[ARTICLE] 2016 A world-wide survey and field study in clinical haemostasis laboratories to evaluate FVIII:C activity assay variability of ADYNOVATE and OBIZUR in comparison with ADVATE
ⓒSoo Young Jung
2017. 7. 26. 11:07
A world-wide survey and field study in clinical haemostasis laboratories to evaluate FVIII:C activity assay variability of ADYNOVATE and OBIZUR in comparison with ADVATE
Discrepancies have been previously reported for one-stage clotting and chromogenic assays for FVIII activity analysis. Inter-laboratory variations in instruments, method of clot detection, assay set-up, reference standard calibration, reagent source and reagent composition all contribute to assay variability.
Aim
To characterise multilaboratory assay variability in measuring ADYNOVATE, OBIZUR and ADVATE FVIII activity in human plasma and survey multinational FVIII activity assay preferences.
Methods
As samples from patients treated with either of the FVIII products are not available in the quantities required for a systematic collaborative study, haemophilia A plasma was spiked in vitro with either ADYNOVATE (PEGylated rFVIII), OBIZUR [Porcine Sequence Antihaemophilic Factor (Recombinant)] or ADVATE at high (0.80 IU or U mL−1), medium (0.20 IU or U mL−1) and low (0.05 IU or U mL−1) FVIII concentrations, based on labelled potencies. Clinical laboratories used their routine FVIII activity assay to determine FVIII activity of each product. Thirty-five data sets using one-stage clotting assay and 11 sets using chromogenic assay were obtained.
Results
A vast majority of laboratories (98%) prefer and rely on the one-stage clotting assay. Mean recoveries across all concentrations were 113%, 120% and 127% for ADYNOVATE, OBIZUR and ADVATE respectively. Assay variation was comparable between ADVATE, ADYNOVATE and OBIZUR with inter-laboratory percent coefficients of variation (%CV) ranging from 11 to 22%. Mean chromogenic assay results were 116%, 51% and 113% for ADYNOVATE, OBIZUR and ADVATE respectively. Inter-laboratory CV's were similar for ADYNOVATE, OBIZUR and ADVATE.
Conclusions
One-stage clotting assays can and will be used with sufficient accuracy and precision for the measurement of ADYNOVATE, OBIZUR and ADVATE in plasma samples from subjects with haemophilia A. Chromogenic assay underestimates OBIZUR potency, particularly at lower concentrations.
Introduction
Haemophilia A, a bleeding disorder defined by reduced endogenous FVIII activity levels (severe disease: FVIII <1%), is best managed by replacement therapy with plasma-derived (pdFVIII) or recombinant FVIII (rFVIII) products that raise circulating FVIII to therapeutic levels [1, 2]. Technological advancements have brought forth a variety of rFVIII products and more are under development. As new products differ from each other and predecessors not only by their novel manufacturing processes (modifying the FVIII structure in some cases) but potentially also by DNA sequence or post-translational modifications, the pharmacokinetic profile and haemostatic efficacy of the products are thereby divergent [3-9]. Therefore, it is of importance to comprehensively characterise new products and accurately determine their haemostatic potency, relative to currently available and well-established FVIII therapy.
Baxalta is developing several new rFVIII products to reduce the burden of frequent injections in severe haemophilia A patients and to address the challenges of haemophilia A patients who develop inhibitors to FVIII. ADYNOVATE is a human, full-length, recombinant factor VIII (rFVIII) modified with polyethylene glycol (PEG). ADYNOVATE is characterised by an extended half-life for on-demand treatment and control of bleeding episodes, and for routine prophylaxis to reduce the frequency of bleeding episodes of patients with haemophilia A. ADYNOVATE is manufactured by covalently binding a branched PEG reagent with a molecular weight of 20 kDa to full-length rFVIII (ADVATE) [9-12]. ADVATE is expressed in Chinese hamster Ovary (CHO) cells by a plasma- and albumin-free cell culture method [13]. Each vial of ADVATE is labelled with the FVIII activity either in actual or nominal IU depending on country-specific regulatory requirements. In the batch release procedure, FVIII activity is both determined with one-stage clotting and chromogenic assays which need to give comparable results within the validated range of accuracy and precision. The potency of ADYNOVATE is assigned in International Units using the one-stage clotting assay.
OBIZUR (antihaemophilic factor recombinant porcine sequence; rpFVIII; Baxalta, Inc.) is a purified glycoprotein produced in Baby hamster Kidney (BHK) cells by recombinant DNA technology [8, 14]. It is an antihaemophilic factor currently licensed in the United States and indicated for the treatment of bleeding episodes in adults with acquired haemophilia A [15]. Acquired haemophilia A patients have circulating autoantibodies that neutralise endogenous FVIII function. OBIZUR is a recombinant porcine sequence FVIII that controls bleeding episodes in the presence of FVIII antibodies [16]. The OBIZUR product insert recommends that the one-stage clotting assay is used to measure FVIII activity levels after OBIZUR dosing. Each vial of OBIZUR is labelled with the actual rpFVIII activity expressed in units following a recent recommendation [17] determined by a one-stage clotting assay, using a reference rpFVIII material calibrated against the World Health Organization (WHO) 8th International Standard for human FVIII concentrates.
The potency of rFVIII products is assigned by FVIII activity assays and some regional regulations may mandate either a one-stage clotting assay or a chromogenic assay for the assignment of potency. ADVATE potency is assigned in IU using the chromogenic assay in the EU or one-stage clotting assay in the United States [13]. Methods and reagents to determine FVIII activity levels are not fully standardised [18]. As a result, inter-laboratory variability in measurements can occur due to different instruments, methods of clot detection, assay set-up, reference standard calibration, reagent source and reagent composition [19-21]. Most clinical laboratories use one-stage clotting assays to analyse postinfusion human plasma samples [22-24]. Chromogenic assays are also in use and have variability resulting from different methods, instruments and assay kits.
B-domain-deleted rFVIII products have historically shown discrepant results when measured using different assay methods [25]. Lower one-stage clotting assay activities were obtained compared to FVIII activities determined using a chromogenic assay. Furthermore, discordant assay results have been reported for a PEGylated BDD-rFVIII product under development (BAY 94-9027) [26]. The source and composition of APTT reagents used for one-stage clotting assays was identified as a source of variability in results for postinfusion FVIII activity tests as well as potency assignment of BAY 94-9027 [27]. The variability in FVIII activity measurements has clinical consequences, as FVIII:C levels determine treatment protocols and dosing regimens [28]. It is therefore important to characterise what, if any, factors contribute towards discrepancies in results across a variety of assay systems with different instruments and reagents.
This study is an international, collaborative study between clinical laboratories to analyse FVIII activity of two new products, ADYNOVATE and OBIZUR in human plasma from patients with haemophilia A. The objective of the study was to determine if variability of assay methods causes discordance of FVIII activity results in the evaluation of the potency of the new compounds, ADYNOVATE and OBIZUR, at high (0.8 IU or U mL−1), medium (0.2 IU or U mL−1) and low (0.05 IU or U mL−1) FVIII concentrations. For measurement of ADVATE as a commercially available product with more than 14 years of experience in clinical use, variability seemed to be in the expected range and recent studies demonstrated no relevant discrepancies when measuring this product with different assay methods and in different laboratories [29, 30]. Thus, ADVATE provided a point of comparison to determine the extent of variability between methods. This field study was preceded by a survey among all invited laboratories to understand assay preferences for FVIII determination in clinical practice. Both for the survey and the study, laboratories were selected from a broad range of centres and laboratories participating in clinical trials with FVIII products or who had shown their respective expertise by publishing and/or participation in collaborative assay studies.
Materials and methods
Survey
The field study was preceded by a survey among all invited laboratories. A questionnaire was sent to 127 laboratories in 25 countries in all continents, 20 in Asia, five in Australia and New Zealand, 45 in Europe, 55 in North America (USA: 53; Canada: 2) and two in South America. The questionnaire asked for details on assay methods, instruments, assay reagents including FVIII-deficient plasma and APTT reagent, reference standard, frequency of preparation of reference curve, number of different dilutions tested per sample, assay controls and further details on patient plasma samples testing. Feedback was collected and the answers were evaluated. (see attached questionnaire in Supporting information).
Assay study design
Participating laboratories were asked to use their routinely established methods for FVIII activity analysis in human plasma samples and report the methods used and the FVIII activity measurements obtained. FVIII activity assays employed could be one-stage clotting assay and/or chromogenic assay according to the routinely used FVIII assay methods available in the respective laboratories. No restriction or requirement for standardisation was imposed on the laboratory methods.
The number of centres enrolled in the study was not limited. Of the 127 laboratories who were sent the survey questionnaire, 56 expressed interest to participate in the study and ultimately 35 data sets were available for evaluation.
Study samples
A commercially available FVIII-deficient plasma pool from donors with haemophilia A was used (George King Biomed, Overland Park, KS, USA) for preparation of the in vitro study samples.
Test articles were ADYNOVATE (human recombinant FVIII modified with polyethylene glycol) and OBIZUR (porcine sequence antihaemophilic recombinant FVIII). ADVATE (antihaemophilic recombinant FVIII, octocog alfa) was used as a reference control.
For each study sample, haemophilic donor plasma was spiked with one of the test articles to achieve one of the following target FVIII concentrations based on labelled potencies: 0.8 IU mL−1, 0.2 IU mL−1 or 0.05 IU mL−1. The positive control was human plasma pool with 0.88 IU mL−1 FVIII concentration.
Clotting assays
Each laboratory received three identical sets of samples and was asked to run each set in an independent assay run, resulting in three results for each study sample which subsequently were averaged. Laboratories which typically employed both one-stage and chromogenic assays performed both assays on the same sample set. Clot detection was mechanical or optical according to routine methods. FVIII chromogenic assay kits used were typically Chromogenic Assay Kit (Siemens), Berichrom FVIII (Siemens), Coamatic FVIII (Chromogenix) and Electrachrome (Instrumentation Laboratory). Instruments used were typically STA-R Evolution/STA compact (Diagnostika Stago), BCS/XP (Siemens), Sysmex, ACL Top, Elite Pro (Instrumentation Laboratory).
Statistical analysis
Data were analysed by Quintiles Inc. (Bloemfontein, South Africa) using appropriate statistical methods. A linear mixed-effects model was used to model the logarithm of FVIII activity measurements per assay. The model consisted of fixed effects for product (ADYNOVATE/OBIZUR/ADVATE), spiked concentration (low/medium/high) and their interaction. Laboratory was included as random effect. The distributions of random effects (inter-laboratory) and residuals (within-laboratory) were permitted to be heterogeneous across products and spiked concentrations. The model was fitted using procedure PROC MIXED of sassoftware 9 (World Headquarters SAS Institute Inc., Cary, NC, USA). Data are presented in table form and in graph form using Prism 5, version 5.04, (GraphPad Software Inc., La Jolla, CA, USA).
Results
Survey
Fifty six of 127 invited laboratories responded to the survey. Of these, 55 (98.2%) reported performing and relying on the one-stage clotting assay, while 41 (73.2%) used the one-stage clotting assay as their sole FVIII activity test system with no other FVIII assay available. One laboratory (1.8%) used only the chromogenic assay, while 14 (25.0%) used both. From the laboratories performing the one-stage clotting assay as their preferred assay system (55), the majority of 32 (58.2%) used silica-based APTT reagents, followed by 16 (29.1%) using ellagic acid reagents and 6 (10.9%) kaolin based reagents. Only one laboratory used a polyphenol-based APTT reagent (1.8%). The method of clot detection was either mechanical (40%) or by optical methods (60%). In general, laboratories are using stored assay reference curves, replacing curves with every change in reagent lots (45%). However, some laboratories run reference curves at regular intervals (31%), with a minority doing daily or before every assay run (22%), 2% of laboratories did not provide information. Almost all laboratories (53 of 55) calibrated their method with a commercially available human plasma reference standard, either from a manufacturer of calibration standards or the manufacturer of their analytical instrument. A total of 65% relied on the potency assignment by the manufacturer, while 33% calibrated in-house, 2% provided no answer. The majority of laboratories analysed patient plasma samples in three dilutions with a range from 1–8 dilutions.
Assay study
Laboratories participating in the field assay study were globally distributed. Approximately, 43% of laboratory results were obtained from USA, with the remaining from Argentina, Canada, France, Germany, India, Lithuania, New Zealand, Spain, Sweden, United Kingdom and Venezuela (Table 1). One-stage clotting assay was used in all countries, whereas chromogenic assay was mainly used in non-US countries (eight laboratories in seven countries). Only three of the 15 US laboratories provided results obtained by a chromogenic method. All participants with the exception of one used automated coagulation analysers for measurements for both one-stage clotting assays and chromogenic assays.
Table 1. Participating countries and laboratory methods
Country
Number of laboratories using one-stage clotting assay
Number of laboratories using chromogenic assay
Argentina
1
0
Canada
1
0
France
6
0
Germany
2
1
India
1
1
Lithuania
1
1
New Zealand
1
0
Spain
2
2
Sweden
1
1
United Kingdom
3
1
USA
15
3
Venezuela
1
1
Twenty-five (71%) laboratories used silica/kaolin-type APTT reagents and 10 (29%) used ellagic acid/polyphenol-type reagents. Clot detection was mechanical for 17 laboratories (49%) and optical for 18 laboratories (51%).
FVIII one-stage clotting assays
The FVIII activities obtained with the one-stage clotting assays were calculated relative to the mathematical target. At the highest target concentration (0.8 IU mL−1 or U mL−1), the geometric mean recovery in vitro was 114.0%, 101.0% and 114.4% of expected target for ADVATE, ADYNOVATE and OBIZUR respectively. In the lower concentrations, a trend of overestimation was seen for all products. At 0.05 IU mL−1, recoveries for ADVATE, ADYNOVATE and OBIZUR were 138.3%, 124.3% and 125.9% respectively. On average across all target FVIII activity levels, recoveries were 127%, 113% and 120% for ADVATE, ADYNOVATE and OBIZUR respectively. Geometric mean recoveries for each product and target concentration are found in Table 2. Overall, laboratories reporting high results for ADVATE also reported high for the other products.
Table 2. One-stage clotting assay: Recovery of ADYNOVATE, OBIZUR and ADVATE in haemophilia A plasma. Intra- and inter-laboratory variance in FVIII measurements
Product (n = 35 laboratories)
Target FVIII activity (IU mL−1; U mL−1)
Absolute FVIII levels geometric means (IU/mL)
95% CI for geometric mean absolute levels
Geometric mean recovery (% of target)
95% CI for geometric mean recovery (% of target)
Intra-laboratory %CV
Inter-laboratory %CV
OBIZUR is labelled in Units (U), ADVATE and ADYNOVATE are labelled in International Units (IU).
ADVATE
0.8
0.912
0.876; 0.949
114.0
109.5; 118.6
6.8
10.9
0.2
0.260
0.247; 0.272
129.8
123.7; 136.1
8.8
13.0
0.05
0.069
0.065; 0.073
138.3
130.2; 146.9
12.9
16.1
ADYNOVATE
0.8
0.808
0.767; 0.852
101.0
95.9; 106.5
7.4
14.7
0.2
0.226
0.212; 0.241
112.9
106.0; 120.4
9.5
17.8
0.05
0.062
0.058; 0.066
124.3
116.5; 132.6
12.4
17.5
OBIZUR
0.8
0.915
0.865; 0.969
114.4
108.1; 121.1
8.9
15.8
0.2
0.239
0.225; 0.254
119.6
112.6; 127
12.3
16.1
0.05
0.063
0.058; 0.068
125.9
116.3; 136.3
15.4
21.6
Intra-laboratory variability (%CV, coefficients of variation) for ADVATE and ADYNOVATE was in the same range between 7% and 13%. OBIZUR had coefficients of variation between approximately 9% and 15%. For all products, the variability increased with decreasing FVIII activity. The inter-laboratory coefficient of variation was comparable for all three products as measured by one-stage clotting assay, ranging from 11% to 22% depending on the FVIII concentration (Figs 1 and 2, and Table 2). There was no difference in geometric mean FVIII activities obtained by either mechanical or optical clot detection methods.
In vitro recoveries of ADVATE, ADYNOVATE and OBIZUR FVIII activity in human FVIII-deficient plasma. Target concentrations were 0.80 IU FVIII mL−1 (a), 0.20 (b) and 0.05 IU FVIII mL−1 (c). Recovery per laboratory (circles, n = 35), mean and standard deviation (black lines) are shown.
In vitro recoveries of ADVATE, ADYNOVATE and OBIZUR FVIII activity in human FVIII-deficient plasma. Target concentrations were 0.80 IU FVIII mL−1 (a), 0.20 (b) and 0.05 IU FVIII mL−1 (c). Recovery per laboratory (circles, n = 11), mean and standard deviation (black lines) are shown.
The diversity of silica-based APTT reagents was higher than the other reagents. Participating laboratories reported the use of nine different brands of silica-based APTT reagents (APTT-SP HEMOSIL, Hemosil Synthasil, IL APTT-SS, STA APTT, STA PTTa, Stago STA- PTT A, STA-PTT A5, SynthASil, TriniClot aPTT S). As the brand names may differ by country, it was not possible to identify redundancies. However, we assume that at least 5–6 silica-based APTT reagents of different composition were used in the study. Two different sources of kaolin-based APTT reagents, CK Prest/STA CK Prest, and one in-house reagent were used. Two brands of an ellagic acid-based APTT reagents were reported (Actin FS and Actin FSL). One polyphenolic-type APTT reagents (Cephascreen) which belongs to the same group as ellagic acid-based APTT reagents was used in the study.
Overall, slightly higher numbers were found with ellagic acid/polyphenolic type APTT reagents compared to silica/kaolin reagents. The effect was similar across all three products (Table 3). Lowest values were found with silica-based APTT reagents from all brands, while results obtained with ellagic acid-based reagents were among those with the highest levels. While this seemed consistent for the majority of the laboratories, there were also some laboratories that presented the lowest values with ellagic acid-based APTT reagents while others were among the highest results with silica-based APTT reagents.
Table 3. Geometric mean recovery of FVIII activity by APTT reagent type
Product
FVIII concentration (IU mL−1)
APTT reagent type
n
FVIII activity (geometric mean) (IU[U] mL−1)
OBIZUR is labelled in Units (U), ADVATE and ADYNOVATE are labelled in International Units (IU).
ADVATE
0.8
Ellagic acid/polyphenol
10
1.0018
Silica/kaolin
25
0.8803
0.2
Ellagic acid/polyphenol
10
0.2877
Silica/kaolin
25
0.2498
0.05
Ellagic acid/polyphenol
10
0.0741
Silica/kaolin
25
0.0678
ADYNOVATE
0.8
Ellagic acid/polyphenol
10
0.9261
Silica/kaolin
25
0.7685
0.2
Ellagic acid/polyphenol
10
0.2643
Silica/kaolin
25
0.2130
0.05
Ellagic acid/polyphenol
10
0.0683
Silica/kaolin
25
0.0603
OBIZUR
0.8
Ellagic acid/polyphenol
10
0.9314
Silica/kaolin
25
0.9151
0.2
Ellagic acid/polyphenol
10
0.2429
Silica/kaolin
25
0.2399
0.05
Ellagic acid/polyphenol
10
0.0598
Silica/kaolin
25
0.0651
Focusing on ADYNOVATE and ADVATE, which contain the same human full-length FVIII molecule, for ADVATE at the highest concentration representing close to normal or desired initial FVIII activity levels upon replacement therapy were 108.3 ± 12.76, 113.6 ± 12.35 and 124.9 ± 9.36 (mean values +/− SD in percent of target values) when measured with silica, kaoline and ellagic acid-based APTT reagents respectively. The values for ADYNOVATE were 92.68 ± 9.67, 110.3 ± 13.89 and 114.7 ± 10.88 respectively. Statistical analysis revealed no significant differences between the different reagents for each of the two products.
FVIII chromogenic assays
Data from a chromogenic FVIII activity assay was reported from 11 laboratories. These laboratories reported the use of the following test kits: Siemens Chromogenic FVIII, Berichrom FVIII (Siemens), Coamatic Factor VIII, Coamactic FVIII, Siemens Sysmex FVIII, Electrachrome FVIII. Again, there was redundancy between test kits due to different brand names used in different geographies. We deduced that these six brands represent four different marketed chromogenic test systems. Across all concentrations, mean FVIII recovery was 116% (range 92–129%), 114% (range 95–124%) and 53% (range 44–61%) for ADVATE, ADYNOVATE and OBIZUR respectively (Table 4), relative to the theoretical or nominal target. Outliers in the group of the chromogenic test systems were related to specific laboratories rather than to a certain manufacturer of a test kit.
Table 4. Chromogenic assay: Recovery of ADYNOVATE, OBIZUR and ADVATE in haemophilia A plasma. Intra- and inter-laboratory variance in FVIII measurements
Product (n = 11 laboratories)
Target FVIII activity (IU mL−1; U mL−1)
Absolute FVIII levels geometric means (IU mL−1)
95% CI for geometric mean absolute levels
Geometric mean recovery (% of target)
95% CI for geometric mean recovery (% of target)
Intra-laboratory %CV
Inter-laboratory %CV
OBIZUR is labelled in Units (U), ADVATE and ADYNOVATE are labelled in International Units (IU).
ADVATE
0.8
1.032
0.965; 1.104
129.0
120.7; 138.0
7.7
9.0
0.2
0.255
0.232; 0.280
127.4
115.9; 140.0
4.8
13.9
0.05
0.046
0.037; 0.058
92.4
74.1; 115.3
15.1
32.5
ADYNOVATE
0.8
0.992
0.921; 1.068
124.0
115.2; 133.5
7.6
10.1
0.2
0.247
0.222; 0.276
123.7
111.0; 137.9
5.4
15.9
0.05
0.048
0.038; 0.06
95.4
75.8; 120.1
17.3
33.7
OBIZUR
0.8
0.487
0.443; 0536
60.9
55.3; 67.0
9.3
13.2
0.2
0.110
0.090; 0.134
54.9
45.2; 66.8
12.0
28.8
0.05
0.022
0.015; 0.031
43.5
30.1; 62.7
29.5
46.8
Intra-laboratory variation, while low, had a tendency to increase with the lowest FVIII concentration (Figs 1 and 2, and Table 4). Between laboratory, variation ranged from 9.0% to 33.7% for ADVATE and ADYNOVATE. Chromogenic assay measurements for OBIZUR demonstrated 46.8% inter-laboratory variation at the lowest concentration (0.05 U FVIII mL−1), likely due to several measurements at or below the lower limit of quantification of the assay used. Two out of the 11 laboratories were unable to measure FVIII activity of 0.05 U mL−1OBIZUR by chromogenic assay due to the low concentration.
Comparison of one-stage clotting and chromogenic assay results
Mean ratio of FVIII recovery results between the two methods is shown in Table 5. The mean ratio of the one-stage clotting assay results/chromogenic assay results at the highest FVIII target activity was 0.89, 0.82 and 1.85 for ADVATE, ADYNOVATE and OBIZUR respectively. Lower concentrations trended towards higher variability between the two assays. The variability between laboratories was high, with ratios ranging from 0.67 to 2.58 for ADVATE, from 0.57 to 2.76 for ADYNOVATE and from 1.04 to 4.41 for OBIZUR.
Table 5. Comparison of one-stage clotting and chromogenic assay results
Calculated from the mean of all one-stage clotting results (n = 35) and the mean of chromogenic data (n = 11).
b
Calculated only from laboratories reporting both one-stage clotting and chromogenic data.
OBIZUR is labelled in Units (U), ADVATE and ADYNOVATE are labelled in International Units (IU).
ADVATE
0.8
0.88
0.89
0.67–1.19
0.2
1.02
1.07
0.79–1.57
0.05
1.44
1.60
0.71–2.58
ADYNOVATE
0.8
0.82
0.82
0.65–1.29
0.2
0.92
0.96
0.65–1.45
0.05
1.25
1.39
0.57–2.76
OBIZUR
0.8
1.89
1.85
1.04–3.00
0.2
2.13
2.25
1.19–4.29
0.05
2.61
3.00
0.78–4.41
Discussion
Data from this international, collaborative study indicated that one-stage clotting assays can be used with sufficient accuracy and precision for FVIII measurement of ADYNOVATE, OBIZUR and ADVATE in plasma samples from subjects with haemophilia A, with a slight trend towards overestimation of small concentrations of FVIII. All participating laboratories reported using one-stage clotting assays. Also, chromogenic assays were used, never exclusively but as the second FVIII assay. Eleven of the 35 participating laboratories reported results obtained with a chromogenic assay demonstrating that this assay can also be used to accurately measure FVIII activity after treatment with ADVATE and ADYNOVATE. However, the chromogenic assay significantly underestimate levels of OBIZUR, particularly at low FVIII concentrations, due to lower potency response with this assay in this product. For this product, the APTT-based clotting assay apparently gave values closer to the target value with an overestimation at the lowest concentration tested. This could be partly due to the labelling method being one-stage clotting assay and the discrepancy between one-stage clotting and chromogenic assays emphasised the difference. Chromogenic assays also exhibited higher inter-laboratory variability compared with the one-stage clotting assays, which is exacerbated at low FVIII concentrations. This is a surprising result as chromogenic FVIII assays have conventionally been assumed to be more sensitive and have greater precision particularly in the low concentration range [24, 31-33]. One explanation of chromogenic assays appearing to be less precise in this study may be because the chromogenic assays were not the laboratories’ routine assay and the imprecision could be due to inexperience. However, laboratories were asked to only use their standard FVIII assay available for daily routine. Another explanation is the lack of dilutional linearity due to poor comparison of the test sample response with the low end of the standard curve. For example, the responses for the low concentration samples may not have been compared with a ‘low concentration range’ standard curve that exhibits the same plasma protein concentration as the test sample. Such testing routine, although more cumbersome, would be recommendable to measure low FVIII concentrations.
Intra- and inter-laboratory variability for the one-stage clotting assay in measuring ADVATE were found to concur with published data [29, 30, 34]. Both ADYNOVATE and OBIZUR demonstrated similar intra- and inter-laboratory variability to ADVATE using the one-stage clotting assay, indicating that assay conditions and reagents had no obvious impact on the results. There was also no impact from the type of instruments and the method of clot detection on assay variability. Assay reagents had a minor effect on FVIII activity variability, with a trend towards slightly higher results when ellagic acid reagents were used. However, this observation was similar for both ADYNOVATE and ADVATE, indicating that both products can be equally well measured with all one-stage clotting reagents. The collaborative study demonstrated that the variability originated mainly from individual laboratories rather than from the type of assay or any specific reagent composition. A general limitation of a study like this is that samples from patients treated with either of the FVIII products could not be used as these would be never available from haemophilia A patients in the amounts needed. In addition, clinical trial restrictions would not allow shipping such patient samples to laboratories widespread over the world. We therefore cannot completely exclude that samples from patients treated with FVIII products might behave differently than ex vivo spiked plasma from a haemophilic patient. However, the study tried to mimic clinical samples as closely as possible.
The field study was preceded by a survey with responses from 56 laboratories, of which 98% noted that the one-stage assay was their preferred system. The majority of laboratories utilising one-stage assay (58%) used silica-based APTT reagents. Previous studies have shown that some of the new FVIII products may not be measured adequately with the one-stage clotting assay, particularly when silica-based APTT reagents are used. Thus, the survey indicates that there is a high risk for a laboratory to obtain an irregular test result. The survey also showed that the one-stage assay currently is the preferred FVIII assay used in clinical practice. This confirmed previously published opinions and surveys [35, 36]. The reluctance to adopt the use of chromogenic assays among the laboratory community participating in the presented study may be attributable to lack of familiarity and perceived higher associated costs compared to one-stage clotting assays [24].
Overall, the data from the collaborative field study show that one-stage clotting assays and chromogenic assays can determine FVIII activity measurement of ADYNOVATE and ADVATE in human haemophilia A plasma, independent of assay set-up, instruments, type of clot detection, source of reagents and standards. Another international comparative laboratory field study with similar design just performed with another full-length rFVIII, which was not modified to extend the half-life, showed comparable results [37]. This indicates that FVIII molecules with intact B-domain irrespective of chemical modification behave as plasma-derived FVIII. Similarly, OBIZUR may also be measured with a variety of one-stage clotting assays with different instruments, detection methods, reagents and standards that are currently in routine use. Some assay variability may always be observed and therefore, ideally, any specific assay should be validated with samples containing the respective FVIII molecule or with plasma from patients treated with a certain FVIII product. The lower FVIII activity for OBIZUR when using chromogenic assay needs to be taken into account when using this assay for treatment monitoring of patients and is in agreement with the description in the prescribing information stating that lower chromogenic potencies are detected with chromogenic assay compared to the one-stage clotting assay. Monitoring of OBIZUR by chromogenic assays is thus not preferred. The results of this study indicate that there is no need for product-specific standards, as deemed necessary for other products [33], due to similar accuracy and assay variation for all new products in comparison with the well-established product, ADVATE.
Acknowledgements
The authors are grateful to the individuals in the participating laboratories enabling performance of the analyses:
Jovan Antovic, Peter Baker, Valdas Banys, Jerome Beltran, Sarah Berberich, Maria Berndtsson, Sarah Bruty, Roger Buchanan, Meera B. Chitlur, Marina Dzhelali, Marion Echenagucia, Audrone Eidukaite, Mary Elaine Eyster, Fabienne Floc'h, Kenneth Friedman, Sara Gay, Laurine Genesta, Paul Giangrande, Lucy Goff, Marc Grimaux, Ralph Gruppo, Sandra Haberichter, Catherine P. M. Hayward, Lourdes Herrera, Tristan Herve, Margareta Holmström, Shawn Jobe, Rashid Kazmi, Richard Ko, Barbara Konkle, John Lazarchick, Petra Linden, Steven Lobel, Anne Lochu, Maria Fernanda López Fernandez, Jonathan Lowe, Paige A. Macy, Janine Martin, Karen Moffat, Jens Müller, Sukesh C. Nair, Johannes Oldenburg, Amparo Santamaria Ortiz, Rafael Parra, Julia Phillips, Steven Pipe, Lina Rageliene, Heesun Rogers, Arlette Ruiz de Sàez, Wade Sandau, Jeff Sanders, Bernard Silver, Alok Srivastava, Andrew Sutherland, Fikre Tadesse, Desiree Tan-Castillo, Michael Tarantino, Gayle Teramura, Stefan Tiefenbacher, Laila Vengal, Gilbert C. White, Anne M. Winkler, Guy Young.
The authors thank Jocelyn Hybiske, PhD, an independent consultant, for providing medical writing services, which was funded by Baxalta, Inc., and Sabine Mandl for editorial assistance.
Author contributions
All authors reviewed and edited the manuscript.
Disclosures
This study was funded by Baxalta, Inc. Peter L. Turecek, Stefan Romeder-Finger, Claudia Apostol, Alexander Bauer and Herbert Gritsch are employees of the sponsor, Baxalta Innovations, GmbH. Alex Crocker-Buque is an employee of Q2 Solutions, a Quintiles Quest Joint Venture, working as a CRO for Baxalta Innovations, GmbH. Divan A. Burger and Robert Schall are employees of Quintiles Biostatistics.
Antihemophilic Factor (Recombinant), PEGylated (Rurioctocog Alfa Pegol) was first approved by the U.S. Food and Drug Administration (FDA) on November 13, 2015, then approved by Pharmaceuticals and Medicals Devices Agency of Japan (PMDA) on April 4, 2016. It was developed and marketed as Adynovate® by Baxalta.
Adynovate® is a PEGylated recombinant full-length human coagulation factor VIII, which replaces the missing clotting factor, thereby enabling a temporary correction of the factor deficiency and correction of the bleeding tendencies. It is indicated in adolescent and adult patients (12 years and older) with hemophilia A.
Adynovate® is available as injection for intravenous use, containing 250 IU, 500 IU, 1000 IU or 2000 IU lyophilized powder in single-use vials. The dose and frequency depend on whether it is used to treat or prevent bleeding.
Intravenous susoctocog alfa is a recombinant, B-domain deleted, porcine sequence antihaemophilic factor VIII (FVIII) product that has recently been approved for the treatment of bleeding episodes in adults with acquired haemophilia A (AHA).
Burness CB, Scott LJ: Susoctocog Alfa: A Review in Acquired Haemophilia A. Drugs. 2016 May;76(7):815-21. doi: 10.1007/s40265-016-0576-1. [PubMed:27098420 ]