Fetomaternal Hemorrhage (FMH) Testing and Related ... - escca

Fetomaternal Hemorrhage (FMH)
Testing and Related Applications of
Red Cell Flow Cytometric Analysis
Bruce H Davis, MD
President, Trillium Diagnostics
Co-Chair, CLSI H-52 RBC Diagnostic Testing Document Committee
Member, CAP Hematology Resource Committee
Past President, International Clinical Cytometry Society
Member, International Council for Standardization of Haematology

Conflict of Interest Disclosure
• President and Founder of Trillium Diagnostics, LLC of
Bangor, Maine, which has developed and markets products
for FMH testing (FMH QuikQuant, FETALtrol, anti-HbF, antipan
Hb)
• Consultant to Beckman Coulter, Becton Dickinson, IQ
Products, Millipore, Horiba Medical and Clavis Pharma
• Trillium Diagnostics has strategic alliances with Beckman
Coulter, IQ Products, Millipore, Life Technologies,
Cedarlane, BioSpecifix, Dartmouth College, Aarhus Univ,
LeukoCare, LeukoDx and Clavis Pharma
• Acknowledgements: Dr Bill Chen, Nancy Bigelow, Kathleen
Thompson, Dr. Steve Olsen, Ben Hunsberger, Linda Wong,
Joost Schuitemaker and others who have furthered the
improvement of FMH testing by flow cytometry.

Agenda for Discussion
• History of FMH Testing
• Kleihauer Betke test: limitations and pragmatic reality
• Flow Cytometric Methods
– Anti-HbF
– Anti-RhD
– Hybrids
• What have recent clinical studies and proficiency testing taught
us?
• Pitfalls and technologic tips of FMH testing
– Fixation effect on both HbF and RhD
– Optimizing antibody concentration
– Effect of nucleated cells on accuracy and how to approach
– Effect of aggregates, how to detect, and how to analyze
• Future improvements in FMH testing:
– FMH QuikQuant (CE IVD in April 2012)
– Automated data analysis
• Questions

History and Clinical Utility of FMH Testing
• Measures the amount of fetal blood entering maternal circulation
so as to provide a basis for adequate dosing of Rh Immune
globulin (RhIg) therapy
• Used to avoid, rather than detect, hemolytic disease of the
newborn by insuring sufficient dosage of RhIg and thereby avoid
maternal anti-D antibody production.
• One of the first “companion diagnostic” test in clinical medicine
• Assay historically performed as manual microscopic counting of
red cells on glass slides after acid elution technique originally
published by Kleihauer, Braun, and Betke in 1957 (KB test)
• Followed by nearly 40 years of attempts to improve or
standardize the KB assay performance with little success
• Flow cytometric methods developed in 1990s that markedly
improved testing performance, but still not universally adopted.
• Level of FMH with therapeutic implications varies with
geographic region and RhIg bottling dosage and accepted
regional medical practice patterns.

Fetal RBC and F Cell Determinations: Clinical Utility
• Fetal-maternal hemorrhage detection
– Rh incompatability basis for Rh immunoglobulin (RhoGam) therapy
– Trauma, placental abruption detection - independent of blood type and
causes circulating fetal RBCs
– Pre-eclampsia found to have increase in circulating fetal RBCs –
diagnostic potential? (Prenat Diagn 21:1022-26, 2001)
• Increase in Adult F Cells
– Therapeutic monitoring of treatments intended to raise Hgb F levels (F cells) in
Sickle Cell Disease (eg. butyrate, hydroxyurea)
– Diagnosis of Hereditary Persistence of Hgb F – elevated F cells
• Homocellular – all F cells
• Heterocellular – mixture of F cells and adult “normal” RBCs
– Evaluation of Hematopoiesis in Anemias and Myelodysplasia – levels of F cells are
predictive of disease severity

Methods for Detection of Fetomaternal Hemorrhage
• Manual Microscopic: Kleihauer Betke or acid elution
• Screening: Microscopic Du test (sensitivity > 0.6%) –
false negative results
• Screening: Rosette Test - false positive results
• Screening: ELISA methods – ELAT – not common
useage
• Contemporary: Antibody based Flow Cytometry
Assays
– Anti-Hemoglobin F (HbF) – CE IVD and FDA cleared
– Anti-D antigen (Rh) – CE IVD and FDA cleared
– Other “neonatal antigens” - i antigen and carbonic
anhydrase

Kleihauer-Betke Stain
Fetal RBCs
Method: slide based acid elution of HbA and microscopic
counting of % fetal cells of 1-2,000 RBCs by visual counting
Problems:
Subjective distinction between fetal cells and F cells;
poor inter-observer reproducibility (CV 50-100%)
Accuracy poor - 1999 College of American
Pathologists HbF Survey:
0.4% sample - 44.5% of labs report >0.6%
0.8% sample - 10.8% of labs report < 0.6%
Adult F Cells (mixture of HbF and HbA)

Kleihauer Betke or Acid Elution Method
• Kleihauer Bettke test: limitations in sensitivity and
precision (CV of 30-100%)
• But pragmatic reality is:
– Used by >85% of patients
– By any criteria would not be acceptable as a companion diagnostic
for Rh Immune globulin Rx
– Labs show no concern for accuracy of testing, particularly if it
requires re-education of clinicians or questioning clinical requests
• Prediction: FMH testing by flow cytometry will require a
24/7 solution, preferably by an automated methodology.

0 50 100 150 200 250
Number
0 50 100 150 200 250
Number
Fetal RBC Detection by Anti-D Antigen Antibody
Adult Cells
Adult Cells
F Cells
Fetal Cells
10 1 10 2 10 3 10 4
Chemicon anti-D
F Cells
0.15% fetal cells
Fetal Cells
10 1 10 2 10 3 10 4
Chemicon anti-D
Nelson M et al, Vox Sang 75:234-241, 1998
Lloyd-Evans P et al, Transfusion 36:432-437, 1996
Lloyd-Evans P et al, Brit J Haematol 104:621-625, 1999
Kumpel BM, Transfusion 41:1059-63, 2001
Useful for detection of
Rh positive fetal cell in
Rh negative maternal
cells, which exists in
minority of
pregnancies (

Fetal RBC Detection by Anti-D on Blood Counter
From Richard Rogers, County Durham and Darlington Acute Hospitals, UK
BH Little et al, International Journal of Laboratory Hematology 27:21-31, 2005

Comparison of Anti-D and Anti-HbF Staining
Anti-D
Anti-HbF
0% Fetal RBCs
0.15% Fetal RBCs 1.45% Fetal RBCs

Flow Cytometric Method of RBC anti-HbF Staining for Detection of Fetal
RBCs and Adult F Cells
BH Davis et al, Transfusion, 38:749-756, 1998
• Fix 0.5 million cells (5-10 mL blood) with 0.05% glutaraldehyde in PBS 0.1% BSA
x 10 min
• Wash 3X - PBS-BSA
• Resuspend - 0.5 mL 0.1% Triton X-100 x 3-5 min
• Wash 2X - PBS-BSA
• Incubate with anti-HbF antibody (Trillium Diagnostics and others)
• Wash 1X - PBS-BSA
• Resuspend in 1% paraformaldehyde in PBS-BSA
• Flow Cytometric analysis of > 50,000 events, collect light scatter and 2
fluorescence signals
• Analyze data by first looking at positive control with >1% Fetal RBCs (either
home brew of cord blood spiked adult normals or stabilized controls (FETALtrol,
Trillium Diagnostics). This defines fetal RBC “positive” region for patient
samples for each batch of specimens.
Commercial Kits and FETALtrol for this assay available from Life Technologies
(Caltag), Trillium Diagnostics, Beckman Coulter, and IQ Products

10 1 10 2 10 3 10 4
SSC-H -->
10 1 10 2 10 3 10 4
autoFL - FL2 -->
Gating strategy important with anti-HbF method
10 1 10 2 10 3 10 4
FSC-H -->
Leukocytes
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
10 1 10 2 10 3 10 4
autoFL - FL2 -->
100 200 300 400 500 600 700
Number
R1
Leukocytes
10 1 10 2 10 3 10 4
SSC-H -->
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
Important Features:
1. Exclude RBC
aggregates by
light scatter
2. Exclude
nucleated cells
(leukocytes)
3. Use positive
control to define
regions

FMH anti-HbF testing: Clinical Experience
Stat Test?
Williams J et al, AJCP
120:470, 2003
•17 year review
•96 of 2,786 (3.45%)
positive at >0.4%
•No clinical
association
•KB had no impact
on clinical decision
•No reason for stat
N u m b e r
100 200 300 400 500 600 700
N u m b e r
100 200 300 400 500 600 700
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
N u m b e r
100 200 300 400 500 600 700
N u m b e r
100 200 300 400 500 600 700
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
N u m b e r
100 200 300 400 500 600 700
N u m b e r
100 200 300 400 500 600 700
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
F Cells
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->
Clinical experience: Chen et al, Cytometry 50:285-90, 2002
•69 of 1248 (5.5%) with detectable FMH
•21 of 1248 (1.7%) with FMH of >0.6% (30 mL)
•7 of 11 with fetal demise

Proper Gating for Anti-HbF FMH QuikQuant Method
Gating Logic: Important to exclude aggregates (rerun if >1%) by gating on singlets
(R1). Also exclude nucleated cells using propidium iodide signal (R3), as
autofluorescence of nucleated cells is similar to fetal RBCs and may cause false
positive results. Thus need boolean gating of R1 and R3.
Nucleated
Cells

0 50 100 150 200
Number
0 50 100 150 200
Number
Adult RBCs
Adult F Cells
Fetal RBCs
10 0 10 1 10 2 10 3 10 4 10 5
Anti- Hb F
Adult RBCs
Adult F Cells
10 0 10 1 10 2 10 3 10 4 10 5
Anti- Hb F
A
C
Fetal RBCs
0 50 100 150 200
Number
0 50 100 150 200
Number
Adult RBCs
Adult F Cells
Fetal RBCs
10 0 10 1 10 2 10 3 10 4 10 5
Anti- Hb F
Adult RBCs
Adult F Cells
Fetal RBCs
10 0 10 1 10 2 10 3 10 4 10 5
Anti- Hb F
Fetomaternal hemorrhage detection using anti-HbF antibody by flow cytometry (FMH
QuikQuant Assay, Trillium Diagnostics, Bangor, Maine). Various samples showing increased
fetal red cells of 1.7% and 6.5% (panels A and C, respectively), increased adult F cells
(panel B), and no fetal red cells (panel D).
B
D

Second Parameter, such as Carbonic Anhydrase, May Facilitate
Improved Separation of Fetal RBCs from Adult F Cells
Slide courtesy of Joost Schuitemaker, IQ Products – Fetal Cell Count Kit Staining

10 1 10 2 10 3 10 4
HbF-TC -->
F RETICULOCYTES (anti-HbF and RNA
staining)
10 1 10 2 10 3 10 4
FL1 LOG -->
10 1 10 2 10 3 10 4
HbF-TC -->
10 1 10 2 10 3 10 4
Thiazole Orange -->
Mundee Y, Bigelow NC, Davis BH, Porter JB: Flow cytometric method for simultaneous assay of
foetal haemoglobin containing red cells, reticulocytes and foetal haemoglobin containing
reticulocytes. Clinical and Laboratory Haematology 23:149-54, 2001

10 30 50 70 90 110 140 170 200 230 260
Number
Quantitation of F-cells
Potential Clinical Utility:
•Sickle Cell Disease Rx with hydroxyurea
•Thalassemia
•Myelodysplasia prognosis?
10 1 10 2 10 3 10 4
ANTI-HGB F -->

18 month old male
HbS and HbF (~40%)
by electrophoresis
Father has 20% HbF
Mother has SS trait
Hgb 12.2 g/dL
MCV 73.4 fL
Hoyer et al.
Am J Clin Pathol
117:857-63, 2002
Hereditary Persistence of HbF
fetal_cells
10 2 10 3 10 4
F - FL1 -->
Number
Number
N u m b e r
100 200 300 400 500 600 700
10 30 50 70 90 110 140 170 200 230 260
10 30 50 70 90 120 150 180 210 240 270
HbF_neg._RBCs
HbF_neg._RBCs
F_cells
Fetal_RBCs
10 1 10 2 10 3 10 4
FL1-H -->
F_cells
Fetal_RBCs
Control
HPFH -Homocellular type
10 1 10 2 10 3 10 4
FL1-H -->
HPFH - Heterocellular type
fetal_cells
10 1 10 2 10 3 10 4
anti-HbF - FL1 -->

F-cell Counting using
autofluorescence gating -
Chen, Bigelow, Davis
Cytometry 42:239, 2000
1. Instrument set-up
requires balanced
fluorescence gains and
color compensation
2. Cursor position
A
B
defined with 99.8% of cells
to left of cursor in FL2
histogram
3. Same channel for
cursor position applied to
FL1 histogram
“Rule-based” F Cell enumeration using
autofluorescence signal to define F cell region
10 1 10 2 10 3 10 4
FL2-H -->
10 1 10 2 10 3 10 4
FL2-H -->
Leuk
RBC
10 1 10 2 10 3 10 4
FL1-H -->
10 1 10 2 10 3 10 4
SSC-H -->
Number
Number
C
100 200 300 400 500
D
100 300 500 700 900 1100
Bkg
10 1 10 2 10 3 10 4
FL2-H -->
HbF-
F_cells
10 1 10 2 10 3 10 4
FL1-H -->

F Cell Counts: Inter-instrument Correlations
F cell percent by Coulter XL
40
35
30
25
20
15
10
5
0
y = 1.041x - 0.49
r 2 = 0.999
0 5 10 15 20 25 30 35 40
F cell percent by FACScan
F cell in 40 normals = 3.8 + 3.6 % - Similar to 3.1 + 2.7 % (Thein and Craig,
Hemoglobin 1998;22:401) and 3.33 + 3.06% (male, monozygotic), 3.19 + 2.58%
(male, dizygotic), 4.19 + 3.04% (female, monozygotic), and 4.07 + 2.78% (female,
dizygotic) (Garner et al, Blood 2000;95:342). From Chen J, Bigelow NC, Davis BH,
Cytometry 42:239, 2000.

F Cell Counts: Imprecision
Intra-Assay Imprecision - 6 observers
Observer F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV
1 35.69 4.07 25.29 3.74 19.83 4.11 11.55 3.38 7.42 7.01 4.72 3.70 3.90 8.92 3.08 20.89
2 34.76 2.04 25.10 2.59 19.80 1.05 11.72 1.87 7.44 2.81 4.76 5.34 3.93 5.49 3.45 12.57
3 34.17 0.49 24.82 1.02 18.53 2.03 10.75 3.50 6.98 6.39 4.27 8.42 3.94 1.93 3.30 11.19
4 34.97 2.08 25.33 3.71 19.72 3.57 11.55 3.38 7.46 5.97 4.66 3.50 3.83 7.08 2.88 17.33
5 35.07 2.22 25.29 3.72 19.69 4.06 11.55 3.38 7.39 6.39 4.72 3.70 3.88 6.79 2.92 18.04
6 35.18 2.50 25.17 3.85 19.83 4.37 11.58 3.32 7.48 6.57 4.72 3.70 3.90 8.26 2.90 17.12
Mean 34.97 2.23 25.17 3.11 19.57 3.20 11.45 3.14 7.36 5.86 4.64 4.72 3.90 6.41 3.09 16.19
Inter-Assay Imprecision - 6 replicates
Replicate F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV F cell % CV
1 34.05 0.51 25.31 1.05 18.88 3.16 11.56 2.25 7.50 1.99 4.61 3.77 3.72 1.62 3.44 1.43
2 35.95 0.63 23.97 1.19 19.86 0.58 11.51 1.32 7.13 1.97 4.76 2.65 3.66 1.50 3.49 8.71
3 35.97 0.99 25.37 1.04 19.23 0.66 11.22 3.47 8.18 3.23 4.88 1.74 4.08 5.03 2.32 6.74
4 34.54 0.37 26.28 4.22 19.93 0.55 11.39 0.55 7.27 0.00 4.46 2.20 4.19 1.99 3.43 3.52
5 35.34 4.30 25.41 1.21 20.69 2.49 11.92 1.73 7.54 1.31 4.70 1.71 4.05 2.34 2.93 12.43
6 34.94 1.08 25.08 0.99 20.05 1.12 11.93 1.87 7.00 2.20 4.88 0.92 3.62 7.24 2.66 20.72
Mean 35.13 1.31 25.24 1.62 19.77 1.43 11.59 1.86 7.44 1.78 4.71 2.17 3.89 3.29 3.05 8.93
Linearity of the F cells quantitation by 6 observers for all the eight
specimens and replicates was 0.9976 +0.0005 with a CV of

Potential Problems with Flow Cytometric Detection of
Fetal RBCs
• Antibody concentration below saturation of all antigen binding sites
• Improper use or concentration of glutaraldehyde or other fixatives
• Improper washing or centrifugation methods
• Instrument setup (gains) and compensation not optimized
• Specimen flow rate on flow cytometer too fast (>5,000 cell/second)
resulting in particle coincidence
• White blood cells not excluded from analysis by gating to remove
autofluorescent events
• Failure to use a positive region for fetal cell enumeration based on a
positive control, such as FETALtrol (Trillium Diagnostics) or artificial mixtures
of cord and adult blood
• Carryover between samples – should run positive control samples last and
flush, rinse, or run a blank (buffer only) tube between samples, if necessary
• RBC aggregates, especially with poor mixing after washes or improperly
high glutaraldehyde concentration and patient samples received posttransfusion:
cause for overestimation of FMH

0 10 20 30 40 50
Number
0 50 100 150 200
Number
Adult RBCs
Effect of Fixative Concentration
Adult F cells
10 0 10 1 10 2 10 3 10 4
Adult RBCs
Adult F cells
anti-HbF
10 microliters
glutaraldehye
Fetal RBCs
10 0 10 1 10 2 10 3 10 4
anti-HbF
50 microliters
glutaraldehye
Fetal RBCs
0 50 100 150 200
Number
0 50 100 150 200
Number
Adult RBCs
Adult F cells
10 0 10 1 10 2 10 3 10 4
Adult RBCs
Adult F cells
anti-HbF
25 microliters
glutaraldehye
Fetal RBCs
10 0 10 1 10 2 10 3 10 4
anti-HbF
75 microliters
glutaraldehye
Fetal RBCs
0 50 100 150 200
Number
0 50 100 150 200 250 300
Number
Adult RBCs
Adult F cells
10 0 10 1 10 2 10 3 10 4
Adult RBCs
From Davis BH et al, Laboratory Medicine 38:365-371, 2007
Adult F cells
anti-HbF
30 microliters
glutaraldehye
Fetal RBCs
RBC
aggregates
100 microliters
glutaraldehye
10 0 10 1 10 2 10 3 10 4
anti-HbF
Fetal RBCs

Effect of Fixative Concentration

CAP HBF Survey: FMH Testing
CAP HBF-A Survey 2000
Methods No. Labs HBF-01 CV HBF-02 CV
Flow Cytometry - Anti-HbF 10 0.53 24.3 1.05 26.2
Kleihauer Assay (home brew) 48 0.70 59.9 1.24 61.9
Kleihauer Assay (Suretech) 151 0.62 45.7 1.09 39.1
Kleihauer Assay (Simmler) 183 0.60 55.3 1.10 50.8
Kleihauer Assay (Sigma) 34 0.71 45.4 1.24 55.3
Gamma Fetal Screen 121 100% + 100% +
Ortho Fetal Screen 245 99.6% + 99.6% +
CAP 1999 HBF Survey: Flow
0.4% sample - 44.5% of labs report >0.6% 0%
0.8% sample - 10.8% of labs report < 0.6% 0%
CAP 2001 HBF Survey:
0.4% sample – 51.2% of labs report >0.6% 12.5%
0.8% sample - 10.6% of labs report < 0.6% 0%

HBF-01
HBF-02
METHOD
2004 CAP HBF Survey for FMH Testing
NO.
LABS MEAN S.D. C.V. MEDIAN
LOW
VALUE
HIGH
VALUE
Fetal Hemoglobin Sigma
Diagnostics 47 0.016 0.048 297.6 0.00 0.00 0.20
Flow Cytometry (anti-HbF) 16 0.034 0.037 108.3 0.02 0.00 0.11
K-B (acid elution) in-house 55 0.016 0.043 269.0 0.00 0.00 0.20
K-B Sure-Tech Diagnostics 274 0.023 0.056 241.7 0.00 0.00 0.32
Simmler Fetal Cell Stain Kit 313 0.019 0.052 278.8 0.00 0.00 0.27
METHOD
NO.
LABS MEAN S.D. C.V. MEDIAN
LOW
VALUE
HIGH
VALUE
Fetal Hemoglobin Sigma
Diagnostics 60 6.416 2.332 36.3 5.74 2.55 12.80
Flow Cytometry (anti-HbF) 17 5.715 0.583 10.2 5.61 4.87 7.29
K-B (acid elution) in-house 63 6.342 2.035 32.1 6.20 0.57 10.50
K-B Sure-Tech Diagnostics 331 5.813 2.335 40.2 5.70 0.05 17.15
Simmler Fetal Cell Stain Kit 390 5.093 2.788 54.7 5.40 0.00 13.50
Conclusions:
•Anti-HbF
method starting
growth
•KB Assay has
lower precision
and accuracy
•Large scale
proficiency
testing for FMH
screening and
quantitation is
feasible

HBF-03
HBF-04
Method
CAP HBF Survey: FMH Testing
HBF-B 2003: Targets: HBF-03 – 0.2%; HBF-04 – 0.6%
No.
Labs
Mean
S.D.
C.V.
Median
Low
Value
High
Value
Eng Scientific (Room Temp.)*
Fetal Hemoglobin Sigma
15 0.398 0.185 46.6 0.35 0.15 0.90
Diagnostics
52 0.308 0.137 44.5 0.30 0.08 0.70
Flow Cytometry (anti-HbF) 29 0.147 0.055 37.3 0.15 0.02 0.24
K-B (acid elution) in-house 62 0.394 0.243 61.6 0.31 0.00 1.10
K-B Sure-Tech Diagnostics 420 0.321 0.161 50.1 0.30 0.00 0.80
Simmler Fetal Cell Stain Kit 453 0.326 0.192 59.1 0.30 0.00 1.00
Method
No.
Labs
Mean
S.D.
C.V.
Median
Low
Value
High
Value
Eng Scientific (Room Temp.)*
Fetal Hemoglobin Sigma
16 1.225 0.569 46.4 1.00 0.74 3.00
Diagnostics
50 0.903 0.269 29.7 0.90 0.40 1.50
Flow Cytometry (anti-HbF) 28 0.579 0.104 18.0 0.60 0.36 0.78
K-B (acid elution) in-house 61 1.163 0.537 46.1 1.00 0.36 3.20
K-B Sure-Tech Diagnostics 422 0.930 0.330 35.5 0.90 0.00 2.01
Simmler Fetal Cell Stain Kit 457 0.996 0.422 42.4 0.93 0.01 2.52

HBF-03
CAP HBF Survey: FMH Testing
HBF-B 2003: Targets: HBF-03 – 0.2%; HBF-04 – 0.6%
Number of
Vials
Therapeutic Decisions for RhoGam Doseage
Number of
Participants
Percent
Number of
Vials
Number of
Participants
Percent
0 11 1.1 0 2 0.2
1 637 62.9 1 90 9.4
2 323 31.9 2 345 35.9
3 25 2.5 3 354 36.8
4 4 0.4 4 113 11.8
5 4 0.4 5 27 2.8
6 2 0.2 6 14 1.5
7 1 0.1 7 3 0.3
10 1 0.1 8 2 0.2
11 5 0.5 9 1 0.1
10 1 0.1
11 4 0.4
12 2 0.2
14 1 0.1
15 2 0.2
HBF-04

2010 CAP HBF B Survey for FMH Testing

2010 CAP HBF Survey for FMH Testing
• CAP now in 2 nd decade of FMH Testing
• Kleihauer Bettke test remains the poorly
performing, but pragmatic reality
• Flow Cytometric Methods
– Anti-HbF clear majority of minority
– Anti-RhD no usage in U.S., but impt in UK
– Hybrids low frequency use, high cost
• Performance unchanged, but variable
due to use of KB assay
• HBF survey by CAP has minimal impact
on clinical practice and has not improve
FMH testing

2000 UKNEQAS Program for FMH Testing
• Program completed pilot and started in April 1998
• By end of 2000 14 participates for screening (10 mL); 199 for acid
elution; 13 for flow cytometry; 16 for both flow and KB
• Testing followed BCSH Guidelines for estimation of fetomaternal haemorrhage.
Transfusion Med 9:87-92, 1999
Survey 0002F included a short questionnaire regarding acid elution techniques. Analysis of
the data revealed the following key points:
• 79% use a commercial kit
• 71% use 80% ethanol (or “alcohol‟) as a fixative
• 55% count the adult cells in each field; 28% count one or more fields but estimate the
rest; 17% assume the number of cells per field based on historical estimates. This
compares with only 26% who counted adult cells in all field in 1997.
• 58% use a graticule (Miller disc) to aid counting adult cells, compared with only 19% in
1997.
Survey Code Target Value %CV Acid Elution %CV Flow Cytometry
9801F 7 53 31
9901F 6 42 12
0003F 6 33 11

2010 UKNEQAS Program for FMH Testing
• Program continue to grow and guidelines modified to adopt flow cytometry as
preferred method if > 2 mL bleed detected (new guidelines in 2009, Transfusion
Med 9:87-92, 1999)
• By end of 2000 14 participates for screening (10 mL); 199 for acid
elution; 13 for flow cytometry; 16 for both flow and KB
• Protective effect of anti-D immunoglobulin is dose-dependent and 125 iu/mL of
packed fetal red cells is recommended when given by the intramuscular route
Calculation Assumptions for reporting:
1800 mL = red cell volume (RCV) of a pregnant woman,
e.g. 6 mL FMH = 0.33% adult RCV
Calculation for 6 mL ‘target’ bleed: X = 0.33 x adult haematocrit / cord haematocrit
where X is the volume of cord blood to be added to each 100 mL of adult blood.
Participation December 2010 by Method UK (n=234) 1 Non-UK (n=34) 2
Quantification by acid elution 190 20
Quantification by flow cytometry 31 16
Screening only 33 5

UKNEQAS Program for FMH Testing: Improved Performance

2009 BCSH Guidelines for Estimation of FMH
• Recognized that the accuracy of counting low frequency events is associated with a high coefficient of
variation (CV) and therefore the cut-off for a significant FMH is now set at 2mL.
• 2006 Guidelines for the use of anti-D immunoglobulin for Rh prophylaxis
– At least 500 iu of anti-D immunoglobulin must be given to every D negative woman with no preformed anti-D within 72 hours
of delivery of a D positive baby - dose will be sufficient to prevent sensitization up to 4mL fetal red cells.
– FMH greater than 4mL is rare but unpredictable; 1% of women have a FMH of greater than 4mL and up to 0.3% greater than
15mL and may not be protected by a 500 iu and 1500 iu dose of anti-D immunoglobulin, respectively.
– Important that the volume of FMH is accurately assessed so that, if necessary, a supplementary dose(s) of anti-D
immunoglobulin can be administered and prevent maternal alloimmunization
• Need for FMH measurement still exists because of the unpredictable nature of large bleeds as review of
134 published cases between 1966 and 1997 where 82% had no demonstrable cause (Giacoia GP. Severe
fetomaternal hemorrhage: a review. Obstet Gynaecol Survey 1997; 52(6):372-380 )
• Flow cytometry tests for a minor D positive population may be required:
– Following a D positive RBC transfusion to a D negative woman of childbearing potential. To estimate or confirm the dose of anti-
D immunoglobulin required, as part of the protocol given in the anti-D guidelines (BCSH 2006a), to prevent sensitisation to the
D antigen.
– In solid organ transplantation when the donor is D positive and recipient D negative with childbearing potential.
• Tests for FMH estimation are not required:
– Before 20 weeks because the fetal blood volume is insufficient to exceed that covered by the minimum anti-D immunoglobulin
dose in standard use.
– When the woman is known to have immune anti-D.
– As it is difficult to distinguish between passive and immune anti-D. In such cases, FMH estimation should be performed.
– When the fetus/baby is known to be D negative.
– When the woman is D positive
– Maternal samples should be tested with saline reacting IgM anti-D reagents that do not detect D
are unlikely to make anti-D that will adversely affect the baby. If there is any doubt, arrange further testing at a reference
laboratory and treat as D negative until results of these tests are available
• In D positive women with unexplained abdominal pain in late pregnancy, FMH tests (by AE) are of limited
diagnostic use. More sensitive and specific tests exist to investigate suspected placental abruption.
• FMH of greater than or equal to 2mL by AE should be confirmed by the flow cytometry method (FC),
using the original sample. If the FC result will not be available within 72 hours, the AE test should be
repeated by a second operator before referral.

Ta
Taken from 2009 BCSH Guidelines

Problems with Current FMH Assay Approach
1. Proficiency testing by College of American Pathologist and UK NEQAS
programs consistently documents inter-laboratory variation between 10 –
30%. While vastly superior to non-flow cytometric methods for FMH
detection, this variation is higher than the intra-laboratory imprecision with
a typical CV of ~5%.
2. Sources of inter-laboratory variation include
• subjectivity of the gating approach
• inconsistency of excluding interfering cell populations (nucleated
cells) and aggregates
• setting regions for the defined subpopulations of Adult RBCs, F cells,
and Fetal RBCs.
3. Education on data analysis is required and often this too can be variable
between products for FMH detection and between technical “experts”.
4. Automated data analysis could remove these variables and reduce the
imprecision.

Benefits of Probability State Modeling
A B C D
E

Study Aims
1. Develop a Probability State Model (PSM) for analysis of flow cytometric listmode data files using
Anti-HbF methods for the detection of fetal red blood cells (RBCs) in maternal blood samples as
used in diagnostic assessment of fetomaternal hemorrhage.
2. Compare PSM model in Gemstone software (Verity Software House, www.vsh.com) to manual
data analysis with WinList software (Verity Software House) as performed by blinded analysis by
an expert (developer of the Anti-HbF method) in >200 data files from analysis of diagnostic
clinical blood samples, artificial mixtures of fetal (cord) blood and adult blood, and stabilized
control product for FMH assays (FETALTrol, Trillium Diagnostics).
3. Examine data from above comparative analysis to determine the correlation between the
methods and determine if any bias is identified between PSM and expert manual analysis.
4. Determine potential for PSM data analysis to provide automated analysis of flow cytometric
data files for clinical fetomaternal hemorrhage testing and provide a more robust alternative to
the current practice of subjective gating and subpopulation region definition that contribute to
the documented imprecision or variation between various diagnostic laboratories performing
flow cytometric fetomaternal hemorrhage testing for pregnant women.

Correlation between Expert “Manual” Analysis
vs Automated Software

Correlation between Two Software Operators

Imprecision of “Manual” Flow Cytometric Data Analysis
Gemstone automated analysis gave a CV of 0% for all 205 samples with triplicate analysis

Conclusions on Gemstone Software to Date
1. Probability State Modeling provides equivalent results to manual data analysis
with standard listmode analysis software by highly experienced individuals for
Fetal RBC analysis, a form of rare event analysis.
2. The Gemstone PSM protocol requires no adjustment of gate or regions,
demonstrating a high likelihood that FMH analysis could be done in an
automated format, thereby removing an important source of inter-individual
variability in the current methods of data analysis.
3. PSM analysis by Gemstone and the improved method of FMH QuikQuant
together promise to further improve the accuracy, precision, and technical
simplicity of FMH detection in clinical flow cytometry.
1. Further study is warranted to validate analysis of clinical samples and
document the degree of improved performance over current manual methods
of flow cytometric data analysis.

Summary of FMH assays:
• Flow cytometric assays (anti-D or anti-HbF) are
superior to KB acid elution in performance.
• FCM HbF assay vs KB cost: higher reagents and
instrumentation, less labor - cost neutral change
• Multiple clinical application of assay:
– Quantitation of fetal maternal hemorrhage
– Evaluation of pre-eclampsia risk?
– Simplified measurement of adult F cell counting for
hemoglobinopathies and MDS
– Evaluation of elevated HbF: homocellular vs.
heterocellular

Questions??