Jour-nul of Vimlogical Methods. 32 (1991) 265-216 0 1991 Elsevier Science Publishers B.V. / 0168-8510/91/$03.50

ADONlS01688.5109100129X

265

VIRMET 01160

Quantitation of hepatitis B surface antibody by an automated microparticle enzyme immunoassay

D.H. Ostrow, Brook Edwards, David Kimes, Jerzy Macioszek, Henry Irace, Linda Nelson, Kevin Bartko, Jan Neva, Christine Krenc and Larry Mimms

Hepatitis R&D. Abbott Laboratories, Abbott Par-k, IL. U.S.A.

(Accepted2 January 1991)

Summary

A fully automated microparticle enzyme immunoassay, IMx AUSAB, was develo- ped for the detection and quantitation of antibody against hepatitis B surface antigen (anti-HBs). The IMx AUSAB assay can complete 24 tests in less than 45 minutes. Anti-HBs concentrations in specimens are calculated automatically by comparison of the specimen rate to values determined from a stored standard curve. IMx AUSAB sensitivity is 2-3 mIU/ml, equivalent in sensitivity to AUSAB RIA or EIA. Speci- mens from blood donors, diagnostic and hospital patients, hepatitis B vaccinees, and individuals with a variety of infectious and autoimmune diseases tested in parallel by IMx AUSAB and AUSAB RIA or IMx AUSAB and EIA gave overall qualitative agreement of 97.8% (1265/1293) and 99.1% (128 l/1293), respectively.

The prevalence of anti-HBs ranged from 5.9% in volunteer blood donors to 47.0% of specimens from a sexually transmitted disease clinic. Most discordant specimens ( 18/34) were low level reactive (< 10 mIU/ml) by AUSAB RIA, but negative by IMx AUSAB and AUSAB EIA. These specimens were also negative for antibodies to hepatitis B core antigen (anti-HBc). Six discordants were low level reactive by IMx but negative by RIA and EIA. Three of these six specimens were also reactive for anti- HBc suggesting that the IMx AUSAB reactivity resulted from the presence of low level anti-HBs. Quantitative agreement between IMx AUSAB and RIA or IMx and EIA for 106 specimens ranging in anti-HBs concentration from 1 to 30 000 mIU/ml gave linear correlation coefficients of 0.91 and 0.96, respectively. The IMx test was useful for monitoring hepatitis B vaccine response and seroconversion levels after hepatitis B infection.

Car-r-espondence to: L. Mimms, Hepatitis R & D, Abbott Laboratories, Abbott Park, IL 60064, U.S.A.

266

Anti-HB; Recombinant HBsAg; Enzyme-linked immunoassay; Hepatitis B serologi- cal marker

Introduction

The antibody to hepatitis B surface antigen (anti-HBs) is associated with conva- lescence and recovery from an acute hepatitis B viral infection. Anti-HBs is detect- able after immunization with hepatitis B vaccine and is a good indicator of protection against hepatitis B virus (HBV) infection. Although HBV infections have been reported in vaccinated individuals with anti-HBs levels above 10 mIU/ml, the risk of infection increases in vaccinees whose anti-HBs levels fall below 10 mIU/ml (Hadler et al., 1986; Stevens et al., 1984; Coursaget et al., 1968). Most of these infections were clinically silent and were only detected by the presence of anti-HBc in sera. Since the kinetics of anti-HBs decay appear similar in all vaccinated individuals, the duration of protection (the time anti-HBs remains above 10 mIU/ml) is dependent on the peak anti-HBs level after the last vaccine dose (Jilg et al., 1988). Jilg et al. (1990) have recommended post vaccine testing to determine peak anti-HBs levels and scheduling of booster doses according to these levels in health care personnel, hemophiliacs, or patients on long term dialysis in developed countries.

Direct solid phase radioimmunoassays (AUSAB RIA) or enzymeimmunoassays (AUSAB EIA) (Mushahwar, 1987; Mushahwar and Spiezia, 1987) have been used routinely to measure anti-HBs concentrations using a reference standard calibrated against the World Health Organization (WHO) anti-HBs standard established in December 1977 (Barkeret al., 1978; Courouce, 1990).

A fully automated, microparticle based immunoassay for the detection and quantit- ation of anti-HBs is described using the IMx instrument. The performance of the auto- mated anti-HBs assay was compared to AUSAB RIA and EIA and showed equivalent sensitivity and detectability. Good quantitative agreement was also observed among assays.

Materials and Methods

Samples

Specimens for in-house and clinical testing were obtained from hospital diagnostic laboratories, reference laboratories, a sexually transmitted disease (STD) clinic and volunteer and commercial blood banks.

Immunoassays

Serum and plasma were tested using kits from Abbott Laboratories (North Chicago, IL): AUSAB RIA and EIA for anti-HBs and CORAB for anti-HBc.

261

Recombinant DNA-derived HBsAg (rHBsAg)

Both major subtypes of HBsAg (adw2 and ayw) were cloned and expressed in mouse L cells and purified by immunoaffinity chromatography. rHBsAg consists of both the glycosylated, gp27, and unglycosylated, ~24, forms of the S gene product and has been extensively characterized as described previously (Mimms et al., 1989).

IMx instrument

The IMx assay for anti-HBs is based on microparticle enzyme immunoassay (MEIA) technology used in the IMx instrument system (Fiore et al., 1988). Briefly, the MEIA technology uses microparticles approximately 0.2 microns in diameter as the solid phase. Separation of bound from unbound material is accomplished by cap- ture of the microparticles on a glass fiber matrix. Microparticles allow increased kin- etics of reaction (total time of assay is less than 45 min for 24 specimens) compared to other solid phases. No operator intervention is required other than pipetting of the samples into the specimen wells, placing the reagent vials into the analyzer and initiat- ing the assay.

IMx A USAB assay

Carboxylated latex particles (0.1 to 0.3 ,um diameter; Seradyn, Indianapolis, IN) were coupled to rHBsAg using l-ethyl-3-3-(dimethylaminopropyl) carbodiimide chemistry (Quash et al., 1978). Affinity purified rHBsAg was biotinylated using Biotin-X-NHS (Boehringer Mannheim, Indianapolis, IN) according to the manufac- turer’s instructions. Affinity purified rabbit anti-biotin was conjugated to alkaline phosphatase using the method of Nakane and Kawaoi (1974).

In the first step of the IMx AUSAB assay, microparticles coated with rHBsAg (sub- types ad and ay) are added to the specimen. An aliquot of the reaction mixture is trans- ferred to the glass fiber matrix. The microparticles bind irreversibly to the matrix and unbound material is washed through the matrix. Biotinylated rHBsAg (subtypes ad and ay) is then added to the matrix to react with captured anti-HBs. Anti-biotinl alkaline phosphatase conjugate is added to the matrix to react with the biotinylated rHBsAg/captured latex microparticle complex. Unbound conjugate is removed by washing the matrix. The bound conjugate complex is detected by incubation with the fluorogenic substrate 4-methylumbelliferyl phosphate. The rate of fluorescence sig- nal generation is proportional to the amount of anti-HBs bound to the microparticle/ matrix solid phase.

Anti-HBs calibrator kit and sensitivity calculation

The anti-HBs calibrator kit consists of six members, (A, B, C, D, E, and F) which have been assigned anti-HBs concentrations by standardization against the World

268

Health Organization (WHO) Anti-HBs Standard. The concentration assignments in terms of mIU/ml or IU/l were 0, 10,50, 100,500,1000 for A, B, C, D, E and F cali- brators, respectively. The calibrators are used to generate a standard curve of fluor- escence rate versus anti-HBs concentration. A four parameter logistic equation best fitting these data is used to calculate the anti-HBs concentration in specimens. Assay sensitivity was calculated as the concentration of anti-HBs at the cutoff value using a line generated between the A and B calibrator rates. The cutoff value equals two times the A calibrator rate.

For AUSAB RIA and EIA, the Abbott AUSAB Quantitation Panel (containing members at 0, 15,40,75 and 150 mIU/ml) was used to determine sensitivity or quan- tify anti-HBs concentration in specimens. In these assays each panel member was run in duplicate and a point to point linear regression is used to define the calibration curve.

Specimens having anti-HBs concentrations greater than 1000 mIU/ml in IMx, were diluted 1 to 11 and 1 to 12 1 (if necessary) into recalcified human plasma nonreactive for anti-HBs and HBsAg (NHP) and retested. Specimens having anti-HBs concentra- tions greater than 150 mIU/ml in the RIA or EIA, were diluted 1 to 11, 1 to 121, and 1 to 133 1 (if necessary) into NHP and retested.

Statistical methods

Within-run, between-run and total standard deviations (SD) and coefficient of vari- ation (CV) were calculated as described in Krouwer and Rabinowitz (1984).

Results

Sensitivity

Calculated sensitivities for IMx AUSAB, AUSAB RIA and EIA determined using the World Health Organization Standard were 1.8-2.8., 1.4-l .7 and 2.9-3.5 mIU/ml, respectively. The sensitivity of the IMx assay was further evaluated by testing twofold serial dilutions of 14 different anti-HBs reactive specimens with titers varying from 4 to 5 12. As shown in Table 1, the IMx assay usually gave twofold better sensitivity than AUSAB EIA and for most specimens was equivalent or twofold less sensitive than AUSAB RIA. The anti-HBs concentration determined by IMx AU-SAB diluted lin- early within the entire calibration range (up to 1000 mIU/ml) (Fig. 1).

Reproducibility

The reproducibility of IMx AUSAB was determined by testing the negative con- trol, positive control and E calibrator (500 mIU/ml) in replicates of four on each of 10 IMx instruments at 7 time points over 15 days. Within run, between run, between instruments and total variability expressed as %CV ranged from 2.74.8, 3.3-8.3, 1.4-7.2 and 5.4-9.7, respectively.

269

TABLE 1

Comparison of IMx AUSAB, AUSAB RIA and AUSAB EIA sensitivity by determining the endpoint

dilution for anti-HBs specimens

Specimen IMx AUSAB AUSAB RIA AUSAB EIA

H7 128 64 64

HI4 128 256 128

H15 512 512 256

R3 256 256 128

R9 256 256 128

HB6-123 32 128 64

HB6-110 32 64 32

89-1460 256 512 64

NIH-20432 8 8 4

HB6-84 512 256 128

HB6-168 64 128 64

HB6-188 256 512 256

HB6-82 128 128 64

HB6-125 64 256 32

- HE

- 93

-1 100 E

- R9

3 - H&-l23

-i - HB6-110

;i - NIH-204.312

!z - HB6-64 .I E - HB6-168 u 10

t n - HB6-168

- HB682

,__’ HB6-125

- HE6 139

IO 100 1

Dilution Factor

30

Fig. 1. Two fold serial dilutions of 14 anti-HBs reactive specimens diluted into NHP were tested in IMx

AUSAB. Anti-HBs concentrations were determined in the IMx assay by comparison of assay signal to a standard curve calibrated against the World Health Organization standard.

Specificity

Random sera and plasma from hospitals, a sexually transmitted disease clinic, a volunteer blood bank, and a plasmaphercsis center (n = 1293) were tested in parallel by IMx AUSAB, AUSAB RIA and EIA. As shown in Table 2, agreement between IMx and RIA and IMx and EIA was 97.8% (1265/1293) and 99.1% (1281/1293), respectively. The AUSAB IMx reactivity rate ranged from 5.9% in blood donors to 47.0% in specimens from a sexually transmitted disease clinic.

Random blood bank and hospital specimens (n = 799) tested by all three AUSAB assays are plotted as frequency histograms (Fig. 2). Frequency histograms showed good separation between anti-HBs negative and positive populations for all three assays with mean S/CO and standard deviations (SD) for IMx, RIA, and EIA of 0.42 rt 0.07,0.40 + 0.09,0.23 + 0.09, respectively. The negative population mean for IMx was 8.3 SD from the assay cutoff compared to 6.7 and 8.5 SD from the cutoff for RIA and EIA, respectively.

Sixty-five hospital specimens containing potentially interfering substances were tested from the following categories: dialysis patients (n = 10) and patients with elev- ated alkaline phosphatase (n = lo), elevated bilirubin (n = lo), elevated liver enzymes (n = lo), rheumatoid factor (n =5), anti-nuclear antibody (n= 5), hypergammaglobuli- nemia (n=lO), or herpes virus infection (n = 5). Ten of these specimens (15.4%) were reactive by IMx and 100% agreement was observed with EIA. Two discordants (IF2- B3, IF2-B8) were low level reactive by RIA but negative for AUSAB EIA, IMx and CORAB (Table 3).

In the blood donor and plasmapheresis populations 100% agreement between IMx and EIA was observed. All seven blood bank specimens giving discrepant results between IMx and RIA were low level reactive (< 5 mIU/ml) by AUSAB RIA (Table

3). Overall, thirty-four specimens gave discordant results between the IMx and bead

assays. These specimens were also tested for anti-HBc and the results are given in Table 3. Eighteen of these specimens were low level reactive by AUSAB RIA but negative by IMx and EIA. All of these specimens were also nonreactive for anti-HBc.

Six discrepant specimens, five from the hospital lab and one from an STD clinic, were low level reactive by IMx (< 5 mIU/ml) but nonreactive by RIA and EIA. Three of these six specimens (HOS-19, HOS-109, HOS-193) were also reactive for anti- HBc.

Six discrepant specimens, all from the STD clinic, were negative by EIA, but low level reactive in RIA and IMx. Only one of these specimens (HB8-111) was also anti- HBc reactive.

Two discordants (VH8-67, VH8- 168) were reactive by EIA and RIA but negative by IMx AUSAB and CORAB. One discordant specimen (VH8- 126) was reactive by EIA and IMx AUSAB, but negative by RIA and CORAB and one specimen (IBB-66) was borderline reactive by AUSAB RIA and CORAB but negative by IMx and EIA. Apparent anti-HBs concentrations in all 4 of these discrepant samples was less than 10 mIU/ml as measured in EIA or RIA.

271

TABLE 2

Population testing by IMx AUSAB, AUSAB RIA and AUSAB ETA

Specimen categories

Blood donors Plasmapheresis donors Hospital patients STDclinic patients

Number IMx AUSAB AUSAB RIA AUSAB EIA Discrepants tested N(%) N(%, N(%) N(%)

493 29 (5.9) 36 (7.3) 29 (5.9) 7(1.4) 100 lO(lO.0) 14(14.0) lO(lO.0) 4 (4.0) 500 48 (9.6) 48 (9.6) 44 (8.8) 12 (2.4) 200 94 (47.0) 97 (48.5) 87 (43.5) 11 (5.5)

Total 1293 181(14.0) 195(15.1) 170(13.1) 34 (2.6)

N = number positive; % = % positive.

250

0”

fi 200

is E IL 150

100

50

0 ‘1 0

q EIA

Fig. 2. Frequency histogram of 799 blood bank and hospital plasma and sera tested in parallel by IMx AUSAB and AUSAB RIA and EIA. In all assays, specimens giving S/CO values greater than or equal to 1 .O

are considered reactive.

272

TABLE 3

Specimens giving discrepant results between AUSAB IMx, RIA and EIA

Specimen ID Specimen source Anti-HBs concentration (mIU/ml)~ CORAB

AUSAB IMx’ AUSAB RIA- AUSAB EIA’

RIA reactive PP6-35 PP6-55 PPG-77 PP6-97 SP3-15 BD-70 BD-80

Blood bank Blood bank Blood bank Blood bank Blood bank Blood bank Blood bank

-- I -- l -- l -- J

_

-J- J --

IBB-22 Plasmapheresis center - IBB-82 Plasmapheresis center - IBB-89 Plasmapheresis center -/- VH8-69 Hospital lab -

VH8- 120 Hospital lab -

HOS-102 Hospital lab -- l HOS-75 Hospital lab -

HB8-87 STD clinic -

HB8-131 STD clinic -- J HB8- 154 STD clinic _

HB8-185 STD clinic _

IF2-B3 Hospital lab -

lF2-B8 Hospital lab

IMx reactive HOS- 19 HOS- 109 HOS-193 HOS-135 VH8-167 HBX-176

Hospital lab Hospital lab Hospital lab Hospital lab Hospital lab STD clinic

3.2J3.6 4.113.3 3.OJ4.1 5.1/4.7 4.0/2. I 1.4/1.6

RIA and IMx reactive HBX-11 I STD clinic HB8-22 STD clinic HB8-IO STD clinic HB8-69 STD clinic HBS-109 STD clinic HB8-156 STD clinic

Other reactives VH8-67 Hospital lab VH8-168 Hospital lab VH8-126 Hospital lab IBB-66 Plasmapheresis center

ELV-12.7 19119 4.2J4.2 1.6J1.8

ELV-12.1 1.911.7

--

: -- 17Jl5

-

1 .Y/2.9/2.7 2.4/1.7JELV-

1.8j2.3 2.5J1.7JELV-

1.5 3.4J2.6 2.6J4.1

-- l -- l -f- -- i -

-- J -I-

3.7 2.8J3.613.7

I l/IO/9 14.4J14.3 211 lJl.3

2.012.2 3.0125

2.8J2.2J2.7 16/10/l 1

8/S 16 6.7/8 8.3J9.8 WI.4

-. - - _ -

-- J -

ELV- -- J

-

- -

-- J -- J --

: -- -- J

ELV-

-- J -- J -- J -- J -- J

-

1 ..sJELY- ELV-JELV- 34124.5 -- J

3.4129 ELV-JELV- 4.9 ELV-JELV-

4.5JS.5 _

5J6.5 -- l

2.9J2.9 1 x12.2

- _.. I ELV-/I .7

2.515 4.4J4.6

818 - t

‘S/CO < 0.7 is indicated by -; ELV- is indicated if 0.7 < S/CO < 1 .O. t Initial and repeat testing rest&s are represented by “i“.

273

Quantitative correlation

Random anti-HBs positive specimens (n = 106) obtained from diagnostic and STD clinics or from hepatitis B vaccinees were tested by IMx AUSAB, AUSAB RIA and EIA. Specimen anti-HBs concentrations ranged from 1 to 30 000 mIU/ml by all three assays. Quantitative agreement was observed between IMx and EIA and between IMx and RIA with linear correlation of 0.955 and 0.9 13, respectively.

On average the AUSAB RIA gave 14% higher anti-HBs concentrations than IMx whereas the EIA gave anti-HBs values 9% lower than the IMx. Anti-HBs concentra- tions determined by IMx and RIA were within twofold of each other in greater than 90% of anti-HBs reactive specimens.

Hepatitis B vaccinees

Sera from eighty-five individuals previously vaccinated with Heptavax or Recom- bivax HB (n = 74) (Merck, Sharp and Dohme), Engerix B (n = 6) (Smith Kline) or Hevac B (n = 5) (Pasteur) were tested by IMx, RIA and EIA. One hundred percent agreement was observed among assays and 8 1(95.3%) were reactive for anti-HBs.

Sequential sera obtained from vaccinees bled at varying times during and after vac- cination were tested by IMx AUSAB and AUSAB RIA. Three representative patterns of immune response to hepatitis B vaccine are shown in Fig. 3. Good quantitative

- Vaccinee I-IMx

- Vaccinee 1 -RIA

- Vaccine? 2.IMx

- Vaccinee 2.RIA

-Tr- Vaccinee 3.IMx

- Vaccinee 3.RIA

1 0 2 4 6 8 10 12 14

Months After First Vaccine Dose

Fig. 3. Sequential bleeds from three vaccinees were tested by IMx AUSAB. Inoculation of Heptavax occurred at 0, 1, and 6 months.

274

agreement between IMx and RIA results was observed for all patients. Vaccinee 1 showed a strong response to the vaccine even after the first dose and had peak anti- HBs concentrations of greater than 60 000 mIU/ml one month after the last booster dose. Vaccinees 2 and 3 had detectable anti-HBs responses only after the second vac- cine dose with peak anti-HBs levels rising to 1000 and 100 mIU/ml, respectively. A significant decay was observed in anti-HBs concentrations in some individuals, (e.g. vaccinee 3) in which anti-HBs levels dropped more than lo-fold in the year following the final booster dose.

Discussion

A fully automated immunoassay for the detection and quantitation of anti-HBs is described for the IMx instrument. The antigen sandwich assay, IMx AUSAB, differs from current assays in using coated microparticles in place of polystyrene beads as a solid phase to increase the kinetics of the reaction. The IMx automatically calculates anti-HBs concentration for a specimen by comparing the specimen rate value to a stored standard curve. Furthermore, the IMx AUSAB uses recombinant HBsAg (Mimms et al., 1989) rather than human plasma derived HBsAg used in current com- mercial bead assays.

Overall qualitative agreement of 97.8% (1265/l 293) and 99.1% (128 l/l 293) was observed between IMx and RIA and IMx and EIA, respectively.

Most discrepant specimens were low level reactive (< 10 mIU/ml) by AUSAB RIA but negative by IMx and EIA and also negative for antibody against hepatitis B core antigen (anti-HBc) suggesting that these RIA reactives may not result from prior HBV infection. These discordants may result from the presence of low level anti-HBs in vaccinees or from IgM or IgA anti-HBs due to unspecific activation of immature B lymphocytes independent of exposure to HBV as has been suggested for some anti- HBc reactive individuals (Sallberg and Magnius, 1989). In this regard, Kessler et al. (1985) have shown that some low level AUSAB RIA specimens become nonreactive after treatment with reductant under conditions expected to destroy IgM reactivity. In the absence of a reliable confirmatory procedure for these Iow level AUSAB RIA reactives, it was not possible to determine whether the reactivity resulted from a non- specific binding component. Furthermore, individuals with low AUSAB RIA reac- tivity (< 10 mIU/ml) may not have protective levels of anti-HBs (Hadler et al., 1986).

On the other hand, three of 6 specimens that were reactive at a low level by IMx AUSAB but negative by AUSAB RIA and EIA were also reactive for anti-HBc, indi- cating probable previous exposure to hepatitis B. IMx AUSAB was apparently more sensitive than AUSAB RIA for these specimens.

Overall good quantitative agreement was observed between IMx and RIA and IMx and EIA with linear correlation coefficients of 0.9 1 and 0.96, respectively. Fewer than 10% of all specimens showed quantitative difference between IMx and RIA of 2-fold or greater. Finding quantitative differences between assays was not surprising since the human polyclonal response to HBsAg is complex, and sera contain a mixture of anti-HBs with varied affinity and epitope reactivity. Some investigators have claimed

275

that more than 12 distinct discontinuous epitopes exist on HBsAg (Pillot et al., 1990). Furthermore, human polyclonal sera have been used to identify at least 9 different HBV subtypes (Courouce et al., 1976). Subtyping of anti-HBs response into anti a, d, or y did not reveal any clear subtype-dependent correlation in anti-HBs detectability among assays (data not shown).

These data taken together indicate that IMx AUSAB has equivalent sensitivity to current anti-HBs bead assays and gives good quantitative and qualitative agreement with EIA and RIA assays. The IMx assay is useful for determining anti-HBs levels in immunized and HBV-infected individuals.

Acknowledgements

We thank Kim Eble and David Vallari for a critical reading of the manuscript and Laurie Metz for all typing.

References

Barker, L.F., Gerety, R.J., Lorenz, D.E., Rastogi, SC. and Seligman, E.B. (1978) Biological Standardiza- tion in Viral hepatitis. In: G. Vyas, S. Cohen, and R. Schmid (Eds), Viral Hepatitis, pp. 58 l-587.

Courouce, A.M., Holland, P., Miller, J. and Soulier, J. (1976). HBs antigen subtypes. Bibl. Haematol. 42, 31-41.

Courouce, A.M. (1990) Quantification of post-vaccination humoral immunity. In: P. Coursaget and M.J. Tong (Eds), Progress in Hepatitis B Immunization, pp. 67-72.

Coursaget, P., Chotard, J., Vincelot, P., Diop-Mar, I., Yvonnet, B., Sarr, M., N’doye, R. and Chiron, J. (1986) Seven year study of hepatitis B vaccine efficacy in infants from an endemic area (Senegal). Lan- cet ii, 1143-I 145.

Fiore, M., Mitchel, J., Doan, T., Nelson, R., Winter, G., Grandore, C., Zeng, K., Haraden, R., Smith, J., Harris, K., Leszynski, J., Berry, D., Safford, S., Barnes, G., Scholnick, A. and Ludington, K. (1988) The Abbott IMx automated benchtop immunochemistry analyzer system. Clin. Chem. 34.17261732.

Hadler, S., Francis, D., Maynard, J. et al. (1986) Long term immunogenicity and efficacy of hepatitis B vac- cine in homosexual men. N. Engl. J. Med. 315,209%214.

Jilg, W., Schmidt, M. and Deinhardt, F. (1988) Persistence of Specific antibodies after hepatitis B vacci- nation. J. Hepatol. 6,201-207.

Jilg, W., Schmidt, M. and Deinhardt, F. (1990) Hepatitis B vaccination strategy for booster doses in high risk population groups. In: P. Coursaget and M.J. Tong (Eds), Progress in Hepatitis B Immunization, 194,pp.419427.

Kessler, H., Harris, A., Payner, J., Hudson, E., Potkin, B. and Levin, S. (1985) Antibodies to hepatitis B sur- face antigen as the sole hepatitis B marker in hospital personnel. Ann. Intern. Med. 103,21-26.

Kitchen, A.D., Harrison, T.J., Meacock, T.J., Zuckerman, A.J. and Harrison, J.F. (1988) Incidence and sig- nificance of hepatitis B core antibody in a healthy blood donor population. J. Med. Virol. 25,69-75.

Krouwer, J.S. and Rabinowitz, R. (1984) How to improve estimates of imprecision. Clin. Chem. 30,29& 292.

Mimms,L.,Goetze, A., Swanson, S.,Floreani,M.,Edwards, B.,Macioszek, J.,Okasinski,G. andKiang, W. (1989) Second generation assays for the detection of antibody to HBsAg using recombinant DNA- derived HBsAg. J. Virol. Methods 25,21 l-232.

Mushahwar, I. (1987) A biotin/avidin solid-phase sandwich enzyme immunoassay for the antibody to hepa- titis B surface antigen (anti-HBs). J. Virol. Methods 16,1-l 1.

Mushahwar, I. and Spiezia, K. (1987) The utilization of biotin/anti-biotin interaction for the detection of

276

antibody to hepatitis B surface antigen (anti-HBs). J. Virol. Methods I6,45-54. Nakane, P.K. and Kawaoi, A. (1974) Peroxidase labeled antibody: a new method of conjugation. J. Histo-

them. Cytochem. 22,1084-1091. Pillot, J., Badur, S., Bertolotto, C. and Bouige, Ph. (1990) Large diversity of discontinuous epitopes on the

HBsAg molecule. In: P. Coursaget and M. Tong (Eds), Progress in Hepatitis B Immunization, 194, pp. 91-92.

Quash, G., Roth, A., Nivekau. A., Grange, J. Keolouanghahot, T. and Huppert, J. (1978) The preparation of latex particles with covalently bound polyamines, IgG, and measles, agglutinins and their use in viral agglutination tests. J. Immunol. Methods 22,165-l 74.

Sallberg, M. and Magnius, L.O. (1989) Enzyme immunoassay for anti-hepatitis B core (HBc) immunoglob- ulin G, and significance of low level results in competitive assays for anti-HBc. J. Clin. Microbial. 27, 8499853.

Stevens, C.E., Taylor, P.E., Tong, M.J., Toy, P.T. and Vyas, G.N. (1984) Hepatitis B vaccine: an overview. In: G.N. Vyas, J.L. Deinstag and J.H. Hoofnagle (Eds), Viral Hepatitis and Liver Disease, pp. 275-291.