Does decreasing the incubation period used in the antibody screen affect its sensitivity?

Pre‐transfusion testing (PTT) encompasses a set of mandatory laboratory tests performed before red blood cell transfusion. The antibody screen, one component of PTT, commonly includes a 10–20 min incubation. The primary aim of this study was to determine if this period can be reduced when using current immunohematology methodologies.

Anti -D, -E, and -K are the most common clinically significant allo-antibodies in Western Australia. 10tibodies other than ABO isoagglutinins are detected using the antibody screen (the 'screen').This test is required as part of pre-transfusion testing to best ensure the provision of compatible blood product. 11It is performed in test tubes (the tube method), with column agglutination technologies (CAT), or with solid phase assays.It includes an incubation at 37 C to facilitate antibody uptake. 12This incubation is 30-60 min when performed with RBCs resuspended in isotonic saline. 13This period can be reduced to 10-20 min with the use of enhancement media such as low ionic strength saline (LISS) 14 or polyethylene glycol (PEG), 15 with the specific time used described in the manufacturer's instructions for use, and dependent on the technique.However, the screen can detect antibodies after only 1 min incubation when performed using the tube method enhanced with LISS. 14[18][19] The hypothesis of this study is that the incubation period used in the antibody screen can be reduced when using Glass or Gel-based CAT, or a solid phase red cell adherence (SPRCA) assay.If so, the time required to complete pre-transfusion testing may be decreased.

| Ethics approval
This study was approved by Royal Perth Hospital (RPH; Project Number RGS0000001018) and Curtin University (Approval Number HRE2018-0470) Human Research Ethics Committees.A PathWest employee that was not a member of the research team de-identified patient samples and assigned them a unique 5-digit code.Only this code and the specificity of the antibody/ies in the sample were provided to the research team.

| Plasma samples
Quality control (QC) samples, containing antibodies of differing specificity, were donated by RPH.They were centrifuged and the plasma separated and stored at 2-8 C.
The protocol used by the Transfusion Medicine Unit at RPH requires antibody-containing patient plasmas to be frozen at À20 C if the sample is not required for an immediate crossmatch.This study aimed to analyse all of these samples available during the period February to August 2018 that were from unique patients.One hundred samples containing single or multiple alloantibodies of differing specificity (Table 1) were analysed in this study.Plasmas of insufficient volume or containing autoantibodies or antibodies of undetermined specificity (AUS) 20 were excluded.Samples were thawed, aliquoted into 12 Â 75 mm plastic test tubes, and testing completed within 8 h.

| Determination of antibody titre
The antibody titre was determined in patient plasmas or QC material containing a single antibody, using a saline-IAT tube method.Briefly, samples (200 μL) were prepared to a 1/2048 dilution using Imidazole Buffered Saline (IBS; 5 mM Imidazole, 150 mM NaCl, pH 7.4) containing 5% Albumin.Fifty microlitre of 3% RBCs (Immulab, Parkville, VIC) with single-dose (presumed heterozygous) antigen expression were added to each tube and incubated at 37 C for 30 min.Samples were washed four times with IBS then centrifuged for 20 s after the addition of anti-IgG AHG reagent (Immulab, Parkville, VIC).All reactions were immediately read and graded on a 0-4+ scale, beginning with the 1/2048 dilution.AHG control cells (Immulab, Parkville, VIC) were added to negative reactions.The titre was reported as the reciprocal of the highest dilution producing a macroscopic 1+ reaction.

| Determination of antibody isotype
Dithiothreitol (DTT) was used to differentiate IgG and IgM anti-M and anti-Le a in patient samples, using the method described in the AABB technical manual. 12Antibody detection was performed using a saline-IAT method in glass tubes, testing 200 μL of treated plasma against 50 μL of 3% RBCs (Immulab, Parkville, VIC).Remaining steps were as outlined in the "Determination of Antibody Titre" method.

| Antibody detection
Antibody screens were completed using three different methods-a Sephadex Gel Matrix CAT system (Gel; Bio-Rad Laboratories, Gladesville, NSW), a Glass Microbead CAT system (Glass; Ortho Clinical Diagnostics, Mulgrave, VIC), and a SPRCA method (Echo Capture-R Ready Screen, Immucor, Norcross, Atlanta USA).LISS-IAT cards and SPCRA microplates were obtained from the relevant manufacturer and used with specialty incubators and centrifuges for each method.
Both CAT systems were performed manually and according to manufacturer's instructions.0.8% RBCs were from Immulab (Parkville, VIC).After incubation, all reactions were read immediately and graded on a 0-4+ scale.
The SPRCA method was performed on the Echo analyser (Immucor, Norcross, Atlanta, USA).Capture-R Indicator Red Cells, Capture-LISS, and Capture-R Ready Screen (3) microplate strips were generously donated by Immulab (Parkville, VIC).After incubation, results were graded automatically by the analyser, with confirmation performed by manually grading the reactions (0-4+ scale) in both the automatically captured well image, and the well itself.

| Optimisation of the incubation period used in each screening method
For each screening method, 10 QC samples and two phenotyping reagents containing antibody of differing specificity were tested using incubation times of 1, 5, 10 and 15 min (all three methods), and 20 min (SPRCA assay only; manufacturer's recommended incubation time).Anti-Fy b and anti-s IgG phenotyping reagents were used as these antibodies were not present in QC samples.To ensure that weak examples of antibodies could be detected, all 12 samples analysed during this phase were tested at the dilution corresponding to their titre.The shortest incubation period producing at minimum a 1+ reaction with all 12 samples was selected as the optimised time (i.e., an incubation period producing a 1+ w (weak) or 0 reaction with any of the 12 samples was considered too short).This was done to best ensure that antibodies with concentrations lower than these 12 diluted samples, when present in clinical samples, could be detected as at least 1+ w reactions.

| Antibody detection in patient samples using optimised incubation periods
Antibody screens were performed on patient samples by all three methods, using the optimised incubation period determined for each method (Gel CAT, 15 min; Glass CAT, 10 min; SPRCA method, 5 min).

| Statistical analysis
All statistical analyses were performed using Stata/IC 14.2 (StataCorp, Texas).Reaction grades and antibody titres were considered categorical variables recorded as 0/1+ w /1+/2+/3+/4+ and 1/2/4/8/16/32/ 64/128/256/512, respectively.Data was analysed to descriptively report the reaction grade generated by each specific antibody in the three screening methods.The following analyses were performed using Fisher's exact tests: the overall association between antibody titre and reaction F I G U R E 1 Determination of the optimal incubation period in the Gel CAT system.Each graph shows the reaction strength of antibodies differing in specificity, tested at a dilution corresponding to their titre.Incubation periods of 15 (A), 10 (B), 5 (C), and 1 (D) minute(s) were used.
grade in each screening method; and, for each antibody, the association between its titre and reaction grade in each screening method.A p-value less than 0.05 was considered statistically significant.

| Optimisation of the incubation period used in each screening method
In the Gel CAT system (Figure 1), all antibodies produced at least a 1+ reaction when using the standard 15-min incubation period (Figure 1A).All antibodies were detected after a 10-min incubation, though a 1+ w reaction was obtained with the anti-Jk b (Figure 1B).At shorter incubation periods, further weak reactions were produced, or antibodies were not detected (Figure 1C, D).Therefore, the standard 15-min incubation (Figure 1A) was used when testing patient samples with this system.
In the Glass CAT system (Figure 2), all antibodies produced at least a 1+ reaction when incubated for 15 or 10 min (Figure 2A, B).
Multiple antibodies were not detected after 5-and 1-min incubation periods (Figure 2C, D).Therefore, a reduced incubation time of 10 min (Figure 2B) was used when testing patient samples with this system.
In the SPRCA method (Figure 3), reaction grades of 1+ or higher were produced using the standard 20-min incubation, and the reduced periods of 15, 10 and 5 min (Figure 3A-D).However, multiple antibodies were not detected using a 1-min incubation (Figure 3E).Therefore, a 5-min incubation (Figure 3D) was used when testing patient samples with this system.It is important to note that the SPRCA method only detects IgG antibodies. 18,21Therefore, the results for the QC sample containing anti-M (IgM in nature) were not considered when determining the incubation period used in further testing.

| Detection of antibodies in patient samples using optimised incubation periods
To assess if these incubation periods could be used in a clinical setting, the sensitivity of each optimised method was determined using 100 patient samples containing antibodies of known specificity.
F I G U R E 2 Determination of the optimal incubation period in the glass CAT system.Each graph shows the reaction strength of antibodies differing in specificity, tested at a dilution corresponding to their titre.Incubation periods of 15 (A), 10 (B), 5 (C), and 1 (D) minute(s) were used.
Results are summarised in Table 1.Each method detected all examples of anti -D, -C, -c, -Fy a , -Jk a , and those containing multiple antibodies, demonstrating that the optimised incubation periods do not affect assay sensitivity for any of these.The Gel CAT system and SPRCA method detected all examples of anti-E (100% sensitivity; Table 1).However, one anti-E, having a titre of 1, was not detected in the Glass CAT system (detected 15/16 samples, 93.75% sensitivity).Both the SPRCA and Gel CAT systems had reduced F I G U R E 3 Determination of the optimal incubation period in the SPRCA method.Each graph shows the reaction strength of antibodies differing in specificity, tested at a dilution corresponding to their titre.Incubation periods of 20 (A), 15 (B), 10 (C), 5 (D), and 1 (E) minutes were used.sensitivity in detecting anti-K (SPRCA: detected 17/18 samples, 94.44% sensitivity; Gel: detected 15/18 samples, 83.33% sensitivity), while the Glass CAT method had 100% sensitivity.The antibody not detected in the SPRCA assay had a titre of 2 and was detected by both CAT methods; those not detected in the Gel CAT had titres of 1, 1, and 2. Of all samples containing anti-K, 5 had a titre of 1 and 3 had a titre of 2.
Anti-M and anti-Le a may be IgG or IgM.Samples containing these antibodies (7 containing anti-M; 8 containing anti-Le a ) were treated with DTT to determine their isotype.Of the samples containing anti-M, two were IgG and the remaining 5 were IgM.The Glass CAT system detected all 7 of these.The Gel CAT system detected 6 (85.71% sensitivity); the antibody missed was IgM in nature with a titre of 2. As expected, the SPRCA method detected only the two IgG anti-M (100% sensitivity).The anti-Le a in all samples was determined to be IgM in nature.The Glass CAT system detected all of these, while the Gel CAT and SPRCA methods did not detect any.
Overall, with the optimised incubation times, the Glass and Gel CAT systems had sensitivities of 98.9% and 96.6%, respectively, and the SPRCA method had a sensitivity of 98.9% (excluding the detection of IgM antibodies).

| Comparison of reaction strengths obtained from antibody screening methods with optimised incubation periods
To determine if any of the optimised systems differentially detected antibody/ies of certain specificity/ies, the reaction grades produced by each antibody in the three systems were compared (Figure 4).-D, -E, -c, or -Fy a were not significantly different between the three methods (Figure 4A-D; p > 0.05).

Reaction grades obtained with samples containing anti
Reaction grades obtained with samples containing anti-K were significantly higher in the SPRCA system than in the Gel CAT system (Figure 4E; SPRCA v Gel, p = 0.028).There was no significant difference in reaction grades between the SPRCA and Glass CAT systems, or the Gel CAT and Glass CAT (Figure 4E; p > 0.05).
Reaction grades obtained with samples containing anti-Jk a were significantly higher in the SPRCA assay than in both the Gel (Figure 4F; p = 0.009) and Glass (Figure 4F; p = 0.012) CAT systems.There was no significant difference in reaction strength between the Gel and Glass CAT methods (Figure 4F; p > 0.05).
T A B L E 1 Sensitivity of antibody screening methods using optimised incubation periods.Note: Described for each antibody is the number of patient samples tested and the number (in brackets, %) that produced a positive result in each screening method.The incubation period used in each method is below the method name.The overall assay sensitivities were calculated with IgG antibodies only, and antibodies that were IgG or IgM.
a Indicates that some samples tested contained antibodies that were IgM in nature and therefore were not detected in the SPRCA method.Note: For all methods, each antibody is grouped by the titre and the reaction strength obtained (in brackets, %).Spaces that are shaded had no samples in the corresponding titre and reaction strength.

| Analysis of the association between antibody titre and reaction strength
Given the higher reaction grades obtained in the SPRCA assay with samples containing anti-K and anti-Jk a , we were interested to investigate the relationship between antibody titre and reaction grade produced in each method.In all methods, the association between antibody titre and reaction strength was significant, though this significance was weaker in the SPRCA method (Table 2; Gel CAT, p = 0.001; Glass CAT, p < 0.001; SPRCA, p = 0.041).Notably, most antibodies with a titre of 1 produced a reaction grade of 1+ or less in both CAT methods, and a reaction grade of 1+ or more in the SPRCA assay.Most with a titre between 2 and 16 produced grades between 1+ and 2+ in the Gel CAT, between 2+ and 3+ in the Glass CAT, and between 3+ and 4+ in the SPRCA.Finally, for antibodies having a titre greater than or equal to 32, a higher proportion of 4+ reactions were produced in the Glass CAT and SPRCA assays compared to the Gel CAT (Gel CAT: 75% vs Glass CAT and SPRCA: 93.3%).
Sub-analyses of these data revealed that, in the Glass CAT and SPRCA assay, the titre of anti-Fy a was significantly associated with the reaction grade produced (Glass CAT, p = 0.008; SPRCA, p = 0.010).Additional significant associations were observed in the Glass CAT with samples containing anti-E ( p = 0.004) and anti-K ( p = 0.013).For all other antibodies and with all methods, there was no significant association between antibody titre and reaction grade obtained ( p > 0.05).

| DISCUSSION
This study investigated if the incubation period used in the antibody screen performed using Gel or Glass CAT, or the SPRCA assay, could be decreased.Using QC samples and phenotyping antibodies, it was found that the incubation periods used in the SPRCA assay and the Glass CAT system could be reduced to 5 and 10 min, respectively.
Decreasing the incubation period in the Gel CAT system was not possible without reducing assay sensitivity, so the standard 15-min incubation period was used.Only these periods were used in subsequent investigations using patient samples.
Antibody screens were performed by all three methods on 100 patient samples.When considering only the detection of IgG antibodies, the Glass CAT system and the SPRCA method had sensitivities of 98.8%, while that of the Gel CAT system was 96.6%.Ideally, for both the Glass CAT system and SPRCA assay, the effect of their shorter incubation period on assay sensitivity would have been determined by comparing the results obtained with those attained using the standard incubation period of the assay.However, this was not possible due to the limited volume of patient sample available.It should be noted that the assay sensitivities obtained in this study are consistent with previous studies that used the standard incubation times. 18,22Given the results of the current study, the SPRCA method is preferred because its shortened incubation period (5 min) is half of that of the Glass CAT (10 min) and a third of that of the Gel CAT (15 min).However, for laboratories using the Glass CAT system, a transition to the 10-min incubation period could be considered should they wish to expedite antibody detection.It must be highlighted that the described optimised incubation periods are outside the manufacturer's instructions for use, and therefore any institution wishing to implement them must first validate their use in their laboratory and provide appropriate staff training and supporting procedural documentation before introducing them into practice.
For each antibody, the reaction grades obtained with each method were compared.In samples containing anti-D and -E, there was no significant difference in reaction grades between the three methods.In samples containing anti-c, there was a non-significant trend to the reaction grades being stronger in the SPRCA assay than with the CAT assays.This lack of significance may be due to the small number of samples tested (n = 3).Therefore, it would be appropriate to repeat this testing on a larger number of samples, because anti-c is incompatible with emergency-issue rr RBCs and has caused a fatal HTR in an emergency situation. 23In these scenarios, rapid identification of this antibody by using a shorter incubation may result in transfusion of these RBCs being ceased sooner and the patient switched to RBCs more likely to be serologically compatible, potentially avoiding a fatal HTR.
Reaction grades obtained with anti-K were significantly greater in the Glass CAT system and the SPRCA method than in the Gel CAT.
Therefore, it is surprising that the SPRCA method did not detect a low-titre anti-K in one sample.It is possible that this was an IgM antibody, 24,25 though this could not be confirmed as there was insufficient sample to treat with DTT.The Gel CAT system did not detect low titre anti-K in three samples.This is an important result to note, given that the incubation period used was that suggested by the manufacturer.[28] Similarly, a low-titre anti-E was not detected by the Glass CAT system but was detected by the Gel CAT system and the SPRCA method.Ideally, this sample would have been re-tested using the standard incubation of the Glass CAT system.However, there was insufficient sample to complete this testing.Other studies, using the same systems used in this study, have also not detected some examples of anti-E. 18,220][31][32] Given that there wasn't a single system that detected all examples of these antibodies, it may be pertinent for laboratories to employ more than one antibody screening method, as has been suggested by others. 22In the context of emergency situations, it is important to note that HTR prevalence due to these antibodies is reduced by administering emergency issue E-and K-RBCs where possible.
Reaction grades obtained with anti-Jk a were significantly higher in the SPRCA assay than in either CAT system.However, all methods had 100% sensitivity in detecting these antibodies.This is in contrast to the findings of Weisbach et al., 18 who failed to detect all examples of anti-Jk a using standard incubation times in all three systems.In a second study, multiple examples of anti-Jk a were detected by the SPRCA assay but not the Gel CAT system. 33The different findings between studies is likely due to biological variation, as anti-Jk a characteristically has low titres, produces weak reactions, and may not be detected. 34It is for these reasons that anti-Jk a continues to be the most common antibody implicated in HTRs, 35 and can be fatal in these situations. 36l examples of Anti-Fy a were detected by, and produced similar reaction strengths in, all three methods.This is consistent with the finding that the anti-Fy a titre was associated with the reaction strength produced in the Glass CAT method and SPRCA assay.Though not immediately comparable to the current study, these results are in contrast with those of Yamada et al, 37 who demonstrated that the SPRCA was less sensitive than the tube PEG-IAT in detecting these antibodies.
The reason for this discrepancy is unknown.
Most anti -M and -Le a are IgM and do not react at 37 C, and therefore are rarely clinically significant.Those that react at 37 C are usually IgG, can cause severe HTRs, and must be detected. 38The Glass CAT system detected all examples of anti-M and anti-Le a , while the Gel CAT detected most examples of anti-M but not anti-Le a .As expected, the SPRCA method detected only the two examples of IgG anti-M.This result highlights one known benefit of the SPRCA method, in that it detects only IgG antibodies. 18,21 seems reasonable to assume that an antibody's titre is directly proportional to the reaction grade it produces.Indeed, our data indicates significant association between antibody titre and reaction grade in both CAT systems, and significant but weaker association between the two in the SPRCA assay.Notably, low-titre antibodies tended to produce higher reaction grades in the SPRCA method.This may explain, at least in part, the early detection of clinically relevant anti-Jk a by the SPRCA assay but not by the Gel CAT. 33is study has several limitations.First, the sample size is relatively small, resulting in the low representation of anti-c and the absence of anti-e, anti-Jk b , anti-Fy b , anti-s, and antibodies against high or low prevalence antigens such as anti-k, anti-Kp a , and anti-Wr a .Several of these have been implicated in HTRs, 35,36 including those fatal. 36However, these antibodies are rarely encountered in the laboratory, consistent with their absence in our sample population.Second, to optimise the incubation period of each assay, reagent samples were used at a dilution corresponding to their titre.
Therefore, it is important to highlight the possibility that antibodies having a concentration lower than these dilutions could produce a 1+ w reaction, or may not be detected, when using the optimised incubation periods described.These antibodies may include antibodies of undetermined specificity (AUS).Approximately one-third of AUS produce 1+ w reactions in antibody detection tests. 20AUS are of mixed clinical significance. 39They may represent antibodies to low prevalence antigens, 20 which rarely cause HTRs 40 ; antibodies to chemicals within reagents or media; or rouleaux. 20They may also be developing alloantibodies, with some being previously evanescent antibodies 41 representing an early anamnestic response post transfusion. 20,39Here, it should be noted that the Western Australian (WA) branch of Australian Red Cross Lifeblood compiles the Antibody Register, a database of all alloimmunised individuals within WA.This allows laboratory staff to identify individuals with previously detected but evanescent antibodies in our State.Evanescent antibodies are otherwise not identified by any antibody detection test. 41In addition, restricting the validity of samples from those who have been recently transfused to 72 h should facilitate the identification of these AUS in subsequent samples, once the anamnestic response has further developed.Further, evidence from the current study demonstrating the detection of weak antibodies is presented in Table 2. Twenty-one clinical samples containing antibodies with titres of 1 or 2 were analysed.Of these, only 1 was missed by each of the Glass CAT or SPRCA assays which used the optimised incubation periods.In contrast, 3 were missed by the Gel CAT performed according to the manufacturer's instructions for use.Finally, others have demonstrated that no single method will detect all antibodies. 18,22However, in light of the described limitations, laboratories wishing to expedite screening results by using the shorter incubation periods could consider limiting their use to clinical scenarios requiring expedited results.Additionally, they could consider re-performing the antibody screen according to the manufacturer's instructions for use once time permits.
It can take up to 60 min to complete pre-transfusion testing. 42is study suggests that the incubation period used in the antibody screen can be decreased to 5 and 10 min when using the SPRCA and Glass CAT methods, respectively, without decreasing assay sensitivity.
This may allow pre-transfusion testing to be expedited in laboratories that validate these findings, resulting in earlier administration of crossmatch-compatible blood products if required.

F I G U R E 4
Comparison of reaction strengths obtained with each antibody screening method using optimised incubation periods.Each graph shows the reaction grade of each antibody in the three testing systems.The line represents the mean reaction grade.Only patient samples containing single antibodies are included.Reaction strengths obtained with samples containing anti-M and anti-Le a were not compared statistically because these antibodies were not detected by all three systems.(A) Anti-D, (B) Anti-E, (C) Anti-c, (D) Anti-Fy a , (E) Anti-K, (F) Anti-Jk a , (G) Anti-M, and (H) Anti-Le a .**p ≤ 0.01, ***p < 0.001.T A B L E 2 Association between antibody titre and reaction strength.