Comparison of four serological methods for the detection of diphtheria anti-toxin antibody

Journal of Immunological Methods 245 (2000) 55–65 www.elsevier.nl / locate / jim

Comparison of four serological methods for the detection of diphtheria anti-toxin antibody J. Walory*, P. Grzesiowski, W. Hryniewicz Department of Immunology and Prevention of Infection, Sera and Vaccines Central Research Laboratory, Warsaw, Poland Received 31 March 2000; received in revised form 30 May 2000; accepted 3 July 2000

Abstract The aim of this study was to compare four serological methods for the detection of Corynebacterium diphtheriae IgG anti-toxin antibodies (IgG–DTAb) in human serum. One hundred serum samples were evaluated for C. diphtheriae IgG–DTAb by four different methods: passive haemagglutination (PHA), latex agglutination test (LA), toxoid enzymelinked immunosorbent assay (Toxoid–ELISA), and toxin-binding inhibition enzyme-linked immunosorbent assay (ToBI– ELISA). As the external standardisation the neutralisation test for C. diphtheriae toxin in Vero cells (TN Vero) was used. For internal standardisation of IgG–DTAb titres, the WHO standard serum of human diphtheria antitoxin was used. The study revealed a poor correlation between the reference test and the PHA (r50.34 Pearson’s correlation coefficient), an acceptable correlation for the LA (r50.74), a good correlation for the Toxoid–ELISA (r50.81) and a very good correlation for ToBI–ELISA (r50.93). The sensitivity measurments of PHA, LA, Toxoid–ELISA and ToBI–ELISA tests, were 14, 100, 94, 96% respectively and the corresponding specificity characteristics were 86, 76, 94, 90 respectively. Of the four evaluated methods, the ToBI–ELISA could be recommended for scientific and precise laboratory assays of diphtheria antibody levels in humans. For screening purposes the Toxoid–ELISA could be used, but the accuracy of antibody titres below 0.1 IU / ml, considered as the limits of protection, is questionable. Both tests offer very useful alternatives to the in vitro diphtheria toxin neutralisation test in Vero cells. Because of their unsatisfactory correlation and sensitivity as compared to the reference method, PHA and LA should be avoided and replaced by one of the two enzyme immunoassays.  2000 Elsevier Science B.V. All rights reserved. Keywords: Corynebacterium diphtheriae; Serological methods; Diphtheria toxin, Diphtheria antitoxin antibody; Seroprotection

1. Introduction Local and systemic toxicity due to Corynebacterium diphtheriae infection is caused by an extracellular toxin with strong antigenic properties. Im*Corresponding author. Present address: Chelmska 30 / 34, 00725 Warsaw, Poland. Tel.: 148-22-841-2949; fax: 148-22-8412949. E-mail address: [email protected] (J. Walory).

munity against the disease depends primarily on the presence of specific IgG antibody (IgG–DTAb) directed against the diphtheria toxin. It is believed that circulating diphtheria specific IgG antibody levels above 0.01 IU / ml determined by neutralisation tests provides seroprotection against the disease. However, neutralisation tests, either in animals or in Vero cells, are time consuming and expensive, and are being replaced in clinical practice by different serological methods (Ipsen, 1946). If serological

0022-1759 / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved. PII: S0022-1759( 00 )00273-8

56

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

techniques are used, the protective level of antibody is defined as equal to or higher than 0.1 IU / ml (Cellesi et al., 1989; Galazka and Kardymowicz, 1989). In many laboratories the most frequently used method for the detection of anti-toxin antibodies is still the haemagglutination test (PHA). A poor correlation between PHA and the reference test, has been reported by several authors and new assays for the detection of diphtheria anti-toxin antibody levels in the population have been introduced. It prompted us to make a comparison between selected methods in order to choose the most reliable assay for clinical and scientific purposes. We have compared four methods for the detection of C. diphtheriae IgG anti-toxin antibodies (IgG– DTAb) in human serum: passive haemagglutination (PHA), a latex test (LA), a toxoid enzyme-linked immunosorbent assay (Toxoid–ELISA) and a toxinbinding inhibition enzyme-linked immunosorbent assay (ToBI–ELISA). As a reference method, a neutralisation test in Vero cells (TN Vero) was used.

2. Materials and methods

2.1. Serum samples and serological methods A total of 100 serum samples from healthy adults, fully vaccinated against diphtheria in childhood, according to the National Vaccination Guidelines, were investigated for diphtheria toxin antibody. After collection, the sera were immediately frozen at 2258C, kept frozen during transport and thawed just prior to testing. In all tests external and internal standardisation was performed. Internal standardisation was based on the reference WHO human diphtheria antitoxin (1.5 IU / ml, standard WHO — National Institute for Biological Standards and control, Hertfordshire, UK), external standardisation included a toxin neutralisation test in Vero cells, as recommended by WHO. Serum IgG–DTAb titres were expressed in International Units per millilitre (IU / ml) according to the diphtheria antitoxin standard. For the clinical interpretation of results, antibody titres were classified into one of the following categories: no protection (,0.1 IU / ml), satisfactory protection (0.1–1.0 IU / ml) and high levels of

protection (.1.0 IU / ml) (Ipsen, 1946; Scheibel et al., 1966; Klouche et al., 1995). Geometric mean titres were used to characterise antibody distributions and differences between correlation and regression tests were compared using Student’s t-test and Fisher exact test with 95% confidence intervals (Bullock and Walls, 1977; Hendriksen et al., 1988; Harlow and Lane, 1988; Nyerges, 1992; Kreeftenberg et al., 1992).

2.2. Toxin-binding inhibition enzyme-linked immunosorbent assay ( ToBI–ELISA) The assay was performed according to the procedure described by Hendriksen et al. (1988, 1989a). The microtitre plates were coated with equine antidiphtheria IgG 1.0 IU / ml in carbonate buffer (pH 9.6). Plates were incubated overnight at 378C in a humid atmosphere, washed three times with phosphate buffered saline (PBS, pH 7.2) containing 0.05% of Tween 80 and unoccupied binding sites blocked with PBS containing 0.5% bovine serum albumin (BSA; Sigma Chemical Company, USA) for 1 h at 378C and washed three times in PBS–Tween 80. After 1.5 h incubation at 378C the plates were washed with PBS–Tween 80 and two-fold dilutions of serum samples prepared in 200 ml quantities were added starting with undiluted serum. Next, aliquots 20 ml of 0.1 Lf / ml diphtheria toxin (115 Lf / ml, Sera and Vaccines Plant, Cracow, Poland) diluted in PBS were added to each well and incubated overnight at 378C in a humid atmosphere. The mixture of toxin and antibody was transferred to the corresponding wells of the antitoxin-coated plates and incubated for 2 h at 378C in a humid atmosphere. After washing, peroxidase labelled rabbit diphtheria specific IgG (0.1 IU / ml, diluted 1 / 3000) was added and incubated for 1.5 h at 378C. The free diphtheria toxin bound to the antitoxin-coated plate was visualised by adding 0.2 mg / ml o-phenylenediamine dihydrochloride (OPD, Sigma Chemical Company, USA) with 30% hydrogen peroxide in 0.05 M citrate buffer pH 5.0. After 20 min incubation at 378C the reaction was stopped by 2 M sulphuric acid. The absorbance was measured at 492 nm using an automatic plate reader (Multiskan Plus, Labsystems, Finland). The absorbance values of the tested serum samples, were compared with the reference serum. The antibody

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

titre was determined by establishing the I 1 of the tested and reference samples. The I 1 dilution was defined as the lowest dilution of the diphtheria antitoxin mixed with 0.1 Lf / ml diphtheria toxin, which incompletely inhibits diphtheria toxin binding to a diphtheria antibody-coated plate. An extinction of 0.1 was used as the arbitrary I 1 level.

2.3. Toxoid enzyme-linked immunosorbent assay ( Toxoid–ELISA) Initially, a comparison between the Toxoid– ELISA and the Toxin–ELISA was performed and indicated an almost exact correlation between both assays (r50.97, P,0.0001). Accordingly, microtitre plates were coated with diphtheria toxoid (3300 Lf / ml, Sera and Vaccines Plant, Cracow, Poland) instead of diphtheria toxin, using a solution containing 2.5 Lf / ml of diphtheria toxoid in 0.05 M carbonate / bicarbonate buffer pH 9.6. The plates were incubated overnight at 378C and washed with PBS pH 7.2 containing 0.05% Tween 20. After blocking the reaction with 1% BSA in PBS for 1 h at 378C, serum samples were diluted 1 / 100 in PBS, containing 0.5% BSA and 0.05% Tween 20, were distributed in triplicate aliquots of 100 ml per-well and incubated for 1 h at 378C. After incubation, peroxidase-labelled rabbit anti-human IgG conjugate (DAKO, Denmark) diluted 1 / 5000 in PBS–BSA– Tween 20 buffer was added. Plates were then incubated for 1 h at 378C and washed with PBS– Tween 20. In the next step, peroxidase substrate (0.2 mg / ml OPD with 30% hydrogen peroxide in citrate buffer pH 5.0) was added to each well. The reaction was stopped after 20 min by adding 2 M H 2 SO 4 . The absorbance values of diluted sera were measured at 492 nm using an automatic multiscan plate reader and compared with the reference WHO standard serum. Only values within the linear part of the standard curve were used to calculate titres. Serum IgG–DTAb titre was expressed in IU / ml according to the diphtheria antitoxin standard.

2.4. Latex agglutination test ( LA) The method according to Martin de Saint et al. (1975) and van Oss and Singer (1966) was modified as follows. Polystyrene latex particles (f 50.74 mm;

57

IKFP PAN, Cracow, Poland), sensitised with diphtheria toxoid (3300 Lf / ml; Sera and Vaccines, Cracow, Poland) were diluted to 40 Lf / ml with 0.05 M carbonate / bicarbonate buffer (pH 9.6). Sensitisation was obtained with purified diphtheria toxoid after incubation for 3 h at 378C with gentle shaking. The serum samples were inactivated for 0.5 h at 568C. A 20 ml aliquot of the latex-sensitised suspension was added to each of serial dilutions of the samples in carbonate / bicarbonate buffer (pH 9.6). After shaking for 3 min, agglutination was observed under bright illumination. The last well with agglutination was taken as the end-point. Antibody titres were expressed in IU / ml according to the reference WHO standard serum.

2.5. Passive haemagglutination ( PHA) The PHA method was performed according to the procedures described by Boyden (1951) and Nyerges (1992). A tannic acid-treated sheep erythrocyte suspension was sensitised by 250 Lf / ml diphtheria toxoid. Serial dilutions of serum samples were prepared in V-shaped microtitre plates according to the method described by Takatsy (1956). Diluted sera and erythrocytes (25 ml) were added to the wells and the plates incubated at room temperature for 1 h. Titration of the diphtheria antitoxin standard was performed with each group of serum samples. Antibody titre was calculated and expressed in IU / ml in comparison with the reference WHO standard serum.

2.6. Reference test As a reference assay, the diphtheria toxin neutralisation test in Vero cells (green monkey renal epithelium) was used. The estimation of diphtheria antitoxin in cell culture was performed according to the procedure described by Kreeftenberg et al. (1985, 1992). Serial two-fold dilutions of serum were mixed with diphtheria toxin (0.0002 Lf) and incubated for 1 h at room temperature. Then 50 ml of suspension containing 2.5310 5 of Vero cells / ml were added to each well. The plates were gently shaken, covered with a polyethylene seal and incubated at 378C for 6 days. Changes in colour from red to yellow indicated

58

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

metabolic inhibition by the toxin due to neutralisation of the toxin by antitoxin antibodies. On the basis of concurrent testing with the WHO reference serum (1.5 IU / ml, National Institute for Biological Standards and Control, Hertfordshire, UK), the antibody titre was expressed in IU / ml.

3. Results From a total of 100 samples used in the study, the reference neutralisation test identified 49 persons as non-protected and 51 as protected.

3.1. Comparison of the ToBI–ELISA with the toxin neutralisation reference test in Vero cells From 49 non-protected persons the ToBI–ELISA incorrectly identified two as protected. This means that for 47 out of a total 100 samples, the ToBI– ELISA test correctly identified the donors as nonprotected. Antibody titres ranged between 0.002 and 9.6 IU / ml, with a mean titre of 0.6861.3 IU / ml (x6standard deviation) (Table 1). The correlation between the ToBI–ELISA and the reference neutralisation test in Vero cells was r50.93 (Pearson correlation coefficient) and the regression line equation

Table 1 Comparison of statistical parameters of methods for the determination of diphtheria antibody levels PHA, LA, Toxoid ELISA, ToBI ELISA test compared to the reference test (the neutralization toxin in Vero cell) Statistical parameters

Numbers of samples (N) Measures of location Arithmetic mean Error arithmetic mean Confidence interval for mean 295% Confidence interval for mean 195% Median Component of variation Mean range x–min. x–max. Standard deviation Variance Intrassay coefficient of variation (CV)

TEST PHA

LA

Toxoid ELISA

ToBI ELISA

NT VERO

100

100

100

100

100

9.68 0.89

0.49 0.09

0.54 0.07

0.68 0.13

0.68 0.13

7.92

0.31

0.40

0.42

0.42

11.43 8.00

0.67 0.03

0.68 0.11

0.94 0.07

0.93 0.15

63.98 0.02 64.00 12.79 163.75

4.80 0.009 4.80 0.91 0.83

3.20 0.005 3.21 0.70 0.503

9.59 0.002 9.6 1.3 1.7

9.59 0.002 9.6 1.29 1.67

132%

184%

130%

190%

190%

Component correlation with NT VERO Correlation r50.34 coefficient (r) p50.00001 Interassay coefficient of variation (CV) CV538%

r50.74 p50.00001

r50.81 p50.00001

r50.93 p50.00001

r51 p50.00001

CV56%

CV560%

CV50%

–

Estimation of accurate tests relative to TN VERO Sensitivity 14% Specificity 86%

100% 76%

94% 94%

90% 96%

100% 100%

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

59

Fig. 1. Comparison of the ToBI ELISA test with reference NT–Vero test.

was: ToBI–ELISA (IU / ml)50.0403810.9412*TN VERO (IU / ml) (Fig. 1). The estimated diagnostic accuracy with reference to the neutralisation test at the break point titre of 0.1 IU / ml expressed as sensitivity and specificity was 96 and 90% respectively (Table 2).

3.2. Comparison of the Toxoid–ELISA with the toxin neutralisation reference test in Vero cells From 49 non-protected persons the Toxoid– ELISA incorrectly identified three as protected. Thus, from a total of 100 samples, the Toxoid– ELISA test correctly identified 46 persons as nonprotected. Antibody titres ranged between 0.005 and 3.21 IU / ml, and the mean titre for all samples was 0.5460.7 IU / ml (Table 1). The correlation between the Toxoid–ELISA and reference neutralisation test in Vero cells expressed as the Pearson correlation

coefficient was r50.81, and the regression line was: Toxoid–ELISA (IU / ml)50.2439910.44419*TN VERO (IU / ml) (Fig. 2). For sera with an antibody titre below 0.1 IU / ml (n549) the correlation coefficient was r50.5 with a regression line Toxoid– ELISA (IU / ml)50.0220910.36224*TN Vero (IU / ml). For sera with an antibody titre above 0.1 IU / ml (n551) the correlation coefficient was r50.74, and the regression line equation was Toxoid–ELISA (IU / ml)50.6097110.32953*TN VERO (IU / ml). The estimated diagnostic accuracy of the Toxoid– ELISA with respect to the reference test at the break point titre of 0.1 IU / ml expressed as sensitivity and specificity was 94% for both (Table 2). Re-evaluation of samples stored at 2258C for 3 months and 8 months, demonstrated a high correlation with primary results of r50.99 and r50.95 respectively. The correlation between the Toxoid–ELISA and the ToBI ELISA test expressed as the Pearson correlation coefficient was also high (r50.79), and the

60

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

Table 2 Estimation of diagnostic accuracy of the ToBI ELISA, Toxoid ELISA, LA and PHA tests at border point titre 0.1 IU / ml ToBI ELISA IU / ml

NT–VERO test IU / ml

.0.1 #0.1 Total Sensitivity Specificity

.0.1 46 5 51 (46 / 51)3100%590.2% (47 / 49)3100%595.9%

#0.1 2 47 49

Total 48 52 100

.0.1 48 3 51 (48 / 51)3100%594.1% (46 / 49)3100%593.9%

#0.1 3 46 49

51 49 100

#0.1 49

.0.1 12 39 51

61 39 100

.0.1 7 44 51

14 86 100

Toxoid ELISA IU / ml .0.1 #0.1 Total Sensitivity Specificity LA IU / ml #0.1 .0.1 Total Sensitivity Specificity

49 (49 / 49)3100%5100% (39 / 51)3100%576.5%

PHA IU / ml #0.1 .0.1 Total Sensitivity Specificity

#0.1 7 42 49 (7 / 49)3100%514% (44 / 51)3100%586%

regression line was: ToBI–ELISA (IU / ml)52 0.117211.4614*Toxoid–ELISA (IU / ml).

3.3. Comparison of the LA with toxin neutralisation reference test in Vero cells The LA identified all non-protected persons correctly. The titre range was between 0.0093 and 4.8 IU / ml, and the mean titre for all samples was 0.4960.91 IU / ml (Table 1). The correlation between the LA and the reference neutralisation test in Vero cells expressed as a Pearson correlation coefficient was r50.74 and the regression line was LA (IU / ml)50.1429210.52769*TN VERO (IU / ml) (Fig. 3). The estimated diagnostic accuracy of the LA with respect to the reference test at the break point titre of

0.1 IU / ml was given by sensitivity and specificity values of 100 and 76% respectively (Table 2).

3.4. Comparison of the PHA with the toxin neutralisation reference test in Vero cells From 49 non-protected persons the PHA incorrectly identified 42 as protected, which means that from total number of 100 samples, the HA test correctly identified seven persons as non-protected. Titres ranged between 0.02 and 64IU / ml, with a mean titre of 9.68612.79 IU / ml. The correlation between the HA and the reference neutralisation test in Vero cells expressed as a Pearson correlation coefficient was r50.34, with a regression line of: HA (IU / ml)5 0.2423410.00819*TN VERO (IU / ml) (Fig. 4). The estimated diagnostic accuracy with respect to the

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

61

Fig. 2. Comparison of Toxoid ELISA test with reference NT–Vero test.

reference test at the break point titre of 0.1 IU / ml was given by sensitivity and specificity values of 14 and 86% respectively (Table 2).

4. Discussion Since routine serological diagnostis of immunity to diphtheria is still based on a variety of different tests, we have compared four different methods for the detection of diphtheria anti-toxin antibody in parallel with the reference diphtheria toxin neutralisation test in Vero cells. Modern immunoassays have shown better correlations with reference to in vivo and in vitro toxin neutralisation tests and immune status of humans than previously used techniques, such as haemagglutination. The selection of the four serological tests and the reference assay used in this study was based on current trends in serodiagnostics. Since Sousa and Evans (Quevillon and Chagnon, 1973) shown in 1957 the high sensitivity and reproducibility of the toxin neutralisation test in Vero cells (Dular, 1993; Gupta and Sigham, 1994a), this assay

has been recommended by WHO as a reference method giving comparable results to toxin neutralisation in cell culture and in guinea pigs. For many reasons, such as assay time and the high costs, the use of reference methods in routine laboratory practice is not practical and the most frequently used in vitro tests are enzyme immunoassays, such as Toxin or Toxoid–ELISA and ToBI–ELISA (Engvall and Perlmann, 1971; Cockeysville, 1973; Miyamura et al., 1974a,b; Bullock and Walls, 1977; Palmer et al., 1983; Kreeftenberg et al., 1985; Kurstak, 1985; Hendriksen et al., 1988; Gupta and Silber, 1994b). However, despite the common usage of such immunoassays, there has been a lack of inter-laboratory comparisons between different tests and the interpretation of results obtained by different enzyme immunoassays because of the utilisation of different procedures, reagents and statistical methods (Bullock and Walls, 1977; Kurstak, 1985; Coulson et al., 1989; Peterman and Butler, 1989; Plikaytis et al., 1991; Reizenstein et al., 1995). In our study, diagnostic accuracy and comparisons between the assays and the reference method were

62

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

Fig. 3. Comparison of latex test (LA) with reference NT–Vero test.

characterised by well defined and generally accepted statistical parameters such as the Pearson coefficient of correlation, intra-assay and inter-assay coefficients of variation, and the sensitivity and specificity of the test at arbitrarily chosen break point titre. The utilisation of these parameters allowed us to perform an objective analysis and comparison of the different methods. High sensitivity and specificity together with the highest correlation with the reference test, was observed using the ToBI–ELISA. From samples with a titre below 0.1 IU / ml, as diagnosed by the reference test, 96% were correctly identified by the ToBI–ELISA. There was no dispersion of differences between the results of ToBI–ELISA and the reference method, which suggests total reproducibility for both tests, despite the fact that, the ToBI– ELISA test showed lower discriminative power. The only practical problem was the relatively high complexity of the procedure with the first phase of neutralisation of the toxin, which was time consum-

ing. This assay could be used in scientific and reference laboratories for precise serological diagnostics and research. Similarly, a high correlation was also obtained for the Toxoid–ELISA, combined with a high accuracy as expressed by sensitivity and specificity. From samples with a titre below 0.1 IU / ml, as diagnosed by reference test, 94% were correctly identified by the Toxoid–ELISA. This method demonstrated a symmetrical distribution of diagnostic accuracy, which is optimal. To exclude differences which resulted from using diphtheria toxin in the reference test and diphtheria toxoid in the Toxoid–ELISA in the first step, a comparative analysis of the Toxin– ELISA and the Toxoid–ELISA was performed. An almost complete correlation between both assays (r50.97) allowed us to use diphtheria toxoid, which is important because of safety reasons. The Toxoid– ELISA test remained reproducible after 3 and 8 months of sample storage suggesting that the assay is not sensitive to conformational changes in immuno-

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

63

Fig. 4. Comparison of haemagglutination test (PHA) with NT–Vero test.

globulins occurring over time. After 8 months of storage, the correlation decreased below 95%. Until recently there were no published data concerning the use of the LA for the determination of diphtheria anti-toxin antibody. Theoretically the LA could serve as a simple, rapid, and inexpensive screening test. However, the relatively high correlation revealed in our study was not associated with a satisfactory level of specificity. For samples with antibody titres below 0.1 IU / ml, as diagnosed by the reference test, the LA identified all of them correctly. It should be stressed, that at the present time, there is no alternative fast and simple assay. With some limitations, it could be recommended in outpatient laboratories with limited resources, in regions requiring rapid diphtheria diagnostis. Additionally, the LA was found to be significantly more sensitive in comparison to the haemagglutination test. Our study showed an unsatisfactory poor sensitivity for the PHA. For samples with titre below 0.1 IU / ml, as identified by the reference test, the PHA correctly diagnosed only 14%. The results we obtained disqualify this test from routine usage.

The classification of protective antibody titres was arbitrarily chosen according to the reference data and trends in clinical practice (Ipsen, 1946; Rose et al., 1997; Jenum et al., 1995). An antibody titre of 0.1 IU / ml as measured by ELISA, LA and PHA is broadly accepted as protective. The enzyme-linked immunosorbent assays are based on binding of an antigen to the microtitre plate. Bacterial toxins and prepared toxoids with a lipophilic moiety coat the plates efficiently (Swenson and Larsen, 1977). If the antibody level is above 0.1 IU / ml, the results of the ELISA correlate strongly with the neutralisation test. However, when the antibody titre is below 0.1 IU / ml, there is a significant decrease in the correlation between the ELISA and the neutralisation test. A titre of 0.001–0.01 IU / ml measured by the neutralisation test, could be 10–100 times higher if measured by ELISA (Knight et al., 1986; Melville-Smith and Balfour, 1988). According to the correlation characteristics, the results of the Toxoid–ELISA could be accepted above the 0.1 IU / ml level and defined as protective. Over this value, the highest correlation and the most accurate results were ob-

64

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65

tained, as observed by other investigators (Galazka and Kardymowicz, 1989; Gupta et al., 1996; Uchimura et al., 1990; World Health Organization, 1990). In neutralisation tests, either in animals or in Vero cells, the defined protection level of antibody is tenfold lower (Lyng, 1988). Generally, as was confirmed by Simonsen et al. (1986), ELISA can be used as a screening test for populations, but samples with antibody levels below 0.1 IU / ml should be re-determined by one of the neutralisation tests. The high performance and accuracy of the ToBI– ELISA has also been observed by other investigators (Hendriksen et al., 1988, 1989a,b). It should be stressed, that the intra-assay coefficient of variation was 0% and there was no difference between the dispersion of results obtained by the ToBI–ELISA and the neutralisation test in Vero cells, which implies a high reproducibility for both methods. Accordingly, the ToBI–ELISA could be recommended as a single test for precise diphtheria antitoxin antibody determinations in medical diagnostics and scientific investigations. However, the procedure is too complicated to be utilised for screening purposes.

References Boyden, S.V., 1951. The adsorption of proteins on erythrocytes treated with tannic acid and subsequent haemagglutination by antiprotein sera. J. Exp. Med. 93, 107. Bullock, S.L., Walls, K.W., 1977. Evaluation of the parameters of the enzyme-linked immunospecific assay. J. Infect. Dis. 136, 279. Cellesi, C. et al., 1989. Immunity to diphtheria in a sample of adult population from central Italy. Vaccine 7, 417. Cockeysville, M.D., 1973. BBL Manual of Products and Laboratory Procedures, 5th Edition. Becton Dickinson Microbiology Systems. Coulson, B.S., Grimwood, K., Bishop, R.F., Barnes, G.L., 1989. Evaluation of end-point titration, single dilution and capture enzyme immunoassays for measurement of antirotaviral IgA and IgM in infantile secretions and serum. J. Virol. Meth. 53, 53. Dular, U., 1993. Comparative studies of the in vivo toxin neutralization and the in vitro Vero cell assay methods for use in potency testing of diphtheria component in combined vaccines / toxoids. 1: Standardization of a modified Vero cell assay for toxin–antitoxin titration of immunized guinea-pig sera. Biologicals 21 (1), 53. Engvall, E., Perlmann, P., 1971. Enzyme-linked immunosorbent

assay (ELISA). Quantitative assay of immunoglobulin G. Immunochemistry 8, 871. Galazka, A., Kardymowicz, B., 1989. Immunity against diphtheria in adults in Poland. Epidemiol. Infect. 103, 287. Gupta, R.K., Sigham, S., 1994a. Suitability of the Vero cell method for titration of diphtheria antitoxin in the United States potency test for diphtheria toxoid. Biologicals 22 (1), 65. Gupta, R.K., Silber, G.R., 1994b. Comparative analysis of tetanus antitoxin titers of sera from immunized guinea pigs determined by toxin neutralization test and enzyme-linked immunosorbent assay. Biologicals 22 (3), 215. Gupta, R.K., Griffin, P.J., Xu, J., Rivera, R., Thompson, C., Siber, G.R., 1996. Diphtheria antitoxin levels in US blood and plasma donors. J. Infect. Dis. 173 (6), 1493. Harlow, E., Lane, D., 1988. Antibodies: A Laboratory Manual Cold Spring Harbor Laboratory. Cold Spring Harbor, New York, USA. Hendriksen, C.F.M., van der Gun, J.W., Nagel, J., Kreeftenberg, J.G., 1988. The toxin binding inhibition test as a reliable in vitro alternative to the toxin neutralization test in mice for the estimation of tetanus antitoxin in human sera. J. Biol. Stand. 16, 287. Hendriksen, C.F.M., van der Gun, J.W., Kreeftenberg, J.G., 1989a. Combined estimation of tetanus and diphtheria antitoxin in human sera by the in vitro Toxin-Binding Inhibition (ToBI) test. J. Biol. Stand. 17, 191. Hendriksen, C.F.M., van der Gun, J.W., Kreeftenberg, J.G., 1989b. The use of the Toxin Binding Inhibition (ToBI) test for the estimation of the potency of the diphtheria component of vaccines. J. Biol. Stand. 17, 241. Ipsen, J., 1946. Circulating antitoxin at the onset of diphtheria in 425 patients. J. Immunol. 54, 325. Jenum, P.A., Skogen, V., Danilova, E., Esklid, A., Sjursen, H., 1995. Immunity to diphtheria in Northern Norway and Northwestern Russia. Eur. J. Clin. Microbiol. Infect. Dis. 14, 794. Klouche, M., Luhmann, D., Kirchner, H., 1995. Low prevalence of diphtheria antitoxin in children and adults in Northern Germany. Eur. J. Microbiol. Infect. Dis. 14, 682. Knight, P.A., Tilleray, J., Queminet, J., 1986. Studies on the correlation of a range of immunoassays for diphtheria antitoxin with the guinea-pig intradermal test. Dev. Biol. Stand. 64, 25. Kreeftenberg, J.G., van der Gun, J., Marsman, I.R., Sekhuis, V.M., Bhandari, S.K., Maheshwari, S.C., 1985. An investigation of a mouse model to estimate the potency of diphtheria component in vaccines. J. Biol. Stand. 13, 229. Kreeftenberg, J.G., van der Gun, J., Hendriksen, C.F.M., 1992, Potency determination in mice of diphtheria toxoid in vaccines by serum neutralization of diphtheria toxin in Vero cell cultures. Laboratory methods for the testing for potency of diphtheria (D), tetanus (T), pertussis (P) and combined vaccines. World Health Organization, Geneva, BLG / 92.1, 23. Kurstak, E., 1985. Progress in enzyme immunoassay. Production of reagents, experimental design and interpretation. Bull. World Health Organiz. 63, 793. Lyng, J., 1988, Toxoids as international reference materials defining LF-units for diphtheria and tetanus toxoids. Report WHO, Geneva, 11–19 Oct.

J. Walory et al. / Journal of Immunological Methods 245 (2000) 55 – 65 Martin de Saint, J., Eyquem, A., Turpin, A., Bizzini, B., 1975. Titration of antibodies to tetanus toxoid by agglutination of purified tetanus sensitized latex particles. Brief Raports. Vox Sang. 28, 238. Melville-Smith, M., Balfour, A., 1988. Estimation of corynebacterium diphtheriae antitoxin in human sera: a comparison of an enzyme-linked immunosorbent assay with the toxin neutralization test. J. Med. Microbiol. 25, 279. Miyamura, K. et al., 1974a. Micro cell culture method for determination of diphtheria toxin and antitoxin titers using Vero cell. I. Studies on factors affecting the toxin and antitoxin titration. J. Biol. Stand. 2, 189. Miyamura, K. et al., 1974b. Micro cell culture method for determination of diphtheria toxin and antitoxin titers using VERO cells. II. Comparison with the rabbit skin method and practical application for seroepidemiological studies. J. Biol. Stand. 2, 203. Nyerges, G., 1992. Passive haemagglutination titration as a secondary method for assessing the antibody response of mice in potency assays of adsorbed diphtheria, tetanus and combined vaccines. World Health Organization, Geneva, BLG / 92.1, 57. van Oss, C.J., Singer, J.M., 1966. The binding of immune globulins and other proteins by polystyrene latex particles. Res. J. Reticuloendothelial Soc. 3, 29. Palmer, S.R., Balfour, A.H., Jephcott, A.E., 1983. Immunisation of adults during outbreak of diphtheria. Br. Med. J. 286, 624. Peterman, H.P., Butler, J.B., 1989. Application of theoretical considerations to the analysis of ELISA data. Biotechniques 7, 608. Plikaytis, B.D., Turner, S.H., Gheesling, L.L., Carlone, G.M., 1991. Comparisons of standard curve fitting methods to

65

qantitate Neisseria Meningitidis group A polysacharide antibody levels by enzyme linked immunosorbent assay. J. Clin. Microbiol. 29, 1439. Quevillon, M., Chagnon, A., 1973. Microtissue culture test for the titration of low concentrations of diphtheria antitoxin in minimal amounts of human sera. Appl. Microbiol. 25 (1), 1. Reizenstein, E., Hallander, H.O., Blackwelder, W.C., Kuhn, I., Ljungman, M., Mollby, R., 1995. Comparison of five calculation modes for antibody ELISA procedures using pertussis serology as a model. J. Immunol. Methods 183 (2), 279. Rose, N.R., Macario, E.C., Folds, J.D., Lane, H.C., Nakamura, R.M., 1997. Mannual of Clinical Laboratory Immunology, 5th Edition. American Society for Microbiology Press, DC 20005, Washington. Scheibel, I., Bentzon, M.W., Christensen, P.E., Biering, A., 1966. Duration of immunity to diphtheria and tetanus after active immunization. Acta Path. Microbiol. Scand. 67, 380. Simonsen, O., Bentzon, M.W., Heron, I., 1986. ELISA for the routine determination of antitoxic immunity to tetanus. J. Biol. Stand. 14, 231. Swenson, S.B., Larsen, K., 1977. An enzyme-linked immunosorbent assay (ELISA) for the determination of diphtheria toxin antibodies. J. Immunol. Methods 17, 249. Takatsy, G., 1956. The use of spiral loops in serological and virological micromethods. Acta Microbiol. Hung. 3, 191. Uchimura, M., Takagi, K., Kitayama, A., Yazaki, H., Horiuchi, Y., 1990. Microtiter enzyme-linked immunosorbent assay for diphtheria antitoxin in human sera-application of parallel line assay method. Kansenshogaku Zasshi 64 (8), 967. World Health Organization, 1990, Requirements for Diphtheria, Tetanus, Pertussis and Combined Vaccines. WHO Technical Report Series No. 800, Annex 2, Geneva.