial antibody screen and panel, SPRCA appears to produce more frequent non-specific reactivity than tube methods.Dwyre et al. reported a 34.5% false-positive rate using SPRCA, Capture-R, to detect RBC antibodies compared to PEG methods as a gold standard [9].The higher false positives led to a higher cost of additional investigations.Bunker et al. described 10.9% of inconclusive results from positive manual SPRCA over negative PEG.While manual SPRCA screening identified the most clinically significant RBC alloantibodies with 80.7% sensitivity and 96.7% specificity at the lowest reagent and labour cost, sequential identification methods using the automated and manual SPRCA of positive screens followed by PEG also had the lowest hands-on time [10].The differential sensitivity between SPRCA or PEG-IAT and SIAT can also impact the WAA-positive rates because of enhancement media used in addition to automated versus manual tube testing, such as LISS in SPRCA and PEG.
When specificity of the autoantibody exists, WAA is most often directed against highly prevalent antigens such as the Rhesus blood group, Kp b and U at the optimal temperature of 37 C [11].
T A B L E 1 Frequencies of categorized antibodies impacted by the new antibody investigation algorithms with the odds ratio (OR) and p-values from Fisher's exact test for two proportions.antigen determination is unreliable in patients with recent transfusions, genotyping to predict recipient RBC antigens plays a significant role in guiding RBC selection, though turnaround times in emergency situations currently limit its usefulness in that setting.
For the above reasons, panreactivity was previously interpreted as potential WAA but is now classified as AUS and will not require unnecessary serological investigations, RBC phenotyping and genotyping.To mitigate the non-specific reactions in SPRCA, Dwyre and Bunker had previously shown that adding a second antibody screening method alleviated false positives [9,10].The use of SPRCA and PEG as primary and secondary methods for antibody screening since 2014 at VGH has shown similar rates of panreactivity, leading to a large number of cases that required further investigations.Many cases with non-specific reactions remain inconclusive even after excluding the antibodies with a known impact on transfusion, where complete antibody determination may not be warranted if crossmatchcompatible RBC units can be provided.The reduction in potential WAA frequencies could be attributed to several factors, such as the prevalence of WAA in particular periods and the reasons for antibody screening, such as routine pre-operative testing or in patients with haemolytic features.However, since the antibody results had to be issued promptly and only one diagnosis could be made, it is impossible to simultaneously perform a direct comparison using old and new AIAs.It is also not feasible or costeffective to refer all specimens with potential WAA, which may have included both true WAA and alloantibodies to high prevalence antigens or AUS, to the reference laboratory for confirmation.
To monitor the patients with AUS for new antibody formation, complete antibody investigations were repeated every 3 months if the blood samples were available.However, if there were significant changes in the antibody screen reactivity by SPRCA, (greater than 1+ of increased reactivity strength on each test cell than the previous findings), reinvestigations would be performed before 3 months.Seventeen patients previously had potential WAA from the old algorithm, and are now classified as AUS when using the new AIA.Among these, only three patients had follow-up antibody investigations based on pre-transfusion testing ordered by clinicians, and all remained as AUS at the follow-up time of 3 and 5 months.One case of AUS later became an unidentified antibody at a follow-up time of 22 months.
Despite the different diagnostic criteria between AUS and unidentified antibodies where the latter's reactions are not panreactive, the policy to provide crossmatch-compatible RBC units is the same.
Four patients were initially identified as having an AUS but later developed specific RBC alloantibodies; two received RBC transfusions between 4 days and 1 month.With the short intervals between AUS identification and subsequent specific alloantibody formation, these scenarios may raise concerns about missing either existing or newly formed antibodies using the new AIA.For two patients for whom anti-E and anti-Wr a were identified without RBC transfusion events, it is possible that the new AIA had missed a low level of antibodies from a previous pregnancy in the first case, or was not able to detect the pre-existing anti-Wr a because the Wr a antigen-positive cells were not used for the SIAT screen in the second case.
A similar phenomenon of panreactive SPRCA accompanied by a negative autocontrol and a lack of specificity but later acquired alloantibodies was described by reflex DATs to reduce the cost of testing [21].
In our study, polyspecific DAT was positive in 40 of 57 patients with AUS (70%).Although a positive DAT enables the diagnosis of WAA, associated with autoimmune haemolytic anaemia, DAT alone does not define WAA since it can be positive in up to 15% of hospitalized patient specimens [22].Positive DAT in AUS can be explained by non-specific RBC agglutination in many conditions, such as drugs, antiphospholipid antibodies, infections and interferences from IgG unassociated with haemolysis [23].The elution study findings were also panreactive in 10 of 11 patients with AUS (91%), ensuring no antibodies of common specificity coating the RBCs.
Limitations of our study were that our retrospective analysis would not have been able to determine any clinical impact of our new AIA.Changes in reinvestigations when comparing our old and new AIAs could not be properly abstracted from our laboratory information system.As our centre is not a reference laboratory and certain immunological techniques are less commonly performed in Canada, we did not perform either enzyme-or chemical-treated RBCs, nor autoadsorption or alloadsorption.
These tests are often referred to an immunohematology reference laboratory through the Canadian Blood Services as they often do not aid transfusion management in a timely fashion [6].Patients' diagnoses and transfusion indications should also be carefully reviewed when handling AUS results, where special patient populations, such as those with haemoglobinopathies, will be at a higher risk for alloimmunization and delayed haemolytic transfusion reactions.Thus, we suggest considering the local patient population for additional investigations and prophylactically matched RBC units in some transfusion-dependent patient populations.
In conclusion, we observed a significant reduction in the frequencies of potential WAA detection post-implementation of the new AIA integrating SIAT results.This approach helped mitigate unnecessary further investigations and phenotypically matched RBC use.However, some patients still developed subsequent RBC alloimmunization, prompting careful evaluation of clinically relevant alloantibodies before determining AUS.Serial antibody investigation tests are also warranted considering forming or evanescent antibodies.
Among 57 AUS cases, other RBC antibodies were previously identified in 21 patients as shown in Table2, including 14 with potential WAA, 3 with combined cold agglutinin and potential WAA, 2 with cold agglutinin and 2 with unidentified antibodies.Twenty-four patients with AUS previously received RBC transfusions.While 16 patients had confirmed transfusion records with a median interval of the most recent transfusion of 4 days (IQR 4-127), 8 patients recalled RBC transfusions but with uncertain dates or documentation.PEG-IAT showed panreactivity with +w reaction in 43 (75%) and 1 to 2+ in 14 patients (25%).Polyspecific DAT was positive in 40 patients (70%), showing +w to 2+ and 3 to 4+ in 29 and 11 cases, respectively.IgG was positive in 32 cases, with +w to 2+ and 3 to 4+ in 29 and 3 cases, respectively.C3d was +w to 2+ in 16 patients.Elutionstudies were performed in 11 cases, showing panreactivity in 10 cases and a negative reaction in 1 case, respectively.
SIAT is still considered the gold standard technique used to detect in vitro sensitization of RBCs by IgG to detect clinically significant IgG antibodies in the serum or plasma of the patient and for crossmatching.Integrating a negative SIAT to help exclude potential WAA in cases of panreactive SPRCA and/or PEG-IAT can simplify the pre-transfusion investigations and prevent transfusion delay.As shown in our study, the frequencies of potential WAA detection were significantly decreased using a new AIA, including the SIAT method.
Antibody results of patients with AUS who previously had RBC antibodies and the transfusion histories.Development of RBC antibodies after AUS identification and the transfusion histories.
[15,16]ations: AIA, antibody investigation algorithm; AUS, antibody of undetermined significance; CI, confidence interval; WAA, warm autoantibody.investigationmethods.In the presence of WAA, underlying alloantibodies are often unable to be excluded, which may lead to an inadvertently haemolytic transfusion reaction if patients are transfused with antigen-positive blood.Patients with WAA are also at risk of developing RBC alloimmunization with a prevalence of 10%-53%[12,13].Avoiding exposure to highly immunogenic antigens, such as the Rhesus and Kell antigens, through transfusing RBCs that are negative for antigens that the patient also lacks, referred to as prophylactic antigen matching, can mitigate the complexity of the pre-transfusion testing[14].Selecting RBC units that are phenotypically or genotypically matched to the patient can prevent alloimmunization, reduce the cost of investigations and avoid delay of RBC provisions in patients who require ongoing transfusion support[15,16].As serological RBC T A B L E 2