Serodiagnosis of Tuberculosis. Evaluation of a Sulpholipid Antigen

Zbl. Bakt. 278, 49-57 (1993) © Gustav Fischer Verlag, StuttgartlNew York

Serodiagnosis of Tuberculosis. Evaluation of a Sulpholipid Antigen COLETTE SAVAGE, PASCAL VINCENT, and HENRI LECLERC Laboratoire de Bacteriologie, CHU - Faculte de Medecine, Lille, France

With 1 Figure· Received December 30,1991 . Revision received August 7,1992 . Accepted September 16, 1992

Summary In order to evaluate the suitability of serology in the diagnosis of tuberculosis the Mycobacterium tuberculosis anti-sulpholipidic IV IgG were assayed, using the ELISA test, in 41 healthy controls, 121 hospitalized but non-tuberculous patients and 142 tuberculous persons including 46 cured patients. At the 150 and 300 O.D. thresholds proposed in previous studies using the same antigen, the specificity was 88.27% and 98.76%, respectively, and the sensitivity 25% and 11.45%.

These results were not related to different potential factors (age, background, associated pathology) but to infection intensity which enhanced the IgG reaction. This may explain discrepancies observed in previous studies made under other epidemiological conditions. Therefore, in this European population from a universitarity hospital, it appears that too many tuberculous patients have such a low IgG reaction that serology cannot be an effective assay in the diagnostic step.

Zusammenfassung Zur Beurteilung des Wertes der Serologie in der Tuberkulosediagnostik wurden mit Hilfe des ELISA IgG-Antikorper gegen das M. tuberculosis Sulfolipid IVan 41 gesunden Kontrollpersonen, 121 Krankenhauspatienten ohne Tuberkulose und 142 Tuberkulosepatienten (davon 46 geheilt) untersucht. Bei Schwellenwerten ftir die optische Dichte (O.D.) von 150 and 300, wie sie in frtiheren Untersuchungen unter Verwendung des gleichen Antigens vorgeschlagen worden waren, betmg die Spezifitat 88,27 bzw. 98,76% und die Empfindlichkeit 25 bzw. 11.45%. Diese Ergebnisse sind nicht an die verschiedenen moglichen Einfluggrogen, wie Alter, Umfeld und Begleitkrankheiten gebunden, sondern an die Intensitat der Infektion, nach der sich das Ausmag der serologischen Reaktion richtet. Damit lassen sich Diskrepanzen zu vorangegangenen Untersuchungen mit anderen epidemiologischen Begleitumstanden erklaren. Es ergab sich, dag bei einer Stichprobe aus einer europaischen Universitatsklinik zuviele Tuberkulosepatienten eine zu schwache humorale Reaktion zeigen, als dag die serologische Untersuchung eine effiziente Rolle in der Diagnostik spielen konnte. 4 Zbl. Bakt. 278/1

50

C. Savage, P. Vincent, and H. Leclerc

Bacteriological diagnosis of tuberculosis either lacks sensitivity (direct examination) or is rather slow (culture). So a great interest has been shown in serological diagnosis. Beside its rapidity, serology would present a greater advantage in extra-pulmonary tuberculosis and in infantile tuberculosis where sampling may be difficult. Although its application appears to be theoretically limited due to the fact that antituberculous immunity is cell-mediated (the macrophages activated by the sensitized-T-Iymphocyte playing an essential role), various tests have been tried as far back as in 1898 (1,5, 7). Recently, different immunological methods have been studied experimentally: agglutination (32, 33, 38) immunofluorescence (27), radioimmunoassay (34) and especially since the early 1980s, the enzyme-linked immunosorbent assay (42). Following the use and description of antigens in Daniel and Debanne's review (12), lipid fractions (16,22,37) have recently raised a very great interest. The immuno-reactivity of several glycolipid fractions described by Papa et a1. (36, 40) has already been assessed. Two sulpholipid antigens, one (SL-I) described by Goren (18, 19,20,31), the other (SL-IV) by Papa (35,36) and Daffe (11), have been studied in tuberculosis patients by Cruaud (9,10). It has been the purpose of our work, following Cruaud's initial results (10) to evaluate the SL-IV sulpholipid antigen known to be more reactive than the SL-I fraction (9) in a greater number of hospitalized patients, including those with pulmonary and extrapulmonary tuberculosis. IgG detection is the sole criterium, since various authors (2, 23, 31, 9) have shown that the IgG response is greater than the IgM response.

Material and Methods

Sera. The antigen response study included 41 blood donors and 263 patients from a University teaching hospital, with sera collected between May 1988 and September 1989. All the samples were stored frozen at -20°C. The studied population can be divided into 4 groups; 1) Healthy subjects: 41 blood donors. Most were young medical students without any hospital contact; the others were members of the administrative staff. 2) Non-tuberculous patients: 121 patients hospitalized for another disease. Among these 121 patients, 4 presented M. kansasii pulmonary infections, and 1 child an M. avium intracellulare lymph-node invasion. 3) Tuberculous patients: 96 patients, whose bacteriological or clinical or radiological diagnosis had been established before or during hospitalization. These patients received no specific treatment at the time of the assay. Two-thirds of them presented pulmonary tuberculosis, including 12 with an additional extrapulmonary localization, the remaining third presented a strictly extrapulmonary tuberculosis: lymph-node (9 cases), osteoarticular (4 cases), meningeal (7 cases), or other types (16 cases) such as pericardial, pleural, perineal, urogenital ... For 19 tuberculous patients, sera samples were taken on several occasions, in order to follow the development of antibody levels during treatment. 4) Former tuberculous patients: 46 cured tuberculous patients, whose treatment had stopped for at least one year. During our study, they presented no signs of tuberculous infection. , The characteristics of these groups were: mean age 24.4 for the first group, 49.8 for nontuberculous patients, 52.4 for tuberculous patients and 56.9 for the former tuberculous. The sex ratio MIF was: 17/24, 81140, 69/27, and 35/11, respectively. Associated diseases were known for all patients. Antigen. The antigen used was the sulpholipid IV fraction of Mycobacterium tuberculosis

Serodiagnosis of Tuberculosis

51

CIPT 1400010059 (Tuberculosis Collection, Pasteur Institute, Paris) isolated and purified by Papa (11) at the Reference Centre for Mycobacteria at the Pasteur Institute, Paris. Method. ELISA was performed as previously described by Torgal-Garcia (40), Papa (36) and adapted for the SL-IV antigen by Cruaud (9). The batch assays were carried out in the chronological order of their receipt all groups together. All sera were tested at a dilution of 11250. The results were expressed in 1000 x OD (optical density) measured at 414 nm. Results The antibody levels were higher in tuberculous patients than in the sick and healthy controls (median = 105, 60 and 46 respectively), while the formerly tuberculous patients on an average presented intermediate levels (median = 76). However, the antibody titers showed a great dispersion in each group, with maximum heterogenicity in tuberculous patients. In these tuberculous patients, both high responses and low levels as in the control group were found (Tables 1,2 and Fig. 1). For a 100% specificity (no false-negatives), the threshold of 499 gave 5% sensitivity to the test. Taking into consideration the sample composition of this study, the probability of a person being non-tuberculous and having an antibody titre below this threshold (negative predictive value) was 0.52. Among tuberculous patients, 19 underwent several antibody assays during treatment. At 1 month after the beginning of the treatment, the assay was performed in 11 patients. The antibody titre was either increased (3 cases) or decreased (8 cases): mean increase = 0.97 ± 0.6, range = 0.57-2.52. At the end of the follow-up period (varying from 2 to 12 months according to the patient), the antibody titre was lower (it decreased in 13 cases among 19 patients): increase = 0.82 ± 0.56, range = 0.19-2.03. All tuberculous patients with a high titre (more than the 150 threshold) before treatment (5 cases) had a low titre « 150) at the end of the follow-up period. In healthy patients with a positive tuberculin test during the previous year (37/41), the antibody level ranged between 6 and 186 (median 46) irrespective of whether it was a recent positive tuberculin test or not. For BCG-vaccinated patients (38/41), the serological results varied between 4 and 78 (median 41). In the group of non-tuberculous patients, two results were high (O.D. 472 and 499). They corresponded to aspergillosis and cholestasis with inflammatory syndrome in a diabetic. None of these 2 patients had a previous history of tuberculosis and neither clinical nor radiological signs of tuberculosis appeared during the following year.

Table 1. Serodiagnosis in the different groups: 1000 x O.D. 414 nm values for SL-IV Ag

Healthy individuals Non-tuberculous patients Tuberculous patients pulmonary extra-pulmonary Cured tuberculous patients

O.D. Values

Number of trials

Median

Range

41 121 96 61 35 46

46 60 105 120 84 76

4- 186 0- 499 0-1700 0-1700 0- 280 2- 320

C. Savage, P. Vincent, and H. Leclerc

52 ODX 1000

414nm

700

· ·

i

·

700

600

··

500

~oo

...

100

0

·

· · ..· ·· ..·

300

200

· ·

.. · · · ..... ....... .... ........ ....... ....

HEALTIIY

CONIROLS 41

....·

. ...· ......

.......... ............ ..•.... •••••........•.. .......... •.•••......•....

....... .... ..... .... ....... ....... ......

......

· .... .... ..... ...... .......... ILL

PATIENTS 121

· ......

... ......

.

Spec. 98.76% Sens. 11,45 %

..

..

.......... ..... ....... .... .. ..... ......

Spec. 88.27 % Sens . 25%

.~

WBERCULOUS FORMER WBERCULOUS

PATIENTS 96

PATIENTS 46

Fig. 1. Distribution of ELISA titres in the different groups (each point indicates one individual).

Serodiagnosis of Tuberculosis

53

If these patients were really some non-diagnosed tuberculosis cases, their exclusion from the control group would reduce the threshold to 266 and would raise the test sensitivity to 15% (at a specificity of 100%) which would still be too low. For the 5 patients presenting M. kansasii pulmonary mycobacteriosis or M. avium intracellulare lymph node mycobacteriosis, serology results ranged between 5 and 220. Discussion One of the aims of the tuberculosis serodiagnosis was to improve the sensitivity of direct diagnosis, whilst maintaining maximal specificity. This result which would be interesting in pulmonary tuberculosis, would be even more so in the extra pulmonary forms, where direct examination is less often positive. In our study, at the 150 and 300 tresholds, as proposed by Cruaud (9), the specificity was 88.27% and 98.76%, respecitvely, and the sensitivity was 25% and 11.45% respectively. This indicates that the method is of little effective value (too many tuberculous patients having a low serological response). Our results do not confirm those recently obtained by Cruaud (10) giving for thresholds of 150 and 300 a specificity of 88 and 100% and a sensitivity of 74.2 and 51.6%, respectively. With serology as a diagnostic step in the laboratory, we could study its additional applicability in the first day diagnosis. The direct examination was positive in 20% of the cases of extrapulmonary tuberculosis and in 50.8% of pulmonary tuberculosis. In the tuberculous patients, serology was less often positive when the sample contained few bacteria (corresponding to a negative direct examination): 20.6% (12/58) versus 52.6% (20/38) at the 150 threshold and 3.4% (2/58) versus 23% (9/38) at the 300 threshold (Table 2). Consequently, in the diagnosis of the difficult forms, especially extrapulmonary cases of tuberculosis with negative direct examination, the contribution of serology remained limited (4 cases out of 28). Overall, the slight interest in serology is due to poor sensitivity (great number of false-negatives). Several hypotheses have been put forward to explain such falsenegatives. For some, the constant release of antigens from the infection site may lead to a chronic stimulation and a massive production of antibodies of low avidity, explaining the hypergammaglobulinemia observed during the illness, and the slight reactivity of certain sera (3). Other authors (4, 6, 8, 29, 39) put foward the possible formation of circulating immune complexes. In our study, as previously noted by Daniel et al. (13), the evolution of the antibodies during antimicrobial therapy showed no increase due to an hypothetical denaturation of such complexes. Likewise, it is a known fact that there exists an individual sensitivity to mycobacterial infections. According to Kaplan (25), an increase of suppressor T-Iymphocytes could be observed during acute bacillosis which would explain, by cell cooperation, the decrease in B-Iymphocyte stimulation and the absence of antibody production. This anarchic immune regulation has been the subject of many explanatory hypotheses (15, 16, 17, 21, 30). Others (3, 14), including Daniel, have put forward the hypothesis of the existence of a "spectrum of immune reactivity" of tuberculosis, according to which healthy adults, probably exposed to tuberculosis bacteria in childhood, possess a well-developed cell-mediated immunity and a very low humoral response, whilst on the contrary the tuberculous patients (miliary tuberculosis) whose cell-mediated immunity has dropped, possess a well-developed humoral response. Between these two extremes, there would be patients presenting localised tuberculosis whose cell immunity is almost normal and humoral response intermediate to the previous cases.

54

C. Savage, P. Vincent, and H. Leclerc

Table 2. Relation between serodiagnosis and bacteriological examinations for the patients with tuberculosis

Direct Examination Cultures Pulmonary tuberculosis

Extra-pulmonary tuberculosis

Whole group

Negative n = 30

Serology (1000 x O.D. at 414 nm) Threshold = 150 Threshold = 300 < 150

>150

<300

>300

Negative n = 7

6

1

7

0

Positive n = 23

16

7

21

1

Positive n = 31

Positive n = 31

14

17

22

9

Negative n = 28

Negative n = 28

14

1

15

0

Positive n = 13

10

3

13

0

Positive n=7

Positive n=7

4

3

7

0

Negative n = 58

Negative n = 22

20

2

22

0

Positive N = 36

26

10

34

2

Positive n = 38

18

20

29

9

Positive n = 38

Another reason of the low sensitivity of serodiagnostic tests in tuberculous patients is the crossreactivity of these test antigens (21). The higher the crossreactivity of the antigen the higher are the antibody titers and the mean antibody level of the patients without tuberculosis. The antibody levels of this control group overlapped with the antibody levels of the tuberculous patients group in a wider range resulting in a higher cut-off point and a lower sensitivity. It is important to detect whether some factors, particular in our study, could explain the low antibody level in the case of tuberculosis. For the tuberculous patients, age, which might be accompanied by antibody decrease, was significantly higher than that of the healthy controls, but close to that of the other non-tuberculous patients (mean age 52.4, 24.4, 49.8 years respectively). So the weak antibody responses observed in tuberculous patients could not be explained by an age-confounding effect. For the same reason, the notion of predisposition could not be retained as a potential bias factor. The associated diseases were alcoholism, diabetes, inflammatory manifestations, cancer, hemopathy, bacterial or viral or parasitic infections, autoimmune disease, renal insufficiency. The proportion of immuno-depressed patients among the tuberculous patients was lower than among the non-tuberculous controls. (25196 and 73/121,

Serodiagnosis of Tuberculosis

55

respectively). Some diseases are indeed accompanied by a decrease in antibodies but they were not observed significantly more often among the number of tuberculous patients. In this last group, the antibody titres for the 14 alcoholic patients had a median rate of 117.5 versus 56 for the 18 alcoholic non-tuberculous patients. A similar result was obtained for the cases of hemopathy: median rate of antibody 76.5 for the 22 cases in the tuberculous group versus 40 for the 11 non-tuberculous controls. Thus they cannot explain an excess of false-negative serological results. Finally, a positive tuberculin test or BCG vaccination in the controls did not influence antibody levels. These observations were in good agreement with those of other authors (2, 13,24, 26, 28). In the absence of obvious bias factors, the lower sensitivity in our study, compared with the previously reported one, could be explained by the patient-sampling. As shown in Table 2, immunological response is related to the infection intensity (bacterial density of the sample). Actually, the false negatives increase with the number of samples with low bacterial counts. The results are thus related to the composition of the studied population. Hence, the disparity in relation to the Cruaud studies (using the same SL-IV antigen and the same methodolgy) could be explained by the differences in patient contamination intensity related to the spread of infection in the countries under study.

Conclusion In this study, the results obtained call into question the immunoreactivity of the antigen fractions used, despite the progress achieved over the last few years and the particularly sensitive immuno-enzymological methods. Taking into account the anarchical nature of the humoral type immune response during tuberculous infection, the sensitivity of the test is very insufficient, at least in hospital populations similar to those encountered in European areas. Acknowledgments. We wish to thank Doctors H. David, F. Papa, P. Cruaud (Unite de la Tuberculose et des Mycobacteries de L'Institut Pasteur, Paris) for providing the antigen and for their advice.

References

1. Arloing, S.: Agglutination du bacille de la tuberculose vraie. C.R. Acad. Sci. 126 (1898) 1398-1400 2. Barrera, L., I. Miceli, V. Ritacco et al.: Detection of circulating antibodies to purified protein derivative by enzyme-linked immunosorbent assay: its potential for the rapid diagnosis of tuberculosis. Pediatr. Infect. Dis. 8 (1989) 763-767 3. Bhatnagar, R., A. N. Malaviya, S. Narayanan, P. Rajgopalan, R. Kumar, and O. P. Bharadwaj: Spectrum of immune response abnormalities in different clinical forms of tuberculosis. Am. Rev. Resp. Dis. 115 (1977) 207-212 4. Bhattacharya, A., S. N. Ranadive, M. Kale, and S. Bhattacharya: Antibody-based enzyme-linked immunosorbent assay for determination of immune complexes in clinical tuberculosis. Am. Rev. Resp. Dis. 134 (1986) 205-209 5. Bordet, j. and O. Gengou: Les sensibilisatrices du bacille tuberculeux. C.R. Acad. Sci 137 (1903) 351-353

56

C. Savage, P. Vincent, and H. Leclerc

6. Brostoff, J., L. Lenzini, P. Rottoli, and L. Rotto/i: Immune complexes in the spectrum of tuberculosis. Tubercle 62 (1981) 169-173 7. Calmette, A. and L. Massol: Sur les conditions d'obtention de la reaction de deviation de l'alexine avec les antigenes et les anticorps tuberculeux. C.R. Soc. BioI. 67 (1909) 528-530 8. Carr, R. 1., A. K. Chakraborty, M. J. Brunda et al.: Immune complexes and antibodies to BCG in sera from patients with mycobacterial infections. Clin. Exp. Immunol. 39 (1980) 562-569 9. Cruaud, P., C. Berlie, J. Torgal Garcia, F. Papa, and H. L. David: Human IgG antibodies immunoreacting with specific sulfolipids from Mycobacterium tuberculosis. Zbl. Bakt. 271 (1989) 481-485 10. Cruaud, P., J. T. Yamashita, N. Martin Casabona, F. Papa, and H. L. David: Evaluation of a noveI2,3-diacyl-trehalose-2'-sulphate (SL IV) antigen for case finding and diagnosis of leprosy and tuberculosis. Res. Microbiol. 141 (1990) 679-694 11. Daffe, M., F. Papa, A. Laszlo, and H. L. David: Glycolipids of recent clinical isolates of Mycobacterium tuberculosis: chemical characterization and immunoreactivity. J. Gen. Microbiol. 135 (1989) 2759-2766 12. Daniel, T. M. and S. M. Debanne: The serodiagnostic of tuberculosis and other mycobacterial diseases by enzyme-linked immunosorbent assay. Am. Rev. Resp. Dis. 135 (1987) 1137-1151 13. Daniel, T. M., S. M. Debanne, and F. Vanderkuyp: Enzyme-linked immunosorbent assay using Mycbacterium tuberculosis antigen 5 and PPD for the serodiagnosis of tuberculosis. Chest 88 (1985) 388-392 14. Daniel, T. M., M. J. Oxtoby, E. Pinto, and E. S. Moreno: The immune spectrum in patients with pulmonary tuberculosis. Am. Rev. Resp. Dis. 1 (1981) 556-559 15. David, H. L. and M. J. Selin: Immune response to Mycobacteria. In: Manual of clinical immunology. (N. R. Rose and H. Friedman, eds.), pp. 520-525. American Society for Microbiology, Washington, DC (1980) 16. Edwards, D. and C. H. Kirpatrick: The immunology of mycobacterial diseases. Am. Rev. Resp. Dis. 134 (1986) 1062-1071 17. Ellner, J. J.: Immune dysregulation in human tuberculosis. J. Lab. Clin. Med. 108 (1986) 142-149 18. Goren, M. B.: Immunoreactive substances of Mycobacteria. Am. Rev. Resp. Dis. 125 (1982) 50-69 19. Goren, M. B.: Biosynthesis and structures of phospholipids and sulfatides. In: The Mycobacteria: a source book, (G. P. Kubica and L. G. Wayne, eds.), pp. 379-415. Marcel Dekker, Inc, New-York (1984) 20. Goren, M. B., o. Brokl, and W. B. Schaefer: Lipids of putative relevance of virulence in Mycobacterium tuberculosis: correlation of virulence with elaboration of sulfatides and strongly acidic lipids. Infect. Imm. 9 (1974) 142-149 21. Grange, J. M.: The humoral immune response in tuberculosis: its nature, biological role and diagnostic usefulness. Adv. Tuberc. Res. 21 (1984) 1-78 22. Grange, J. M., J. Gibson, and A. Batty: The specificity of the humoral immune response to soluble mycobacterial antigens in tuberculosis. Tubercle 61 (1980) 153-156 23. Grange, J. M., J. Gibson, E. Nassau, and T. Kardjito: Enzyme-linked immunosorbent assay (Elisa): A study of antibodies to Mycobacterium tuberculosis in the IgG, IgA and IgM classes in tuberculosis, sarcoIdosis and Crohn's disease. Tubercle 61 (1980) 145-152 24. Kalish, S. B., R. C. Radin, J. P. Phair, D. Levitz, C. R. Zeiss, and E. Metzger: Use of an enzyme-linked immunosorbent assay technique in the differential dignosis of active pulmonary tuberculosis in humans. J. Infect. Dis. 147 (1983) 523-530 25. Kaplan, M. K. and M. W. Chase: Antibodies to Mycobacteria in human tuberculosis. I. Development of antibodies before and after antimicrobial therapy. J. Infect. Dis. 142 (1980) 825-834

Serodiagnosis of Tuberculosis

57

26. Kardjito, T., I. Handoyo, and J. M. Grange: Diagnosis of active tuberculosis by immunological methods. 1 - The effects to tuberculin reactivity and previous BCG vaccination on the antibody levels determinated by ELISA. Tubercle 63 (1982) 269-274 27. Kiran, U.: SHRINIWAS. Soluble antigen fluorescent antibody test in serodiagnosis of pulmonary tuberculosis. Eur. J. Resp. Dis. 67 (1985) 204-208 28. Krambovitis, E.: Detection of antibodies to Mycobacterium tuberculosis plasma membrane antigen by enzyme-linked immunosorbent assay. J. Med. Microbio!' 21 (1986) 257-264 29. Krambovitis, E., M. Harris, and D. T. D. Hughes: Improved serodiagnosis of tuberculosis using two assay test. J. Clin. Patho!' 39 (1986) 779-785 30. Lefford, M. S., D. D. McGregor, and G. B. Mackaness: Immune response to Mycobacterium tuberculosis in rats. Infect. Immun. 8 (1973) 182-189 31. Middlebrook, G., C. Coleman, and W. B. Schaefer: Sulfolipid from virulent tubercle bacilli. Proc. Nat. Acad. Sci. (USA) 45 (1959) 1801-1804 32. Middlebrook, G. and R. J. Dubos: Specific serum agglutination of erythrocytes sensitized with extracts of tubercle bacilli. J. Exp. Med. 88 (1948) 521-529 33. Nassau, E. and A. E. Nelstrop: Specific tubercle antigen. Tubercle 57 (1976) 197-201 34. Nassau, E. and E. R. Parsons: Detection of antibodies to Mycobacterium tuberculosis by solid-phase radioimmune-assay. J. Immuno!. Meth. 6 (1975) 261-271 35. Papa, F., P. Cruaud, and H. L. David: Antigenicity and specificity of selected glycolipid fractions from Mycobacterium tuberculosis. Res. Microbio!. 140 (1989) 569-578 36. Papa, F., A. Laszlo, and H. L. David: Specificity of Mycobacterium tuberculosis phenolic glycolipid (PGL-Tb1) antiserum. Ann. Inst. Pasteur/Microbio!' 139 (1988) 535-545 37. Reggiardo, Z., E. Vazquez, and L. Schnaper: Elisa tests for antibodies against mycobacterial glycolipids. J. Immun. Meth. 34 (1980) 55-60 38. Riska, H.: Assessment of the tuberculosis agglutination test. J. Resp. Dis. 61 (1980) 310-314 39. Samuel, A. M., M. D. Ashtekar, and R. D. Ganatra: Significance of circulating immune complexes in pulmonary tuberculosis. Clin. Exp. Immunol. 58 (1984) 317-324 40. Torgal Garcia, J., H. L. David, and F. Papa: Preliminary evaluation of a Mycobacterium tuberculosis phenoglycolipid antigen in the serologic diagnosis of tuberculosis. Ann. Inst. Pasteur/Microbiol. 139 (1988) 284-294 41. Viljanen, M. K., J. Eskola, and E. Tala: Enzyme-linked immunosorbent assay (Elisa) for antibodies to purified protein derivative of tuberculin (PPD). IgM, IgA and IgG anti. PPD antibodies in active pulmonary tuberculosis. Eur. J. Resp. Dis. 63 (1982) 257-262 42. Voller, A., D. Bidwell, and A. Bartlett: Enzyme-linked immunosorbent assay. In: Manual of clinical immunology (N. R. Rose and H. Friedman, eds.), pp. 359-371. American Society for Microbiology, Washington/DC (1980) Colette Savage, Laboratoire de Bacteriologie A, Faculte de Medecine, Place de Verdun, F-59045 Lille Cedex, France