THROMBOSIS RESEARCH 65; 699-708,1992 0049-3848/92 $5.00 + .OO Printed in the USA. Copyright (c) 1992 Pergamon Press Ltd. All rights reserved.

PHARMACOKINETICS OF TWO PASTEURIZED DIFFERENT AND MULTICENTER ASSAYS

FACTOR VIII CONCENTRATES OF FACTOR VIII ACTIVITY

BY

A.

Messori', K. Schimpf*,

M. Morfini' S. C'nottil, 6. ongol, M. Blomback3, U. Delvos 5 and K. Schumacher*, A. N vakova-Banet 6 , H. Kjellman s , From the 'Hematolog & Hemophilia Center, USL 10/D, Florence, I aly, the Pharmaceutical Service, USL 10/D, Florence, Italy, the 5 Depart. sofDepart. Clinical Chemistry & Blo d Coagulation, the a Rehabilitation Karolinska Hospital, Stockholm, Sweden, Hospital and Hemophi ia Center Heidelberg, Rehabilitation Foundation, FRG, and the 5 Research Laboratories of Behringwerke AG, Marburg, FRG. (Received

27.9.1991;

accepted

in revised form 8.1.1992

by Editor K. Lechner)

ABSTRACT

We assessed the pharmacokinetic characteristics of a new high-purity pasteurized FVIII concentrate in comparison with an intermediate purity pasteurized concentrate, produced by the same manufacturer. The study was designed as a cross-over single-dose pharmacokinetic investigation in 8 non-bleeding patients with severe hemophilia A. All patients were given 25 IU/kg of each of the two concentrates, with an interval of at least one week between the two administrations. Decay curves were assessed by collecting 10 serial blood samples over 36 hours following the end of infusion. The concentration of Factor VIII in blood samples was determined in triplicate in three different laboratories using each of the following assay methods: a one-stage clotting assay, a two-stage clotting assay, and a two-stage chromogenic-peptide substrate assay. All pharmacokinetic parameters were calculated by model-independent methods. The two products were found to differ significantly both in the clearance, which was on average 13.8% lower for Haemate P, and in the in-vivo recovery, which was 11.7% lower for Factor VIII:C P on the average. In comparison with previous pharmacokinetic data obtained from other heated Factor VIII concentrates, the clearance of Haemate P was found to be significantly slower, while the half-life of both products was longer. No differences were observed in the Vd-area. These findings indicate that the purification procedures to which both products are subjected do not increase the in-vivo rate of plasma disappearance of Factor VIII. Key words: Factor VIII, Pharmacokinetics, 699

Pasteurization

700

FACTOR VIII CONCENTRATES

Vol. 65, No. 6

INTRODUCTION The present investigation was undertaken to compare the pharmacokinetics of a new high-purity pasteurized Factor VIII concentrate (Factor VIII:C P or Beriate P, Behringwerke Marburg, FRG) with that of an intermediate purity pasteurized product (Haemate P), previously produced by the same manufacturer. The study was designed as a cross-over single-dose pharmacokinetic investigation in non-bleeding hemophilia A patients. All the patients were enrolled from a single institution, and the clinical part of the study was therefore conducted only in this center. Using the same clinical material, another study [the results of which are reported elsewhere (l)] was carried out to assess the reproducibility of three Factor VIII assays (one-stage clotting, two-stage clotting, and two stage chromogenic methods) across three different laboratories (Florence, Heidelberg, and Stockholm) which agreed to use the same laboratory techniques. This inter-laboratory investigation showed that both the onestage clotting and the chromogenic methods for Factor VIII assay were well reproducible between the three centers whereas the reproducibility of the two-stage clotting assay was poor (1). The pharmacokinetic evaluation of the two products, presented herein, reflects the fact that each sample taken from the patients was assayed three times in each of the three laboratories. Hence, the availability of this large body of experimental material has allowed to expand the purposes of the present study beyond a simple product evaluation and to carry out an original multivariate analysis to assess the reproducibility of the pharmacokinetic results of Factor VIII concentrates [when using the same samples and the same laboratory methodology (1) in different centers]. All pharmacokinetic calculations were performed according to the 'so-called' model-independent approach (2,3), which has previously been shown to be advantageous to standardize the methodology for evaluating the pharmacokinetics of Factor VIII/IX concentrates. PATIENTS AND METHODS Patients Eight adult hemophilia-A patients with Factor VIII less than 10 IU/L were included in the study. The patients' age ranged from 14 to 48 years, their body weight from 49 to 75 kg. None of the All patients were patients had antibodies against Factor VIII. selected from a group of subjects who had previously been treated The patients received no with other Factor VIII concentrates. Factor VIII administration during the two weeks before this study. The patient enrollment and the kinetic study were completed within 9 months. VIII concentrates Factor a Factor VIII Haemate P (Behringwerke, Marburg, FRG) is by various salt precipitation concentrate which is prepared steps followed by pasteurization (virus-inactivation by wet-heat treatment at 60°C for 10 h in aqueous solution) (4-6). Factor VIII:C P or Beriate P (Behringwerke, Marburg, ","G,h,', high-purity Factor VIII concentrate, obtained by a new matographic procedure after cryoprecipitate has been pasteurized. Both preparations are stabilized with albumin before lyophiliza-

Vol. 65, No. 6

FACTOR VIII CONCENTRATES

701

tion. Their specific activity before addition of albumin ranges, IU/mg protein (Behring Laboratories, for Haemate P, from 26-66 from 152-161 unpublished observation) and, for Factor VIII:C P, IU/mg protein (7); in the final formulation it ranges from 2.7 to 3.5 IU/mg of total protein (5). of concentrates and collection of post-infusion Administration blood samples All patients were given a single dose of each of the two concentrates, with an interval of at least one week between the two administrations. A cross-over Latin square design was adopted: thus, four patients were first infused with Haemate P and then with Factor VIII:C P, while the remaining four received the two products in the opposite sequence. The dose was 25 IU/kg b.w. for both Haemate P and Factor VI_II:C P and was calculated from the Factor VIII potency declared on the label. The exact dose actually infused was determined from the amount of injected product (weight of syringe before and after infusion) taking into account that the mean potency of both Factor VIII concentrates was 33 kU/L (range 32 to 34 kU/L) according to the Quality Control Department of Behringwerke. The infusion of either product was performed over 10 minutes (range 7 to 13 minutes) by one and the same nurse. In each patient, blood samples were drawn before infusion and at 0.08, 8, 12, 24, 28 and 36 h after the end of the 0.25, 0.5, 1, 4, infusion. To obtain each plasma sample, blood (11 ml) was collected from a forearm vein: 2 ml were added to a vacutainer tube with EDTA for hematocrit determination by S Plus Coulter Counter; 3.8% 9 ml were added to a tube with 1 ml of trisodium-citrate (w/v) and centrifuged at 4OC and 1200 g for 20 min; plasma aliquots were then transferred into 12 plastic tubes appropriately coded (tubes Al to A6, Bl to B3, and Cl to C3 containing each 0.4 ml of plasma) and immediately frozen at -7OOC. Tubes coded Al to A3 were assayed at the Florence laboratory within 14 days of the collection; this part of the study [reported in detail elsewhere (l)] demonstrated that no loss in Factor VIII activity occurred during the storage mentioned below. No pharmacokinetic evaluation was performed using tubes Al to A3. Tubes coded A4 to A6, Bl to B3, and Cl to C3 were stored in frozen state at the Florence Hemophilia Center. Samples were then dispatched in dry ice to the Heidelberg (Bl to B3) and the Stockholm (Cl to C3) laboratories, where samples were again stored at -70OC. At these two laboratories, the Factor VIII determinations were performed 1 to 6 months (Heidelberg) or 2 to 8 months (Stockholm) after the collection of blood samples. In Florence, too, an assay was again performed (using tubes coded A4 to A6) 3 to 11 months after the original collection of samples. In the statistical evaluations presented below, the results of the assays of tubes coded A4 to A6, Bl to B3, and Cl to C3 were compared with each other despite the fact that the length of the storage at -7OOC for individual tubes could range from 1 to 11 months. Calculation of pharmacokinetic parameters from factor VIII decay curves In the model-independent (noncompartmental) approach (2,3), three pharmacokinetic parameters are used: 1. The total body clearance (CL), which expresses the overall rate at which the plasma is cleared from the substance: 2. The volume of distribu-

702

FACTOR VIII CONCENTRATES

Vol. 65, No. 6

tion ('area' method, Vd-area), which is the ratio between the amount of substance that is present in the body at any time and the corresponding concentration in the plasma; 3. The mean residence time (MRT), the practical function of which is similar to that of the well-known half-life (t1/2). Estimation of CL, Vdarea, and MRT from single-dose experimental data is based on the calculation of the area under the curve (AUC) and the area under the moment curve (AUMC). The assessment of the monophasic or biphasic shape of individual decay curves was performed by compartmental methods, as previously described (3). Data analysis and statistical evaluations The comparison between the three assay methods performed in the three laboratories using this large series of samples demonstrated a good agreement between one-stage and chromogenic assays although the correlation of the respective Factor VIII activities was not perfectly linear (1). In contrast, the twostage method was shown to be poorly reproducible in our hands (1). In the present work, a preliminary statistical evaluation (not reported in detail) was therefore conducted to assess whether the pharmacokinetic parameters CL, MRT, and Vd-area were influenced by the assay method. Since this statistical evaluation indicated the presence of a highly significant effect of the assay method on the pharmacokinetic parameters, all the subsequent statistical evaluations were performed separately for each of the assay methods to better address the two following issues: 1. The presence of any pharmacokinetic differences between the two products within a single assay method; 2. The reproducibility of the products' pharmacokinetic parameters between the three centers and within a single assay method. In this way, it was easier to distinguish these two problems from the overwhelming statistical effect of the differences between the assay methods. For each of the three assay methods, a multivariate statistics was separately conducted to assess the two problems mentioned above. All calculations were performed using the MANOVA subroutine of the SPSS-PC Plus statistical package (SPSS Inc., Chicago, Illinois). In the Results section, the three laboratories participating in the multi-center assay of the denoted as Centers A, B, and C, while the samples are identity of the individual laboratories is kept confidential. Factor VIII assay Plasma samples were assayed at each of the three laboratories by using: (i) a one-stage clotting assay: (ii) a two-stage clotsubstrate ting assay; and (iii) a two-stage chromogenic-peptide assay. Irrespective of the method used, the 11 samples per patient were always assayed on the same day. The three assays were performed as previously described (1). Details on the standards used for each assay are given in the previous paper (1). RESULTS Regardless of the assay method, the shape of the 144 disappearance curves of the two concentrates was nearly equally distributed between biphasic and monophasic decays. Thirty-two (44.4%) and 40 (55.5%) biphasic; for curves of Haemate P were monophasic Factor VIII:C P 40 curves (55.5%) were monophasic and the remaining 32 (44.4%) biphasic. This finding confirms the appropriate-

Vol. 6.5, No. 6

FACTOR VIII CONCENTRATES

703

ness of our choice of a model-independent pharmacokinetic analyand sis, which is known to be equally suitable for monophasic biphasic curves (2). Pharmacokinetic assessment using one-stage clotting assay The two products (Table 1) were found to differ significantly both in the CL (Haemate P = 2.82 ml/h/kg; Factor VIII:C P = 3.27 ml/h/kg, p=O.O12), which was on average 13.8% lower for Haemate P, and in the in-vivo recovery (Haemate P: 110.3%; Factor VIII:C P: 97.4%, p=O.O04), which was 11.7% lower for Factor VIII:C P on the average. The three pharmacokinetic parameters (CL, MRT, and Vd-area) remained virtually unchanged between the three centers. TABLE 1 Pharmacokinetic assessment using one-stage centers A, B AND C. Mean+SD; N=8.* CLEARANCE (ml/h/kg)

MRT-(h)

assay,

Vd-Area (ml/kg)

in the three

IN-VIVO RECOVERY(%)

Hamate P PVIII:C P Haemate P WIII:C P

Haemate

A B

2.7520.98 2.87+1.89

3.3521.59 3.4222.59

21.727.9

C

2.84~1.55

3.0321.54

Haemate

P

FvIII:C

P

PVIII:C P 91.bL26.6 98.bL25.9

27.bz22.7

57.4zl4.9 63.4228.6

63.5220.2

104.3+24.6 107.7~24.4

23.128.2

21.5~12.2 -

60.6+20.3

54.1z17.6

118.9231.3 101.9z25.3

P-0.86'"

64.9z25.7

P

2O.bz7.8 23.2+10.9

p-0.47**

P-0.17+*

p=o.o3**

Pairwise comparisons between products were not significant for both MRT and Vd-area. Significance was found for CL (P-0.012, see text) and in-viva recovery (P-0.004. see text). ** These p-values refer to differences between centers irrespective of differences between ??

products. As regards in-viva recovery, pairwise comparisons showed significant differences between Centers A and C. and between Centers B and C, not between Centers A and B.

In contrast, a significant difference in the in-vivo recovery was found depending on the centers: Center C, in fact, provided consistently the highest recovery estimates while Center A the lowest ones. Phannacokinetic assessment using two-stage clotting assay The CL did not change between the centers (Table 2), while the MRT differed significantly.and the Vd-area was close to significance. In contrast, the two products differed in no pharmacokinetic parameter. Since the previous laboratory study (1) Pharmacokinetic

TABLE 2 assessment using two-stage assay. MeantSD, N=8.*

CLEARANCE (ml/h/kg)

A B C

Haemate P

PVIII:C P

1.9320.50

2.1421.03 2.6721.51 2.5051.20

2.2620.96

2.03~0.83

MRT Haemate P 25.7~5.4 24.4z9.0 18.9+ 7.1

(h)

Vd-Area (ml/kg)

IN-VIVO RECOVERY (%)

FVIII:C P

Haemate P

PVIII:C P

Haemate P

WIII:C

28.3~10.8 24.927.1 19.3+11.1

48.7~13.2 53.1Llb.2 36.9Llb.5

51.729.2 59.7,lg.E 41.7512.5

130.4z37.0 115.0~31.7 154.5+70.0

113.8L26.8 107.2z36.2 157.1~82.7

p-0.21** P=O.O04"f P=O.b2** p=o.o7** * All pairwise comparisons between products were not significant. ** These p-values refer to differences between centers irrespective of differences between products. Post-hoc pairwise comparisons showed significant differences in HRT between Centers A and C. and between Centers B and C. not between Centers A and B. Significant differences between Centers B and C were also found for Vd-area and in-viva recovery.

P

FACTOR VIII CONCENTRATES

704

Vol. 65, No. 6

showed a poor reproducibility of the two-stage assay and an overestimating tendency of Center C using this assay, our results reflect essentially these previous findings. Pharmacokinetic assessment using two-stage chromogenic assay Most of the pharmacokinetic parameters of the two products (namely CL, Vd-area, in-vivo recovery) differed depending on the centers (Table 3): for example, Center C provided significantly lower estimates of CL and of Vd-area and higher values of in-vivo recovery. The results of Centers A and B were however more homogeneous. The two products were found to differ significantly in the CL (p=O.O16), which was on average 7.8% lower for Haemate P compared with Factor VIII:C P. This result is similar to the one obtained from the pharmacokinetic assessment using one-stage assay. Pharmacokinetic N=8."

assessment

CLEABANCE (ml/h/kg)

A B C

Baamate P 2.8220.95 2.83zO.91 2.48*0.81

FVIII:C P 3.18+1.08 3.06zl.19 2.60~0.73

MBT Haemate P 17.5z7.7 17.024.5 16.026.2

TABLE 3 using chromogenic

(h) FVIII:C P 16.324.8 17.824.8 15.724.3

Vd-Area Haemate P 46.1+8.9 47.1+6.3 36.4~4.0

(ml/kg)

FVIII:C P 47.9z7.9 56.7z25.9 38.327.1

assay.

Mean+SD,

IN-VIVO RECOVERY (a) Heemate P 111.2~15.4 106.8+12.8 132.5+21.6

FVI11:C P 98.3+10.4 98.02X.6 133.5235.1

P-0.01"'

P-0.17" P-0.02** P-0.001'" comparisons between products were not significant with the exception of CL (see text. P.O.016). ** These p-values refer to differences between centers. Pairwise comparisons showed significant differences in each of the parameters CL, Vd-area, and in-viva recovery between Centers A and C and between Centers B and C, not between Centers A and B. A significant difference was also found for MBT between Centers B and C. *

All

pairwise

Comparison between the pharmacokinetic parameters obtained using the one-stage clotting and the two-stage chromogenic assay Our previous laboratory study (1) revealed that the ratios between the one-stage and the chromogenic assays were not homogeFactor VIII activineously distributed at different values of less than concentrations of Factor VIII were When the ty. IU/L, the ratio one-stage/chromoto 200 Or equal genic averaged 1.22 indicating that the one-stage assay tends to provided by the give higher Factor VIII activity than that when Factor chromogenic method (or viceversa). On the contrary, VIII concentrations were above 200 IU/L, the mean ratio was 0.88 suggesting that, at these activity values, the one-stage assay VIII concentrations gives a slightly lower estimate of Factor than the chromogenic method (or vi&versa). This fact had profound consequences on the estimation of most pharmacokinetic parameters. In particular, the MRT of both products was much longer when estimated with the one-stage assay (Haemate P = 24.1 h; Factor vIII:C P = 21.8 h) than with the chromogenic one (Haemate P = 16.8 h: Factor VIII:C P = 16.6 h). The effects on CL were less marked (CL estimated by One-Stage method: Haemate P = 2.82 ml/h/kg; Factor VIII:C P = 3.27 ml/h/kg: CL estimated by chromogenic method: Haemate P = 2.71 ml/h/kg:

FACTOR VIII CONCENTRATES

Vol. 65, No. 6

705

Factor VIII:C P = 2.95 ml/h/kg) and also the differences in the in-vivo recovery were slight (recovery estimated by one-stage method: Haemate P = 110.3%; Factor VIII:C P = 97.4%;.reCOVery estimated by chromogenic method: Haemate P = 116.8%; Factor VIII:C P = 109.9%). In contrast, the Vd-area underwent wide variations related to the assay method (Vd-area estimated by one-stage method: Haemate P = 60.5 ml/kg; Factor VIII:C P = 60.8 ml/kg; Vd-area estimated by chromogenic method: Haemate P = 43.2 in the ml/kg; Factor VIII:C P = 47.6 ml/kg). These variations Vd-area can be explained in pharmacokinetic terms because the volume of distribution is calculated from the equation: Vd-area = identical equations) in CL x MRT (or from other, mathematically which Vd-area and MRT are directly proportional: hence, a prolongation in the MRT, which possibly is not paralleled by a decrease in the CL, can result in an increase in the Vd-area. Comparison with traditional concentrates Table 4 shows a comparison of the two products studied in the present investigation with a large reference group reported in a previous study (including 1 unheated Factor VIII and 3 heattreated Factor VIII concentrates ). The clearance of Haemate P was 27% lower in comparison with the reference group, while the half-life of both products was longer. No differences were observed in the Vd-area. The comparison was limited to the kinetic findings of one-stage Factor VIII:C assay because the previous population data are based only on this method (8). Comparison

TABLE 4 previous pharmacokinetic one-stage assay.

with

CLEARANCE (ml/h/kg)

HALF-LIFE (h)

data

based on

Vd-AREA (ml/kg)

Haemate P (N=24): 2.82+1.45

16.7+9.8 -

60.5221.2

Beriate P (N=24): 3.2721.89

15.127.0

60.8221.0

Historical data* (N=87)

11.0+4.9 -

58.2221.3

:

3.8521.94

* (From ref.8) DISCUSSION The present study has provided original information on two issues: 1. The pharmacokinetic characteristics of Haemate P and Factor VIII:C P; 2. The reproducibility of the pharmacokinetic results in investigations of Factor VIII concentrates. These two issues are discussed separately. The two Factor VIII concentrates evaluated herein are characterized by relatively low values of CL (Haemate P: 2.82 3.27 ml/h/kg using one-stage data). ml/h/kg; Factor VIII:C P: This finding indicates at least that the purification procedures to which both products are subjected do not increase the rate of plasma disappearance of Factor VIII in-vivo. Interestingly

706

FACTORVWCONCENTRATES

Vol. 65,No.6

enough, a comparison with historical data shows that Haemate P has a lower clearance (and a longer half-life) than other Factor VIII products evaluated in previous studies (7). The volume of distribution of Haemate P and of Factor VIII:C P is in the range of the values previously reported for other products. This result is a proof of the reliability of the labeled potency of the two concentrates; additionally, it witnesses the usefulness of Vdarea as an estimate of the in-vivo potency of Factor VIII concentrates when using the one-stage assay. Small but significant differences were observed in some pharmacokinetic characteristics of Haemate P and Factor VIII:C P: in particular, the former had a lower clearance (as measured by one-stage and chromogenic methods) and a higher in-vivo recovery (as measured by one-stage) than the latter. As far as the Factor VIII specific activity of both products is concerned, it is quite evident from other reports (6) that Haemate P and Beriate P do not substantially differ from each other. The most evident difference is the low Content of Ristocetin cofactor activity (RiCof) in Factor VIII:C P (722 IU/ml) compared to Haemate P (95230 IU/ml) (7). This qualitative difference does not seem to affect the half-life of Factor VIII:C, even though Haemate P showed a trend toward a better behavior. The second issue addressed by the present study is whether, in general, the pharmacokinetic investigations on Factor VIII concentrates are reproducible when assays are performed in different centers and using different assay methods (e.g. one-stage, twostage, and chromogenic methods). In the first place, the combined results of the present investigation and of the study reported previously (1) reveal a poor performance of the two-stage assay in terms of reproducibility between different centers and, consequently, a poor reliability of the pharmacokinetic parameters calculated using this method. For example, the in-vivo recovery shows a definitely larger variability between centers when estimated using the two-stage clotting method (mean in-vivo recovery ranging from 107.2% to 157.1% for Factor VIII:C P and from 115.0% to 154.5% for Haemate P; variations between the highest and the lowest mean: 1.46 for Factor VIII:C P, 1.34 for Haemate P) in comparison with the one-stage method (mean in-vivo recovery ranging from 91.6% to 101.9% for Factor VIII:C P and from 104.3% to 118.9% for Haemate P; variations between the highest and the lowest mean: 1.11 for Factor VIII:C P, 1.14 for Haemate P) and the chromogenic method (mean in-vivo recovery ranging from 98.3% to 133.5% for Factor VIII:C P and from 106.8% to 132.5% for Haemate P; variations between the highest and the lowest mean: 1.36 for Factor VIII:C P, 1.24 for Haemate P). Regrettably, some centers were not accustomed to doing two-stage assays routinely and this fact probably contributed to the poor performance of this assay. On the other hand, if one excludes the results of the two-stage of the pharmacokinetic parameters assay, the reproducibility calculated using the other two assay methods appears to be satisfactory provided that the comparisons are made between Centers and within a single assay method (not between the one-stage and the chromogenic assays, which is an issue that is specifically discussed below). Although significant differences between centers were found in the estimates of the in-vivo recovery (using one-stage assay) and of CL, Vd-area, and in-vivo recovery (using

Vol. 65, No. 6

FACTOR VIII CONCENTRATES

707

chromogenic assay), the magnitude of these differences was always less than 15% for CL, which is the basic pharmacokinetic parameCL in ter. This fact emphasizes the importance of determining pharmacokinetic studies of Factor VIII. The differences between centers in Vd-area and in its more traditional counterpart (invivo recovery) were of greater magnitude and deserve therefore The results obtained using the one-stage specific comments. method (Table 1) indicate that, while the in-vivo recovery differed significantly between the centers, the variations in the Vd-area were far from statistical significance. This finding confirms that the Vd-area is more stable (within a single assay method) than the in-vivo recovery. As regards the chromogenic the pharmacokinetic estimates provided by Cenassay (Table 3), ters A and B were rather homogeneous, whereas Center C showed a consistent trend towards an underestimation of clearance and Vdarea and an overestimation of in-vivo recovery. The latter terms, the finding outlined result reflects, in pharmacokinetic in our previous report (1) that Center C showed a slight though systematic bias towards an overestimation of the Factor VIII activity [the slope of the regression line of activity (y) measured in Center C versus activity (x) measured in Center A was: 1.14 (significantly different from 1, p

Pharmacokinetics of two pasteurized factor VIII concentrates by different and multicenter assays of factor VIII activity.

We assessed the pharmacokinetic characteristics of a new high-purity pasteurized FVIII concentrate in comparison with an intermediate purity pasteuriz...
860KB Sizes 0 Downloads 0 Views