development, technical and clinical validation of a...

Clinical and Vaccine Immunology

23 March 2009 1

Development, technical and clinical validation of a quantitative 1

ELISA for the detection of human antibodies to Hepatitis B surface 2

antigen (HBsAg) in recipients of recombinant HBV vaccine 3

4

Running Title: Quantitative ELISA for detection of anti-HBsAg 5

6

Authors: Pierre Cambron*, Jeanne-Marie Jacquet, Bernard Hoet, Marc Lievens 7

GlaxoSmithKline Biologicals, Rue de l’institut 89, Rixensart, Belgium. 8

9

*Corresponding author 10

Pierre Cambron 11

GlaxoSmithKline Biologicals, 12

Rue de l’institut 89, Rixensart, 13

Belgium. 14

E-mail: [email protected] 15

Phone: +32-2-656-61-86 16

Fax: +32-2-656-91-44 17

18

Key words: hepatitis B, HBsAg, ELISA, validation study 19

Copyright © 2009, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.Clin. Vaccine Immunol. doi:10.1128/CVI.00431-08 CVI Accepts, published online ahead of print on 24 June 2009

on July 9, 2018 by guesthttp://cvi.asm

.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 2

ABSTRACT (244 words, max allowed 250): 20

Pending removal from the market of a commercial assay (AUSAB™EIA) for the 21

determination of antibodies to Hepatitis B Surface Antigen (HBsAg), a new in-house 22

quantitative ELISA to measure antibodies against HBsAg (anti-HBs) was developed 23

(anti-HBs in-house). Specific anti-HBs were sandwiched between pre-coated HBsAg 24

ad and ay subtypes purified from plasma from HBV human carriers and recombinant 25

HBsAg adw2 subtype tagged with horseradish peroxidase. The assay was calibrated 26

against the 1st-IRP for anti-hepatitis B Immunoglobulin (1977-W1042). Analytical 27

sensitivity and the Limit of Quantitation were estimated at 0.43mIU/ml and 28

2.0mIU/ml respectively. Overall reproducibility was 11.86% and accuracy was 29

estimated at 94.89%. More than 4000 samples from 7 clinical trials were tested with 30

the anti-HBs in-house and compared to results generated with AUSAB™EIA and 31

AUSAB™RIA. During the technical validation, anti-HBs in-house was compared to 32

AUSAB™RIA as reference (n=919). Overall assessment of concordance and 33

Deming’s regression analysis were performed. The coefficient of correlation between 34

AUSAB™RIA and anti-HBs in-house was 0.9815 with slope 0.9187. The overall 35

agreement between anti-HBs in-house and AUSAB™RIA was 97.61% considering the 36

clinical cutoff at 3.3mIU/ml and 1.0mIU/ml for the respective assays. On a clinical 37

perspective, seroprotection rates and anti-HBs geometric mean antibody 38

concentrations for individual studies calculated with either the in-house assay or the 39

references assays were similar. Conclusions of individual studies were confirmed. The 40

performance characteristics of the in-house assay are acceptable. There is no evidence 41

that use of the new assay would lead to different clinical conclusions compared to the 42

reference method. 43


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 3

INTRODUCTION 44

Hepatitis B infection is a global health problem but most acutely affects developing 45

countries (16). Currently there is no effective therapy against hepatitis B, whose 46

disease spectrum ranges from asymptomatic disease to chronic liver diseases 47

including cirrhosis and hepatocellular carcinoma. Prevention of the illness through 48

vaccination remains the method of choice for its control and eradication. Active 49

immunization against hepatitis B infection can be achieved using vaccines containing 50

either inactivated Hepatitis B Virus (HBV) surface protein (HBsAg) physico-51

chemically purified from plasma from HBV human carriers, or, recombinant surface 52

antigen produced by transfer of the S gene of HBV coding for HBsAg to an 53

appropriate plasmid that is then inserted into the desired expression vector. The 54

recombinant vaccine manufactured by GlaxoSmithKline Biologicals (GSK) is 55

produced in yeast and is antigenically similar to plasma derived HBsAg (3). Clinical 56

studies of this recombinant vaccine either formulated as single component 57

(Engerix™-B), or formulated in combination with other antigens such as hepatitis A 58

vaccine (Twinrix™) or pediatric diphtheria-tetanus-pertussis-based vaccines (such as 59

Infanrix hexa™ and Tritanrix™ HepB) have proven its efficacy and immunogenicity 60

(4). 61

To date, commercial assays from Abbott laboratories were used at GSK to quantify 62

the immune response to HBV vaccines in terms of antibodies against HBsAg (anti-63

HBs). However, since these assays are no longer commercially available in Europe, 64

GSK has developed an in-house assay with adequate technical and clinical 65

performance to ensure long-term supply of an assay with consistent quality. This 66

paper describes the development, the technical validation and the clinical assessment 67

of the new anti-HBs in-house assay. 68


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 4

MATERIAL AND METHODS 69

Abbreviations 70

4PL: 4 Parameters Logistic, BSA: bovine serum albumin, CV: coefficient of variance, 71

FBS: fetal bovine serum, GMC: geometric mean concentration, HRP: horseradish 72

peroxidase, IRP: International Reference Preparation, OD: optical density, PBS: 73

phosphate buffered saline, SD: standard deviation, SSD: standard and sample diluent 74

Selection of the immunological ELISA format 75

A sandwich format was selected whereby specific anti-HBs were sandwiched between 76

pre-coated HBsAg passively adsorbed onto the microplate and HBsAg tagged with 77

HRP. Because the purpose of HBV vaccination is to protect against possible future 78

contact with the wild type virus, native HBsAg ad and ay subtypes (10), purified from 79

plasma from HBV human carriers were used as the solid phase. Quantification of 80

antibodies raised by vaccination used HRP conjugate prepared from recombinant 81

adw2 HBsAg subtype, similar to that contained in the GSK vaccine (7), tagged with 82

HRP. 83

Preparation of the HBsAg pre-coated plates 84

Microplates were passively coated with 100µl of a mix (w/w) of native human 85

HBsAg ad and ay subtypes (14) (The Binding Site) at 1µg/ml (w/v) in carbonate 86

buffer pH 9.5 for 24 hours at room temperature. The plates were then washed four 87

times with distilled water supplemented with NaCl (9g/liter) and Tween-20 (Merck) 88

and blocked with 200µl of PBS (0.0095M pH 7.4 (Cambrex) supplemented with 89

enzymatically hydrolyzed casein 0.1% (Sigma) and proclin-300 0.5ml/litre (Supelco) 90

as preservative for 24 hours at room temperature. Plates were washed 4 times with 91

distilled water with Tween-20 (0.05%) and saccharose 10g/l (Sigma) then dried 92


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 5

overnight in a laminar flux. Each microplate was individually packed in a sealed 93

aluminum bag that included desiccant. 94

Preparation of HBsAg-HRP conjugate 95

The recombinant HBsAg adw2 subtype was conjugated to sodium-periodate-oxidized 96

HRP (type VI RZ 3.0, Sigma) by a modification of the original method of Nakane et 97

al (6). Half a milliliter of a freshly prepared aqueous solution of 0.14 M sodium 98

periodate solution (Aldrich) was added to 1 ml of 50 mg/ml HRP solution in distilled 99

water and gently stirred for 30 minutes in the dark in order to generate active aldehyde 100

functions on the hydroxyl groups of the several carbohydrate moities. Then the HRP-101

aldehyde solution was chromatographically separated from periodate by gel filtration 102

(Sephadex G-25) and quantified by absorbance at 403nm (1mg/ml of Oxidized HRP 103

displays an OD of 1.6 at 403nm). The molar ratio between oxidized HRP after 104

separation and the HBsAg to be coupled was 2.0. 105

The HBsAg solution previously dialyzed overnight at 4-8°C against sodium carbonate 106

buffer 0.1mol/l pH 9.5 was mixed with the HRP-aldehyde solution for 2 hours at 107

room temperature in the dark in order to generate Schiff’s bases between the primary 108

amine groups on HBsAg and the aldehyde groups present on the oxidized HRP. 109

Remaining reactive aldehyde groups were blocked by adding 2M Tris solution at a 110

rate of 22µl per mg of HRP-aldehyde and incubated for 30 minutes at room 111

temperature. Subsequently the Schiff bases were reduced for 30 minutes at 4°C by the 112

addition of 0.21M sodium borohydrure (Aldrich) at a rate of 10µl/mg of HRP-113

aldehyde engaged in the conjugation process. The HBsAg-HRP was 114

chromatographically separated by gel filtration on ACA-34 (Amersham 115

Biosciences) equilibrated with Tris buffer 0.05M pH 7.5 containing Tween-20 116

0.5ml/L. Finally, the fractions containing the HBsAg-HRP conjugate were pooled, 117


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 6

supplemented with BSA (Cohn fraction V, Serologicals) 10g/l and proclin-300 2ml/l 118

and stored at 4-8°C after filtration on a 0.2 micron filtering device. 119

Preparation and calibration of anti-HBs Abs standard 120

A pool of sera from HBV vaccine recipients with high levels of anti-HBs was diluted 121

in FBS (Hyclone) containing thymol 0.2g/liter (Sigma) and gentamicin-sulphate 122

100µg/ml (Sigma) as preservatives. After adjustment to around 190 mIU/ml, referring 123

to the 1st IRP (International Reference Preparation) for anti-hepatitis B 124

immunoglobulin (15), the standard preparation was dispensed in 1 ml aliquots and 125

freeze-dried. To conduct an assay, one freeze-dried vial was reconstituted with 1 ml of 126

distilled water and then serially diluted (6 times) with SSD (FBS containing proclin-127

300 2ml/l as preservative). SSD was used as zero standard. FBS was used instead of 128

recalcified human plasma free of anti-HBs (human plasma) in order to prevent future 129

availability issues. The standard was calibrated against the anti-HBV Immunoglobulin 130

reference preparation 1st IRP, 1977-W1042. For calibration purposes, 4 in-house anti-131

HBs standard vials were reconstituted with 1 ml of distilled water then serially diluted 132

(neat plus 6 steps) with FBS and incubated on 4 different lots of pre-coated plates. At 133

the same time, serial dilutions in human re-calcified plasma free of anti-HBs antibody 134

ranging between 1.58 mIU/ml and 200.00 mIU/ml of one reference ampoule (1st IRP, 135

1977-W1042) freshly reconstituted according to instructions were incubated. 136

Consequently, 96 measures based on serial dilutions were obtained. After correction 137

for their respective dilution factor, the average concentration of the 96 individual 138

values was determined. The values obtained before correction for dilution were 139

extrapolated from the standard curve generated by the WHO preparation using a 4PL 140

fitting algorithm. 141


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 7

Preparation of assay controls 142

Four anti-HBs controls were prepared from separated pools of recalcified human 143

plasma. The negative control (CN-1) was non-reactive for anti-HBs antibody. CB1-1 144

was spiked with serum obtained from a recipient of HBV vaccine and had an antibody 145

level of 6.7 mIU/ml. CM-1 had a level of 18.1 mIU/ml and was spiked with serum of 146

hospital origin. CM1-100 was unspiked and had a level of 6436 mIU/ml. CN1-1, 147

CB1-1 and CM1-1 were freeze-dried and stored at 4-8°C whereas CH-100 was liquid 148

and stored at –20°C. At the time of use CN1-1, CB1-1 and CM1-1 were reconstituted 149

with 0.5 ml of distilled water. Control CH1-100 was thawed and prediluted 1 to 100 150

with SSD before use. 151

Collection, preparation and storage of test samples 152

Whole blood from healthy recipients of HBV vaccine was collected after informed 153

consent was obtained. Blood collection was approved by relevant ethics committees. 154

Sera were stored at –20°C until testing. Undiluted samples with responses that were 155

higher than the highest anti-HBs Abs standard were diluted by successive factors of 156

10 until the response was quantifiable. 157

ELISA procedure 158

50µl each of standards, controls and specimen were added in duplicate to the pre-159

coated microplate. The plate was covered and incubated overnight at room 160

temperature, then washed four times with 350µl of distilled water containing Tween-161

20 (0.05%) and NaCl (0.154M). 100µl of HBsAg-HRP conjugate at a working 162

dilution of 1: 1000 in the HRP diluent (PBS 0.0095M pH 7.4 supplemented with BSA 163

10% and proclin-300 1ml/liter as preservative) was added across the entire plate. 164

Plates were covered and gently shaken for 2 hours at room temperature. Unbound 165


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 8

conjugate was removed by washing as previously described. 100µl of the single 166

component TMB- peroxidase substrate (Biorad) was added to all wells. After 30 167

minutes incubation in the dark, the color reaction was stopped by the addition of 168

100µl of diluted sulphuric acid 0.36N in distilled water. Plates were read in an ELISA 169

reader at 450nm against a reference filter at 630nm within 30 min after stopping. 170

The calibration curve and its fitting algorithm 171

The dose response curve expressed in terms of the response y (OD's on a linear scale) 172

versus the dose x (mIU/ml on a logarithmic scale) commonly display a pronounced 173

sigmoidal shape. The 4PL function fits these data with a high degree of accuracy (2, 174

8, 9). An unweighted 4PL function was used to fit the standard curve for anti-HBs in-175

house. In order to estimate the accuracy of the standard curves, r2 and the percentage 176

of the relative error or the relative bias: % R.E. (also called percentage of backfit dose 177

error) was determined for each standard point. 178

The backfit dose is the recalculated concentration of each standard obtained by 179

extrapolating its response (OD’s) through the fitted curve. The % R.E of each 180

standard point (zero standard is excluded from the analysis) is as follows: 181

% R.E. = (recalculated value – assigned value)/assigned value x 100%. To evaluate 182

the overall quality of the fitting of the standard curve, 167 standard curves collected 183

during the development phase of the assay were assessed for r2 and % R.E. 184

The literature and current guidelines recommend an absolute value of R.E. ≤ 20% and 185

r2 >0.99 (1,12). 186

Analytical sensitivity 187

Sixty four replicates of the SSD were incubated on one plate along with one set of 188

anti-HBs standards. The average response and its SD were calculated in terms of ODs. 189


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 9

Then the upper limit of the 1-sided 95% CI was determined, injected into the standard 190

curve and interpolated in terms of mIU/ml from the fitted standard curve. 191

Limit of Detection (LOD) 192

One hundred and forty serum samples collected from healthy adults and shown to be 193

negative in AUSAB™ RIA were tested in the assay. Average and standard deviation 194

were calculated after logarithmic transformation of the OD responses. Then the upper 195

limit of the 1-sided 95% CI was determined, injected into the standard curve and 196

interpolated in terms of mIU/ml from the fitted standard curve. 197

Limit of Quantitation (LOQ) 198

Serial 2-fold dilutions (6 consecutive ones) were made for 56 samples (post-vaccinees 199

and hospital subjects). If required, samples were pre-diluted with SSD to around 10 200

mIU/ml in order to limit the number of dilutions per samples. For each sample, the 201

lowest calculated concentration that did not differ no more than 2-fold from the 202

expected concentration (the concentration obtained at the first dilution was considered 203

as the most accurate) was considered as still reliable. Data were logarithmically 204

transformed before statistical analysis. The average lowest reliable concentration was 205

calculated with its SD. LOQ was set at the upper limit of the 1-sided 95% CI. 206

Comparison of FBS and human-plasma as standard and sample diluent 207

Ninety eight samples covering a large range of anti-HBs concentrations were assayed 208

either with FBS or with human plasma as SSD and the results compared by linear 209

regression analysis after logarithmic transformation of the data. In addition a paired-t 210

test was performed on the geometric mean concentrations obtained with each diluent. 211


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 10

Accuracy-recovery 212

In order to assess accuracy, one ampoule of the 1st IRP for anti-HBs immunoglobulin 213

(1977-W1042) was reconstituted according to instructions and diluted to 10,000 214

mIU/ml in human plasma. Further, a 25µl aliquot was added to different volumes of 215

human plasma free of anti-HBs antibodies in order to cover a concentration range 216

from 3.3 to 100,000 mIU/ml. Then each sample was measured in the assay. The 217

accuracy was expressed as the geometric mean recovery percentage calculated on the 218

whole set. 219

Assessment of repeatability and reproducibility of anti-HBs in-house in routine 220

conditions 221

Precision was evaluated by a repeatability and reproducibility experiment made on 12 222

sera covering a broad range of anti-HBs concentrations observed in routine 223

conditions. To measure repeatability (precision under minimum variable condition), 224

the same dilution of serum was tested twice by the same certified operator in the same 225

day in the same run on the same plate with the same lot of reagents. To estimate the 226

reproducibility (precision obtaining under maximal variable conditions), all sera were 227

tested twice on the same plate at 4 different days (within a period of 2 weeks) by 2 228

certified operators. A random 3-way ANOVA with day, operator and sample as 229

random factor was carried out (using mixed procedure from statistical SAS® 230

software) on log-10 transformed concentrations. Computations were done on 231

transformed values but CVs expressed from non-transformed concentrations. In 232

conclusion 192 observations were generated (2 operators x 4 days x 2 repeats x 12 233

samples). 234


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 11

Linearity and parallelism 235

Ten samples displaying concentrations between 100 and 51,895 mIU/ml were pre-236

diluted depending on their initial concentration to reach the upper limit of the 237

analytical range. Then 6 serial 2-fold dilutions were applied to each sample. Each 238

measured concentration was corrected for its dilution. For the analysis, samples were 239

gathered into 4 groups depending on the first pre-dilution applied: neat for group 1, 1 240

to 10 for group 2, 1 to 100 for group 3 and finally 1 to 1000 for group 4. A statistical 241

analysis was performed on the logarithmically transformed data in order to answer the 242

following questions: is the model globally linear, are the slopes equal (parallel lines) 243

from one group to another? Are the slopes equal from one sample to the other within 244

each group? And finally, are the measured concentrations proportional to the applied 245

successive dilutions? 246

Specificity 247

In order to establish specificity of the assay, 3 samples displaying a concentration 248

below the cutoff and 10 samples displaying a very large range of concentrations were 249

pre-incubated (vol/vol) with an excess of either a mix of native ad and ay HBsAg 250

subtypes (the combination used for the solid phase), or adw2 HBsAg subtype (the one 251

used for the preparation of the HRP conjugate) for 2 hours at 37°C. The excess of 252

antigen corresponded to 40 times the concentration of antigen used for the coated 253

plate preparation. The reference condition was similar to the experimental one but 254

with PBS-BSA buffer instead of HBAg moities. Then samples were tested with anti-255

HBs in-house but also with AUSAB™ EIA as specificity control. Specificity was 256

expressed as the difference in concentration observed before and after inhibition 257

expressed as percentage of the corresponding concentration measured without 258

inhibition. 259


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 12

Assessment of high dose hook-effect 260

In order to rule out underestimation of the antibody concentration in samples with 261

high levels of the analyte, three sera highly concentrated for anti-HBs were tested 262

undiluted and at 5 consecutive 10-fold serial dilutions. There is no hook-effect if the 263

neat sample reports a concentration higher than the upper limit of the analytical range. 264

Clinical cutoff and correlation with AUSAB™ RIA 265

The assay cutoff has to be equal or higher than the LOQ. In order to determine the 266

most appropriate cutoff for seropositivity, different cutoffs from 2 to 3.9 mIU/ml were 267

compared during the technical bridge of the new assay compared to AUSAB™ RIA. 268

In this respect the concentrations obtained with the new assay for 919 samples 269

including pre-vaccination (7%) and post-vaccination samples were compared to the 270

historical results observed with the reference method. The cutoff was elected at a level 271

for which the best balance was observed in terms of overall agreement, sensitivity and 272

specificity. In addition Deming’s regression was performed between double positive 273

samples in order to evaluate the correlation between both assays. 274

Clinical validation of the assay 275

The accepted serological correlate of protection of ≥10 mIU/ml against HBV infection 276

was determined from studies that employed the AUSAB™ RIA. Therefore, the clinical 277

validation included several studies having used AUSAB™ RIA. Three additional 278

studies having used AUSAB™ EIA were also considered. In total, 4180 samples from 279

seven clinical studies were re-tested. These studies had been conducted in infants, 280

adolescents, adults and the elderly (>50 years of age). Subjects had received 281

vaccination with various HBV vaccine-containing products including vaccine 282


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 13

combinations such as combined hepatitis A-HBV vaccine and combined diphtheria-283

tetanus-acellular pertussis-HBV-inactivated poliovirus vaccine (Table 1). 284


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 14

RESULTS 285

Preparation of the anti-HBs standard and its calibration to anti-HBs IRP lot 286

17.02.1977 287

The distribution of the 96 individual concentrations obtained from the serial dilutions 288

was Gaussian and the coefficient of variation of the recalculated concentrations was 289

4%. The final concentration of the freeze-dried in-house standard was expressed in 290

mIU/ml. The calibration curve of the assay consists of one zero standard and 7 291

positive standards ranging from 2.9 to 187.0 mIU/ml namely 187.0, 93.5, 48.8, 23.4, 292

11.7, 5.8, and 2.9 mIU/ml. The loss of immunoreactivity after freeze-drying was 1.5% 293

on average for two successive anti-HBs standard batches and therefore was 294

considered negligible. 295

The calibration curve and its fitting algorithm 296

For all the 167 standard curves the r2

was greater than 0.998. The 0.95 percentile of 297

the absolute values of % R.E. were determined (Table 2). Accordingly we conclude 298

that the accuracy of the standard curve within the analytical range set from 3.3 299

mIU/ml to 187 mIU/ml meets current guidelines (1,12). 300

Analytical sensitivity 301

The individual OD of the 64 replicates of the zero standard were normally distributed. 302

The average response and the SD of all replicates were calculated. Then the upper 303

limit of one-sided 95% CI of the mean OD was injected into the standard curve 304

generated at the same time and the corresponding concentration was extrapolated 305

from the curve. Accordingly, the analytical sensitivity was estimated at 0.43 mIU/ml. 306


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 15

Limit of Detection (LOD) 307

The OD corresponding to the upper limit of the one-sided 95% CI of the average OD 308

response of the 142 negative samples was interpolated from the standard curve 309

allowing the determination of an equivalent concentration expressed in mIU/ml. 310

Accordingly, LOD was set at 1.0 mIU/ml. 311

Limit of Quantification (LOQ) 312

The average lowest measured concentration was 0.68 mIU/ml with a SD of 0.26 The 313

LOQ of the in-house test was set at 2.0 mIU/ml, equivalent to the one-sided upper 314

limit of the 95% CI. 315

Comparison of FBS and human-plasma as standard and sample diluent 316

The standard curves generated with either FBS or human-plasma were 317

superimposable (data not shown). The concentrations of the 98 samples measured 318

with either FBS or human-plasma were statistically compared by linear regression 319

analysis. The interval for the slope was 1.0055 (95% CI: 0.9948; 1.0162) and the 320

intercept -0.0051 (95% CI: -0.0244; 0.0143) comprising 1 and zero, respectively, 321

indicated no significant difference between both diluents (Figure 1). Moreover, a 322

paired t-test showed no significant difference between means (the observed difference 323

in log was 0.0036 for which the 95% CI included zero [-0.0060; 0.0132]). 324

Consequently FBS can be used instead of human-plasma free of anti-HBs as sample 325

and standard diluent without affecting sample concentration determination. 326

Accuracy-Recovery 327

The spiking experiments were performed on two different days. Recovery was 328

calculated by expressing the measured concentrations in % of the theoretical 329

concentration. Recovery ranged from 79.3% to 101.2%. The geometric mean recovery 330


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 16

was 94.89% (95% CI: 92.24%; 97.62%, Table 3. Accordingly anti-HBs in-house is 331

considered as accurate vis-à-vis the 1st IRP for anti-HBV Immunoglobulin (1977-332

W1042). 333

Precision 334

A random 3-way ANOVA with day, operator and sample as random factor was 335

carried out (using proc mixed of SAS®)

on log10-transformed concentrations. 336

Computations were done on transformed values but CVs were expressed from non-337

transformed concentrations. The CV for repeatability of the assay was 4.72%. The 338

global CV R&R for reproducibility and repeatability in routine conditions was 339

11.86%. Precision is in agreement to what is currently displayed by ELISA (5) 340

Linearity and parallelism 341

A mixed model was used with the factors log dilution (fixed and continuous), group 342

(fixed, 4 groups) and samples (random, nested into the groups, 10 samples). The test 343

for curvature of the log dilution effect was not statistically significant (p-value 344

0.5940). Consequently a linear model was fit on the data (Figure 2). The test of the 345

interaction of the log dilution and the group factors gave a non significant p-value of 346

0.8903 meaning that the log dilution effect (slope) can be considered equal from one 347

group to another. The p-value of the interaction between the log dilution and the 348

sample factors was non significant (0.24526) meaning that the slope of one sample is 349

equal to the slope of another within each group. Finally, the estimated slope with its 350

95% CI was calculated for each group (Table 4). None of the slopes were significantly 351

different from -1, supporting the proportionality between dilutions. Consequently 352

linearity and parallelism is demonstrated over the analytical range. 353


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 17

Specificity 354

A shown in Table 5, the specificity of anti-HBs in-house is around 100%, similar to 355

specificity displayed by AUSAB™ EIA. 356

Hook-effect 357

The response of the 3 undiluted samples gave a higher response than that displayed by 358

the upper standard point, for which the response was 3.0 ODs in this experiment 359

(upper standard specifications: 2.5–3.3 ODs) (Figure 3). In addition, CV% calculated 360

on dilution 10,000 and 100,000 was as good as 2% (Table 6). Concentrations as high 361

as 823,976 mIU/ml can be reliably measured with the new assay without displaying 362

any hook-effect. 363

Clinical cutoff and correlation with AUSAB™ RIA 364

Table 7 displays the overall agreement, specificity and sensitivity for the comparison 365

of anti-HBs in-house and AUSAB™ RIA during the technical validation of the assay. 366

The cutoff of anti-HBs in-house was fixed at a value above the LOQ which gave the 367

best balance in respect to overall agreement, sensitivity and specificity vis a vis the 368

reference assay while keeping a non significant p value for the Mc Nemar test of 369

dissymmetry in concordance. Accordingly, the cutoff was fixed at 3.3 mIU/ml. At 370

this level, the overall agreement was 97.61% (CI: 96.40 to 98.49) with a sensitivity 371

and a specificity calculated at 98.04% (CI: 96.84 to 98.88) and 94.06% (CI: 87.52 to 372

97.79) which were considered as good. Deming’s regression was performed on the 373

double positive samples (n = 802) considering a CV% at 11.86% as determined 374

during the repeatability and reproducibility experiment (Figure 4). Finally, regression 375

was considered as good between anti-HBs in-house and AUSAB™ RIA. The Fuller’s 376

coefficient of correlation (simple correction on AUSAB™ RIA) was calculated to be 377


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 18

0.9815 which can be considered as high. The estimation of the slope and its 95% CI 378

limit was established at 0.9187 (0.9060; 0.9315). However in Table 8, the test of the 379

bio-equivalence of the slope was demonstrated because the 95% CI was within the 380

0.90-1.11 equivalence range (11,13). 381

Analytical range 382

The lower limit of the analytical range usually corresponds to the LOQ of the assay. 383

However we elected to use the cutoff for positivity in order to increase confidence in 384

the assay. Thus, the lower limit of the analytical range was set at 3.3 mIU/ml. The 385

upper analytical range was fixed at 187.0 mIU/ml, level at which the accuracy is still 386

effective (see section The calibration curve and its fitting algorithm). 387

Clinical validation of the assay 388

Seroprotection rates observed with the anti-HBs in-house and reference assays in 7 389

clinical trials are summarized in Table 9. Figure 5 shows anti-HBs GMCs using the 390

in-house and reference assays in three studies representative of different age groups. 391

Study B investigated immune responses to HBV vaccination in non-responders 392

subjects having received at least 4 doses of the Engerix™ HBV vaccine and was 393

included in order to verify that those non-responders would not have been considered 394

as protected by the anti-HBs in-house assay, which would be an expression of a lack 395

of specificity of the new assay. Out of 96 subjects assessed as seronegative at study 396

start with AUSAB™ RIA, 9 were assessed as seropositive with anti-HBs in-house 397

assay. Accordingly, the specificity was calculated at 90.6% at study start (data not 398

shown). However, out of these 9 samples, 7 were confirmed as seropositive with 399

AUSAB™ EIA. At this stage, this apparent sub-optimal specificity could be linked to 400

a better sensitivity of anti-HBs in-house and AUSAB™ EIA compared to AUSAB™ 401


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 19

RIA. These non-responders then participated in a new trial (mentioned as study B in 402

Table 9). The final analysis performed on study B after the complete vaccine course 403

showed similar seroprotection rates and similar GMT determined with either anti-HBs 404

in-house or AUSAB™ RIA. Overall, the specificity of anti-HBs in-house compared to 405

AUSAB™ RIA determined on a normal population was determined at 94.06% (see 406

previous section Clinical cutoff and correlation with AUSAB™ RIA) which was 407

considered as good. 408

In the re-analysis of 7 studies in which different recombinant HBsAg-containing 409

vaccines were used, a very good agreement in terms of seroprotection rate was 410

observed after full schedule and also one month post-vaccination. Antibody GMCs 411

were similar across the in-house and reference assays, with a trend to overestimation 412

using the in-house assay in the lower concentration range. Importantly, there was no 413

evidence that the use of the anti-HBs in-house assay would lead to different clinical 414

conclusions compared to the reference assays. 415 on July 9, 2018 by guesthttp://cvi.asm

.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 20

DISCUSSION 416

The new anti-HBs in-house assay has been developed to replace use of a commercial 417

assay soon to be removed from the market. The performance characteristics of the 418

new in-house assay have been thoroughly tested and are acceptable. Clinical 419

application of the in-house assay has been documented by re-analysis of 4180 samples 420

from individuals of all ages participating in clinical trials of recombinant HBsAg-421

containing vaccines. In these studies seroprotection rates and anti-HBs GMCs for 422

individual studies were similar using either the in-house assay or the references 423

assays. In all studies the immunogenicity conclusions determined by the reference 424

assay were confirmed using the in-house assay. 425

The specificity of the anti-HBs in-house assay determined in Study B was 90.6% but 426

this figure was calculated at study start and in non-responder subjects having received 427

at least 4 doses of hepatitis B vaccine. After the full schedule, seroprotection rates and 428

GMT were similar for anti-HBs in-house and the reference assay 429

In conclusion, the new in-house anti-HBs antibody assay has been technically and 430

clinically validated and displays adequate performance characteristics. There is no 431

indication that use of the new assay would lead to different clinical conclusions 432

compared to the currently used reference methods. 433


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 21

TRADEMARKS 434

Engerix™-B, Ambirix™, Twinrix™, Tritanrix™, Infanrix hexa™ and Infanrix™ 435

penta are trademarks of the GlaxoSmithKline group of companies. Recombivax™ is a 436

trademark of Merck & Co. AUSAB™ is a trademark of Abbott Laboratories, North 437

Chicago. 438

ACKNOWLEDGEMENTS 439

The assay was developed and supported by GlaxoSmithKline Biologicals. On behalf 440

of GSK Biologicals, the authors thank Dr. Joanne Wolter, Dr. Veronique Delpire, 441

Vidya Virajith, and Manjula K for assistance with manuscript preparation, as well as 442

ensuring manuscript coordination. 443

CONFLICT OF INTEREST 444

All authors (P. Cambron, J. Jacquet, B. Hoet and M. Lievens) are employed by the 445

commercial entity which has developed the in-house assay. J. Jacquet is also a 446

shareholder of GlaxoSmithKline Biologicals. 447


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 22

REFERENCES 448

1. Braggio, S., R. J. Barnaby, P. Grossi, and M. Cugola. 1996. A strategy for 449

validation of bioanalytical methods. J. Pharm. Biomed. Anal. 14:375–388. 450

2. Findlay, J.W.A., W.C. Smith, J.W. Lee, G.D. Nordblom, I. Das, B.S. DeSilva, 451

M.N. Khan, and R.R. Bowsher. 2000. Validation of immunoassays for 452

bioanalysis: a pharmaceutical industry perspective. J. Pharm. Biomed. Anal. 453

21:1249–1273. 454

3. Hauser, P., P. Voet, E. Simoen, H. C. Thomas, J. Pêtre, M. De Wilde, and 455

J. Stephenne. 1987. Immunological properties of recombinant HBsAg produced 456

in yeast. Postgrad. Med. J. 63:83–91. 457

4. Keating, G.M., and S. Noble. 2003. Recombinant hepatitis B vaccine 458

(Engerix™-B): a review of its immunogenicity and protective efficacy against 459

hepatitis B. Drugs. 63:1021–1051. 460

5. Miller, K.J., R. R. Bowsher, A. Celniker, J. Gibbons, S. Gupta, J. W. Lee, S. 461

J. Swanson, W. C. Smith, R. S. Weiner. 2001. Workshop on bioanalytical 462

methods validation for macromolecules: summary report. Pharm. Res. 18:1373–463

1383. 464

6. Nakane, P.K., and A. Kawaoi. 1974. Peroxidase-labeled antibody. A new 465

method of conjugation. J. Histochem. Cytochem. 22:1084–1091. 466

7. Pêtre J., Van Wijnendaele F., De Neys B., Conrath K., Van Opstal O., 467

Hauser P., Rutgers T., Cabezon T., Capiau C., Harford N., et al. 1987. 468

Development of a hepatitis B Vaccine from transformed yeast cells. Postgrad. 469

Med. J. 63 Suppl 2:73–81. 470


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 23

8. Plikaytis, B.D., S. H. Turner, L. L. Gheesling, and G. M. Carlone. 1991. 471

Comparisons of standard curve fitting methods to quantitate Neisseria 472

meningitidis group A polysaccharide antibody level by enzyme-linked-473

imunosorbent assay. J. Clin. Microbiol. 29:1439–1446. 474

9. Rodbard, D., and G. R. Frazier. 1975. Statistical analysis of radioligand assay 475

data. Methods Enzymol. 37:3–22. 476

10. Soulier, J.P., and A. M. Couroucé-Pauty. 1973. New determinants of hepatitis 477

B antigen (Au or HB antigen). Vox Sang. 25:212–234. 478

11. Tan, C. Y., and B. Iglewicz. 1999. Measurement-methods comparisons and 479

linear statistical relationship. Technometrics. 41:192–201. 480

12. U.S. Department of Health and Human Services Food and Drug 481

Administration Center for Drug Evaluation and Research (CDER) Center 482

for Veterinary Medicine (CVM). 2001. Guidance for Industry. Bioanalytical 483

Method Validation. 484

13. U.S. Department of Health and Human Services Food and Drug 485

Administration Center for Drug Evaluation and Research (CDER). 2001. 486

Guidance for Industry. Statistical approaches to establish bioequivalence. 487

14. Vnek, J., and A. M. Prince. 1976. Large-scale purification of hepatitis B surface 488

antigen. J. Clin. Microbiol. 3:626–631. 489

15. World Health Organization (WHO). Expert Committee on Biological 490

Standardization Geneva 6-12th

Dec 1977. Study of a proposed international 491

reference preparation for anti-hepatitis B Immunoglobulin. 492

16. World Health Organization (WHO). 2004. Hepatitis B vaccines WHO position 493

paper. W.E.R. 79:255-263. 494


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 24

Table 1: Clinical studies retested with the anti-HBs in-house assay 495

Study Population Group (vaccine)

Reference assay

N/group to be tested

TimePoint Rationale

A 50-70 years ● Experimental HepB formulation ● Engerix™-B

RIA 72 (subset)

PI-m1 PII-m2 PII-m6 PIII-m7

Verify concordance with AUSAB™ RIA in an elderly population

which reaches lower seroprotection rate and GMT than adults 18-40 years

B adult non-responders****

● Experimental HepB formulation ● Engerix™-B

RIA 58 (all)

Pre*** PI-m1 PII-m2 PII-m6 PIII-m7

Verify concordance with AUSAB™ RIA in a difficult population,

check specificity of the anti-HBs in-house assay on pre-vaccination samples

C 11-18 years ●Engerix™-B ●Recombivax™

RIA 72 (subset)

PI-m1 PII-m2 PII-m6 PIII-m7

Verify concordance with AUSAB™ RIA after vaccination with a

vaccine from another manufacturer

D 18-40 years ●Engerix™ old 1* ●Engerix™ old 2** ●Engerix™

EIA 72 (subset)

p2m6 p3m7

Verify concordance with AUSAB™ EIA following different

formulations of the same vaccine

E 12-15 years ●Ambirix™ 720/20 ●Twinrix™ 360/10

RIA, EIA 140 (all)

PIII-m7 Verify concordance with AUSAB ™ RIA and AUSAB™ EIA

F Children < 1 year

●Infanrix hexa™ ●Infanrix™ penta + Hiberix™

RIA 50 (subset)

Pre PII-m6 PII-m8 PIII-m9

Verify concordance with AUSAB™ RIA after vaccination with

HBV in combined pediatric vaccines

G children < 1 year

● Infanrix hexa™ ●Infanrix™ penta + Hiberix™

EIA 62 (all)

Pre-m1.5 PIII-m7 PIII-m15 PIV-m6

Verify concordance with AUSAB™ EIA after vaccination with

HBV in combined pediatric vaccines

* Old 1 : previous formulation, containing thiomersal as a preservative, **Old 2 : previous formulation, using thiomersal during the production process, *** pre= sample taken prior to vaccination, PI/ (II, III, IV) = post dose 1 (2, 3 or 4)m m1-15 = study month, **** non-responder = healthy adults with anti-HBs concentration < 10 mIU/ml 2 to 5 months after having received at least 4 doses of HBV


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 25

Table 2: Calibration curve - distribution of the percentage of the absolute 496

values of the relative bias (% R.E.) for the complete set of standard points (0.95 497

percentile) 498

mIU/ml 2.92 5.84 11.69 23.38 46.75 93.50 187.00

0.95 percentile 14.41 8.59 6.31 4.51 4.11 3.12 1.60

% R.E. = (recalculated value - assigned value)/assigned value x 100% 499


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 26

Table 3. Accuracy of anti-HBs in-house in terms of human anti-HBV 500

Immunoglobulin 501

Theoretical(mIU/ml) Measured (mIU/ml) Recovery %

3.3 3.4 100.4

6.3 5.6 90.4

6.3 5.8 93.4

10.4 9.8 94.2

10.4 9.4 90.3

20.8 19.9 95.5

20.8 18.9 90.8

41.5 42.0 101.2

41.5 41.7 100.4

82.6 81.1 98.2

82.6 80.1 96.9

163.9 129.9 79.3

163.9 150.0 91.5

476.2 446.0 93.7

909.1 840.6 92.5

2000.0 1967.9 98.4

3333.3 3304.4 99.1

5000.0 5048.4 101.0

10000.0 9885.2 98.9

Geometric mean recovery (%) 94.9

CI 95% lower limit (%) 92.2

CI 95% upper limit (%) 97.6


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 27

Table 4: Demonstration of assay linearity. Estimated slope for each pre-502

dilution group with its 95% CI using a mixed model with log dilution, group and 503

samples as factors 504

95% confidence interval

Group Slope Lower limit Upper limit

1 -0.9990 -1.0318 -0.9662

2 -0.9910 -1.0348 -0.9472

3 -0.9802 -1.0180 -0.9423

4 -0.9976 -1.0722 -0.9230

10 samples with concentrations between 100 and 51,895 mIU/ml were grouped into 4 groups 505

according to the pre-dilution applied to reach the upper limit of the analytical range 506


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 28

Table 5: Specificity of anti-HBs in-house compared to AUSAB™ EIA 507

Before preincubation HBsAg preincubation: ad + ay HBsAg preincubation: adw2

AUSLAB®EIA a-HBs in-house AUSLAB ®EIA a-HBs in-house AUSLAB ®EIA HBs in-house

mIU/ml mIU/ml mIU/ml %

specificity

mIU/ml %

specificity

mIU/ml %

specificity

mIU/ml %

specificity

20 15 <3.3 100 <3.3 100 <3.3 100 <3.3 100

94 61 <3.3 100 <3.3 100 <3.3 100 <3.3 100

119 125 <3.3 100 <3.3 100 <3.3 100 <3.3 100

346 352 <3.3 100 <3.3 100 <3.3 100 <3.3 100

382 428 <3.3 100 <3.3 100 <3.3 100 <3.3 100

1556 1111 <3.3 100 <3.3 100 <3.3 100 <3.3 100

1957 2410 <3.3 100 <3.3 100 <3.3 100 <3.3 100

2237 2575 <3.3 100 <3.3 100 <3.3 100 <3.3 100

49680 42660 <3.3 100 <3.3 100 86 99.8 49 99.9

54490 61065 <3.3 100 <3.3 100 651 98.8 103 99.8

<3.3 <3.3 <3.3 NA <3.3 NA <3.3 <3.3

<3.3 <3.3 <3.3 NA <3.3 NA <3.3 <3.3

<3.3 <3.3 <3.3 NA <3.3 NA <3.3 <3.3

Preincubation with either an excess of a mix of native ad and ay HBsAg subtypes or adw2 HBsAg subtype for 508

2 hours at 37°C. Specificity= difference in concentration observed before and after inhibition expressed as 509

percentage of the corresponding concentration measured without inhibition. 510


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 29

Table 6: Calculation of the average concentration and coefficient of variance 511

(CV%) in three samples with high anti-HBs concentrations 512

Sample Dilution OD mUI/ml mUI/ml * CV (%)

1 1 > 3.000 >187

10000 1.652 772731 754411 3

100000 0.200 736091

2 1 > 3.000 >187

10000 1.438 644921 649409 1

100000 0.180 654897

3 1 > 3.000 >187

10000 1.705 833075 823976 2

100000 0.215 814878

* average concentration determined at 10,000 and 100,000 dilutions; OD=optical density;

CV=coefficient of variance


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 30

Table 7: Contingency table between anti-HBs in-house and the AUSAB™ 513

RIA with different cutoffs (n= 919) 514

Cut-off (mIU/ml) 2 2.6 3.3 3.9

Agreement (95% CI) 97.71 (96.53,98.58) 97.82 (96.66,98.67) 97.61 (96.40,98.49) 96.74 (95.37,97.79)

Sensitivity (95% CI) 99.02 (98.08,99.58) 98.53 (97.45,99.24) 98.04 (96.84,98.88) 96.94 (95.52,98.01)

Specificity (95% CI) 87.13 (79.00,92.96) 92.08 (84.99,96.52) 94.06 (87.52,97.79) 95.05 (88.82,98.37)

Kappa 0.881 (0.830,0.931) 0.891 (0.843,0.938) 0.883 (0.835,0.931) 0.847 (0.793,0.900)

Mc-Nemar p value* 0.3833 0.5034 0.0525 0.0003

*: test of dissymmetry in concordance 515


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 31

Table 8: Bio-equivalence of the slope for anti-HBs in-house compared to 516

AUSAB™ RIA 517

Deming Regression (Ultrastructural Model) with CV (RIA) =11.86% and CV ( anti-HBS in-

house)=11.86%: Test of Bio-equivalence on the Slope

Equivalence range

Lower specification Upper specification Equivalence

0.50 2.00 Yes

0.67 1.50 Yes

0.80 1.25 Yes

0.90 1.11 Yes


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 32

518

Table 9 Seroprotection rates (%) per time point and study: Anti-HBs in-house versus reference assays 519

Vaccine Group 1 Vaccine Group 2 Vaccine Group 3 Study N Time point in-house RIA EIA in-house RIA EIA in-house EIA

72 PI-m1 15.3 20.8 - 0.0 2.7 - - - 72 PII-m2 80.6 81.9 - 23.6 26.4 - - - 72 PII-m6 95.8 93.1 - 59.7 56.9 - - -

A

72 PIII-m7 98.6 98.6 - 81.9 83.3 - - -

51 PI-m1 47.1 49.0 - 63.6 63.6 - - - 51 PII-m2 60.8 60.8 - 81.8 81.8 - - - 50 PII-m6 48.0 42.0 - 85.2 81.5 - - -

B

47 PIII-m7 72.3 72.3 - 96.2 96.2 - - -

100 PI-m1 2.0 2.0 - 9.7 12.5 - - - 72 PII-m2 80.6 68.1 - 68.1 69.4 - - - 72 PII-m6 93.1 93.1 - 95.8 95.8 - - -

C

72 PIII-m7 98.6 97.2 - 100.0 100.0 - - -

D 72 PIII-m7 97.2 - 95.8 97.2 - 97.2 97.2 97.2

134 PI/II-m2 53.0 40.3 41.8 90.4 86.8 86.8 - - E 145 PII/III-m7 99.3 98.6 99.3 100.0 100.0 100.0 - -

50 Pre 13.8 16.9 - 12.2 12.2 - - - 50 PII-m6 98.0 98.0 - 96.0 96.0 - - - 50 PII-m8 94.0 94.0 - 100.0 100.0 - - -

F

50 PIII-m9 100.0 100.0 - 100.0 100.0 - - -

34 Pre-m1.5 52.9 - 49.0 59.6 - 68.4 - - 47 PIII-m7 100.0 - 100.0 95.9 - 95.0 - - 53 PIII-m15 96.6 - 96.6 93.0 - 96.0 - -

G

57 PIV-m6 100.0 - 100.0 98.3 - 98.3 - -

N = number of subjects per group. If the number was different in the 2 groups, lowest number was taken, pre= sample taken prior to vaccination, PI/ (II, III,

IV) = post dose 1 (2, 3 or 4)m m1-15 = study month, in-house - anti-HBs in-house, RIA = AUSAB™ RIA, EIA= AUSAB™ EIA


.org/D

ownloaded from

http://cvi.asm.org/


23 March 2009 33

FIGURE LEGENDS 520

Figure 1: Linear regression between FBS and human-plasma anti-HBs free as diluent 521

Figure 2: Observed values and the model predictions for linearity assessment; for group 522

1 and 4, prediction line for each sample is superimposed with the prediction line for mean 523

sample within each group 524

Figure 3: Optical density response of serially dilutions of three high concentrated anti-525

HBs samples 526

Figure 4: Deming’s regression of anti-HBs in-house and AUSAB™ RIA with CV(RIA) = 527

11.86% and CV (anti-HBs in-house) = 11.86% 528

Figure 5: GMCs (mIU/ml) at each blood sampling time point in Study A (elderly), Study 529

E (adolescents) and Study F (infants): anti-HBs in-house versus reference assays 530

531

532


.org/D

ownloaded from

http://cvi.asm.org/


.org/D

ownloaded from

http://cvi.asm.org/

development, technical and clinical validation of a...

Documents