Keywords
Activated partial thromboplastin time (aPTT) - coagulation factor IX - haemophilia
B - coagulation tests - blood - standardisation
Introduction
Intravenous injection of replacement factor IX (FIX) is the mainstay of treatment
for haemophilia B worldwide ([1]). BeneFIX® (BeneFIX® [package insert]. Pfizer Inc., Philadelphia, PA; 2011) is a recombinant factor IX
(rFIX) product that is approved for the treatment of haemophilia B. Measurement of
FIX levels for diagnosis of haemophilia B, or for the assessment of FIX activity in
post-injection samples, is typically performed in clinical haemostasis laboratories
using the FIX one-stage clotting assay ([2]) with commercially available pooled normal plasma standards calibrated against the
World Health Organization (WHO) FIX plasma International Standard ([3]).
The FIX one-stage clotting assay is associated with significant variability, in particular
when measuring recombinant FIX products ([4]–[7]). A number of elements contribute to the observed variability in one-stage assay
performance. The lower limit of quantitation for most one-stage clotting assays is
approximately 1 % factor activity levels (nearly equivalent to the cutoff for severe
haemophilia), making accurate assessment of low activity levels difficult ([8]). Moreover, the wide selection of commercially available reagents and instruments
leads to inter-laboratory variability ([5]–[7]). Differences between the characteristics of FIX International Standards (plasma
or concentrates) and recombinant products further complicate activity determinations
([9]).
As new factor products are developed for the treatment of haemophilia, the accuracy
and precision of commonly used assays with these products will need to be determined.
While many clinicians currently prefer to treat empirically rather than based on laboratory
results, adoption of longer-acting products may be accompanied by increased interest
in individualised pharmacokinetic (PK) assessments to guide development of an initial
dosing regimen ([10]). Failure to identify and recognise analytical variability can potentially lead
to misinterpretation of PK parameters and inaccurate dosing ([11]).
The use of conventional rFIX products in the management of haemophilia B is limited
by their short half-life, ranging from 14–36 hours (h) depending on the length of
the sampling time period ([12]–[15]). Fc fusion technology is a well-established technology for extending the half-life
of therapeutic proteins ([16]), and seven Fc fusion proteins are currently approved by the US Food and Drug Administration
(FDA) ([17]). This technology has been used for the development of a long-acting recombinant
factor IX Fc fusion protein (rFIXFc) approved for the treatment of haemophilia B ([18], [31]). rFIXFc is a recombinant protein composed of a single molecule of rFIX covalently
fused to the Fc domain of human immunoglobulin G1 (IgG1) with no intervening sequence ([19]–[21]). Fc fusion technology utilises an endogenous pathway that is responsible for the
long half-life of human IgG. During circulation, rFIXFc is internalised by endothelial
cells lining the vasculature via pinocytosis. Upon binding to the neonatal Fc receptor
(FcRn), rFIXFc is transported out of the cell and cycled back into the bloodstream.
This interaction delays lysosomal degradation, thus prolonging the circulating half-life
of the protein ([19], [22]).
Biochemical characterisation of rFIXFc demonstrated appropriate gamma-carboxylation,
which is required for full activity, and indicates that other post-translational modifications
were similar to those of BeneFIX and plasma-derived FIX ([20]). However, non-human glycan structures (i. e. N-glycolylneuraminic acid, galactose-alpha-1,3-galactose),
which were detected in Chinese hamster ovary-derived BeneFIX ([23]), were not found in rFIXFc. This was as expected, since rFIXFc is produced in a
human cell line, HEK 293H. Furthermore, the kinetics of in vitro tenase complex assembly
and activated factor X (FXa) generation using purified components appear unaffected
by the Fc fusion ([24]).
The objective of this field study was to evaluate the suitability of a variety of
commercially available one-stage clotting assay reagents and instruments for the measurement
of rFIXFc activity in plasma samples. Haemophilia B plasma samples spiked with rFIXFc
were tested by clinical haemostasis laboratories using their routine procedures and
plasma calibrators traceable to the WHO FIX plasma International Standard. A rFIX
comparator (BeneFIX) was included to evaluate general laboratory variability when
measuring conventional rFIX products.
Materials and methods
Field study kits
Field study kits were prepared and distributed to the participating laboratories (see
Suppl. Information, available online at www.thrombosis-online.com). Human haemophilic donor plasma with no detectable FIX activity (< 0.5 %) was spiked
with either rFIXFc or a rFIX comparator, BeneFIX, at nominal concentrations of 0.80
IU/ml, 0.20 IU/ml, or 0.05 IU/ml, based on the label potency for each product. The
potency of rFIXFc was assigned using the Actin® (Siemens, Erlangen, Germany) activated
partial thrombo-plastin time (aPTT) reagent against an rFIXFc working standard that
was calibrated to the WHO FIX concentrate standard by the same method. Kits contained
3 frozen 1 ml aliquots of each of the three concentrations for both BeneFIX- and rFIXFc-spiked
plasma to permit assays to be conducted in triplicate.
Study design
Field study kits were distributed to 30 laboratories in seven countries in North America,
Europe, South America, and Australia. All clinical laboratories were blinded with
respect to the drug product and concentration in each vial, and were instructed to
assay three sets of samples on different occasions using their routine one-stage assay
procedure and their in-house FIX plasma standard. Each laboratory provided procedural
data on the type and source of reagents and substrate plasma used, instrument employed,
number of dilutions performed in each assay, source of calibrators, and laboratory
certification and type of proficiency testing conducted. Laboratories provided raw
data and FIX activities calculated from in-house standard curves, and supplied the
final results to Biogen Idec for statistical analysis.
In-house verification of field study results
To discriminate between reagent variability and inter-laboratory variability, the
one-stage assay was performed in-house on samples spiked with either 0.80 IU/ml BeneFIX
or 0.80 IU/ml rFIXFc, based on label potency. The aPTT reagents PTT-A® (Stago, Parsippany, NJ, USA) and C. K. Prest® (Stago) were evaluated on a Compact® coagulation analyser (Stago), while Actin and Actin FSL (both from Siemens), SynthAFax®
(Instrumentation Laboratory, Bedford, MA, USA), SynthASil® (Instrumentation Laboratory),
APTT-SP® (Instrumentation Laboratory), and Auto-APTT® (Tcoag, Parsippany, NJ, USA) were tested on a Sysmex® CA-1500 system (Siemens), each in three independent runs. The Sysmex instrument and
the Stago analyser were calibrated using the WHO International Standard for FIX plasma
(07/192), except for the evaluations of APTT-SP and Auto-APTT, which used reference
plasma from Precision Biologic (Dartmouth, NS, Canada) for calibration.
In-house chromogenic assay
The field study samples were also assayed for chromogenic activity using a Biophen
FIX chromogenic kit (Hyphen Biomed, Neuville, France). The WHO plasma standard (07/192)
was used as the reference standard. All field study samples were tested with multiple
dilutions in duplicate in a 96-well plate on a SpectraMax M5 microplate reader (Molecular
Devices, Sunnyvale, CA, USA) using an endpoint reading method.
Comparison of WHO 4th FIX concentrate and plasma standards
Potency assignment for FIX products typically employs a standard that is calibrated
against the WHO FIX concentrate standard, while clinical haemostasis laboratories
generally use a normal pooled plasma standard, which has an activity assigned against
the WHO FIX plasma standard. To assess potential differences in FIX activity between
the concentrate and plasma standards and determine whether this may have contributed
to differences between measured and nominal FIX activity observed in the field study,
the WHO 4th FIX concentrate standard was directly compared against the WHO 4th FIX
plasma standard using six different aPTT reagents at three dilutions in triplicate
experiments. The aPTT reagents included Actin, Actin FSL, SynthAFax, SynthASil, PTT-A,
and C. K. Prest.
Data and statistical analysis
Manual data entries were cross-checked against source documents. The 0.80 IU/ml and
0.20 IU/ml BeneFIX samples were apparently swapped in one laboratory, resulting in
outliers for the corresponding results, and these two test results were excluded from
the analysis. Two results from different laboratories for the 0.05 IU/ml rFIXFc sample
were excluded as outliers because the results were reported as ≤ 0.01 IU/ml. Results
were analysed for intra-laboratory and inter-laboratory variation and accuracy measurements
were compared with nominal or consensus, i. e. laboratory mean, activity. Relative
variation and accuracy for rFIXFc compared with BeneFIX was evaluated based on type
of activating reagent (e. g. silica, kaolin, ellagic acid) and instrumentation used.
Statistical evaluations were performed by unpaired t-test or analysis of variance
(ANOVA) methods to assess whether assay-related differences, including particular
assay reagents, standards, instrumentation, or methodology, affected FIX activity
measurements.
Results
Reagent and instrument use
An array of one-stage assay reagents and critical steps in methodology were employed
by the 30 participating laboratories (Suppl. Table 1, available online at www.thrombosis-online.com). aPTT activating reagents used included ellagic acid (8 laboratories), silica (17
laboratories), kaolin (4 laboratories), and polyphenols (1 laboratory). Assay instrumentation
varied among laboratories, but typically a combination of analyser and reagents from
the same vendor was used. Overall, the range of aPTT reagents and instruments used
by the 30 participating laboratories was representative of the distribution of reagents
and coagulation analysers used among all major specialty haemostasis laboratories,
based on the 2012 FIX proficiency testing administered by the College of American
Pathologists (CAP; Northfield, IL, USA, http://www.cap.org) and the External Quality Control of Diagnostic Assays and Tests (ECAT) foundation
(Leiden, The Netherlands, http://www.ecat.nl).
One-stage clotting assay
The FIX activity in the BeneFIX or rFIXFc samples was calculated as a percentage of
nominal (label) activity (►[Table 1], ►[Figure 1]). There were 90 test results per dose level of BeneFIX or rFIXFc across the 30 clinical
laboratories. At 0.80 IU/ml, the mean spike recovery by the one-stage clotting assay
was 0.966 IU/ml for Bene-FIX and 0.707 IU/ml for rFIXFc. This corresponds with 121
% of expected for BeneFIX and 88 % of expected for rFIXFc. At the lower concentrations
of 0.20 IU/ml and 0.05 IU/ml, the mean spike recovery for BeneFIX was 0.289 IU/ml
and 0.084 IU/ml, respectively, which corresponds with 144 % and 168 % of expected.
For rFIXFc, the mean spike recovery at 0.20 IU/ml and 0.05 IU/ml was 0.214 IU/ml and
0.066 IU/ml, respectively, which corresponds with 107 % and 132 % of expected. The
higher degree of overestimation of FIX activity at the 0.20 IU/ml and 0.05 IU/ml levels
relative to the 0.80 IU/ml results was observed for both BeneFIX and rFIXFc in nearly
all participating laboratories (►[Figure 1]). A small number of laboratories observed good dilution linearity for the three
samples; however, there was no obvious trend in the methodology that correlated with
laboratories that performed well in this respect. The median intra-laboratory coefficient
of variation (CV) for the one-stage assay (n=3 independent tests per level) was below
10 % for both products (►[Table 1]). Comparing mean results between laboratories, the inter-laboratory CV ranged from
12 % to 30 % for BeneFIX and from 26 % to 44 % for rFIXFc (►[Table 1]).
Table 1
rFIX one-stage clotting assay results.
|
rFIX
product
|
Label activity
(IU/ml)
|
Mean spike
recovery
(IU/ml)
|
% of expected
(label activity)
|
Intra-laboratory
CV (n=3)
|
Inter-laboratory
CV (n=30)
|
|
Median
|
Range
|
|
BeneFIX
|
0.80
|
0.966
|
121
|
5.6 %
|
0 %-25.0 %
|
12.1 %
|
|
0.20
|
0.289
|
144
|
3.6 %
|
0 %-21.4 %
|
19.7 %
|
|
0.05
|
0.084
|
168
|
7.3 %
|
0 %-39.8 %
|
29.8 %
|
|
rFIXFc
|
0.80
|
0.707
|
88
|
6.7 %
|
0 %-18.4 %
|
26.3 %
|
|
0.20
|
0.214
|
107
|
6.7 %
|
1.4 %-20.9 %
|
35.5 %
|
|
0.05
|
0.066
|
132
|
8.4 %
|
0 %-36.7 %
|
44.1 %
|
rFIX, recombinant factor IX; IU, International Units; CV, coefficient of variation;
rFIXFc, recombinant factor IX Fc fusion protein.
Figure 1: Mean (SD) FIX activity by one-stage clotting assay (n = 30 laboratories). SD, standard deviation; FIX, factor IX; rFIXFc, recombinant factor IX Fc fusion
protein; IU, International Units. Dashed line represents the expected recovery. Dotted
lines indicate the degree of non-linearity observed by the average laboratory across
the three concentrations of BeneFIX or rFIXFc.
Differences were observed between laboratories, depending on the activating reagent
used in the clotting assay (►[Figure 2]). Ellagic acid (8 laboratories) generally resulted in the highest observed activities,
with mean ± SD percent spike recovery and interlaboratory variation (%CV) of 124 ±
15 % (12 % CV), 162 ± 31 % (19 % CV), and 193 ± 67 % (35 % CV) for BeneFIX and 115
± 14 % (12 % CV), 150 ± 27 % (18 % CV), and 182 ± 75 % (41 % CV) for rFIXFc at the
0.80 IU/ml, 0.20 IU/ml, and 0.05 IU/ml concentrations, respectively. Silica (17 laboratories)
typically resulted in lower observed recoveries. For silica, mean ± SD percent spike
recovery and %CV were 122 ± 15 % (13 % CV), 142 ± 27 % (19 % CV), and 162 ± 44 % (27
% CV) for BeneFIX and 85 ± 15 % (17 % CV), 98 ± 27 % (27 % CV), and 120 ± 39 % (32
% CV) for rFIXFc at the 0.80 IU/ml, 0.20 IU/ml, and 0.05 IU/ml concentrations, respectively.
Observed recoveries were also generally lower with kaolin (4 laboratories). For kaolin,
mean ± SD percent spike recovery and %CV were 115 ± 7 % (6 % CV), 125 ± 16 % (13 %
CV), and 148 ± 22 % (15 % CV) for BeneFIX and 53 ± 5 % (10 % CV), 61 ± 8 % (14 % CV),
and 80 ± 14 % (18 % CV) for rFIXFc at the 0.80 IU/ ml, 0.20 IU/ml, and 0.05 IU/ml
concentrations, respectively. For BeneFIX activity in this field study, these differences,
based on the type of aPTT reagent, were not statistically significant (p > 0.05).
However, there were also no obvious correlations between other methodological differences
(e. g. instrument, source of calibration plasma, calibration frequency, and number
of dilutions per test) and relative accuracy among the 30 laboratories for BeneFIX
or rFIXFc. Laboratories that trended high or low for rFIXFc also tended to report
correspondingly higher or lower than expected results for BeneFIX at all concentration
levels studied (►[Figure 3]).
Figure 2: Median (25 %/75 %) FIX activity in the one-stage clotting assay by type
of aPTT activating reagent for BeneFIX and rFIXFc at concentrations of 0.80 IU/ml
(A), 0.20 IU/ml (B), and 0.05 IU/ml (C). FIX, factor IX; aPTT, activated partial thromboplastin time; rFIXFc, recombi-nant
factor IX Fc fusion protein; IU, International Units; EA, ellagic acid. *p < 0.05.
Figure 3: Individual laboratory one-stage assay results for Bene-FIX and rFIXFc at
concentrations of 0.80 IU/ml (A), 0.20 IU/ml (B), and 0.05 IU/ml (C). rFIXFc, recombinant factor IX Fc fusion protein; IU, International Units; FIX, factor
IX; P, polyphenols. Dashed line represents the expected recovery.
In-house verification of field study results
The results of the in-house evaluation of BeneFIX and rFIXFc are shown in ►[Figure 4]. BeneFIX spike recovery with the eight reagents ranged from 83 % (SynthAFax) to
125 % (PTT-A); the spike recovery of rFIXFc ranged from 51 % (C. K. Prest) to 121
% (SynthAFax). The percent spike recovery of rFIXFc averaged 105.5 ± 3.4 % of expected
with the Actin reagent. For the Actin FSL reagent, the percent spike recovery of rFIXFc
averaged 104.1 ± 1.5 % of expected. SynthAFax, an aPTT reagent composed of ellagic
acid and a synthetic mix of phospholipids, resulted in underestimation of BeneFIX
with a spike recovery of 83.4 % and an overestimation of rFIXFc with a spike recovery
of 120.7 %. Of note, SynthAFax was not used by any of the 30 laboratories in the field
study. The four silica-based aPTT reagents evaluated in the in-house study averaged
81.0 ± 3.3 % spike recovery of rFIXFc. The evaluation using the kaolin reagent C.
K. Prest and the Stago coagulation analyser showed a spike recovery of 51 ± 0.7 %
for rFIXFc. Overall, the in-house results agreed well with the trends observed in
the field study.
The in-house chromogenic substrate assay results for BeneFIX and rFIXFc are shown
in ►[Table 2]. When calibrated using the WHO International Standard for FIX plasma (07/192), the
chromogenic assay recovered 101 %, 93.5 %, and 84 % of the nominal rFIXFc activity
at the 0.80 IU/ml, 0.20 IU/ml, and 0.05 IU/ml levels, while BeneFIX activity ranged
from 95 % to 62 % at the three concentrations.
Figure 4: In-house confirmation of one-stage assay reagent-specific differences. rFIXFc, recombinant factor IX Fc fusion protein; IU, International Units. Dashed
line represents the expected recovery. The Stago reagents PTT-A and C. K. Prest were
evaluated on a Stago Compact analyser, while all other reagents were evaluated on
a Sysmex CA-1500 system.
Table 2
In-house chromogenic substrate assay results (n = 3 independent assays).
|
rFIX product
|
Nominal
concentration
(IU/ml)
|
Mean ± SD FIX activity (IU/ml)
|
|
BeneFIX
|
0.80
|
0.763 ± 0.031
|
|
0.20
|
0.154 ± 0.012
|
|
0.05
|
0.031 ± 0.007
|
|
rFIXFc
|
0.80
|
0.808 ± 0.072
|
|
0.20
|
0.187 ± 0.020
|
|
0.05
|
0.042 ± 0.005
|
rFIX, recombinant factor IX; IU, International Units; SD, standard deviation; rFIXFc,
recombinant factor IX Fc fusion protein.
Comparison of WHO 4th FIX concentrate and plasma standards
In the field study, BeneFIX activity was overestimated by an average of 121 % at the
0.80 IU/ml concentration. The potency of the BeneFIX product used in this study was
evaluated against the WHO 4th FIX concentrate standard and was demonstrated to have
an average FIX activity within 5 % of the label claim (data not shown). Thus, the
higher apparent FIX activity for BeneFIX observed in the field study was not due to
overfill of the BeneFIX vials. Results of the evaluation of the WHO 4th FIX concentrate
standard directly against the WHO 4th FIX plasma standard demonstrated that the concentrate
standard had 10–20 % higher FIX activity than the plasma standard with most aPTT reagents
used (Suppl. Figure 1, available online at www.thrombosis-online.com). At the 1.0 IU/ml concentration, the exact difference ranged from –0.6 % to +18
%, depending on the aPTT reagent used. The SynthAFax reagent produced different results
from the other five reagents in that it showed comparable activity or slightly lower
activity for the concentrate standard compared with the plasma standard. The SynthAFax
results are also consistent with the lower apparent activity observed for BeneFIX
in the in-house assays when using this reagent compared to the other reagents evaluated
(►[Figure 4]). These differences between the two standards may in part account for the higher
than expected activity seen for Bene-FIX, and for rFIXFc using ellagic acid activators,
in the field study and the in-house assays (►[Figure 2] and ►[Figure 4]).
Discussion
The primary aim of this field study was to determine whether clinical haemostasis
laboratories will be able to accurately monitor the activity of rFIXFc in people with
haemophilia using their routine one-stage assay procedures and standards. Thirty laboratories
in seven countries participated in the study, which demonstrated considerable inter-laboratory
variability for both products in the assessment of FIX activity by the one-stage assay.
Elements contributing to this finding include the large range of reagents and instrumentation
used as well as differences in laboratory technique ([5]–[7]).
Most laboratories in the field study overestimated the FIX activity of BeneFIX at
all concentrations. rFIXFc was similarly overestimated in laboratories that used ellagic
acid activators in the one-stage clotting assay. At the 0.80 IU/ml level, the overestimation
may in part be explained by a difference in the standards used for potency assignment
and clinical testing. The cause for the progressively higher overestimation of activities
at the two lower FIX levels in the majority of laboratories is, however, unclear and
was not observed in all laboratories. Accurate preparation of the test samples was
confirmed by an in-house chromogenic FIX assay, which showed excellent dilution linearity
for the three test concentrations against the WHO FIX plasma standard.
The measurement of rFIXFc activity was found to be dependent on the aPTT reagent used
in the one-stage assay. In the eight laboratories using ellagic acid-based aPTT reagents,
the measured rFIXFc activity averaged 115 % (12 % CV) of the label activity at the
0.80 IU/ml level. This modest overestimation of rFIXFc activity could be due to higher
activity in the concentrate standard used for potency assignment compared with the
plasma standard used in clinical haemostasis laboratories, as discussed above. Compared
to the results with ellagic acid-based aPTT reagents and compared to label potency,
rFIXFc activity was sometimes modestly underestimated by silica-based reagents. For
the 17 laboratories using silica reagents, on average there was a 15 % underestimation
of rFIXFc activity at 0.80 IU/ml versus the nominal value. However, compared with
the average ellagic acid result for rFIXFc, this resulted in an apparent difference
of 30 %. Approximately onefourth of all laboratories performing silica-based assays
produced results for rFIXFc that were in the same range as the results from laboratories
using ellagic acid. At the lower concentrations, the under estimation of rFIXFc by
silica-based reagents was compensated by a consistent overestimation of FIX activity
by an average of 20 % to 30 % relative to the 0.80 IU/ml sample. Thus, it is not possible
to conclude that silica-based FIX assays will necessarily produce less accurate results
for rFIXFc than ellagic acid reagents, since this may depend largely on the particular
laboratory, as clearly shown with BeneFIX.
Four of 30 laboratories in this field study used a kaolin-based aPTT reagent (C. K.
Prest), with average results of 53 % (10 % CV) of the label activity for the 0.80
IU/ml rFIXFc samples. Similar results were found with the in-house studies using the
same kaolin-based reagent on a Stago coagulation analyser, which showed a spike recovery
of 51 ± 0.7 % for the 0.80 IU/ml rFIXFc sample. The use of the kaolin reagent thus
appears to result in more substantial and consistent underestimation of rFIXFc activity
than the silica reagents. The underestimation of rFIXFc by the kaolin reagent versus
the nominal 0.05 IU/ml was on average only 20 %, which would not be expected to be
clinically relevant. However, when compared to the average result obtained by ellagic
acid at 0.05 IU/ml, there was still a 50 % discrepancy. These points should be taken
into account by treaters of haemophilia when using the kaolin reagent to measure rFIXFc
activity.
What is known about this topic?
-
The FIX one-stage clotting assay is associated with significant variability, in particular
when measuring recombinant FIX products.
-
The wide selection of commercially available reagents and instruments leads to inter-laboratory
variability.
-
Modified FIX products may encounter additional assay-related variabilities when activity
is measured in the one-stage clotting assay at different laboratories.
What does this paper add?
-
This paper describes results of a field study to evaluate the suitability of commercially
available one-stage clotting assay reagents and instruments for the measurement of
rFIXFc activity in plasma samples.
-
Results revealed generally high inter-laboratory variability for both BeneFIX and
rFIXFc.
-
Despite these assay-related issues, the activity of rFIXFc was measured with acceptable
accuracy and reliability in the majority of laboratories using their existing methods
and reagents.
Potency assignments for most FIX products are derived from a standard calibrated against
the WHO FIX concentrate standard, whereas most clinical laboratories use a standard
calibrated against the WHO FIX plasma standard. Potential differences between the
concentrate standard and the plasma standard were investigated in the collaborative
study that established the 4th International Standard for FIX concentrate ([25]). In that study, the 3rd International Standard for FIX plasma was found to be 6
% lower than its assigned value when assayed against the 3rd International Standard
for FIX concentrate. The 4th International Standard for FIX plasma was calibrated
against the 3rd International Standard for FIX plasma, and this discrepancy between
the plasma and concentrate standards was likely carried forward. When the WHO 4th
FIX concentrate and plasma standards were compared in this study, a 10 % to 20 % lower
FIX activity was seen with the plasma standard when compared to the concentrate standard
for most of the six aPTT reagents tested. However, despite the differences between
the two standards, the distribution of reagent-specific effects was similar when rFIXFc
activity was assayed against the concentrate standard (data not shown). Therefore,
discrepancies between the nominal and measured activities could not be entirely explained
by the differences in the calibration standard.
The clinical significance of the observed differences in assaying FIX activity has
not been fully established. In the phase 3 clinical study, B-LONG, all measurements
were performed utilising a central laboratory ([15]). The central laboratory assay was developed and validated for plasma FIX, BeneFIX,
and rFIXFc prior to conducting this field study, and used a particular silica-based
aPTT reagent and a commercially available plasma standard, which proved to accurately
recover BeneFIX and rFIXFc at all concentrations when included in the blinded field
study (Lab 11 in ►[Figure 3]).
Inaccurate FIX activity measurements are a concern during surgical procedures, when
local laboratory measurements are utilised in order to maintain sufficient FIX activity
to allow for adequate recovery without increased risk of thrombosis ([26]). Marked variability in the measurement of FIX activity during the perioperative
period might lead to unnecessary dose corrections and, in extreme cases, potential
safety issues. However, it is unlikely that dose adjustments based on the range of
laboratory measurements seen in this field study would be large enough to result in
risk to the person with haemophilia from either underdosing or overdosing rFIXFc.
Overall, the results of this field study confirmed that the large variety of aPTT
reagents used for FIX determinations by the onestage assay and differences in methodologies
and reference standards continue to produce large inter-laboratory variability, with
considerable inaccuracies in many cases ([27]). The results of this field study suggest that, at some sites, physicians may not
be able to rely on local laboratory results for accurate PK determinations for either
drug because the high assay variability and non-linearity observed in measurements
for both BeneFIX and rFIXFc may render PK calculations misleading. In particular,
the relative overestimation at lower FIX activity levels may significantly impact
the prediction of terminal half-life and projected trough levels, resulting in overestimation
of the amount of circulating BeneFIX by many laboratories and subsequent risk of undertreatment
of the person with haemophilia. While current procedures in the majority of laboratories
appear adequate for routine monitoring of BeneFIX and rFIXFc, further progress toward
harmonisation of reference standards and method and reagent uniformity could potentially
improve the overall accuracy when measuring recombinant FIX products.
A limitation of this study is the use of human haemophilic donor plasma spiked with
rFIXFc or BeneFIX to evaluate FIX activity levels. It is possible that assay results
using samples taken from people with haemophilia receiving treatment with recombinant
FIX products would be different. However, haemophilic samples spiked with concentrate
have been utilised by previous field studies comparing the accuracy and consistency
of the onestage clotting assay across clinical laboratories ([28]–[30]). In addition, it would not be feasible to conduct a field study on this scale using
blood samples from people with haemophilia due to the volume of sample that would
be required.
Conclusions
In this field study, significant inter-laboratory variability was observed for both
BeneFIX and rFIXFc at 30 representative clinical haemostasis laboratories. Specific
reagent dependency was observed for rFIXFc, with ellagic acid-based aPTT reagents
yielding comparable results to BeneFIX at equivalent nominal concentrations. While
BeneFIX was consistently overestimated in most laboratories, marginally lower FIX
measurements were observed for rFIXFc in some laboratories using silica-based activators,
and notably lower results were observed for rFIXFc in the four laboratories using
kaolin-based activators. Overall, 80 % of laboratories measured rFIXFc within an accuracy
of ±30 % at the 0.80 IU/ml level, compared to 70 % of laboratories for BeneFIX. Not
including the laboratories that used the kaolin-based activator, over 90 % of laboratories
reported rFIXFc results within ±30 % of the expected 0.80 IU/ml value. We thus conclude
that the activity of rFIXFc was measured with acceptable accuracy and reliability
in the majority of laboratories using existing one-stage assay methods and current
plasma standards.
