Collaborative study for the calibration of a replacement international standard for diphtheria toxoid adsorbed

Biologicals 38 (2010) 529e538

Contents lists available at ScienceDirect

Biologicals journal homepage: www.elsevier.com/locate/biologicals

Collaborative study for the calibration of a replacement international standard for diphtheria toxoid adsorbed Paul Stickings a, *, Peter Rigsby b, Laura Coombes a, Kiran Malik c, Paul Matejtschuk c, Dorothea Sesardic a a

Division of Bacteriology, National Institute for Biological Standards and Control, Health Protection Agency, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, United Kingdom b TDI Biostatistics, National Institute for Biological Standards and Control, Health Protection Agency, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, United Kingdom c TDI Standardisation Science, National Institute for Biological Standards and Control, Health Protection Agency, Blanche Lane, South Mimms, Potters Bar, Hertfordshire EN6 3QG, United Kingdom

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 February 2010 Accepted 7 April 2010

We present the results of a collaborative study for the characterization of a preparation of diphtheria toxoid adsorbed, and its calibration in terms of the 3rd International Standard (IS) for Diphtheria Toxoid Adsorbed. Calibration was performed using established World Health Organization (WHO) and European Pharmacopoeia (Ph. Eur.) protection models. Two candidate toxoid preparations were included in the study, one of which was adopted as a replacement Ph. Eur. Biological Reference Preparation (BRP, batch 4) in February 2009. The second candidate preparation was found to have a unitage of 213 IU/ampoule based on the calibration by in vivo bioassay in 19 laboratories in 16 countries, and was established as the 4th IS for Diphtheria Toxoid Adsorbed by the WHO Expert Committee on Biological Standardization (ECBS) in October 2009. The study also assessed performance of the replacement standard in mouse and guinea pig serological assays which are used as alternative procedures for diphtheria potency testing. Participants tested both candidate preparations and potency was expressed in relative terms only. Results suggest that the replacement standard is suitable for use as the reference vaccine in serological assays and that the Vero cell assay may be suitable for calibration of future replacement standards. Ó 2010 The International Association for Biologicals. Published by Elsevier Ltd. All rights reserved.

Keywords: Diphtheria Vaccine Standardization Potency Collaborative study

1. Introduction Diphtheria is an acute, often fatal bacterial disease caused by toxigenic strains of Corynebacteria. Diphtheria toxin (DTx) is the major virulence factor of these organisms and the clinical manifestations of the disease are due mainly to the presence of circulating toxin in the bloodstream of infected individuals [1]. Active immunization against diphtheria is based on the use of a formaldehyde-detoxified preparation of DTx, diphtheria toxoid (DTxd) to induce protective antibody responses. The introduction of routine immunization programs against diphtheria in the 1940s and 1950s led to almost complete eradication of this disease from many countries by the 1980s. However, diphtheria is still endemic in certain parts of the world and several outbreaks have occurred, the most serious of which occurred in Russia and the Newly

* Corresponding author. Tel.: þ44 1707 641 447; fax: þ44 1707 641 054. E-mail addresses: [email protected] (P. Stickings), peter.rigsby@nibsc. hpa.org.uk (P. Rigsby), [email protected] (L. Coombes), kiran.malik@ nibsc.hpa.org.uk (K. Malik), [email protected] (P. Matejtschuk), [email protected] (D. Sesardic).

Independent States (NIS) of the former Soviet Union where the World Health Organization (WHO) registered 47 000 cases of disease with more than 1700 fatalities [2]. It is clear that disease can re-emerge in previously low-prevalence countries under particular conditions e for example gaps in childhood vaccination coverage combined with waning immunity in adult populations. A number of population immunity studies has identified sizeable proportions of adults with diphtheria immunity levels below the putative protective threshold, even in countries with good childhood immunization coverage [3e5], and in most countries it is recommended that adults should receive booster vaccinations [6]. Good childhood vaccination coverage and appropriate booster immunization of adults is essential to maintain protection against diphtheria in the population, and the supply of effective vaccines is dependent on confirmation of vaccine potency. The WHO and European Pharmacopoeia (Ph. Eur.) impose recommendations to confirm efficacy of every new batch of diphtheria vaccine manufactured [7,8]. The specifications for these vaccines are therefore dependent on the use of the International Standard (IS) or material calibrated against it, with vaccine potency expressed in IU. The 3rd WHO IS for Diphtheria Toxoid Adsorbed (coded 98/560) was

1045-1056/$36.00 Ó 2010 The International Association for Biologicals. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.biologicals.2010.04.001

530

P. Stickings et al. / Biologicals 38 (2010) 529e538

established by the WHO Expert Committee on Biological Standardization (ECBS) in 1999 [9]. This material, with an assigned unitage of 160 IU/ampoule was used extensively as a working reference preparation and for calibration of in-house or secondary reference preparations according to the intention of the WHO and Ph. Eur. Commission. The demand for this standard preparation increased significantly in the last decade as a result of the introduction of new combination vaccines containing diphtheria toxoid. As a result, and on request by the WHO, a project was initiated at the National Institute for Biological Standards and control (NIBSC), UK to calibrate and establish a replacement standard in terms of the 3rd WHO IS so that continuation and traceability of the IU can be maintained. Two candidate materials for the replacement standard were provided to NIBSC for stabilization and filling prior to freezedrying. Preliminary studies performed at NIBSC confirmed that the freeze-dried products were suitable for evaluation as candidate replacement standards. A collaborative study (NIBSC code CS 357) was initiated with the primary aim of calibrating these materials in terms of the 3rd WHO IS for Diphtheria Toxoid Adsorbed, using established WHO/Ph. Eur. protection methods in guinea pigs. In total, 30 laboratories in 20 countries (Australia, Belgium, Canada, China, Croatia, Denmark, Egypt, France, Germany, India, Indonesia, Italy, Japan, Republic of Korea, The Netherlands, Norway, Russian Federation, South Africa, United Kingdom and Vietnam) participated in the collaborative study and 21 of these performed protection assays used for calibration of the replacement standard. The participating laboratories are listed in the Appendix and are referred to throughout this report by a code number, allocated at random, and not related to the order of listing. The WHO/Ph. Eur. challenge models in guinea pigs are the gold standard methods to confirm a protective response to diphtheria vaccination. Serological methods based on the measurement of antibody responses in vitro after immunization and bleeding of mice or guinea pigs are alternative methods to the direct challenge models. The WHO and Ph. Eur. allow the use of the Vero cell toxin neutralization test (TNT) to measure functional antibody responses after immunization and bleeding of animals, and allow the use of ELISA subject to validation for a particular product [10,11]. In each case, the serological response to the test vaccine is compared to that of the reference preparation and vaccine potency is expressed in IU. Therefore, a secondary aim of the collaborative study was to assess performance of the candidate replacement standards in serological assays. Participants were asked to perform a serological assay in mice (Vero cell assay) or in guinea pigs (Vero cell assay or ELISA) according to in-house protocols. 14 laboratories participated by performing the mouse Vero cell assay and 8 laboratories participated by performing serology in guinea pigs. Participants were asked to test both candidate preparations and potency was expressed in relative terms only. 2. Materials and methods 2.1. Bulk material and processing Adsorbed diphtheria toxoid preparations were provided by two different manufacturers and were identified as Candidate A and Candidate B. Full composition details of the two candidate materials provided to NIBSC are summarized in Table 1. Preliminary studies were performed at NIBSC to optimize the formulation for freeze-drying. A sugar stabilizer (Trehalose), which had been used for stabilization of the 3rd WHO IS [12], was compared with a protein stabilizer (Polygeline). Results showed that, compared to trehalose, the protein stabilizer gave rise to a more robust freezedried product with lower moisture content and this formulation was chosen for production (data not shown). Five litres of each

batch of diphtheria toxoid was stabilized by the addition of an equal volume (5 L) of sterile HaemaccelÒ solution (3.5% w/v degraded gelatin - Polygeline, KoRa Healthcare, Dublin) before freeze-drying. Filling (1 ml per 5 ml DIN ampoule) was performed on a Bausch & Strobel AVF 5090 filling line (Bausch & Strobel, Ilshofen, Germany) at room temperature with constant stirring within NIBSC’s Standards Processing Division on 16th August 2007 (Candidate B) and 10th September 2007 (Candidate A). For freeze-drying, filled ampoules were loaded into a pre-cooled (50  C) freeze-dryer (Serail CS100, Serail, Argentueil, France) with gentle shaking to ensure homogeneity of filled material. The freeze-drying programme was set as follows: the products were frozen at 50  C for 26 h followed by sublimation at 35  C for 30 h at 100 mbar vacuum. This was followed by a ramp over 10 h to a secondary drying temperature of 30  C and a continuation of secondary drying for 23 h at 30 mbar vacuum. Ampoules were back-filled with low moisture nitrogen followed by flame sealing under boil-off gas from high purity liquid nitrogen. The finished products were coded 07/216 (Candidate B) and 07/218 (Candidate A) respectively and were stored at 20  C in the dark at NIBSC. 2.2. Post-fill characterization of candidate toxoids Freeze-dried candidate toxoids were examined for appearance, residual moisture content, oxygen head space, total antigen content and potency. The precision of fill was determined by weighing ampoules after fill. Representative ampoules were weighed at 1 min intervals throughout the production run. A total of 374 ampoules were weighed for 07/216 and a total of 443 ampoules were weighed for 07/218. Measurement of the mean oxygen head space after sealing served as a measure of ampoule integrity. The mean oxygen head space was measured using an Orbisphere Pharmapack 3600 analyser (Hach Ultra Ltd., Chesterfield, UK). Residual moisture content was measured using the coulometric Karl Fischer method with total moisture expressed as a percentage of the mean dry weight of the ampoule contents. 2.3. Capture ELISA for measurement of diphtheria antigen content An in-house capture ELISA was used for measurement of diphtheria antigen content in liquid bulk toxoid and freeze-dried toxoid preparations. Toxoid samples were desorbed using 10% w/v sodium citrate to elute the antigen from the adjuvant, and centrifuged to remove any remaining adjuvant. In a duplicate preparation, the toxoid sample was centrifuged to pellet the adjuvant and adsorbed antigen so that the supernatant could be used to measure the amount of non-adsorbed antigen. In the ELISA, test samples were titrated along with a purified diphtheria toxoid reference (NIBSC 02/176, 1100 Lf/ml) on microtitre plates coated with a monoclonal antibody directed against the binding domain of diphtheria toxin/ toxoid. The amount of antigen bound to the monoclonal antibody was visualized by successive incubations with polyclonal antibody against diphtheria toxoid, detection-labelled antibody, and suitable substrate. Antigen content was calculated relative to the reference toxoid using parallel line analysis and expressed in Lf/ml. The percent adsorption in each preparation was calculated by subtracting the non-adsorbed antigen value from the total antigen value, and expressing this as a percentage of the total. Full details of the assay have been described previously [13]. 2.4. Intradermal challenge assay for measurement of potency of diphtheria vaccine The potency of the bulk liquid toxoids and freeze-dried toxoids was determined at NIBSC using the intradermal challenge method

P. Stickings et al. / Biologicals 38 (2010) 529e538

531

Table 1 Details of bulk material and filled product for candidate replacement standards. Details of bulk material

Candidate A

Candidate B

Diphtheria Toxoid Contenta Adjuvant Buffer Inactivator Preservative Diphtheria Toxoid Contentb Potencyc

150 Lf/ml Aluminium Hydroxide Gel 2.5 mg/ml (aluminium) Na2HPO4.2H2O 0.12 mg/ml KH2PO4 0.60 mg/ml Formaldehyde 0.06 mg/ml None 134.5  7.8 Lf/ml (n ¼ 2) 302 (226e398) IU/ml

200 Lf/ml Aluminium Phosphate 1.5 mg/ml (aluminium) Na2HPO4 3.49 mg/ml NaH2PO4. 2H2O 0.56 mg/ml Formaldehyde 0.05 mg/ml Thiomersal <0.005 mg/ml 223.5  6.4 Lf/ml (n ¼ 2) 566 (410e798) IU/ml

Details of filled product Stabilizer Nominal fill volume NIBSC Product Code Collaborative Study Code No. Ampoules Filled Appearance Mean fill mass Mean dry weight Mean residual moisture Mean oxygen head space Diphtheria Toxoid Contentb Potencyc

Candidate A 3.5% w/v polygeline (1:1) 1 ml 07/218 Preparation A 9970 Robust homogenous cake 1.005 g (CV 0.18%) (n ¼ 443) 0.029 g (CV 0.57%) (n ¼ 6) 0.68% (CV 18.08%) (n ¼ 12) 0.28% (CV 88.74%) (n ¼ 12) 69.0  4.4 Lf/ml (n ¼ 5) 114 (76e160) IU/ampoule

Candidate B 3.5% w/v polygeline (1:1) 1 ml 07/216 Preparation B 9977 Robust homogenous cake 1.005 g (CV 0.21%) (n ¼ 374) 0.031 g (CV 0.77%) (n ¼ 6) 0.40% (CV 12.17%) (n ¼ 9) 0.04% (CV 99.54%) (n ¼ 12) 113.4  5.9 Lf/ml (n ¼ 5) 268 (212e337) IU/ampoule

a b c

Manufacturer’s reported diphtheria toxoid content. Determined at NIBSC using in vitro antigen ELISA. Determined at NIBSC by intradermal challenge in guinea pigs.

in guinea pigs [11]. Animals, 8 per dose, were immunized with 3 dilutions of one of the candidate liquid toxoids or the 3rd WHO IS for Diphtheria Toxoid Adsorbed (98/560) as the reference vaccine. After 28 days, animals were challenged by intradermal injection of 6 dilutions of diphtheria toxin (Ph. Eur. BRP batch 1 [14]). After 48 h each animal was given an intradermal score based on the presence or absence of specific erythema at the site of toxin challenge. The intradermal challenge scores were transformed (score2) and used to obtain an estimate of potency for the test vaccines relative to the scores obtained in animals immunized with the reference vaccine by pair wise parallel-line quantitative analysis.

preparation of dilutions. Twenty-one laboratories from 17 different countries performed protection assays used for calibration of the replacement standard. A summary of participating laboratories, preparations tested and methods performed is listed in Table 2. The majority of participants performed two independent challenge tests as requested. Four laboratories (7, 18, 20 and 26) performed a single assay. Laboratories 2 and 22 performed 3 assays for both candidate preparations. Laboratory 13 performed 3 assays for Preparation B and 2 assays for Preparation A. The number of animals used per vaccine dilution ranged from 8 to 16 and the number of doses used per vaccine ranged from 3 to 6 (Table 2).

2.5. Collaborative study design for the calibration of the replacement standard An International collaborative study was organized by NIBSC for calibration of the candidate diphtheria toxoids in International Units. Assays were performed using established WHO/Ph. Eur. challenge methods in guinea pigs [7,11] and potency was expressed in terms of the 3rd WHO IS for Diphtheria Toxoid, Adsorbed. As stocks of the 3rd WHO IS are very low, the Ph. Eur. BRP batch 3 for Diphtheria Vaccine Adsorbed, which is identical to the 3rd WHO IS but which was kept at the European Directorate for the Quality of Medicines & Healthcare (EDQM, Strasbourg, France), was kindly donated by the EDQM and used for the calibration. The BRP had been sent to NIBSC before the study and was distributed together with the two candidate materials by NIBSC. The 3rd WHO IS and BRP batch 3 were tested at NIBSC using the in vitro capture ELISA to determine diphtheria toxoid Lf content. Results confirmed that there was no difference between these two materials in terms of their antigen content: 3rd WHO IS 98/560 had an antigen content of 114 (110e119) Lf/ml (n ¼ 4) and BRP batch 3 had an antigen content of 111 (107e115) Lf/ml (n ¼ 4). Diphtheria toxoids 07/218 (Candidate A) and 07/216 (Candidate B) were labelled as Preparations A and B respectively for the collaborative study. The 3rd WHO IS/Ph. Eur. BRP (batch 3) was coded as Preparation C (defined potency 160 IU/ampoule). These samples were sent to participants with instructions for storage and use. Potency estimates for Preparations A and B determined at NIBSC after fill (Table 1) were provided to participants to help in the

Table 2 Summary of laboratories performing challenge methods for calibration. Lab Code

Preparations Tested

Challenge Method

No. of Assays

No. Vaccine Dilutions

No. Animals Per Dilution

1 2 3 4 26

A, B, C A, B, C B, C A, B, C B, C

Intradermal Intradermal Intradermal Intradermal Intradermal

2 3 2 2 1

4 3 4 3 4

8 9 10 8 14

6 7 8 9 10 11 12 13

A, B, A, B, A, B, A, B, A, B, A, B, A, B, A, B,

C C C C C C C C

Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic

3 3 3 3 4 4 3 4

16 10 12 16 14 12 16 10

14 17 18 19 20 21 22 23

A, A, A, A, A, A, A, A,

C C C C C C C C

Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic

2 1 2 2 2 2 2 2 (A,B,C); 1 (B,C) 2 2 1 2 1 2 3 2

3 3 3 4 4 3 4e6 3

12 16 10 8 8 16 10 16

B, B, B, B, B, B, B, B,

Labs 5, 15 and 16 withdrew due to time/workload pressures or problems with animal supply. Lab numbers 24, 25 and 27e33 performed serology assays only.

532

P. Stickings et al. / Biologicals 38 (2010) 529e538

2.5.1. Reporting of data and statistical analysis All raw data together with assay details (preparation of candidate standard dilutions and challenge toxin) were provided to NIBSC to permit independent analysis. Participants were asked to provide details of response and scores measured, and data from all assays were analyzed using the principles of parallel-line bioassay analysis comparing transformed assay responses to log vaccine dose. For systemic (lethal) challenge assays, a probit transformation was used. For the intradermal challenge assays, the (score)2 transformation was used. All analysis was performed using CombiStats (Version 4.0, EDQM). Note that for laboratory 3 a probit analysis was performed because guinea pigs were challenged intradermally with one dilution of toxin only. For all assays, the data for each candidate preparation were analyzed separately against the 3rd WHO IS for Diphtheria Toxoid, Adsorbed (98/560, 160 IU/ampoule) and the resulting potency estimates are therefore based on direct pair-wise comparisons with the sample coded Preparation C. Some data were excluded from the systemic challenge assays before statistical analysis: where a maximum (100%) response was recorded for the two largest vaccine doses included in the assay, the largest dose was omitted from the analysis as the maximum response had clearly been achieved at the smaller of the two doses. Similarly, if a zero (0%) response was recorded for the two smallest vaccine doses, the smallest dose was omitted from the analysis. All other data were included in the analysis unless otherwise stated. For laboratory 2 the smallest vaccine dose was omitted for all preparations in all 3 assays performed as the vaccine was too dilute and all animals gave an intradermal score of zero. 2.6. Stability studies Representative samples (40 ampoules) for Preparation B were stored at each of þ4  C, þ20  C, þ37  C, þ45  C and þ56  C for accelerated thermal degradation studies initiated after freeze-drying on 09 October 2007. After storage at the above temperatures for 1 year, samples from each temperature group were removed and shipped to participating laboratories for potency testing relative to an identical sample of the appropriate preparation stored at 20  C. Each participant was asked to perform a single potency test using a WHO/Ph. Eur. challenge model. Laboratories 14, 17 and 19 performed a single potency test using the guinea pig systemic challenge model. Data obtained for each degradation sample were analyzed separately against the appropriate preparation stored at 20  C using the parallel-line bioassay analysis methods described previously. Representative samples of Preparation B stored at the elevated temperatures listed above were also tested at NIBSC using the inhouse capture ELISA to measure total antigen content, as described in Section 2.3. This in vitro assay can be used as a predictor of antigen integrity [13] and was used as an alternative to the in vivo potency assays to predict long term stability of the freeze-dried toxoid with analysis performed on degradation samples after 1 and 2 years of storage at elevated temperatures.

essentially as described previously [15] or with minor modifications [16]. Participants were asked to test both candidate preparations (Preparations A and B) and potency was expressed in relative terms only. Potency estimates for Preparations A and B determined at NIBSC after fill (Table 1) were provided to participants to help in the preparation of dilutions. 14 laboratories participated by performing the mouse Vero cell assay and 8 laboratories participated by performing serology in guinea pigs. The majority of laboratories performed two independent serology assays. Laboratory 6 performed a single serology assay in guinea pigs and laboratory 24 performed a single serology assay in mice and in guinea pigs. Laboratory 22 performed 3 independent serology assays in mice and in guinea pigs. For the guinea pig serology assays, 4 laboratories performed analysis using the Vero cell assay and 3 laboratories performed analysis using ELISA. One laboratory analyzed results using both the Vero cell assay and ELISA. See Table 3 for a summary of analysis methods and responses used in each laboratory. 2.7.1. Reporting of data and statistical analysis All raw data together with assay details (preparation of candidate standard dilutions and details of in vitro assays) were provided to NIBSC to permit independent analysis. For the Vero cell assay, some laboratories reported results as negative or positive scores, representing presence or absence of a cytotoxic effect respectively. Some laboratories expressed results for individual serum samples in the Vero cell assay relative to a positive control serum as described elsewhere [11]. All ELISA results were expressed relative to a positive control serum included in the in vitro assay. Data from all assays were analyzed using the principles of parallel-line bioassay analysis comparing assay responses (score or log antibody level) to log vaccine dose, as described previously. As the data do not allow a more complex model (e.g. four parameter logistic) to be fitted, relative potency estimates were calculated using the linear section only. A visual inspection of the data was used to detect any doses that did not lie on the linear section of the dose response curve. Where necessary, these doses were excluded from the analysis: in five assays, no significant decrease in response was observed between the two lowest vaccine doses and the smaller of these doses was excluded from further analysis (the affected assays were mouse Vero cell assays in laboratories 22, 28 and 32 and a guinea pig Vero cell assay in laboratory 22); in one assay (laboratory 22, guinea pig Vero cell), the highest two doses (of six) showed no further significant increase in response and these were excluded from the analysis. Removal of these doses allowed a linear dose response relationship to be fitted and statistically valid estimates of relative potency to be calculated. In all cases, the significance of non-linearity and nonparallelism was determined prior to calculation of relative potency. The assays in question are highlighted in Table 9.

3. Results 3.1. Characterization of candidate diphtheria toxoids

2.7. Serological assays For serology studies, participants were asked to perform a serological assay in mice (Vero cell assay) or in guinea pigs (Vero cell assay or ELISA). The Vero cell assay is an in vitro toxin neutralization test which measures the ability of serum samples from individual animals to protect Vero cells from a cytotoxic dose of diphtheria toxin. ELISA and Vero cell assays were performed according to in-house protocols and an example of the ELISA assay for determination of diphtheria antibodies in guinea pig serum has been described in detail elsewhere [11]. The general principle of the Vero cell assay performed by participating laboratories was

The diphtheria antigen content and potency of the liquid bulk materials was confirmed at NIBSC prior to stabilization, fill and freeze-drying. Of the two bulk toxoid materials, Candidate B had the higher diphtheria toxoid content when determined by in-house capture ELISA (Table 1). This is consistent with the manufacturers’ reported values for diphtheria toxoid content of the two preparations. Both toxoids were highly adsorbed to the aluminium adjuvant with approximately 98% of the antigen being adsorbed. This was also the case for the freeze-dried toxoid products post-fill (not shown). The higher diphtheria antigen content of Candidate B was reflected in a higher in vivo potency of 566 (410e798) IU/ml, compared to

P. Stickings et al. / Biologicals 38 (2010) 529e538

533

Table 3 Summary of serological assays performed by laboratories participating in the extended collaborative study (Preparation A vs. B). Lab Code

Species

4 6 11 14 19 22 24 25

Guinea Guinea Guinea Guinea Guinea Guinea Guinea Guinea

2 9 12 14 20 22 24 27 28 29 30 31 32 33

Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse Mouse

Pig Pig Pig Pig Pig Pig Pig Pig

Sex

Strain

Weight/age

No. of Assays

No. Vaccine Doses

No. Animals Per Dilution

Bleed time Weeks

Serology Method

Response

F, Ma M/F F M/F F F M/F F

DH DH Hartley DH DH Hartley DH Hartley

250e350 250e350 250e350 250e350 250e350 300e400 250e350 250e350

2 1 2 2 2 3 1 2

3 3 3 4 4 4e6 4 3

8 10 10 9 6b 10 8 10

5 6 5 6 6 4 5 5

ELISA and VERO VERO ELISA VERO ELISA VERO VERO ELISAc

Log Ab Score Log Ab Score Log Ab Log Ab Score Log Ab

F M F F F F F M F F M M F F

CD-1 ddY ICR (CD-1) ICR NIH ddY NIH (RIVM) ddY ddY CD-1 ddY ddY ddY ddY

12e14 g 12e14 g 14e16 g 18e24 g 12e14 g 5 weeks 10e14 g 12e14 g 26 g 17e24 g 5 weeks 13e15 g 5 weeks 24 g

2 2 2 2 2 3 1 2 2 2 2 2 2 2

4 4 4 4 4 4e5 5 6 4 3 5 4 4 5e6

12 8 12 12 8 10 8 8 10 16 10 12 10 12

5 5 5 5 5 4 5 5 4 5 4 5 4 4

VERO VERO VERO VERO VERO VERO VERO VERO VERO VERO VERO VERO VERO VERO

Score Score Score Score Score Log Ab Score Score Log Ab Rel score Log Ab Score Log Ab Log Ab

g g g g g g g g

M ¼ Male, F ¼ Female. a female guinea pigs assay 1, male guinea pigs assay 2. b sera was pooled for ELISA analysis. c 3 ELISA assays were performed for each assay and the antibody titres were combined for relative potency calculation.

Candidate A which had an in vivo potency of 302 (226e398) IU/ml, determined by intradermal challenge in guinea pigs (Table 1).

may have resulted in the lower moisture content compared to that seen for 07/218.

3.2. Post-fill analysis of freeze-dried candidate toxoids

3.3. Calibration of replacement standard

After filling and freeze-drying, the candidate toxoids were examined for appearance, residual moisture content, oxygen head space, total antigen content and potency. The lyophilized product for both toxoids was of very good appearance, giving rise to robust and homogenous cakes. For 07/216 the mean fill mass was 1.005 g with a coefficient of variation (CV) of 0.21% (n ¼ 374), and for 07/218 the mean fill mass was 1.005 g with a CV of 0.18% (n ¼ 443). The mean oxygen head space was determined as 0.04% for 07/216 (n ¼ 12) and 0.28% for 07/218 (n ¼ 12). Residual moisture content was determined to be 0.40% for 07/216 (n ¼ 9) and 0.68% for 07/218 (n ¼ 12). The details of the freeze-dried candidate toxoids are summarized in Table 1. Both candidate preparations fulfil the WHO requirements for reference preparations regarding precision of fill, residual moisture and oxygen head space. The antigen content and potency of 07/216 was found to be 51% and 47% of the value determined for the bulk liquid starting material respectively. This is consistent with a 1:1 dilution of the toxoid after stabilization with HaemaccelÒ, and suggests near complete recovery of biological activity after freeze-drying. For 07/218, the diphtheria antigen content of the freeze-dried material was 51% of the value determined for the starting material but the diphtheria potency was only 38% of that determined for the starting material. As a result, the recovery of biological activity after freeze-drying was estimated to be approximately 75% for 07/218, which is comparable to the recovery of biological activity for the 3rd WHO IS/Ph. Eur. BRP batch 3 which was prepared using diphtheria toxoid of similar formulation [12]. Due to a technical fault on the freeze-drying run for 07/ 216, the 50  C freezing stage and secondary drying stage were extended. Results of the post-fill analysis suggest that this did not adversely affect the finished product and the extended drying time

For calibration, potency estimates were calculated against material coded Preparation C which was the Ph. Eur. BRP batch 3 for diphtheria toxoid adsorbed, donated for the study by the EDQM. This material was identical to the 3rd WHO IS but was kept at the EDQM. Preliminary analysis performed at NIBSC confirmed that this material and the 3rd WHO IS (coded 98/560) did not differ in terms of their diphtheria antigen content. In addition, laboratory 13 included the 3rd WHO IS for diphtheria toxoid, adsorbed (98/560) in all three systemic challenge assays performed for collaborative study. The results for this vaccine were also returned to NIBSC which allowed for relative potency calculation. The potency of BRP batch 3 relative to the 3rd WHO IS was 1.06 (0.86e1.30), weighted geometric mean of 3 challenge assays. This supports the conclusion from the antigen assay and confirms that there was no difference in diphtheria potency between the material coded 98/560 and stored at NIBSC and the BRP batch 3 stored at EDQM. 3.3.1. Assay validity In general, challenge assays satisfied the requirements for validity of potency tests described in the Ph. Eur. monograph for diphtheria vaccine adsorbed [11] and by the WHO [17]. The results from all assays valid at the 1% level of inference are included in this report. Assays where deviations from the model were significant at the 5% level (0.01 < p < 0.05) are clearly indicated in the relevant figure (Fig. 1) and table of results (Table 4). In laboratory 18 (assay 1) and 23 (assay 2) only 40% protection was achieved at the largest dose of Preparation B although estimates of potency were calculated and included in the final calibration. For some assays no potency estimates have been calculated: in systemic challenge assays, 20% protection was observed at the largest dose of

534

P. Stickings et al. / Biologicals 38 (2010) 529e538

10 9 Systemic Intradermal

Number of laboratories

8 7 6

4

5

2

26

4 3 2 1

22

23

17

1

19

14

10

3

18

13

8

20

12

11

6

9

21

0 50

Fig. 1. Individual laboratory estimates for potency of Preparation B as determined by intradermal or systemic challenge in guinea pigs. Data are IU/ampoule with 95% confidence limits. *denotes non-linearity of the dose response for preparation B (0.01 < p < 0.05); xdenotes non-parallelism between preparations B and C (0.01 < p < 0.05).

Preparation B in laboratory 22 (assay 1), and in laboratory 7 (assay 1), 80% protection was observed at the smallest dose of Preparation B. 3.3.2. Potency of Preparation A Material coded Preparation A was adopted by the Ph. Eur. Commission as Ph. Eur. BRP batch 4 for Diphtheria Vaccine (Adsorbed) in February 2009, with an assigned potency of 97 IU/ ampoule. The report for calibration and establishment of this preparation was published in full previously [18]. 3.3.3. Potency of preparation B Potency estimates (IU/ampoule) for Preparation B are summarized in Table 4 and Figs. 1 and 2. Most estimates within individual

Table 4 Potency estimates (IU/ampoule) for Preparation B against Preparation C (3rd WHO IS/Ph. Eur. BRP Batch 3). Challenge Method

Assay 1

Assay 2

1 2 3 4 26 6 7 8 9 10 11 12 13 14 17 18 19 20 21 22 23

Intradermal Intradermal Intradermal Intradermal Intradermal Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic

156.1 253.0b 315.5 268.0 177.2b 249.1 Invalidc 237.4 298.6b 222.1 212.9 160.7 162.6 156.2 167.4 146.2 130.0 346.9 558.9 Invalidc 185.2

280.2 287.1 369.4 236.4a

a b c

Assay 3

254.6b

255.1 251.2 280.2 283.5 216.7 188.9 239.9 271.5 244.4

157.1

187.2 582.4 83.6 151.2

127.5

denotes non-linearity of the dose response for preparation B (0.01 < p < 0.05). denotes non-parallelism between preparations B and C (0.01 < p < 0.05). see text for details of invalid tests.

400

800

laboratories were homogeneous and a weighted geometric mean was calculated for the lab mean. Estimates from laboratory 14 were significantly heterogeneous (p < 0.05) and a semi-weighted geometric mean was calculated for the lab mean (Table 4). For intradermal challenge assays (n ¼ 5) the unweighted geometric mean potency was 249.0 (186.6 e 332.3) IU/ampoule and the between-laboratory GCV was 26.2%. Laboratory 21 performed the systemic challenge assay and was found to be a significant outlier for the analysis of Preparation A [18]. Although the result returned by laboratory 21 was not a statistical outlier for Preparation B, this laboratory observed an unusual death pattern in animals immunized with the 3rd WHO IS (coded Preparation C) which resulted in high potency estimates for Preparations A and B compared to other laboratories. The Potency of Preparation A relative to Preparation B returned by laboratory 21 was consistent with the results returned by all other laboratories performing challenge methods in guinea pigs (Fig. 3). Excluding laboratory 21 from the calibration of Preparation B, the unweighted geometric mean potency estimate (with 95% confidence limits) for all laboratories (n ¼ 19) was 213.4 (185.7

Lab Mean 237.6 261.5 338.8 256.6 177.2 251.3 e 244.4 288.6 252.8 214.2 175.8 185.2 203.7 199.4 146.2 154.2 346.9 572.8 113.6 167.2

200 Potency (IU/ampoule)

Fig. 2. Histogram showing the estimates of potency (lab mean), expressed in IU/ ampoule for Preparation B, as determined by intradermal challenge (open boxes) and systemic challenge (grey boxes) against Preparation C (3rd WHO IS/Ph. Eur. BRP Batch 3). Boxes represent individual laboratories and are code labelled.

Number of laboratories

Lab Code

100

14 13 12 11 10 9 8 7 6 5 4 3 2 1 0

22 0.25

6

6 29 27 9 2 1 21 17 14 13 10 9 8

4 14 33 28 22 20 14 2 4 19 0.50

Systemic Intradermal Mouse vero cell Guinea pig vero cell Guinea pig ELISA 24 4 32 30 23 12 11

19 31

25 11 1.00

2.00

Relative Potency Fig. 3. Histogram showing the estimates of potency (lab mean), expressed in relative terms for preparation A against preparation B, as determined by intradermal challenge (open boxes), systemic challenge (grey boxes), mouse Vero cell serology (horizontal lines), guinea pig serology Vero cell assays (diagonal lines) and guinea pig serology ELISA assays (vertical lines). Boxes represent individual laboratories and are code labelled. Note that estimates obtained by guinea pig serology ELISA assays are significantly different from the estimates obtained in the other methods (p < 0.05, ANOVA with Tukey’s test).

P. Stickings et al. / Biologicals 38 (2010) 529e538 Table 5 Final potency estimates calculated for Preparation B. Assay Method

Intradermal Challenge Systemic Challenge Without lab 21 All Challenge Assays Without lab 21

Table 7 Results of in vitro stability assays for Preparation B (after 1-2 year storage).

Preparation B Potency IU/ampoule (95% confidence limits)

n

GCV %

249 (187e332) 217 (174e269) 202 (170e240) 224 (190e265) 213 (186e245)

5 15 14 20 19

26.2 47.9 34.3 43.0 33.5

e 245.4) IU/ampoule. For systemic challenge assays (n ¼ 14) the unweighted geometric mean potency was 202.0 (170.4 e 239.5) IU/ ampoule. The between-laboratory GCV was 33.5% for all assays and 34.3% for systemic challenge assays. One-tailed t-test using log transformed values for lab means show that there is no significant difference in estimates for intradermal or systemic challenge assays (p ¼ 0.09). See Table 5 for a summary of potency estimates calculated for each assay method. 3.4. Stability studies For in vivo potency, the results from all assays were valid showing no deviations from the model at the 5% significance level. Potency estimates relative to samples stored at 20  C are shown in Table 6. Laboratory 17 performed two assays and the weighted geometric mean of their assays is shown. No significant change in the relative potencies was observed over the range of temperatures used and the data do not allow predictions of stability to be made. However, the data obtained for Preparation B suggest that this preparation will be highly stable when stored at the recommended storage temperature of 20  C. The antigen content results obtained for samples stored at elevated temperature were compared to those obtained for identical samples of the appropriate preparation stored at 20  C. Results obtained using this assay are summarized in Table 7. The relative contents of the accelerated thermal degradation samples were used to fit an Arrhenius equation relating degradation rate to absolute temperature assuming first-order decay [19] and hence predict the degradation rates when stored at 20  C. A predicted loss of toxoid antigen content of 0.09% per year when stored at 20  C was obtained for Preparation B, supporting the conclusion from the in vivo study which suggests that this preparation will be highly stable when stored at the recommended temperature. 3.5. Relative potency of candidate toxoids Preparation A and B 3.5.1. Challenge assays The results from all assays valid at the 1% level of inference are included in this report. Assays where deviations from the model were significant at the 5% level (0.01 < p < 0.05) were observed for some assays and these are clearly indicated in the results shown in

Table 6 Results of in vivo stability assays for Preparation B (after 1 year storage). 

Storage temperature

Lab Code

Potency relative to 20 C sample (with 95% confidence limits)

þ4  C þ20  C þ37  C þ37  C þ45  C þ56  C

14 17 19 14 17 19

0.84 1.00 1.09 1.08 1.01 0.99

(0.59e1.18) (0.82e1.21) (0.68e1.76) (0.76e1.54) (0.82e1.23) (0.59e1.66)

Data are from systemic challenge assays in guinea pigs.

535

Storage temperature

Antigen content relative to sample stored at 20  C Geometric mean (GCV) n ¼ 3 Year 1

Year 2

þ4  C þ20  C þ37  C þ45  C þ56  C

1.02 0.98 0.84 0.79 0.74

0.99 0.99 0.97 0.85 0.78

(23.7%) (19.4%) (4.4%) (15.3%) (11.6%)

(4.8%) (5.6%) (9.5%) (10.5%) (4.3%)

Data are expressed as the geometric mean for 3 ampoules of Preparation B stored at elevated temperature for up to 2 years. The geometric mean antigen content (Lf/ml) of 3 ampoules of Preparation B stored at 20  C was used for calculation of the relative antigen content of degradation samples.

Table 8. In laboratory 23 (assay 2) only 40% protection was achieved at the largest dose of Preparation B although estimates of potency were calculated and included. For some systemic challenge assays no potency estimates have been calculated: in laboratories 6 (assay 1) and 18 (assay 1) 20% protection was achieved at the largest dose of Preparation A, while in laboratory 22 (assay 1) no protection was observed at all doses of this preparation. In laboratory 7 (assay 1), 80% protection was observed at the smallest dose of Preparation B. In addition, significant non-parallelism (p < 0.01) was observed in laboratory 22 (assays 2 and 3). The relative potency of Preparation A against Preparation B in challenge assays was calculated for comparison with results obtained in serological assays. Results from individual laboratories are shown in Table 8. The unweighted geometric mean relative potency estimate for all challenge assays (n ¼ 15) was 0.44 (0.39 e 0.50) with a GCV of 25.3%. 3.5.2. Serological assays The results from all assays valid at the 1% level of inference are included in this report. Assays where deviations from the model were significant at the 5% level (0.01 < p < 0.05) were observed for some assays and are clearly indicated in the results shown in Table 9.

Table 8 Summary of relative potency estimates (A vs. B) obtained in challenge assays. Lab Code

Assay Method

Assay 1

Assay 2

1 2 4 6 7 8 9 10 11 12 13 14 17 18 19 21 22 23

Intradermal Intradermal Intradermal Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic Systemic

0.39 0.49 0.44 Invalidc Invalidc 0.43 0.37 0.40b 0.64a 0.59 0.36 0.39 0.48 Invalidc 0.55 0.40a Invalidc 0.49b

0.36 0.33 0.46 0.32

Assay 3 0.38

0.43 0.38 0.32 0.61 0.69 0.54 0.39 0.39 0.56 0.36a Invalidc 0.86

Invalidc

Lab Mean 0.37 0.38 0.45 0.32 e 0.43 0.37 0.36 0.63 0.65 0.42 0.39 0.44 e 0.55 0.38 e 0.62

All lab means are weighted geometric means with the exception of the estimate for lab 23 which is semi-weighted due to the significant heterogeneity between the assay estimates. a denotes non-linearity of the dose response for Preparation A or B (0.01 < p < 0.05). b denotes non-parallelism between preparations A and B (0.01 < p < 0.05). c see text for details of invalid assays.

536

P. Stickings et al. / Biologicals 38 (2010) 529e538

For some assays no potency estimates have been calculated: in mouse Vero cell assays, significant non-parallelism (p < 0.01) was observed in laboratories 12 (assay 1), 22 (assay 3) and 24 (assay 1). Significant non-linearity (p < 0.01) for Preparation B was observed in laboratories 2 (assay 2) and 12 (assay 2). This was also the case for Preparation A in laboratory 30 (assay 2). For guinea pig ELISA, significant non-parallelism (p < 0.01) was observed in laboratory 4 (assay 2). For guinea pig Vero cell assay 1 by laboratory 22, the antibody responses were below the limit of detection and no further analysis could be carried out. The relative potency estimates for Preparation A against Preparation B from serological assays are summarized in Table 9. For mouse serology assays, all estimates within individual laboratories were homogenous and a weighted geometric mean was calculated for the laboratory mean. For guinea pig serology assays, the estimates from laboratory 22 were significantly heterogeneous (p < 0.05) and a semi-weighted geometric mean was calculated for the laboratory mean. The unweighted geometric mean relative potency estimate (A against B) for serological assays (n ¼ 21) was 0.56 (0.47 e 0.65). For mouse Vero cell serology assays (n ¼ 12) the unweighted geometric mean relative potency estimate was 0.52 (0.46 e 0.59). For guinea pig serology assays (n ¼ 9) the unweighted geometric mean relative potency estimate was 0.61 (0.42 e 0.88) with an unweighted geometric mean relative potency estimate of 0.45 (0.30 e 0.69) for the assays analyzed using the Vero cell assay (n ¼ 5) and 0.88 (0.47 e 1.63) for the assays analyzed by ELISA (n ¼ 4). The between-laboratory GCV was 42.0% for all serology assays; 22.7% for mouse Vero cell serology assays; 62.3% for all guinea pig serology assays (40.4% for assays analyzed by Vero cell assay and 47.6% for assays analyzed by ELISA). A comparison of relative potency estimates for Preparation A against Preparation B for all methods (challenge and serological assays) is shown in Fig. 3. One way analysis of variance (ANOVA)

Table 9 Summary of relative potency estimates (A vs. B) obtained in serological assays. Lab Code

Assay

Assay 1

Assay 2

2 9 12 14 20 22 24 27 28 29 30 31 32 33 4 6 14 22 24 4 11 19 25

Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Mouse Vero cell Guinea pig Vero cell Guinea pig Vero cell Guinea pig Vero cell Guinea pig Vero cell Guinea pig Vero cell Guinea pig ELISA Guinea pig ELISA Guinea pig ELISA Guinea pig ELISA

0.52b 0.48 Invalidc 0.42 0.42 0.47a Invalidc 0.36 0.45 0.45 0.69 0.73 0.55a 0.60 0.59 0.43 0.51 Invalidc 0.59 0.56 1.60a 0.79 1.47a

Invalidc 0.39 Invalidc 0.56a 0.54 0.45b

Assay 3

Invalidc

0.39 0.50 0.44 Invalidc 0.73 0.73 0.50a Invalidc 0.49a 0.31a Invalidc 0.93 0.65 0.98

0.22

Lab Mean 0.52 0.44 e 0.52 0.49 0.46 e 0.38 0.47 0.44 0.69 0.73 0.69 0.54 0.59 0.43 0.50 0.26 0.59 0.56 1.17 0.72 1.26

All lab means are weighted geometric means with the exception of the estimate for lab 22 (Guinea pig Vero cell assay) which is semi-weighted due to the significant heterogeneity between the assay estimates. a denotes non-linearity of the dose response for Preparation A or B (0.01 < p < 0.05). b denotes non-parallelism between Preparations A and B. c see text for details of invalid assays. Relative potency estimates in shaded boxes were calculated after exclusion of one or more vaccine doses e see text for details.

comparing log relative potency estimates for Preparation A against Preparation B obtained in all assay methods used in the study (intradermal challenge, systemic challenge, mouse Vero cell serology, guinea pig Vero cell serology and guinea pig ELISA serology) showed a significant difference between the methods (p < 0.001). Tukey’s test to compare all groups showed that results obtained in the guinea pig ELISA serology assays were significantly different to those obtained using the other 4 methods (p < 0.05). 4. Discussion Two candidate standards were included in the collaborative study in order to provide separate standards for WHO and Ph. Eur. Both bulk toxoids were analyzed at NIBSC for antigen content (in vitro) and potency (in vivo) and confirmed to be suitable to act as candidate replacement standards for diphtheria toxoid adsorbed. Both toxoids were stabilized with degraded gelatine (HaemaccelÒ) after initial trial fills at NIBSC suggested that this stabilizer gave rise to a more robust freeze-dried product compared to trehalose which was used in formulation of the 3rd WHO IS/Ph. Eur. BRP batch 3 (data not shown). The sample coded Preparation A (NIBSC code 07/ 218) was adopted as the 4th Ph. Eur. BRP by the European Pharmacopoeia Commission in February 2009 and full details on calibration of this standard have been published elsewhere [18]. Ampoules coded Preparation B (NIBSC code 07/216) were tested and confirmed to fully comply with WHO recommendations for precision of fill, residual moisture content and integrity. Preliminary in vivo potency assay at NIBSC after fill showed a recovery of biological activity for the toxoid of approximately 94% which is superior to that achieved for the previous standard [12]. Based on the results of valid in vivo challenge assays in guinea pigs returned from 19 laboratories it was proposed to recommend sample coded B (07/216) as a suitable replacement WHO International Standard for Diphtheria Toxoid Adsorbed with an assigned potency of 213 IU/ampoule. This material was adopted as the 4th WHO IS for Diphtheria Toxoid Adsorbed by the WHO ECBS in October 2009 [20]. The between-laboratory GCV (%) for all assays was 34% for Preparation B (with data from lab 21 excluded) which is comparable to results obtained in calibration of the previous WHO IS [12]. The data from lab 21 was excluded for calculation of the final potency estimate for Preparation B since the result from this lab had been an outlier for the potency estimate obtained for Preparation A [18]. The result obtained for Preparation B was not significantly different to those returned by the other labs but was the highest of all the potency estimates returned and the laboratory in question did comment that an unusual death pattern was observed for the sample coded Preparation C. This was supported by the result for relative potency of Preparation A vs. Preparation B which was no different to results returned from all other labs performing challenge assays. After review of all available data it was decided that the performance of the sample coded Preparation C was questionable in laboratory 21 and that results for samples expressed against it should be excluded from the final potency estimate. From in vivo degradation studies included as part of the collaborative study it was not possible to make formal predictions of long term stability, because no significant loss of activity was observed after storage for 1 year at any of the elevated temperatures used. However, the in vitro antigen ELISA was used to make a prediction of long term stability after 2 years storage at elevated temperature and results obtained suggest that the replacement WHO IS will be highly stable when stored as indicated. It is proposed to continue monitoring stability of the replacement standard both in-use as the reference vaccine in potency assays and by testing the antigen content of accelerated degradation samples after long term storage at NIBSC. Historically, in vivo

P. Stickings et al. / Biologicals 38 (2010) 529e538

potency assays have been used to predict long term stability of WHO International Standards for diphtheria toxoid adsorbed. In this study however, no predictions of long term stability could be made from in vivo data since no significant loss of vaccine potency was observed at any temperature after 1 year storage. At the same time point, the in vitro antigen ELISA revealed a reduction in antigen content with increasing storage temperature relative to ampoules stored at 20  C. This allowed for a prediction of long term stability to be made (which was ultimately based on data obtained at 1 and 2 year time points). The fact that the in vitro assay was able to detect temperature induced changes after 1 year while none were detected in vivo may simply be due to the relative precision of the two assay methods in question, although it may be the case that the changes in antigen binding highlighted by ELISA were related to non-functional epitopes that had little impact on vaccine potency. However, previous work in our laboratory has suggested that the antigen ELISA can predict immunogenicity of diphtheria vaccine after exposure to elevated temperature [13], and for material such as the freeze-dried toxoid vaccines which have well established stability profiles it may be appropriate to consider replacement of the in vivo potency assay with the more sensitive and reproducible antigen ELISA to make predictions of long term stability. Some laboratories performed additional serological studies in mice or guinea pigs following in-house protocols and experience. Participants were asked to test Preparations A and B and results were expressed as relative potency of A vs. B. Preparations A and B were confirmed as suitable for use in these serological assays with both preparations showing significant regression of the dose response with no significant deviations from linearity in the majority of assays. In terms of the relative potency estimates obtained, results from Vero cell assays used with both mouse and guinea pig serology methods were comparable to the results obtained in challenge assays. The results obtained using ELISA with guinea pig serology were significantly different from the results obtained in challenge assays. Although only a small number of laboratories performed this assay as part of the collaborative study the results suggest that this method will not be suitable for calibration of replacement standards. The results do suggest however that Vero cell serology assays may be suitable for calibration of replacement standards. It is important to make the distinction between calibration of replacement (i.e. monovalent adsorbed toxoid) and secondary standards which includes product specific reference preparations. From these studies it was not possible to predict performance in relative terms with complex combined vaccine products. In vivo potency assays are subject to a high degree of variability and the relative performance of reference and test vaccine preparations may not be consistent between different labs even for the same vaccine product. As a result, further collaborative studies should be performed using a range of combined vaccine products to determine performance relative to the WHO IS in challenge assays as well as alternative serological assays.

Acknowledgements We especially thank the manufacturers for arranging and donating the two candidate materials. We are extremely grateful to the study participants listed in the Appendix for contributing data. We also thank Dr Jean-Marc Spieser and Dr Karl-Heinz Buchheit of the EDQM for arranging donation of the BRP batch 3 for use as the collaborative study reference vaccine. The authors also acknowledge Michelle Stanley and staff in the Standards Processing Division at NIBSC for running the production fills.

537

Appendix. List of participants (alphabetical order by country). AUSTRALIA

Dr Adam Smith Immunobiology Section Therapeutic Goods Administration PO Box 100, Woden ACT 2606

BELGIUM

Dr Denis Lambrigts & Ir. Valérie Decauwert GSK Biologicals QC Department e B32 Rue de l’Institut, 89 B-1330 Rixensart

CANADA

Dr Don Kemp Quality Control Sanofi Pasteur Ltd. 1755 Steeles Ave. West Toronto, Ontario M2R 3T4

CANADA

Dr Sushama Sontakke & Maria Baca-Estrada Center for Biologics Evaluation Biologics & Genetic Therapies Directorate 100 Eglantine Driveway, Ottawa Ontario K1A 0K9

CHINA

Dr Zhang Shumin NICPBP SFDA, Temple of Heaven 100050-Beijing

CROATIA

Dr Lea Lupret Goles Institute of Immunology, Inc. Quality Control Department Rockefellerova 2 HR-10000 Zagreb

DENMARK

Dr Ellen Sloth Wilhelmsen QC Bacterial Vaccines Statens Serum Institut B. 50/421 Artillerivej 5 DK-2300, Copenhagen

EGYPT

Dr Hoda Gama El Din QC Department Bacteriology Laboratory VACSERA 51 Wezaret El-Zeraa St., Agoza

FRANCE

Dr Anne Lefebvre QC Department, Building I15 Sanofi Pasteur 1541 Avenue Marcel Merieux F-69280 Marcy l’Etoile

FRANCE

Dr Sonia Prieur, Dr Didier Sauvaire, Monique Parès & Dr Dominique Garcia Afssaps, 321 Avenue Jean Jaures F-69007 Lyon

GERMANY

Dr Stefan Knapp Novartis Vaccines and Diagnostics GmbH&Co.KG Emil-von-Behring-Strasse 76 PO Box 1630, D-35041 Marburg

GERMANY

Dr Ute Rosskopf Paul Ehrlich Institut 51-59 Paul Ehrlich Strasse D-63225 Langen

INDIA

Dr Sunil Gairola Quality Control Serum Institute of India Ltd. 212/2, Hadapsar PUNE - 411 028

INDIA

Dr Surinder Singh Central Drugs Laboratory Central Research Institute Kasauli-173204 Himachal Pradesh (continued on next page)

538

P. Stickings et al. / Biologicals 38 (2010) 529e538

Appendix (continued) INDONESIA

Dr Siam Subagyo NQCLDF National Agency of Drug & Food Control Jl. Percetakan Negara No. 23 Jakarta Pusat 10560

INDONESIA

Dr Iin Susanti Bio Farma Pasteur 28, Bandung 40161

ITALY

Dr Marco Martelli & Dr Giacomo Matteucci Animal Resources Centre Novartis Vaccines & Diagnostics srl Via Fiorentina 1 I-53100 Siena

ITALY

Dr Christina Von Hunolstein, Dr L. Ciceroni, Dr F. Scopetti & Dr A. Pinto Istituto Superiore di Sanità Viale Regina Elena, 299 I-00161 Roma

JAPAN

Dr Toshiyuki Onishi Kanonji Institute The Research Foundation for Microbial Diseases, Osaka University 2-9-41 Yahata-Cho, Kanonji Kagawa 768-0061

JAPAN

Dr Akihiro Suehara Technical Biological Products Department Takeda Pharmaceutical Company Ltd. 4720 Takeda Mitsui, Hikari, Yamaguchi

JAPAN

Dr Motohide Takahashi Department of Bacterial Pathogenesis & Infection Control National Institute of Infectious Diseases 4-7-1, Gauken Musashimurayama-shi Tokyo, 208-0011

JAPAN

Dr Eiji Tokunaga Quality Control Department Kikuchi Research Centre Kaketsuken, Kyokushi, Kikuchi Kumamoto 869-1298

JAPAN

Dr Tetsuya Watanabe Denka Seiken Co. Ltd. 1-2-2, Minami Honcho Gosen-shi Niigata

REPUBLIC OF KOREA

Dr Yong Gi Jang & BooSun Kim Berna Biotech Korea Corp. 227-3 Gugal-dong, Giheung-gu Yongin-si Gyeonggi-do 449-903

THE NETHERLANDS

Dr Johan van der Gun Quality Control Support Netherlands Vaccine Institute Antonie van Leeuwenhoeklaan 9-11 PO Box 457, 3720 AL, Bilthoven

NORWAY

Dr Randi Winsnes Statens Legemiddelverk Norwegian Medicines Agency Sven Oftedals vei 8 N-0950, Oslo

RUSSIAN FEDERATION

Professor N. V. Medunitsin Tarasevic Scientific Research Institute for Standardization and Control of Medical Biological Preparations Ministry of Health, Sivcev Vrazek 41 119002, Moscow

SOUTH AFRICA

Dr W.P. Vergeer & Mrs D vd Merwe National Control Laboratory University of the Free State Swot Street Bloemfontein 9300

UNITED KINGDOM

Dr Dorothea Sesardic & Dr Paul Stickings Division of Bacteriology National Institute for Biological Standards and Control (NIBSC) Health Protection Agency Blanche Lane, South Mimms Potters Bar, Hertfordshire, EN6 3QG

VIETNAM

Professor Le Van Hiep Institute of Vaccines and Medical Biologicals (IVAC) 9-Pasteur St., Nha Trang Khanh Hoa

References [1] Galazka AM. The immunological basis for immunization series, module 2: diphtheria. World Health Organization; 1993. [2] Hardy I, Dittman S, Sutter R. Current situation and control strategies for resurgence of diphtheria in newly independent states of the former Soviet Union. Lancet 1996;347:1739e44. [3] Maple PA, Efstratiou A, George RC, Andrews NJ, Sesardic D. Diphtheria immunity in UK blood donors. The Lancet 1995;345:963e5. [4] Edmunds WJ, Pebody RG, Aggerback H, Baron S, Berbers G, Conyn-van Spaendonck MA, et al. The sero-epidemiology of diphtheria in Western Europe. ESEN project. European Sero-Epidemiology Network. Epidemiol Infect 2000;125(1):113e25. [5] Yuan L, Lau W, Thipphawong J, Kasenda M, Xie F, Bevilacqua J. Diphtheria and tetanus immunity among blood donors in Toronto. Can Med Assoc J 1997;156 (7):985e90. [6] Galazka AM, Robertson SE. Immunization against diphtheria with special emphasis on the immunization of adults. Vaccine 1996;14(9):845e57. [7] World Health Organization Expert Committee on Biological Standardization. Requirements for diphtheria vaccine, adsorbed. Technical report series 800. Geneva: WHO; 1990. 91 e 108. [8] European Pharmacopoeia Monograph Diphtheria Vaccine (Adsorbed). 6th ed. 01/2008:0443. [9] Sesardic D, Winsnes R, Rigsby P, Gaines Das R. Calibration of replacement International standard and European Pharmacopoiea biological reference preparation for diphtheria toxoid (Adsorbed). WHO Report BS/99; 1915. [10] World Health Organization Expert Committee on Biological Standardization. Recommendations for diphtheria, tetanus, pertussis and combined vaccines (Amendments 2003). Technical report series 927. Geneva: WHO; 2005. Annex 5. [11] European Pharmacopoeia Chapter 2.7.6. Assay of Diphtheria vaccine (adsorbed). 6th ed. 01/2008:20706. [12] Sesardic D, Winsnes R, Rigsby P, Gaines-Das R. Calibration of replacement International standard and European Pharmacopoeia biological reference preparation for diphtheria toxoid, adsorbed. Biologicals 2001;29:107e22. [13] Coombes L, Stickings P, Tierney R, Rigsby P, Sesardic D. Development and use of a novel in vitro assay for testing of diphtheria toxoid in combination vaccines. J Immunol Methods 2009;350:142e9. [14] Sesardic D, Prior C, Daas A, Buchheit KH. Collaborative study for establishment of the European Pharmacopoeia BRP batch 1 for diphtheria toxin. Pharmeuropa Bio 2003;1:7e21. [15] Kreeftenberg JG, van der Gun JW, Marsman FR, Sekhius VM, Bhandari SR, Maheswari SC. An investigation of a mouse model to estimate the potency of the diphtheria component in combined vaccines. J Biol Stand 1985;13(3):229e34. [16] Aggerbeck H, Heron I. Improvement of a Vero cell assay to determine diphtheria antitoxin content in sera. Biologicals 1991;19:71e6. [17] World Health Organization. Manual of laboratory methods for testing of vaccines used in the WHO Expanded Programme on Immunization 1997; WHO/VSQ/97.04. [18] Stickings P, Rigsby P, Buchheit K-H, Sesardic D. Calibration of European Pharmacopoeia biological reference preparation for diphtheria vaccine (Adsorbed) batch 4. Pharmeuropa Bio and Scientific Notes 2009;1:1e9. [19] Kirkwood TBL. Predicting the stability of biological standards and products. Biometrics 1977;33:736e42. [20] Stickings P, Rigsby P, Sesardic D. Collaborative study: calibration of replacement WHO International standard for diphtheria toxoid adsorbed. Geneva: World Health Organization; 2009. WHO/BS/09.2123.