Human Immunology xxx (xxxx) xxx–xxx

Contents lists available at ScienceDirect

Human Immunology
journal homepage: www.elsevier.com/locate/humimm

Rapid optimized flow cytometric crossmatch (FCXM) assays: The Halifax

and Halifaster protocols
⁎
Robert S. Liwskia, , Anna L. Greenshieldsa, David M. Conrada, Cathi Murpheyb, Robert A. Brayc,
Jorge Neumannd, Howard M. Gebelc
a
Department of Pathology, Dalhousie University, Halifax, Nova Scotia B3H 1V8, Canada
b
Southwest Immunodiagnostics Inc., San Antonio, TX 78229, USA
c
Department of Pathology, Emory University Hospital, Atlanta, GA 30322, USA
d
Laboratory of Transplantation Immunology, Santa Casa Hospital, Porto Alegre, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: The flow cytometric crossmatch (FCXM) assay, which detects the presence of donor specific HLA antibodies in
Flow cytometry cross match patient sera, is a cornerstone of HLA compatibility testing. Since relatively long FCXM assay turnaround times
HLA antibodies may contribute to transplant delays and increased graft ischemia time, we developed and validated two modified
Assay optimization crossmatch procedures, namely the Halifax and Halifaster FCXM protocols. These protocols reduce FCXM assay
Transplantation
time > 60% and simplify their set-up without compromising quality or sensitivity. Optimization of the FCXM
(the Halifax protocol) includes a 96-well tray platform, reduced wash times, increased serum to cell suspension
volume ratio, shortened incubations and higher incubation temperature. The Halifaster protocol is a further
modification, employing methods that improve lymphocyte purity compared to density gradient centrifugation
(96 ± 2.63% vs 69 ± 19.06%), reduce cell isolation time (by ∼40%) and conserve FCXM assay reagents.
Importantly, linear regression analysis of the median channel fluorescence shift (MCFS) values revealed excellent
concordance (R2 of 0.98–0.99) among all three FCXM protocols (standard vs Halifax vs Halifaster). Finally, a
retrospective review of 2013 crossmatches performed using the Halifax protocol demonstrated excellent cor-
relation with the virtual crossmatch (95.7% and 96.8% specificity and sensitivity, respectively) regarding the
identification of donor specific antibodies (HLA-A/B/DR) assigned based on the single antigen bead (SAB) assay
testing with a 2000 mean fluorescence intensity (MFI) cutoff. Implementation of the Halifax or Halifaster pro-
tocols will expedite pre-transplantation work-up and improve patient care.

1. Introduction
The flow cytometric crossmatch (FCXM) assay is used by the ma-
jority of histocompatibility laboratories in North America as part of pre-
transplant risk assessment to detect donor specific anti-HLA antibodies
[1–4]. The FCXM assay is significantly more sensitive than cytotoxic
crossmatch assays [5,6] allowing better detection of low level HLA
antibodies, thereby improving assessment of pre-transplant im-
munological risk [7–9].
The standard three color FCXM protocol currently used by the
majority of laboratories was described > 25 years ago [2,10] and can
be divided into two parts. The first part consists of donor lymphocyte
isolation [10] and treatment (e.g., pronase; DNase) to reduce Fc re-
ceptor expression and remove dead and dying cells [11]. The second
part is the crossmatch assay (set up, testing and acquisition) per se [10].
The entire protocol is time intensive, typically requiring four to five
hours to complete, and can lead to significant transplant delays and
compromise organ quality [12,13].
In this study, we developed an expeditious FCXM procedure without
compromising quality or sensitivity of the current assay. In the first part
of the study, we investigated the impact of several assay parameters on
FCXM results. Subsequently, these parameters were modified resulting
in a rapid optimized FCXM procedure (henceforth referred to as the
Halifax protocol), which is completed within 40 min (a 60% decrease

Abbreviations: 3SD, 3 standard deviations; CDC, complement-dependent cytotoxicity; DSA, donor-specific HLA antibodies; FCXM, flow cytometry crossmatch; FWB, flow wash buffer;
HLA, human leukocyte antigen; MCF, median channel fluorescence; MCFS, median channel fluorescence shift; MFI, mean fluorescence intensity; NC, negative control; PBMC, peripheral
blood mononuclear cell; PBS, phosphate buffered saline; PC, positive control; RT, room temperature; SAB, single antigen bead; SD, standard deviation; SFCXM, standard flow cytometry
crossmatch; SWID, Southwest Immunodiagnostics Inc.; TAT, turnaround time; VXM, virtual crossmatch; XM, crossmatch
⁎
Corresponding author at: Dr. D.J. Mackenzie Building, HLA Typing Laboratory, 5788 University Avenue, Room 625, Halifax, Nova Scotia B3H 1V8, Canada.
E-mail address: robert.liwski@nshealth.ca (R.S. Liwski).

https://doi.org/10.1016/j.humimm.2017.10.020
Received 28 July 2017; Received in revised form 25 October 2017; Accepted 31 October 2017
0198-8859/ © 2017 American Society for Histocompatibility and Immunogenetics. Published by Elsevier Inc. All rights reserved.

Please cite this article as: Liwski, R.S., Human Immunology (2017), http://dx.doi.org/10.1016/j.humimm.2017.10.020
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

compared to standard protocol) and maintains excellent assay sensi-
tivity. Next, we performed a retrospective analysis of 2013 cross-
matches performed using the Halifax FCXM and correlated the results to
a virtual crossmatch (VXM) based on the single antigen bead (SAB)
LABScreen assay testing. Finally, we incorporated the EasySep Direct
lymphocyte isolation technique [14] to improve lymphocyte purity
(> 90%), isolation time, and further optimize the FCXM assay (Hali-
faster protocol).
2. Materials and methods
2.1. Flow cytometric crossmatch assay reagents
Flow wash buffer (FWB) composed of phosphate buffered saline
(PBS; Life Technologies Inc., Burlington ON, Canada) with 2% (v/v)
fetal calf serum (Life Technologies Inc.) was used for all FCXM washes.
Anti-CD3-PerCP (clone SK7) and anti-CD19-PE (clone SJ25C1) mono-
clonal antibodies were purchased from BD Biosciences (Mississauga,
ON, Canada). Fluorescein (FITC) conjugated F(ab’)2 fragment goat anti-
human IgG, Fcγ specific polyclonal antibody (IgG-FITC), 1.0 mg/mL
stock solution, was purchased from Jackson ImmunoResearch
Laboratories Inc. (West Grove, PA).
2.2. Control sera
A pool of 10 non-sensitized patient sera and 20 highly-sensitized
patient sera served as the negative control (NC) and positive control
(PC), respectively, for all FCXM optimization and validation experi-
ments performed by the Halifax HLA laboratory.
2.3. Virtual crossmatching
A positive virtual crossmatch (VXM) was defined as having at least
one HLA DSA (HLA-A/B/C/DR/DQ) with ≥1000 mean fluorescence
intensity (MFI) identified by the LABScreen SAB luminex assay
(OneLambda, Canoga Park, CA). A negative VXM was assigned when all
HLA DSAs had MFI values < 1000. All patient sera were treated either
with 6.0 mM EDTA (Halifax and SWID HLA laboratories) [15,16] or
heat (56 °C; Santa Casa HLA Laboratory) prior to SAB testing to prevent
complement mediated interference.
2.4. Donor cells
Cells from volunteer donors were used for all FCXM optimization
experiments. Donor cells for the Halifax FCXM and Halifaster FCXM
protocol validation experiments were obtained from whole blood
samples collected from volunteers as well as live and deceased donors.
For the Halifax protocol evaluation studies performed by the Santa Casa
HLA laboratory cells were isolated from deceased donor spleen or
lymph node samples.
2.5. Cell isolation
2.5.1. Peripheral blood mononuclear cell enrichment
Peripheral blood mononuclear cells (PBMC) were enriched by
density gradient centrifugation. Briefly, 12 mL of acid dextrose citrate
anti-coagulated donor blood was mixed 1:1 with PBS. This blood mix-
ture was layered over 18 mL of Lympholyte-H density medium
(Cedarlane, Burlington, Ontario, Canada) in a 50 mL Falcon tube and
centrifuged at 400×g for 30 min with no brake. The mononuclear cell
layer was transferred into a fresh 15 mL Falcon tube and washed three
times in 10 mL PBS at 400×g for 10 min.
2.5.2. Lymphocyte isolation from peripheral blood samples
Lymphocyte purification from acid dextrose citrate anti-coagulated
donor whole blood was performed by immunomagnetic bead negative
selection method using the EasySep™ Direct Human Total Lymphocyte
Isolation Kit (EasySep™ Direct; STEMCELL Technologies Inc.,
Vancouver, BC, Canada) according to the manufacturer’s instructions.
After isolation, lymphocytes were washed twice with PBS at 400×g for
5 min.
2.6. Cell treatment with pronase and DNase
Following isolation, all donor cells (PBMC or lymphocytes) were
treated with pronase (4.7 kuntz units/mL; Sigma-Aldrich, St Louis, MO)
for 15 min, followed by DNase (11,000 u/mL; Sigma-Aldrich) for 2 min
at 37 °C. The cells were then washed twice in 1 mL PBS at 400×g for
1 min, and resuspended in 1 mL of FWB. Total cell number and cell
purity was determined using Sysmex XN10 analyser (Sysmex Canada
Inc. Mississauga, ON). The cell concentration was adjusted to either
8.3 × 106 cells/mL for the standard FCXM or 1.0 × 107 cells/mL for
both the Halifax and Halifaster protocols.
2.7. Flow cytometry crossmatch protocols
2.7.1. Standard FCXM, tube method
The standard FCXM (SFCXM), tube method (Table 1), was per-
formed as previously described [10]. Briefly, 30 µL of test or control
sera and 30 µL of PBMC (2.5 × 105 cells) were added to 5 mL poly-
styrene Falcon tubes, mixed by vortexing, and incubated for 30 min at
4 °C. The cells were washed three times with 1 mL FWB at 500×g for

Table 1
Comparison of the standard, Halifax, and Halifaster FCXM protocols.

Parameter Standard FCXM Tube Method Standard FCXM Tray Method Halifax FCXM Protocol Halifaster FCXM Protocol

Assay platform 5-mL, 12 × 75 mm tubes 96-well tray 96-well tray 96-well tray
Serum (µL) 30 30 50 30
Cell isolation method, cell preparation Lympholyte-H, PBMC Lympholyte-H, PBMC Lympholyte-H, PBMC EasySep™ Direct, Lymphocytes
Pronase/DNase treatment time (min) 15/2 15/2 15/2 15/2
Cell preparation time (min) 90 90 90 55
Cell suspension (µL) 30 30 25 15
Cell number 2.5 × 105 2.5 × 105 2.5 × 105 1.5 × 105
First incubation (min) 30 30 20 20
Washes (1st set) 3 × 5 min at 500×g 3 × 1 min at 500×g 3 × 1 min at 500×g 3 × 1 min at 500×g
Wash buffer (µL) 1000 200 200 200
Antibody cocktail (µL) 100 100 100 50
PBS/CD3/CD19/IgG-FITC 89.75/5/5/0.25 89.75/5/5/0.25 94.75/3/2/0.25 46.88/2/1/0.125
Second incubation (min) 30 30 10 5
Washes (2nd set) 2 × 5 min at 500×g 2 × 1 min at 500×g 2 × 1 min at 500×g 2 × 1 min at 500×g
Final suspension µL) 500 500 400 400
FCXM assay time (min) 85 65 35 30
Total time, including cell preparation (min) 175 155 125 85

2
R.S. Liwski et al.
5 min. Antibody cocktail (5 µL of anti-CD3-PerCP, 5 µL of anti-CD19-PE,
and 0.25 µL of anti-IgG-FITC, made up to 100 µL with PBS) was added
to the cells, mixed by vortexing, and incubated at 4 °C in the dark for
30 min. Cells were washed twice in 1 mL FWB at 500×g for 5 min and
resuspended in 0.5 mL of FWB before acquisition by flow cytometry.
2.7.2. Standard FCXM, tray method
The standard FCXM, tray method (Table 1), was performed in 96-
well U-bottom BD Falcon Microplate trays (BD Biosciences). Briefly,
30 µL of test or control sera and 30 µL of enriched donor PBMC
(2.5 × 105 cells) were added to a reaction well, dry vortexed, and in-
cubated at 4 °C for 30 min. The cells were washed three times with
200 µL FWB at 500×g for 1 min. Antibody cocktail (5 µL of anti-CD3-
PerCP, 5 µL of anti-CD19-PE, 0.25 µL of anti-IgG-FITC, made to a final
volume of 100 µL with PBS) was added to the cells, dry vortexed, and
incubated at 4 °C in the dark for 30 min. Cells were then washed twice
with 200 µL FWB at 500×g for 1 min, resuspended in 250 μL FWB, and
transferred into 5 mL polystyrene Falcon tubes containing an additional
250 μL of FWB before acquisition on the flow cytometer.
2.7.3. Halifax FCXM protocol
The Halifax FCXM protocol (Table 1) was performed in 96-well U-
bottom BD Falcon Microplate trays. Briefly, 50 µL of test or control sera
and 25 µL of enriched donor PBMC (2.5 × 105 cells) were added to each
reaction well in a 96 well tray, dry vortexed, and incubated at RT for
20 min. Cells were washed three times with 200 µL FWB at 500×g for
1 min, after which antibody cocktail (3 µL of anti-CD3-PerCP, 2 µL of
anti-CD19-PE, 0.25 µL of anti-IgG-FITC, made up to 100 µL with PBS)
was added to the cells, dry vortexed, and incubated for 10 min at RT in
the dark. Cells were washed twice more with 200 µL FWB at 500×g for
1 min, resuspended in 150 μL FWB, and transferred into 5 mL poly-
styrene Falcon tubes containing an additional 250 μL of FWB before
acquisition on the flow cytometer.
2.7.4. Halifaster FCXM protocol
The Halifaster FCXM protocol (Table 1) was performed in 96-well
U-bottom BD Falcon Microplate trays. Briefly, 30 µL of test or control
sera and 15 µL of purified (EasySep Direct) lymphocytes (1.5 × 105
cells) were then added to each reaction well in a 96-well U-bottom BD
Falcon Microplate tray, dry vortexed, and incubated at RT for 20 min.
Cells were washed three times with 200 µL FWB at 500×g for 1 min,
after which antibody cocktail (2 µL of anti-CD3-PerCP, 1 µL of anti-
CD19-PE, 0.125 µL anti-IgG-FITC, made up to 50 µL with PBS) was
added to the cells, dry vortexed, and incubated for 5 min at RT in the
dark. Cells were washed twice more with 200 µL FWB at 500×g for
1 min, resuspended in 150 μL FWB, and transferred into 5 mL poly-
styrene Falcon tubes containing an additional 250 μL of FWB before
acquisition on the flow cytometer.
2.8. Halifax and Halifaster FCXM protocol validation/evaluation
2.8.1. Halifax protocol validation
The Halifax protocol was compared and validated against the
SFCXM tray method in the Halifax HLA Laboratory. A total of 144
crossmatches were performed in parallel using cells from 12 live and 3
Table 2
Comparison of FCXM performance using different study protocols.
Parameter T cell FCXM
Standard
Negative control, mean MCF (SD) 168 (21)
Negative crossmatch patient sera, mean MCF (SD) 161 (26)
3SD cutoff value, MCFS 79
Mean crossmatch cutoff, MCF 247
Human Immunology xxx (xxxx) xxx–xxx
deceased donors. Crossmatch sera were selected using VXM (Section
2.3). Of the 144 FCXMs, 90 were VXM negative and 54 were VXM
positive.
2.8.2. Halifaster FCXM protocol validation
The Halifaster FCXM protocol was validated in two separate HLA
laboratories. In the Halifax laboratory, the Halifaster FCXM was vali-
dated against the Halifax FCXM protocol by parallel testing of patient
sera against cells from 5 live and 3 deceased donors. Sera were selected
using VXM (Section 2.3) results. Of the 101 parallel FCXM performed,
85 were VXM negative and 16 were VXM positive. The SWID HLA la-
boratory validated the Halifaster FCXM protocol against the standard
FCXM tray method. The standard FCXM tray method was performed
essentially as described in Section 2.7.2 with the following exceptions:
1) sera (25 μL/reaction) were added to dry cell pellets (250,000 cells/
reaction), 2) 10 μL/reaction of each anti-CD3-PerCP and anti-CD19-PE
were used in the antibody cocktail, and 3) 0.5 μL/reaction of anti-IgG-
FITC were used in the antibody cocktail. 28 FCXMs were performed in
parallel using the two XM protocols, of these 8 were VXM positive and
20 were VXM negative (Section 2.3).
2.8.3. Halifax protocol performance evaluation
2843 deceased donor FCXMs were performed over a 1 year period
(March 2015–March 2016) at the Santa Casa Hospital HLA laboratory
in Porto Alegre, Brazil using the Halifax protocol (Section 2.7.3). In all
cases, donor lymphocytes were isolated from either the spleen or lymph
node by Ficoll density gradient centrifugation and were treated with
pronase (4.7 kuntz units/mL) and DNase (11,000 units/mL). VXMs
(Section 2.3) were performed retrospectively and correlated with the
FCXM results. All FCXM with DSA directed against low expression loci
(HLA-C, DQ or DP; n = 830) were excluded from analysis. FCXM that
were either VXM negative (n = 1859), or were VXM positive with DSA
exclusively targeting one or more high expression locus antigen (HLA-
A/B/DR) (n = 154) were included in the study. Cumulative DSA MFI
values were used for correlation with the FCXM results.
2.9. FCXM acquisition and analysis
In all three laboratories participating in the study, flow cytometry
data were acquired using the FACSCanto II flow cytometers (BD
Biosciences) and analyzed with BD FACSDIVA™ software (BD
Biosciences) on the 1024 channel scale. Optimization experiment
crossmatch data was expressed as the median channel fluorescence shift
(MCFS) from the NC serum. Validation experiment crossmatch data
were expressed as MCFS from the 3 standard deviation (3SD) cutoff,
determined using the negative patient sera FCXM median channel
fluorescence (MCF) values. Halifax HLA laboratory 3SD cutoffs for the
SFCXM, Halifax FCXM and Halifaster FCXM protocols are shown in
Table 2. The positive FCXM cutoffs used by the SWID HLA laboratory
were > 77 and > 89 MCFS for T and B cells, respectively. The positive
cutoffs used by the Santa Casa Laboratory were > 77 and > 113 MCFS
for T and B cells, respectively.
B cell FCXM
Halifax Halifaster Standard Halifax Halifaster
168 (18) 176 (16) 216 (25) 215 (25) 259 (26)
158 (25) 170 (24) 213 (32) 203 (34) 237 (33)
74 71 97 101 98
242 247 313 316 357
3
R.S. Liwski et al.
2.10. Statistical analysis
Pearson product-moment correlation coefficients, paired Student’s t-
test, and chi-square analyses were computed using GraphPad Instat
software (GraphPad Software Inc., La Jolla, Ca).
3. Results
3.1. FCXM assay optimization
3.1.1. Experimental design
In the optimization phase of the study, assay platform (tube method
vs 96-well tray method), incubation time with serum (3, 5, 10, 15 and
30 min) and with anti-IgG-FITC secondary antibody (5, 10, 15 and
30 min), incubation temperature (4 °C vs RT), serum volume (30 vs
60 μL), and cell number (1.25 × 105–1.0 × 106/per reaction) were
modified to investigate the effects of each assay parameter on FCXM
results. Serial dilutions of PC sera (ranging from 1:50 to 1:400) were
chosen to react within the spectrum of relevant FCXM reactivity, ran-
ging from weakly positive (≤50 channels above the 3SD cutoff for T
cells and ≤100 channels above the 3SD cutoff for B cells) to unequi-
vocally positive (150–250 channels above the 3SD cutoff for T cells and
200–300 channels above the 3 SD cutoff for B cells). Crossmatch results
are presented as MCFS values, relative to the NC serum.
3.1.2. The SFCXM tray platform reduces assay time compared to the tube
method
The results of FCXM testing performed by the standard method
using tubes or in 96 well trays are summarized in Fig. 1, which shows
that the PC MCFS values obtained with the two methods were essen-
tially equivalent for both T cell (Fig. 1A) and B cell (Fig. 1B) FCXMs at
all dilutions of PC tested (1:50–1:400). NC MCF values obtained with
the two methods were also not significantly different (data not shown).
Standard FCXMs performed in 96 well trays took on average 25–30 min
less time than when performed in tubes (data not shown). As the tray
method was more time efficient and yielded equivalent crossmatch
results, the remainder of this validation study was carried out using 96
well U-bottom trays.
3.1.3. Effects of serum and anti-IgG-FITC incubation time on FCXM results
There was a time-dependent increase in PC MCFS values when
serum incubation time was increased from 3 min to 30 min
(Fig. 2A and B); however, the majority of antibody binding occurred
within the first 3–5 min of serum incubation regardless of the con-
centration of PC sera. Additional increases in MCFS values obtained
with longer serum incubation times were minimal beyond the 15 min
time point (Fig. 2A and B).
Interestingly, the maximal FCXM reactivity for T cells (Fig. 2C) and
B cells (Fig. 2D) was attained within the first 5 min of incubation with
anti-IgG-FITC, and no further increase in PC MCFS values were gained
A T cell XM B B cell XM
300 300
250 250
MCFS from NC
MCFS from NC
200 200
150 150
100 100
50 50
0 0
1/50 1/100 1/200 1/400
PC dilution
Human Immunology xxx (xxxx) xxx–xxx
by extending the anti-IgG-FITC incubation time beyond 5 min
(10–30 min), regardless of PC serum concentration.
3.1.4. Reduced cell number, increased incubation temperature and
increased serum to cell suspension volume ratio increase FCXM sensitivity
Increasing the incubation temperatures from 4 °C to RT simplified
the FCXM assay in addition to providing a slight increase in PC MCFS
values of approximately 15–35 channels for both T cell (Fig. 3A) and B
cell (Fig. 3B) crossmatches. Likewise, there was a similar increase in PC
MCFS values (15–30 channels) when the serum to cell suspension vo-
lume ratio was increased from 1:1 to 2:1 (Fig. 3C and D). Lastly, when
the number of cells per reaction was reduced from 1 × 106 cells/well to
1.25 × 05 cells/well, PC MCFS values also increased (Fig. 3E and F).
Notably, changes in reaction cell number (between 2.5 × 105 and
1.0 × 106), the range used for clinical testing by most HLA laboratories,
influenced PC MCFS values more than any other parameter tested, with
up to 120 channel differences observed among experimental groups
(Fig. 3E and F).
3.2. Validation and evaluation of optimized FCXM assay protocols
3.2.1. Validation of the rapid and optimized FCXM assay: the Halifax
FCXM protocol
The results of the optimization studies were used to develop a rapid
and optimized FCXM assay, i.e., the Halifax FCXM protocol. The per-
formance of the Halifax FCXM protocol was then compared to the
standard FCXM in a validation study. Differences among assay para-
meters for the SFCXM and the Halifax protocol are listed in Table 1;
these include: decreased serum incubation time (20 vs 30 min), de-
creased IgG-FITC incubation time (10 vs 30 min), increased incubation
temperature (RT vs 4 °C), and increased serum to cell suspension vo-
lume ratio (2:1 vs 1:1). While the 15 min (first incubation) and 5 min
(second incubation) times were predicted to be sufficient based on our
optimization studies, we decided to extended each incubation by 5 min
(to 20 and 10 min, respectively) in the Halifax FCXM protocol to
minimize the possibility of missing weak positive reactions. Thus, the
FCXM assay time (not including cell preparation) for the Halifax pro-
tocol is 35 min, which represents approximately 60% time reduction
compared to the assay time for the SFCXM performed in tubes
(Table 1). Finally, based on titration studies (data not shown) we re-
duced the amount of anti-CD3-PerCP and anti-CD19-PE antibodies to
3 μL and 2 μL per reaction, respectively. This contributed to a reduction
in the FCXM assay costs (Table 3).
144 crossmatches (90 VXM negative and 54 VXM positive) were
performed in parallel using the Halifax FCXM and SFCXM protocols. All
VXM positive sera selected for these experiments had a cumulative DSA
MFI > 5000 to ensure a high likelihood of positive FCXM reaction.
There was excellent correlation between the MCFS values obtained with
the two protocols in both the T cell (R2 = 0.977; Fig. 4A) and B cell
(R2 = 0.975; Fig. 4B) crossmatches, and the MCFS values for each
Fig. 1. Standard FCXMs performed in tubes or trays yield
comparable results. T cell (A) and B cell (B) flow cytometric
crossmatches were performed with different dilutions of
positive control (PC) and negative control (NC) sera by the
SFCXM method carried out in tubes (dashed line) or trays
(solid line). Data are expressed as median channel fluores-
cence shifts (MCFS) from the negative control. Square (■)
and circle (●) Two independent experiments are shown.
1/50 1/100 1/200 1/400
PC dilution
4
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

A T cell XM B B cell XM
300 Serum incubation time 400 Serum incubation time
3 min 5 min 10 min 3 min 5 min 10 min
15 min 30 min 350 15 min 30 min
250
MCFS from NC

MCFS from NC
300
200
250
150 200
150
100
100
50
50
0 0
1/50 1/100 1/200 1/400 1/50 1/100 1/200 1/400
PC dilution PC dilution

C T cell XM D B cell XM
300 IgG-FITC incubation time 400 IgG-FITC incubation time
5 min 10 min 350 5 min 10 min
250
MCFS from NC

MCFS from NC
15 min 30 min 300 15 min 30 min
200
250
150 200
150
100
100
50
50
0 0
1/50 1/100 1/200 1/400 1/50 1/100 1/200 1/400
PC dilution PC dilution
Fig. 2. Time course experiments showing effects of serum and IgG-FITC incubation time on FCXM results. T cell (A and C) and B cell (B and D) flow cytometric crossmatches were
performed with different dilutions of positive control (PC) sera and negative control (NC) sera using the tray method. In panels A and B, donor cells were incubated with sera for 3, 5, 10,
15 or 30 min and then with IgG-FITC for 30 min. In panels C and D, donor cells were incubated with sera for 30 min and then with IgG-FITC for 5, 10, 15 or 30 min. Data are expressed as
median channel fluorescence shifts (MCFS) from the negative control. One representative experiment is shown.

crossmatch obtained by the two methods were not statistically dif-
ferent. In addition, the mean NC serum MCF values, mean negative
patient crossmatch MCF values, and calculated 3SD cutoffs were very
similar for both protocols (Table 2). One unexpected positive (T+/B−
crossmatch) reaction out of 90 predicted negative (VXM) crossmatches
occurred with both the Halifax FCXM (26 MCFS above the 3SD cutoff)
and SFCXM (34 MCSF above the cutoff; Fig. 4C) protocols. All 54 of the
VXM positive sera gave positive reactions in the SFCXM and Halifax
FCXM protocols. Chi-square analysis showed that the Halifax FCXM and
SFCXM protocols are equivalent for crossmatch interpretation
(p < 0.001, Fig. 4C).
3.2.2. Halifax FCXM protocol results strongly correlate with the virtual
crossmatch
In order to assess the performance of the Halifax FCXM protocol we
conducted a retrospective analysis of 2843 FCXM performed by the
Santa Casa HLA laboratory between 03/2015 and 03/2016 and corre-
lated the FCXM with VXM results. Samples that were either VXM po-
sitive with DSA directed exclusively against one or more of the high
expression HLA locus antigens (HLA-A, B, DR; n = 154), or VXM
negative (n = 1859), were included in the analysis (n = 2013). Samples
with DSA directed against low expression HLA locus antigens (HLA-C,
DQ and/or DP) were excluded. Of 1859 VXM negative samples only 35
were FCXM positive confirming a low rate of unexpected positive XM
for the Halifax FCXM protocol (1.9%). Of 154 VXM +ve samples, 119
were FCXM positive and 35 were negative. In the majority of VXM
positive/FCXM negative cases (31/35, 89%) the DSA were weak with
cumulative MFI between 1000 and 1999. The correlation between the
VXM and the Halifax FCXM results improved greatly as the cumulative
MFI of DSA increased (Fig. 5). For T cells (HLA-A and/or B DSA), Ha-
lifax FCXM protocol results correlated with a positive VXM in 41.9%
(DSA MFI 1000–1999), 85.0% (DSA MFI 2000–2999), 93.4% (DSA MFI
3000–4999), and 100% (DSA MFI ≥ 5000) of cases (Fig. 5A). For the B
cell XM (HLA-A and/or B and/or DR), the correlation was slightly im-
proved: 48.3% (DSA MFI 1000–1999), 86.9.0% (DSA MFI 2000–2999),
98.0% (DSA MFI 3000–4999), and 100% (DSA MFI ≥ 5000) of cases
(Fig. 5B). Thus with 1000 MFI VXM cutoff, the overall Halifax FCXM
sensitivity was 77.3% and the specificity was 98.1%. When 2000 MFI
VXM cutoff was applied, the Halifax FCXM sensitivity improved to
95.7% and the specificity was 96.6%.

5
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

Fig. 3. Effects of incubation temperature, serum to cell

A T cell XM B B cell XM suspension volume ratio and cell number on FCXM results. T
cell (A, C and E) and B cell (B, D and F) flow cytometric
400 Incubation temperature 400 Incubation temperature
MCFS from NC

MCFS from NC
crossmatches were performed with different dilutions of
350 4C RT 350 4C RT
positive control (PC) sera using the tray method. In panel A
300 300 and B, crossmatches were performed either at 4 °C or at
250 250 room temperature (RT). In panel C and D, sera were in-
200 200 cubated with donor cells either at a 1:1 or 2:1 volume ratio.
150 150 In panel E and F, varying numbers of donor cells were used.
100 100 Data are expressed as median channel fluorescence shifts
50 50 (MCFS) from the negative control (NC). One representative
0 0 experiment is shown.
1/50 1/100 1/200 1/400 1/50 1/100 1/200 1/400
PC dilution PC dilution
C T cell XM D B cell XM
350 Serum:Cell suspension volume ratio 400 Serum:Cell suspension volume ratio
MCFS from NC

300 1:1 2:1 MCFS from NC 350 1:1 2:1

250 300
250
200
200
150
150
100 100
50 50
0 0
1/50 1/100 1/200 1/400 1/50 1/100 1/200 1/400
PC dilution PC dilution

E T cell XM F B cell XM
400 Cell number per reaction 350 Cell number per reaction
MCFS from NC

MCFS from NC

350 125K 250K 500K 1000K 300 125K 250K 500K 1000K
300 250
250
200
200
150
150
100 100
50 50
0 0
1/50 1/100 1/200 1/400 1/50 1/100 1/200 1/400

PC dilution PC dilution

3.2.3. Design and validation of the enhanced rapid and optimized FCXM
assay: the Halifaster FCXM protocol
An important time limitation of FCXM is the cell isolation step. Many
HLA labs use either Ficoll or Lympholyte density gradient centrifugation to
obtain PBMC. This procedure is time consuming and yields variable
lymphocyte purity. To improve the isolation time and the purity of lym-
phocytes, we evaluated the EasySep Direct lymphocyte purification
method, which utilizes an immunomagnetic bead negative selection
principle. As shown in Fig. 6E, EasySep Direct method greatly enhanced
the purity of lymphocytes and the consistency of lymphocyte purity
(mean = 96.77 ± 2.63%; median = 96.8%; range = 88.3–99.1%) when
compared to the Lympholyte density medium centrifugation technique
(mean = 68.83 ± 19.06%; median = 63.9%; range = 23.9–88.8%). Im-
portantly, there was no difference in the lymphocyte yield between the
two cell isolation methods (10.2 ± 5.0 × 106 for EasySep Direct vs
11.1 ± 5.6 × 106 for Lympholyte; Fig. 6F). Furthermore, the application
of the EasySep Direct procedure increased the ease of identifying the
lymphocyte population during acquisition and analysis by flow cytometry
(Fig. 6A–D) and reduced the time required for cell preparation from 90 to
55 min (including pronase/DNase treatment; Table 1). Interestingly, the
cell isolation costs were similar between the Lympholyte and EasySep
Direct methods when the cost reduction associated with a decrease in
technologist time was considered (Table 3).
Increased donor cell lymphocyte purity achieved with the EasySep
Direct purification method allowed us to redesign and further optimize
the Halifax FCXM protocol. This was possible for two main reasons: 1)
fewer total cells could now be used to reach the same number of lym-
phocytes per reaction, and 2) fewer contaminating cells were present to
compete with the lymphocytes for binding to HLA antibodies and anti-
IgG-FITC antibody. The resulting Halifaster FCXM method (Table 1)
features a reduction in total donor cell number per reaction (from
2.5 × 105 to 1.5 × 105), cell suspension volume (from 25 µL to 15 µL),
serum volume (from 50 µL to 30 µL), and antibody cocktail volume
(from 100 µL to 50 µL) compared to the Halifax FCXM protocol. In
addition, the anti-IgG-FITC secondary antibody incubation time was
reduced from 10 to 5 min, based on the original optimization studies
(Section 3.1.3). Finally, based on antibody titration experiments (data
not shown), anti-CD3-PerCP and anti-CD19-PE secondary antibody
volumes were reduced to 2 μL and 1 μL per reaction, respectively, which
contributed to a further reduction in FCXM assay costs (Table 3).
The Halifaster FCXM protocol was validated against the Halifax
FCXM method by performing a total of 101 crossmatches (85 VXM
negative and 16 VXM positive) in parallel against target cells from 5
live and 3 deceased donors. All VXM positive sera selected for these
experiments had a cumulative DSA MFI > 5000 to ensure a high
likelihood of positive FCXM reaction. There was excellent correlation
between MCFS values obtained with the Halifaster FCXM and Halifax
FCXM protocols for both T cell (R2 = 0.982; Fig. 7A) and B cell
(R2 = 0.978; Fig. 7B) crossmatches. The MCFS values obtained for each
crossmatch by the two methods were not statistically different. Table 2
shows that the mean NC serum MCF values, mean negative patient
crossmatch MCF values, and calculated 3SD cutoffs were similar for

6
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

Table 3
Cost comparison for the standard, Halifax, and Halifaster FCXM protocols.

Parameter Standard FCXM Tube Standard FCXM Tray Halifax FCXM Halifaster FCXM Halifaster FCXM
Method Method Protocol Protocol Protocol

Cell Isolation
Cell isolation reagent type Lympholyte-H Lympholyte-H Lympholyte-H EasySep™ Direct Whole EasySep™ Direct Buffy
blood Sample Coat sample

Cell isolation reagent* cost

Live donor, 10 mL of blood $5 $5 $5 $35 $17.5
Deceased donor, 20 mL of blood $10 $10 $10 $70 $35

Cell preparation time 90 min 90 min 90 min 55 min 60 min

Technologist time isolation cost

Live donor, regular shift time $52.5 $52.5 $52.5 $32 $35
Deceased donor, on call time $105 $105 $105 $64 $70

Total cell isolation cost

Live donor $57.5 $57.5 $57.5 $67 $52.5
Deceased donor $115 $115 $115 $134 $105

FCXM assay
FCXM antibody cocktail CD3/CD19** volume 5 μL/5 μL 5 μL/5 μL 3 μL/2 μL 2 μL/1 μL 2 μL/1 μL
and cost/reaction $3.8 $3.8 $1.8 $1.0 $1.0

FCXM antibody cocktail cost

Live donor XM (per 12 rxn) $46 $46 $22 $12 $12
Deceased donor (per 30 rxn) $114 $114 $54 $31 $31

FCXM assay time 85 min 65 min 35 min 30 min 30 min

FCXM assay technologist time cost

Live donor, regular shift time $50 $38 $20 $17.5 $17.5
Deceased donor, on call time $99 $76 $41 $35 $35

Total FCXM assay cost

Live donor $95 $84 $42 $30 $30
Deceased donor $214 $190 $95 $66 $66

Total cost***
Total cost (cell isolation + FCXM)
Live donor $153 $141 $100 $97 $82
Deceased donor $329 $305 $210 $200 $171

All costs are expressed in Canadian dollars and are based on approximate reagent list prices and average Halifax Medical Laboratory Technologist remuneration rates.
* The cost of pronase/DNase is not included as it is similar across the protocols and negligible.
** The cost of anti-IgG-FITC is not included as it is similar across the protocols and negligible.
*** The costs of consumables are not included as they are similar across the protocols.

Fig. 4. The Halifax and standard FCXM protocols show ex-

A T cell XM B B cell XM cellent results concordance. T cell (A and C) and B cell (B
600 600 and C) flow cytometric crossmatches were performed in
MCFS from cutoff

MCFS from cutoff

parallel using the standard FCXM (SFCXM) protocol and the

500 500 Halifax FCXM protocol. Patient sera predicted by the SAB
SFCXM

SFCXM

400 400 assay to give a negative FCXM (n = 90) or a positive FCXM

(n = 54) against selected donor cells were used. In panel A
300 300 and B the data are expressed as median channel fluores-
200 200 cence shifts (MCFS) from the 3SD cutoff. In panel C, chi-
square analysis of FCXM interpretation is shown.
100 y = 0.9256x - 9.3858 100 y = 0.9799x - 1.6227
0 R² = 0.977 0 R² = 0.9753
-200 0 200 400 600 -200 0 200 400 600
-100 -100
-200 MCFS from cutoff -200 MCFS from cutoff
Halifax FCXM Halifax FCXM

C
Chi-square Halifax FCXM Halifax FCXM
p < 0.001 negative positive
SFCXM negative 88 1

SFCXM positive 1 54

7
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

Fig. 5. The Halifax FCXM correlates well with the virtual

A T cell XM B B cell XM crossmatch results. T cell (A) and B cell (B) flow cytometric
crossmatches (FCXM) were performed using the Halifax
HLA-A+/-B DSA HLA-A+/-B+/-DR DSA
100 100 protocol and correlated with the virtual crossmatch results
based on the presence of DSA directed against one or more
80 80 high expression HLA locus antigens (HLA-A/B for T cell and
Positive XM %

Positive XM %
HLA-A/B/DR for B cell XM) with cumulative MFI ≥ 1000 by
60 60 SAB assay. Data shown depict the percentage of positive
Halifax FCXM reactions as a function of increasing VXM
40 40 DSA MFI ranges (K = 1 × 103).

20 20

0 0
1-2K 2-3K 3-5K >5K 1-2K 2-3K 3-5K >5K
SAB DSA MFI range SAB DSA MFI range

both protocols. The Halifaster FCXM protocol yielded 2 unexpected
positive reactions (VXM negative), while, 1 unexpected positive (VXM
negative) reaction occurred with the Halifax protocol (Fig. 7C). These
reactions were all positive for T cell and negative for B cell crossmatch
with less than 10 MCFS above the 3SD cutoff, and were interpreted as
false positive and not due to HLA antibodies. Finally, both protocols
gave 2 negative reactions for sera with positive VXM. The two protocols
were equivalent for crossmatch interpretation as shown by the chi-
square analysis (p < 0.001, Fig. 7C).
The Halifaster FCXM was also compared to the SFCXM protocol in
the SWID HLA laboratory by performing 28 crossmatches (8 VXM −ve
and 20 VXM +ve). There was excellent correlation between MCFS va-
lues obtained with the two protocols for both T cell (R2 = 0.976;
Fig. 8A) and B cell (R2 = 0.983; Fig. 8B) crossmatches. The MCFS va-
lues obtained for each crossmatch by the two methods were not sta-
tistically different. There were 2 discordant crossmatch reactions ob-
served where the Halifax FCXM reported a positive crossmatch, while
the SFXM was negative (Fig. 8C). Both crossmatches were predicted

Fig. 6. EasySep Direct lymphocyte isolation improves lym-

A Lympholyte SSC/FSC B Lympholyte CD19/CD3 phocyte purity while maintaining lymphocyte yield. Donor
Serum incubation time PBMCs and lymphocytes were enriched from whole blood
using the Lympholyte density medium centrifugation and
52.1%
CD19-PE

EasySep Direct Isolation Kit (EasySep Direct), respectively.

Representative side scatter vs forward scatter dot plots (A,
CD19-PE

C) and CD19 vs CD3 dot plots (B, D) are shown for lym-
MCFS
SSC

pholyte (A, B) and EasySep Direct (C, D) isolated cells.

Lymphocyte purity (E) and total cell/lymphocyte yield (F)
were determined using a Sysmex analyzer. Data show the
mean and SD of 20 independent cell isolations.

FSC CD3-PerCP

C EasySep SSC/FSC D EasySep CD19/CD3

97.8%
CD19-PE
SSC

FSC CD3-PerCP

E Cell Purity F Cell Yield

100 20
90
population (%)

80
15
106

70
60
Cell # x

50 10
40
30
5
20
10
0 0
Neutrophils Lymphocytes Monocytes Total Cells Lymphocytes

Lympholyte EasySep Direct Lympholyte EasySep Direct

8
R.S. Liwski et al. Human Immunology xxx (xxxx) xxx–xxx

Fig. 7. The Halifaster and Halifax FCXM protocol results

A 600
T cell XM B 600
B cell XM show excellent concordance. T cell (A and C) and B cell (B
and C) flow cytometric crossmatches were performed in

MCS from cutoff

MCS from cutoff
Halifaster FCXM

Halifaster FCXM
500 500 parallel using the Halifax FCXM protocol and the Halifaster
FCXM protocol. Patient sera predicted to give a negative
400 400
FCXM (n = 85) or a positive FCXM (n = 16) against se-
300 300 lected donor cells were used. In panel A and B the data are
200 200 expressed as median channel fluorescence shifts (MCFS)
y = 1.0786x + 4.0398 y = 1.1521x + 17.776 from the 3SD cutoff. In panel C, a chi-square analysis of
100 100 R² = 0.9789 FCXM interpretation is shown.
R² = 0.9819
0 0
-200 -100 0 200 400 600 -200 -100 0 200 400 600

-200 -200
Halifax FCXM Halifax FCXM
MCS from cutoff MCS from cutoff

C
Chi-square Halifax FCXM Halifax FCXM
p < 0.001 negative positive
Halifaster FCXM 84 1
negative
Halifaster FCXM 2 14
positive

Fig. 8. The Halifaster FCXM protocol results show excellent

A T cell XM B B cell XM concordance with those of the standard FCXM. T cell (A
600 600 and C) and B cell (B and C) flow cytometric crossmatches
MCS from cutoff
MCS from cutoff

were performed in parallel using the Halifax FCXM pro-

SFCXM, SWID

SFCXM, SWID

500 500 tocol and the Halifaster FCXM protocol. Patient sera pre-
400 400 dicted by the single antigen bead assay to give a negative
FCXM (n = 10) or a positive FCXM (n = 18) against se-
300 300 lected donor cells were used. In panel A and B the data are
200 200 expressed as median channel fluorescence shifts (MCFS)
from the 3SD cutoff. In panel C, a chi-square analysis of
100 y = 0.9231x - 8.1032 100 y = 0.9637x - 5.1998
FCXM interpretation is shown.
R² = 0.976 R² = 0.9829
0 0
-200 -100 0 200 400 600 -200 0 200 400 600
-100
-200
Halifaster FCXM, SWID -200 Halifaster FCXM, SWID
MCS from cutoff MCS from cutoff

C
Chi-square Halifaster FCXM Halifaster FCXM
p < 0.001 negative positive
SFCXM negative 10 2

SFCXM positive 0 16

positive by the VXM. One of the discordant reaction was T−/B+ for
the SFXM (the T cell value was 8 channels below the 3SD positive
cutoff) while it was T+/B+ with the Halifaster FCXM protocol. The
virtual XM demonstrated cumulative class I HLA DSAs of approximately
5500 MFI, suggesting that the T cell SFXM was false negative. The
second discordant reaction was T+/B- for the SFXM and T+/B+ for
the Halifaster FCXM. In this case, clear class I HLA DSAs (cumulative
MFI of approximately 4100) were observed in the VXM indicating that
the B cell SFXM reaction was false negative. There was also one false
T+/B− reaction reported with both FCXM protocols in the context of a
negative VXM. In addition, there were 2 false negative B cell reactions
reported with both FCXM protocols with a weak HLA-DQ locus specific
DSA (MFI of approximately 3000) seen in both cases. Finally, the chi-
square analysis demonstrated that the two protocols were equivalent
for crossmatch interpretation (p < 0.001, Fig. 8C).
4. Discussion
The FCXM remains a cornerstone of transplantation compatibility
testing and for 25 years, the procedure to perform this assay has

9
R.S. Liwski et al.
remained unchanged. The aim of this study was to optimize the FCXM
protocol in order to reduce turnaround time (TAT) without compro-
mising assay performance. As a first step to optimize the FCXM assay,
the platform was changed from test tubes to a 96 well U-bottom tray.
This single modification saved 25–40 min per FCXM, depending on the
number of tests performed. Time savings were attributed to decreased
centrifugation time (1 min vs. 5 min per spin), the ability to add re-
agents to multiple reactions at a time with a multichannel pipette, as
well as rapid removal of supernatants from all reactions en bloc (using
the tray flick technique). Assay time was reduced an additional
30–35 min by decreasing serum and anti-IgG-FITC incubation times. To
compensate for possible decreases in MCF values associated with
shorter incubation times, the antibody/antigen reaction was enhanced
by increasing the incubation temperature and serum volume. Finally, in
the Halifaster FCXM protocol, designed for use with pure lymphocytes,
we reduced the number of target cells per reaction. This approach
served two purposes, namely to increase both the sensitivity of the
FCXM and the likelihood to have sufficient lymphocytes available for
crossmatches where cell yield is low. The Halifaster FCXM protocol is a
substantial time saver, particularly with the EasySep Direct lymphocyte
purification procedure. Combined with cell isolation, the entire cross-
match, from start to finish can be performed in less than two hours. This
represents a significant time savings when compared to the standard
assay, which typically takes 4–5 h including the cell isolation.
During the optimization phase of the study we identified a number
of assay parameters that affect FCXM results, most notably the number
of donor cells per reaction. The cell number was inversely proportional
to crossmatch reactivity. For example, crossmatches performed with
positive control sera against 1.25 × 105 cells per reaction yielded MCF
shifts that were higher by 50–120 channels compared to crossmatches
performed using 1.0x106 cells per reaction. Thus, the number of target
cells used in a crossmatch assay may have a significant impact on
crossmatch interpretation, especially in the case of low level DSA. In
our experience, the number of donor cells used per reaction vary sig-
nificantly between labs. Indeed, in some laboratories, the number of
cells used in a crossmatch depends in large part on the total number of
donor cells and the number of recipient sera that need to be tested. Such
heterogeneity likely contributes to/ explains the high inter- and intra-
laboratory variability of crossmatch results [17]. Our study underscores
that protocol optimization and standardization leads to improved
testing quality. This should be an integral part of any test development
and validation.
The Halifax protocol was validated against the SFCXM by com-
paring results from 144 parallel crossmatches. The two methods pro-
vided virtually identical results with an R2 value of 0.98 for both T cell
and B cell crossmatches. Importantly, there was a 98.6% concordance
in crossmatch interpretation between the two protocols. There was 1
false positive reaction with each protocol, both were for the same
sample and were borderline positive for T cell but negative for B cell
crossmatch and would have been interpreted as negative in the clinical
setting, given the absence of DSA on SAB testing.
A retrospective revue of 2013 crossmatches performed at the Santa
Casa HLA laboratory in Porto Alegre, Brazil, displayed excellent cor-
relation between the Halifax FCXM protocol and VXM results. The
unexpected positive XM rate (in the context of a negative VXM) for the
Halifax FCXM was less than 2%, lower than the 3–16% rates for the
standard FCXM assay reported by several recent studies [18–20]. When
a 2000 MFI DSA (HLA-A/B/DR) cutoff was applied, the concordance
between VXM and the Halifax FCXM results was 92.9% for T cell and
95.7% for B cell XM with excellent assay sensitivity (95.7%) and spe-
cificity (96.6%), which is comparable to or better than what is reported
for the standard FCXM assay [19,21]. Furthermore, with DSA
MFI ≥ 5000, a 100% concordance was observed between the VXM and
Halifax FCXM for both T and B cell crossmatches. These results de-
monstrate that the Halifax FCXM is a safe and effective protocol for use
in clinical transplant laboratories.
10
Human Immunology xxx (xxxx) xxx–xxx
An important limiting factor in crossmatch TAT is the time required
to isolate cells. Density gradient centrifugation is still used by many labs
for PBMC isolation. This procedure is labor-intensive and time-con-
suming. Furthermore, depending on the age and quality of the blood
sample, cell yield can vary significantly from donor to donor. The purity
of lymphocytes in enriched PBMC preparations is often poor and highly
variable, which may affect crossmatch results [22]. In contrast, lym-
phocyte isolation by the EasySep Direct technique is rapid and con-
sistently yields > 90% pure lymphocytes. We incorporated this lym-
phocyte isolation procedure into the Halifaster FCXM protocol to
increase the purity and speed of lymphocyte isolation. Purification by
this technique occurs by negative selection, which was an important
feature in the choice of this technology as antibody-mediated positive
selection could interfere with the FCXM assay.
The correlation between the Halifax and Halifaster FCXM protocols,
based on 101 parallel crossmatches, was excellent (R2 = 0.98) with a
97% concordance in crossmatch results. The 3 unexpected positive re-
actions were borderline T cell crossmatch-positive and B cell cross-
match-negative, and would not affect crossmatch interpretation clini-
cally. Comparison of 28 parallel crossmatches using the SFCXM and
Halifaster protocols by the SWID HLA laboratory confirmed that ex-
cellent result correlation (R2 = 0.98) was also retained between these
two methods.
We have been using the Halifax FCXM protocol for routine clinical
testing since May 2011, moving to the Halifaster FCXM method in
September 2014. The most obvious benefit of these protocols is reduced
TAT, which is particularly critical during deceased donor work-ups on
call. This allows the clinical team to finalize recipient selection and
organize transplant logistics such as patient transport and OR time
more efficiently. Indeed, a shorter TAT may translate into a reduction in
cold ischemia time where prospective crossmatches are required as part
of an import donor work-up [12]. Importantly for the laboratory, this
shorter FCXM assay has been associated with a reduction in on-call staff
fatigue. Implementation of the EasySep Direct cell purification had a
very positive impact on our cell isolation and FCXM assay set-up time.
Furthermore, we observed an improvement in events acquisition as well
as time and ease of gating and analysis based on the improved lym-
phocyte purity. The 96 well tray platform also allows us to use a more
robust set of controls, improving the assessment of assay performance.
Finally, as a result of optimizing other parameters, our Halifax and
Halifaster protocols now require much less antibody per reaction than
the SFCXM method. Indeed, a reduction of ∼70% total antibody vo-
lume per reaction (Halifaster vs. SFCM) equates to significant reagent
cost savings (Table 3) and offsets the increased cost of cell isolation
reagents associated with using the Easysep Direct kits. In addition, the
cost of Easysep Direct reagent can be reduced by at least 50% if lym-
phocytes are isolated from pooled buffy coat samples, as per manu-
facturer instructions, bringing the total cell isolation costs (reagents and
technologist time) to below the lympholyte method (Table 3). Coupled
with a reduction in the technologist time, the overall cost savings as-
sociated with implementing the Halifax/Halifaster FCXM can be sig-
nificant (Table 3), especially given the large number of samples pro-
cessed in a typical HLA laboratory.
The Halifax protocol has been tested in a multicenter evaluation
study involving 13 HLA labs in Canada [23]. The results showed that
the Halifax protocol compares favorably with other methods, and im-
proves FCXM results by augmenting the signal to noise ratio and inter-
laboratory concordance of the results [23]. Based on these findings,
HLA labs in Canada have validated and adopted the Halifax protocol for
routine clinical practice with some laboratories using the Halifaster
assay. Several labs in the United States, United Kingdom, Brazil and
other countries have since either adopted the Halifax or Halifaster
protocols or are in the process of doing so. We predict that broader
application of these protocols will improve pre-transplant risk assess-
ment, facilitate transplant allocation, and improve patient care.
R.S. Liwski et al.
Author contributions
RL was responsible for the study concept, study design, data ac-
quisition, analysis and interpretation. He also drafted and revised the
manuscript. AG and DC participated in data analysis, data interpreta-
tion, manuscript drafting, and revision. CM and JN participated in data
acquisition, data analysis, and manuscript revision. RB and HG parti-
cipated in study design, data analysis and manuscript revision.
Conflicts of interest
The authors declare they have no conflicts of interest.
Acknowledgments
The authors gratefully acknowledge the technical assistance of the
Halifax HLA Laboratory technologists: Geoff Peladeau (MLT), Geoff
Adams (MLT) and Kelly Heinstein (BSc, MLT, HLA lab supervisor); the
Santa Casa HLA Laboratory: Juliana Montagner; the SWID HLA la-
boratory: Steve Chang, Cody Murray and Shannon Mesa. This research
did not receive any specific grant from funding agencies in the public,
commercial, or not-for-profit sectors.
References
[1] M.R. Garovoy, M.A. Rheinschmidt, M. Bigos, H. Perkins, B. Colombe, N. Feduska,
O. Salvatierra, Flow cytometry analysis: a high technology crossmatch technique
facilitating transplantation, Transpl. Proc. 15 (1983) 1939–1944.
[2] R.A. Bray, L.K. Lebeck, H.M. Gebel, The flow cytometric crossmatch: dual-color
analysis of T cell and B cell reactivities, Transplantation 48 (1989) 834–840.
[3] R.S. Liwski, H.M. Gebel, Of cells and microparticles: assets and liabilities of HLA
antibody detection, Transplantation (2017), http://dx.doi.org/10.1097/TP.
0000000000001818.
[4] R.S. Liwski, A.L. Greenshields, R.A. Bray, H.M. Gebel, Becoming a chef in the
human leukocyte antigen kitchen : interpretation and modification of human leu-
kocyte antigen – antibody assays, (2017) 1–8. http://dx.doi.org/10.1097/MOT.
0000000000000423.
[5] R.A. Bray, C. Tarsitani, H.M. Gebel, J. Lee, Clinical cytometry and progress in HLA
antibody detection, Methods Cell Biol. 103 (2011) 285–310, http://dx.doi.org/10.
1016/B978-0-12-385493-3.00012-7.
[6] R.A. Bray, P.W. Nickerson, R.H. Kerman, H.M. Gebel, Evolution of HLA antibody
detection technology emulating biology, Immunol. Res. 29 (2004) 41–53.
[7] R.J. Mahoney, K.A. Ault, S.R. Given, R.J. Adams, A.C. Breggia, P.A. Paris,
G.E. Palomaki, S.A. Hitchcox, B.W. White, J. Himmelfarb, D. Leeber, The flow cy-
tometric crossmatch and early renal transplant loss, Transplantation 49 (1990)
527–535.
[8] K. Ogura, P.I. Terasaki, C. Johnson, R. Mendez, J.T. Rosenthal, R. Ettenger,
D. Martin, E. Dainko, L. Cohen, T. Mackett, T. Berne, L. Barba, E. Lieberman, The
significance of a positive flow cytometry crossmatch test in primary kidney
11
Human Immunology xxx (xxxx) xxx–xxx
transplantation, Transplantation 56 (1993) 294–298.
[9] M. Karpinski, D. Rush, J. Jeffery, M. Exner, H. Regele, S. Dancea, D. Pochinco,
P. Birk, P. Nickerson, Flow cytometric crossmatching in primary renal transplant
recipients with a negative anti-human globulin enhanced cytotoxicity crossmatch,
J. Am. Soc. Nephrol. 12 (2001) 2807–2814.
[10] C. Hamrick, L. Lebeck, Flow cytometry T cell and B cell crossmatching, ASHI Lab.
Man. (2000).
[11] P.I. Lobo, C.E. Spencer, W.C. Stevenson, T.L. Pruett, The use of pronase-digested
human leukocytes to improve specificity of the flow cytometric crossmatch,
Transpl. Int. 8 (1995) 472–480.
[12] C.J. Taylor, V. Kosmoliaptsis, L.D. Sharples, D. Prezzi, C.H. Morgan, T. Key,
A.N. Chaudhry, I. Amin, M.R. Clatworthy, A.J. Butler, C.J.E. Watson, J.A. Bradley,
Ten-year experience of elective omission of the pretransplant crossmatch test in
deceased donor kidney transplantation, Transplantation 89 (2010) 185–193,
http://dx.doi.org/10.1097/TP.0b013e3181c926f2.
[13] B.C. Eby, R.R. Redfield, T.M. Ellis, G.E. Leverson, J.S. Odorico, Virtual HLA cross-
matching as a means to safely expedite transplantation of imported pancreata,
Transplantation 100 (2016) 1103–1110, http://dx.doi.org/10.1097/TP.
0000000000001125.Virtual.
[14] Stem Cell Technologies Inc, EasySep direct human total lymphocyte isolation kit,
Prod. Inser. (n.d.). http://www.stemcell.com.
[15] R.S. Liwski, A.L. Greenshields, C. Murphey, R.A. Bray, H.M. Gebel, It’s about time :
the development and validation of a rapid optimized single antigen bead (ROB)
assay protocol for LABScreen, Hum. Immunol. (2017) 1–11, http://dx.doi.org/10.
1016/j.humimm.2017.05.001.
[16] M. Schnaidt, C. Weinstock, M. Jurisic, B. Schmid-horch, A. Ender, D. Wernet, HLA
antibody specification using single-antigen beads — a technical solution for the
prozone effect, Transplantation 92 (2011) 510–515, http://dx.doi.org/10.1097/TP.
0b013e31822872dd.
[17] R.S. Liwski, D. Pochinco, K. Tinckam, P. Campbell, P. Nickerson, Going with the
flow, Canadian crossmatch standardization, Hum. Immunol. 73 (supplement)
(2012) 28, http://dx.doi.org/10.1016/j.humimm.2012.07.065.
[18] A.R. Tambur, D.S. Ramon, D.B. Kaufman, J. Friedewald, X. Luo, B. Ho, A. Skaro,
J. Caicedo, D. Ladner, T. Baker, J. Fryer, L. Gallon, J. Miller, M.M. Abecassis,
J. Leventhal, Perception versus reality?: virtual crossmatch — how to overcome
some of the technical and logistic limitations, Am. J. Transplant. 9 (2009)
1886–1893, http://dx.doi.org/10.1111/j.1600-6143.2009.02724.x.
[19] T. Bachelet, C. Martinez, A. Del Bello, L. Couzi, S. Kejji, G. Guidicelli, S. Lepreux,
J. Visentin, N. Congy-Jolivet, L. Rostaing, J.-L. Taupin, N. Kamar, P. Merville,
Deleterious impact of donor-specific anti-HLA antibodies toward HLA-Cw and HLA-
DP in kidney transplantation, Transplantation 100 (2015) 159–166, http://dx.doi.
org/10.1097/TP.0000000000000821.
[20] C.P. Johnson, J.J. Schiller, Y.R. Zhu, S. Hariharan, A.M. Roza, D.C. Cronin,
B.D. Shames, T.M. Ellis, Renal transplantation with final allocation based on the
virtual crossmatch, Am. J. Transplant. 16 (2016) 1503–1515, http://dx.doi.org/10.
1111/ajt.13606.
[21] T.M. Ellis, J.J. Schiller, A.M. Roza, D.C. Cronin, B.D. Shames, C.P. Johnson,
Diagnostic accuracy of solid phase HLA antibody assays for prediction of cross-
match strength, Hum. Immunol. 73 (2012) 706–710, http://dx.doi.org/10.1016/j.
humimm.2012.04.007.
[22] R. Liwski, G. Adams, G. Peladeau, K. Heinstein, The impact of lymphocyte purity on
flow cytometry crossmatch (FCXM) assay. It’s not purely theoretical, Hum.
Immunol. 77 (2016) 110–111, http://dx.doi.org/10.1016/j.humimm.2016.07.164.
[23] R. Liwski, D. Pochinco, K. Tinckam, H. Gebel, P. Campbell, P. Nickerson, Canada-
wide evaluation of rapid optimized flow crossmatch (ROFCXM) protocol, Hum.
Immunol. 73 (2012) S26, http://dx.doi.org/10.1016/j.humimm.2012.07.062.