Vol. 29, No. 2

JOURNAL OF CLINICAL MICROBIOLOGY, Feb. 1991, p. 272-276

0095-1137/91/020272-05$02.00/0 Copyright © 1991, American Society for Microbiology

Evaluation of a Commercial Enzyme-Linked Immunosorbent Assay for Johne's Disease M. T. COLLINS,'* D. C. SOCKETT,1 S. RIDGE,2 AND J. C. COX3 School of Veterinary Medicine, University of Wisconsin, Madison, Wisconsin 53706,1 and Department ofAgriculture and Rural Affairs, Veterinary Research Institute, Attwood,2 and Commonwealth Serum Laboratories, Parkville, Victoria, Australia Received 23 July 1990/Accepted 28 October 1990

A new commercial kit for diagnosis of bovine paratuberculosis (Johne's disease), called the Johne's Absorbed EIA (enzyme immunoassay; Commonwealth Serum Laboratories, Parkville, Victoria, Australia), was evaluated by using serum specimens from the National Repository for Paratuberculosis Specimens. The evaluation was specifically designed to measure test sensitivity and specificity for detection of dairy cattle with subclinical paratuberculosis. The case definition of subclinical bovine paratuberculosis was isolation of Mycobacterium paratuberculosis from fecal samples or internal organs of cattle without diarrhea or chronic weight loss. Animals designated as free of the disease originated exclusively from four herds in Wisconsin that were certified to be free of disease. The kit had a sensitivity of 47.3% for serum specimens from 150 infected cattle. The test detected 59.7% of animals that shed M. paratuberculosis in their feces, as defined by conventional fecal culture, at the time of serum collection. Testing of 196 serum specimens from cattle without paratuberculosis yielded two false-positive results; the test specificity was thus 99.0%. Decision analysis procedures on the economics of using the kit in a test-and-cull disease control program indicated it would be cost-effective in any herd with a true paratuberculosis prevalence of .3%. Comparison of the sensitivity and specificity of the Johne's Absorbed EIA with those of other tests for detection of subclinical paratuberculosis indicated that it may be the most accurate commercially available test at present and better than the standard complement fixation test used in the United States.

For the past decade, the majority of the research effort on Johne's disease (paratuberculosis) has been directed at development of improved diagnostic tests for the disease. Several new techniques have recently been described that improve culture detection of the causative agent, Mycobacterium paratuberculosis (9, 11, 16, 21, 48), use gene probes (15, 22, 44), or are based on serological methods (25, 39, 43, 51). Among the serological techniques, enzyme-linked immunosorbent assays (ELISAs) that use absorption of sera to reduce nonspecific reactions have shown considerable promise (24-26, 49-51). This technique was commercialized by Commonwealth Serum Laboratories, Parkville, Victoria, Australia, and was released in March 1990 as a kit called the Johne's Absorbed EIA (enzyme immunoassay) for diagnosis of bovine paratuberculosis. In this report we describe the evaluation of the diagnostic accuracy of this kit by using serum specimens from the National Repository for Paratuberculosis Specimens.

Paratuberculosis Specimens housed at the University of Wisconsin-Madison (39a). An extra effort was made to identify animals with the full spectrum of M. paratuberculosis infection severity. At the time of serum collection fecal specimens were collected and cultured both by conventional methods (46) and by a new radiometric technique (9). In addition, serum specimens from all animals were recultured three or more times during the next 18 months by conventional methods at the Wisconsin Animal Health Laboratory. In Wisconsin, for the conventional method we used 0.25% hexadecylcetylpyridinium chloride and two tubes of Herrold egg yolk (HEY) agar with mycobactin and one tube without mycobactin. Cultures were incubated and observed for 3 months. To recover tissues for histopathology and culture, surgery was performed on animals that were culture negative at every test but that were positive for antibodies to M. paratuberculosis by an absorbed ELISA (Allied Laboratories, Inc.). A fullthickness (1 by 2 cm) piece of terminal ileum, taken 10 to 20 cm proximal to the ileocecal junction, and an ileocecal lymph node were obtained through a right-flank laparotomy incision. Histopathology with both hematoxylin-eosin and acid-fast stains was done on half of each tissue, and the remainder was homogenized and cultured for M. paratuberculosis by the radiometric method (9). A total of 856 serum specimens (0.5-ml aliquots) from the repository were shipped on dry ice to the Australian Animal Health Laboratory, Geelong, Victoria. All serum specimens were those collected concurrently with fecal specimens from individual animals at the time of first sampling at each farm. Testing was performed blind by technical staff of the Veterinary Research Institute. Johne's Absorbed EIA. Assays were performed with the commercial kits according to the directions of the manufac-

MATERIALS AND METHODS Specimens. Serum and fecal specimens were collected concurrently from 856 dairy cattle in 13 Wisconsin herds. Nine of the herds, containing a total of 660 dairy cattle, were infected with M. paratuberculosis and had prevalence rates that ranged from 5 to over 60%. Four of the herds, containing a total of 196 cattle, were certified to be free of paratuberculosis by the state of Wisconsin Department of Agriculture, Trade and Consumer Protection based on a regular program of annual whole-herd testing for paratuberculosis by the conventional fecal culture method (10). These specimens constituted a part of the National Repository for *

Corresponding author. 272

VOL. 29,

1991

EVALUATION OF AN ELISA FOR JOHNE'S DISEASE

turer. Negative, low-positive, and high-positive control sera are included in the kit, and each has a specified range of optical density values (450-nm filter) to call an assay valid. The cutoff value for valid assays is the mean of duplicate negative controls plus 0.100. Sera with discrepant results between wells or with borderline values were retested. Results were reported as (optical density of test serum cutoff value) x 100. All values greater than zero were considered positive. Data analysis. Vaccinated animals were excluded from analysis. All animals from herds certified to be free of paratuberculosis were used as the negative control population for test specificity analysis. Herds in Wisconsin certified to be free of paratuberculosis are those in which all animals .20 months of age have been tested and found to be negative for paratuberculosis by the conventional fecal culture method performed on at least three consecutive samples taken not less than 6 months but not more than 1 year apart. Herd certification is renewed annually (10). The diagnostic difficulty in paratuberculosis is detection of disease (infection) in subclinically affected animals. Thus, only such animals were evaluated, and the case definition for subclinical paratuberculosis was isolation of the causative agent, M. paratuberculosis, from any fecal sample or biopsy tissues. Disease status was determined without reference to or knowledge of the Johne's Absorbed EIA results. Only infected animals were used for diagnostic test sensitivity analysis. Of 856 serum specimens tested, 196 were from nonvaccinated animals in herds certified to be free of paratuberculosis, and 150 were from animals defined as having subclinical Johne's disease. Test results for the sera from the other 510 cattle were not used for test specificity determination, since the animals resided in infected herds and their M. paratuberculosis infection status could not be unequivocally proven, and thus, inclusion of the results in specificity determination is inappropriate (33). Test sensitivity and specificity estimates with 95% confidence limits were calculated by standard methods (35). The Johne's Absorbed EIA was compared with conventional fecal culture (first sampling only) by using the McNemar test (34) and the kappa statistic (19). Decision analysis. Decision analysis was performed to assess the short-term benefit-cost of using the Johne's Absorbed EIA in a test-and-cull disease control program since this is the principal application of diagnostic tests for paratuberculosis (lOa). Standard methods were used (5, 6, 13, 27, 29, 31, 32, 36). Pretest Johne's disease prevalence was the independent variable. Standard equations for calculating the apparent prevalence from the true prevalence (18, 37) and the predictive value model equations (12, 45) were used to calculate the probabilities of getting true-positive, truenegative, false-positive, or false-negative results. Costs associated with each of these four possible diagnostic test outcomes were obtained from the literature (1, 2, 4, 47). The cost of paratuberculosis to a dairy was conservatively estimated as a 6% decrease in total milk production for each infected cow for the next lactation, although studies have reported that the effect of paratuberculosis on milk production is generally higher (1, 2, 4, 47). Market prices for replacement of culled cows, slaughter of adult cows, and milk sold were based on 1987 prices in Wisconsin (41). EIA sensitivity and specificity values used were those defined in the present study, i.e., 47.3 and 99.0%, respectively. Essentially, the decision tree calculated the net economic benefitcost for a dairy farmer's decision to use the Johne's Absorbed EIA and cull all test-positive animals rather than not

273

doing anything. The analysis was performed at paratuberculosis prevalence rates from 0 to 100%. Herd characteristics typical of the average herd in Wisconsin were used: 55 adult cows, culling rate of 30%, and 6,266 kg (13,786 lb) of milk per cow production average (41). A cost to farmers of $5 to test each cow was also used.

RESULTS Test specificity. Among the 196 serum specimens from herds certified to be free of paratuberculosis, only 2 were positive by the Commonwealth Serum Laboratories test, giving a specificity estimate of 99.0% ± 1.4%. The two positive serum specimens gave low test values of 7 and 16 (high values were .185); the latter originated from an animal that was negative for paratuberculosis by biopsy evaluation. Both cows from which the sera came were fecal culture negative on all three samplings, in addition to being from herds certified to be free of paratuberculosis. Test sensitivity. Among the 150 serum specimens from cows with paratuberculosis, 71 were positive by the Johne's Absorbed EIA, giving a sensitivity of 47.3% ± 8.2%. Positive ELISA results [(optical density of test serum - cutoff value) x 100] ranged from 7 to .186 (mean, 42; standard deviation, 62). The age of test-positive cows was 3.7 ± 1.6 years (mean ± standard deviation), and for test-negative cows the mean age was 4.6 ± 1.9 years, which were not significantly different (the Student t test). Correlation of Johne's Absorbed EIA with conventional culture. For 143 animals, both Johne's Absorbed EIA (EIA) and conventional fecal culture (on HEY agar slants) results were available for specimens collected simultaneously. The distribution of test results for these paired fecal and serum specimens was as follows: positive by both tests, 46 cows (32.2%); positive by HEY only, 31 cows (21.7%); positive by EIA only, 21 cows (14.7%); and negative by both tests, 45 cows (31.5%). Application of the McNemar test (34) to these data indicated no significant difference between the two tests in the diagnosis of subclinical paratuberculosis (P > 0.05). The kappa statistic was 0.276, however, which is considered a relatively low value (20). Decision analysis. For typical Wisconsin dairy herds, decision analysis indicated that on an annual basis it would be economical to use the Johne's Absorbed EIA in a test-andcull paratuberculosis control program whenever the prevalence of the disease in the herd was .3%.

DISCUSSION The Johne's Absorbed EIA was highly specific. Only 2 of 196 cows in four herds certified to be free of paratuberculosis tested positive, and the resulting test specificity estimate of 99.0% was not significantly different from the value of 99.8% found when the kit was used to test 1,000 serum specimens from cattle in Western Australia, where paratuberculosis does not occur (lOb), nor from the absorbed ELISA specificity of 98.9% ± 0.6% reported by Yokomizo (49). For specificity estimates of serological tests for paratuberculosis, cattle that are unequivocally free of the disease have often not been used. Instead, animals residing in infected herds have been used and three negative fecal cultures obtained at intervals of .6 months have been relied upon to classify animals as noninfected (30, 38, 40, 42). Problems surrounding the use of an imperfect standard for the diseasefree classification have been described previously (33) and were illustrated in the present study. Among 488 animals in

274

COLLINS ET AL.

the M. paratuberculosis-infected herds that were negative on all three conventional fecal culture attempts, 24 were found to be infected either by radiometric culture of feces or by radiometric culture of surgically collected tissues. Had those 488 animals been used for test specificity determination, the result would have been 95.1%. This underestimation of test specificity would have adversely affected adoption of the test by laboratories because of the concern about the economic consequences to farmers associated with falsepositive test results. The stage of M. paratuberculosis infection significantly affects diagnostic test sensitivity for paratuberculosis. As a case definition of subclinical paratuberculosis, many investigators have used animals without signs of diarrhea and weight loss from which M. paratuberculosis was isolated from fecal specimens collected two or three times over a period of roughly 18 months (8, 14, 15, 38). The case definition in the present study was more inclusive, as indicated by the fact that 25 of the 150 infected animals classified as having paratuberculosis were negative on three or more fecal culture attempts (conventional methods). The diagnosis was established in 11 of these animals by radiometric culture of filter concentrated fecal specimens (9) and in the remaining 14 animals by culture of biopsy tissues. The diagnostic "gold standard" that we used ensured a more rigorous definition of diagnostic test sensitivity than has previously been attempted. The sensitivity of the Johne's Absorbed EIA for detection of both fecal culturepositive and -negative cattle was 47.3% ± 8.2%. Alternatively, if only animals that were culture positive by conventional methods at the time of serum collection were defined as having subclinical paratuberculosis, the test "sensitivity" would be 59.7%. With clinically affected animals, the sensitivity of the Johne's Absorbed EIA was 87.5% (unpublished

data).

Detection of conventional culture-positive cattle by the Johne's Absorbed EIA (59.7%) was the same as the 60% organism excreter detection rate found by Yokomizo (49) when 58 culture-positive cattle were tested by an absorbed ELISA. Using a positive fecal culture as the diagnostic criterion for Johne's disease, Sherman et al. (38) reported the sensitivity and specificity of the agar gel immunodiffusion kit (Rapid Johne's Test; ImmuCell Corp., Portland, Maine) to be 18.9 and 99.4%, respectively. They also reported the sensitivity and specificity of the standard complement fixation test (28) to be 10.8 and 97.4%, respectively, when titers of .1:32 were classified as positive, or 31.1 and 90.3%, respectively, when titers of .1:16 were considered positive. The greater sensitivity of ELISA-based techniques over those of gel diffusion methods for the serodiagnosis of Johne's disease was also reported by Tsai et al. (43). While the specificity measures reported by Sherman et al. (38) may be in error for the reasons described above, absorbed ELISA techniques appear to achieve greater diagnostic sensitivity without compromising diagnostic specificity, making them more accurate than the other available sero-

logical tests.

The McNemar test (34) indicated that conventional fecal culture and the Johne's Absorbed EIA are not significantly different in diagnostic accuracy. The kappa statistic, however, was relatively low, suggesting that the two tests detected different groups of animals. Fecal culture detects animals that are excreting viable M. paratuberculosis in their feces. Serological tests, such as the Johne's Absorbed EIA, potentially detect animals prior to initiation of the infectious

J. CLIN. MICROBIOL.

stage (organism excretion). Some fecal excreters were missed by the serological test, however. Detection of fecal excreters of M. paratuberculosis may be of more importance to paratuberculosis control efforts than is detection of serologically positive cattle. It is not known whether all seropositive cattle progress to organism excretion, clinical disease, or both (7). The stage of Johne's disease in which economically significant reduction of animal productivity occurs has been shown to correspond with detection of M. paratuberculosis in fecal specimens (2), providing an economic argument for organism detectionbased tests. Long-term prospective studies are needed to determine whether cattle that become infected and serologically positive always progress to the infectious stage of disease, shedding M. paratuberculosis in their feces, and then clinical disease and whether serologically positive but fecal culture-negative animals have reduced productivity. Johne's disease has been shown in several studies to cause significant reductions in the productivity of dairy cattle (1, 2, 4, 23, 47). Decision analysis indicated that use of the Johne's Absorbed EIA as the basis for culling of presumptively diseased cows would be economical whenever the true prevalence of Johne's disease in the herd was -3%. That is, the benefits of eliminating infected cattle from a herd outweigh the cost of testing and the costs associated with test errors. The test's high specificity would minimize the cost of culling animals because of false-positive test results. By using the test sensitivity and specificity reported here, the kit can be used to estimate the true Johne's disease prevalence from the observed (apparent) prevalence rate of test-positive animals in a population (18, 37). This would make the serological survey data obtained by use of the kit more meaningful. For example, Braum et al. (3) recently reported that 17.1% of dairy cattle in Florida tested positive by an absorbed ELISA (apparent prevalence) but made no estimate of the accuracy of their test nor the true prevalence of Johne's disease. If their test had the same sensitivity and specificity as those of the Johne's Absorbed EIA kit, the true prevalence of Johne's disease in Florida's dairy cattle is 34.8%. Estimation of true prevalence is also a useful clinical tool for advising farmers on the infection status of their herds and the economics of paratuberculosis control. The kit from Commonwealth Serum Laboratories is the first commercially available kit for serodiagnosis of Johne's disease by using absorbed ELISA techniques. It differs from other absorbed ELISAs that have been reported, in that absorption of cross-reactive antibodies is effected by using soluble Mycobacterium phlei antigens simply by dilution of test sera in M. phlei antigen-containing diluent and incubation for 30 min before the assay is performed. This eliminates the need for overnight absorption and centrifugation of serum specimens (25, 49). A standardized paratuberculosis serodiagnostic assay is important for development of state or national control programs and the international trade of cattle. Use of absorbed ELISA procedures without standardized reagents can result in significantly different diagnostic accuracies between laboratories, as shown by Lopez et al. (17), and thus should be evaluated carefully before adoption. ACKNOWLEDGMENTS This work was funded in part by a grant from the U.S. Department of Agriculture Special Grants Program. We thank the Wisconsin dairy farmers who participated in development of the repository and the Australian Animal Health Laboratory for use of their facilities. We also thank Carolyn Smith and

VOL. 29, 1991

Ann Pengelly for technical assistance and Ian Morgan for critical review of the manuscript. REFERENCES 1. Abbas, B., H. P. Reimann, and D. W. Hird. 1983. Diagnosis of Johne's disease (paratuberculosis) in northern California cattle and a note on its economic significance. Calif. Vet. 8:20-24. 2. Benedictus, G., A. A. Dijkhuizen, and J. Stelwagen. 1987. Economic loss due to paratuberculosis in dairy cattle. Vet. Rec. 121:142-146. 3. Braum, P. K., C. D. Buergelt, R. C. Littell, S. B. Linda, and J. R. Simpson. 1990. Use of an enzyme-linked immunosorbent assay to estimate prevalence of paratuberculosis in cattle of Florida. J. Am. Vet. Med. Assoc. 196:1251-1254. 4. Buergelt, C. D., and J. R. Duncan. 1978. Age and milk production data of cattle culled from a dairy herd with paratuberculosis. J. Am. Vet. Med. Assoc. 173:478-480. 5. Carpenter, T. E. 1986. Epidemiologic programs for computers and calculators. Decision-tree analysis using a microcomputer. Am. J. Epidemiol. 124:843-850. 6. Carpenter, T. E., S. L. Berry, and J. S. Glenn. 1987. Economics of Brucella ovis control in sheep: computerized decision-tree analysis. J. Am. Vet. Med. Assoc. 190:983-987. 7. Chiodini, R. J., J. J. Van Kruiningen, and R. S. Merkal. 1984. Ruminant paratuberculosis (Johne's disease): the current status and future prospects. Cornell Vet. 74:217-262. 8. Colgrove, C. S., C. 0. Thoen, B. 0. Blackburn, and C. D. Murphy. 1989. Paratuberculosis in cattle: a comparison of three serologic tests with results of fecal culture. Vet. Microbiol. 19:183-187. 9. Collins, M. T., K. B. Kenefick, D. C. Sockett, R. S. Lambrecht, J. McDonald, and J. B. Jorgensen. 1990. Enhanced radiometric detection of Mycobacterium paratuberculosis by using filterconcentrated bovine fecal specimens. J. Clin. Microbiol. 28: 2514-2519. 10. CoUins, M. T., and A. R. McLaughlin. 1989. Experience in Wisconsin in control and accreditation of Johne's disease infected herds, p. 67-73. In A. R. Milner and P. R. Wood (ed.), Johne's disease. Current trends in research, diagnosis and management. Commonwealth Scientific and Industrial Research Organization Publications, Melbourne, Australia. 10a.CoUins, M. T., and I. R. Morgan. 1990. Abstr. 33rd Annu. Meet. Am. Assoc. Vet. Lab. Diagnosticians 1990, p. 62. 10b.Cox, J. C., D. Drane, S. Ridge, and A. R. Milner. 1990. Abstr. Annu. Meet. Aust. Soc. Microbiol. 1990, S14.4, p. 241. 11. Damato, J. J., and M. T. Collins. 1990. Growth of Mycobacterium paratuberculosis in radiometric, Middlebrook and eggbased media. Vet. Microbiol. 22:31-42. 12. Dierksheide, W. C. 1987. Medical decisions: interpreting clinical tests. ASM News 53:677-680. 13. Fetrow, J., J. B. Madison, and D. Galligan. 1985. Economic decisions in veterinary practice: a method for field use. J. Am. Vet. Med. Assoc. 186:792-797. 14. Goudswaard, J., N. J. L. Gilmour, R. G. Dijkstra, and J. J. Van Beek. 1976. Diagnosis of Johne's disease in cattle: a comparison of five serological tests under field conditions. Vet. Rec. 98:461462. 15. Hurley, S. S., G. A. Splitter, and R. A. Welch. 1989. Development of a diagnostic test for Johne's disease using a DNA hybridization probe. J. Clin. Microbiol. 27:1582-1587. 16. Kim, Y. G., S. Bech-Nielsen, J. Gordon, R. D. Selmons, and E. Spangler. 1989. Comparison of two methods for isolation of Mycobacterium paratuberculosis from bovine fecal samples. Am. J. Vet. Res. 50:1110-1113. 17. Lopez, B., V. Ritacco, I. N. de Kantor, R. Debenedetti, A. Nader, and A. Bernardelli. 1990. Evaluation of an ELISA for serodiagnosis of bovine paratuberculosis. Paratuberc. Newsl. 2:5-8. 18. Marchevsky, N. 1974. Errors in prevalence estimates in population studies: a practical method for calculating real prevalence. Zoonosis 16:85-109. 19. Martin, S. W. 1977. The evaluation of tests. Can J. Comp. Med. 41:19-25.

20. Martin, S. W., A. H. Meek, and P. Willeberg. 1987. Measure-

EVALUATION OF AN ELISA FOR JOHNE'S DISEASE

275

ment of disease frequency and production, p. 48-76. In Veterinary epidemiology-principles and methods. Iowa State Uni-

versity Press, Ames. 21. McDonald, J., M. T. Collins, and R. S. Lambrecht. 1989. Radiometric detection of Mycobacterium paratuberculosis from clinical specimens, p. 146-152. In A. R. Milner and P. R. Wood (ed.), Johne's disease. Current trends in research, diagnosis and management. Commonwealth Scientific and Industrial Research Organization Publications, Melbourne, Australia. 22. McFadden, J. J., E. Green, and J. Hermon-Taylor. 1988. DNA probes to identify and detect Mycobacterium paratuberculosis in clinical and veterinary samples, p. 201-205. In R. S. Merkal and M. F. Thorel (ed.), Proceedings of the Second International Colloquium on Paratuberculosis. Laboratoire Central de Recherches Veterinaires, Maisons-Alfort, France. 23. Merkal, R. S., A. B. Larsen, and G. D. Booth. 1975. Analysis of the effects of inapparent bovine paratuberculosis. Am. J. Vet. Res. 36:837-838. 24. Milner, A. R., A. W. D. Lepper, W. N. Symonds, and E. Gruner. 1987. Analysis by ELISA and Western blotting of antibody reactivities in cattle infected with Mycobacterium paratuberculosis after absorption of serum with M. phlei. Res. Vet. Sci.

42:140-144. 25. Milner, A. R., W. N. Mack, K. Coates, P. R. Wood, P. Sheldrick, J. Hill, and I. Gill. 1989. The absorbed ELISA for the diagnosis of Johne's disease in cattle, p. 158-163. In A. R. Milner and P. R. Wood (ed.), Johne's disease. Current trends in research, diagnosis and management. Commonwealth Scientific and Industrial Research Organization Publications, Melbourne, Australia. 26. Milner, A. R., W. N. Mack, K. J. Coates, J. Hill, I. Gill, and P. Sheldrick. Vet. Microbiol., in press. 27. Mohammed, H. O., S. Loefler, and J. Shearer. 1990. Financial comparison of three testing strategies for detection of estrus in dairy cattle. J. Am. Vet. Med. Assoc. 196:865-869. 28. National Veterinary Services Laboratory. 1986. Microtitration complement fixation techniques, p. 1-18. National Veterinary Services Laboratory, Ames. 29. Ngategize, P. K., J. B. Kaneene, S. B. Harsh, P. C. Bartlett, and E. L. Mather. 1986. Decision analysis in animal health programs: merits and limitations. Prev. Vet. Med. 4:187-197. 30. Norris, M. J., and T. L. Spencer. 1989. The efficacy of the complement fixation test for the detection of Mycobacterium paratuberculosis infection in cattle, p. 153-157. In A. R. Milner and P. R. Wood (ed.), Johne's disease. Current trends in research, diagnosis and management. Commonwealth Scientific and Industrial Research Organization Publications, Melbourne, Australia. 31. Parsons, T. D., G. Smith, and D. T. Galligan. 1986. Economics of porcine parvovirus vaccination assessed by decision analysis. Prev. Vet. Med. 4:199-204. 32. Pauker, S. G., and J. P. Kassirer. 1987. Decision analysis. N. Engl. J. Med. 316:250-258. 33. Ransohoff, D. F., and A. R. Feinstein. 1978. Problems of spectrum and bias in evaluating the efficacy of diagnostic tests. N. Engl. J. Med. 299:926-930. 34. Remington, R. D., and M. A. Schork. 1970. Chi-square tests for frequency data, p. 229-252. In Statistics with applications to the biological and health sciences. Prentice-Hall, Inc., Englewood Cliffs, N. J. 35. Sackett, D. L., R. B. Haynes, and P. Tugwell. 1985. The interpretation of diagnostic data, p. 59-138. In Clinical epidemiology. A basic science for clinical medicine. Little, Brown & Co., Boston. 36. Salman, M. D., D. A. Dargatz, C. V. Kimberling, and R. P. Ellis. 1988. An economic evaluation of various treatments for contagious foot rot in sheep, using decision analysis. J. Am. Vet. Med. Assoc. 193:195-204. 37. Schwabe, C. W., H. P. Riemann, and C. E. Franti. 1977. The

mathematical approach, p. 66-97. In Epidemiology in veterinary practice. Lea and Febiger, Philadelphia. 38. Sherman, D. M., J. M. Gay, D. S. Bouley, and G. H. Nelson. 1990. Comparison of the complement fixation and agar gel

276

J. CLIN. MICROBIOL.

COLLINS ET AL.

immunodiffusion tests for the diagnosis of subclinical bovine paratuberculosis. Am. J. Vet. Res. 51:461-465. 39. Sherman, D. M., R. J. F. Markham, and F. Bates. 1984. Agar gel immunodiffusion test for diagnosis of clinical paratuberculosis in cattle. J. Am. Vet. Med. Assoc. 185:179-182. 39a.Sockett, D. C. 1990. Abstr. 33rd Annu. Meet. Am. Assoc. Vet. Lab. Diagnosticians 1990, p. 90. 40. Spangler, E., S. Bech-Nielsen, and L. E. Heider. 1988. A prospective evaluation of an enzyme-linked immunosorbent assay (ELISA) and agar gel immunodiffusion (AGID) relative to fecal culture in the diagnosis of clinical and subclinical paratuberculosis in cattle in central Ohio. Acta Vet. Scand. 84:243245. 41. Spencer, C. D., J. E. Ramey, M. Heiser, and N. Harker. 1988. 1988 Wisconsin dairy facts, p. 1-48. Wisconsin Agricultural Statistics Service, Madison. 42. Sugden, E. A., A. H. Corner, B. S. Samagh, B. W. Brooks, C. Turcotte, K. H. Nielsen, R. B. Stewart, and J. R. Duncan. 1989. Serodiagnosis of ovine paratuberculosis, using lipoarabinomannan in an enzyme-linked immunosorbent assay. Am. J. Vet. Res. 50:850-854. 43. Tsai, S. J., L. J. Hutchison, and A. Zarkower. 1989. Comparison of a dot immunobinding assay, enzyme-linked immunosorbent assay and immunodiffusion for serodiagnosis of paratuberculosis. Can. J. Vet. Res. 53:405-410. 44. Vary, P. H., P. R. Andersen, E. Green, J. Hermon-Taylor, and J. J. McFadden. 1990. Use of highly specific DNA probes and the polymerase chain reaction to detect Mycobacterium paratuberculosis in Johne's disease. J. Clin. Microbiol. 28:933-937. 45. Vecchio, T. J. 1966. Predictive value of a single diagnostic test in unselected populations. N. Engl. J. Med. 274:1171-1173. 46. Whipple, D. L., and R. S. Merkal. 1985. Procedures for the field and laboratory processing of fecal specimens for the isolation of Mycobacterium paratuberculosis, p. 239-245. In Proceedings of

47.

48.

49.

50.

51.

the American Association of Veterinary Laboratory Diagnosticians, 28th Annual Meeting. American Association of Veterinary Laboratory Diagnosticians, Milwaukee, Wis. Whitlock, R. H., L. T. Hutchison, R. S. Merkal, L. T. Glickman, C. Rossiter, S. Harmon, P. Spencer, J. Fetrow, J. Bruce, C. E. Benson, and J. Dick. 1985. Prevalence and economic consideration of Johne's disease in the northeastern U.S., p. 484-490. In Proceedings of the 89th Annual Meeting of the U.S. Animal Health Association. U.S. Animal Health Association, Richmond, Va. Whitlock, R. H., A. E. Rosenberger, and P. A. Spencer. 1990. Laboratory culture techniques for Johne's disease: a critical evaluation of contamination and incubation times, p. 382-386. In Proceedings of the 93rd Annual Meeting of the U.S. Animal Health Association. U.S. Animal Health Association, Richmond, Va. Yokomizo, Y. 1986. Evaluation on an enzyme-linked immunosorbent assay (ELISA) using Mycobacterium phlei-absorbed serum for the diagnosis of bovine paratuberculosis in a field study. JARQ 20:60-67. Yokomizo, Y., K. Nishimori, M. Kishima, and H. Yugi. 1988. Evaluation of an enzyme-linked immunosorbent assay (ELISA) for the diagnosis of bovine paratuberculosis. A proposal of replacing the complement fixation test with the ELISA as the official diagnostic test for paratuberculosis in Japan, p. 206-214. In R. S. Merkal and M. F. Thorel (ed.), Proceedings of the Second International Colloquium on Paratuberculosis. Laboratoire Central de Recherches Veterinaires, Maisons-Alfort, France. Yokomizo, Y., H. Yugi, and R. S. Merkal. 1985. A method of avoiding false-positive reactions in an enzyme-linked immunosorbent assay (ELISA) for the diagnosis of bovine paratuberculosis. Jpn. J. Vet. Sci. 47:111-119.

Evaluation of a commercial enzyme-linked immunosorbent assay for Johne's disease.

A new commercial kit for diagnosis of bovine paratuberculosis (Johne's disease), called the Johne's Absorbed EIA (enzyme immunoassay; Commonwealth Ser...
1MB Sizes 0 Downloads 0 Views