Assessment of performance of selected serological tests for diagnosing brucellosis in pigs.

P.M. Munoz, J.M. Blasco, B. Engel, M.J. de Miguel, C.M. Marín, L. Dieste, R.C. Mainar-Jaime

Research output: Contribution to journalArticleAcademicpeer-review

12 Citations (Scopus)

Abstract

Swine brucellosis due to Brucella suis is considered an emerging zoonotic disease whose control is based on serological testing and the subsequent culling of seropositive animals or the full depopulation of affected flocks. Here we assessed the performance of several serological tests (Rose Bengal Test [RBT], indirect ELISA [i-ELISA], blocking ELISA [b-ELISA], and two competitive ELISAs [c-ELISA]) for diagnosing swine brucellosis caused by B. suis biovar 2. Both frequentistic and Bayesian statistical inference were used. A frequentistic analysis, using sera from known gold standard (GS) populations (i.e., from truly infected or brucellosis free animals), resulted in maximum (100%) diagnostic sensitivity (Se) and specificity (Sp) in the RBT, i-ELISA and b-ELISA tests. However, c-ELISAs resulted in lower diagnostic Se (ranging from 68.5% to 92.6%, according to the different cut-offs selected). A Bayesian analysis of tests yielding the best diagnostic performance with GS sera (RBT, i-ELISA and b-ELISA), but using a large collection of field sera, resulted in similar Se among tests but markedly lower (˜80%) than that resulting from the frequentistic analysis using the GS serum populations. By contrast, the estimated Sp in the Bayesian analysis was only slightly lower than 100%, thus similar to that obtained frequentistically. Our results show that adequate diagnostic tests for brucellosis in swine are available, but also emphasize the need for more extensive validation studies before applying these tests under field conditions.
Original languageEnglish
Pages (from-to)150-158
Number of pages9
JournalVeterinary Immunology and Immunopathology
Volume146
DOIs
Publication statusPublished - 2012

Fingerprint

Brucellosis
brucellosis
Serologic Tests
immunologic techniques
Swine
Enzyme-Linked Immunosorbent Assay
enzyme-linked immunosorbent assay
swine
Rose Bengal
Brucella suis
testing
Brucella melitensis biovar Suis
gold
diagnostic sensitivity
Bayes Theorem
Bayesian theory
Serum
diagnostic specificity
culling (animals)
Validation Studies

Keywords

  • fluorescence polarization assay
  • melitensis infection
  • swine brucellosis
  • validation
  • suis
  • specificity
  • sensitivity
  • dependence
  • antigens
  • efficacy

Cite this

Munoz, P.M. ; Blasco, J.M. ; Engel, B. ; de Miguel, M.J. ; Marín, C.M. ; Dieste, L. ; Mainar-Jaime, R.C. / Assessment of performance of selected serological tests for diagnosing brucellosis in pigs. In: Veterinary Immunology and Immunopathology. 2012 ; Vol. 146. pp. 150-158.
@article{e493d4a269144e3bad5d247c7a03f1ef,
title = "Assessment of performance of selected serological tests for diagnosing brucellosis in pigs.",
abstract = "Swine brucellosis due to Brucella suis is considered an emerging zoonotic disease whose control is based on serological testing and the subsequent culling of seropositive animals or the full depopulation of affected flocks. Here we assessed the performance of several serological tests (Rose Bengal Test [RBT], indirect ELISA [i-ELISA], blocking ELISA [b-ELISA], and two competitive ELISAs [c-ELISA]) for diagnosing swine brucellosis caused by B. suis biovar 2. Both frequentistic and Bayesian statistical inference were used. A frequentistic analysis, using sera from known gold standard (GS) populations (i.e., from truly infected or brucellosis free animals), resulted in maximum (100{\%}) diagnostic sensitivity (Se) and specificity (Sp) in the RBT, i-ELISA and b-ELISA tests. However, c-ELISAs resulted in lower diagnostic Se (ranging from 68.5{\%} to 92.6{\%}, according to the different cut-offs selected). A Bayesian analysis of tests yielding the best diagnostic performance with GS sera (RBT, i-ELISA and b-ELISA), but using a large collection of field sera, resulted in similar Se among tests but markedly lower (˜80{\%}) than that resulting from the frequentistic analysis using the GS serum populations. By contrast, the estimated Sp in the Bayesian analysis was only slightly lower than 100{\%}, thus similar to that obtained frequentistically. Our results show that adequate diagnostic tests for brucellosis in swine are available, but also emphasize the need for more extensive validation studies before applying these tests under field conditions.",
keywords = "fluorescence polarization assay, melitensis infection, swine brucellosis, validation, suis, specificity, sensitivity, dependence, antigens, efficacy",
author = "P.M. Munoz and J.M. Blasco and B. Engel and {de Miguel}, M.J. and C.M. Mar{\'i}n and L. Dieste and R.C. Mainar-Jaime",
year = "2012",
doi = "10.1016/j.vetimm.2012.02.012",
language = "English",
volume = "146",
pages = "150--158",
journal = "Veterinary Immunology and Immunopathology",
issn = "0165-2427",
publisher = "Elsevier",

}

Assessment of performance of selected serological tests for diagnosing brucellosis in pigs. / Munoz, P.M.; Blasco, J.M.; Engel, B.; de Miguel, M.J.; Marín, C.M.; Dieste, L.; Mainar-Jaime, R.C.

In: Veterinary Immunology and Immunopathology, Vol. 146, 2012, p. 150-158.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Assessment of performance of selected serological tests for diagnosing brucellosis in pigs.

AU - Munoz, P.M.

AU - Blasco, J.M.

AU - Engel, B.

AU - de Miguel, M.J.

AU - Marín, C.M.

AU - Dieste, L.

AU - Mainar-Jaime, R.C.

PY - 2012

Y1 - 2012

N2 - Swine brucellosis due to Brucella suis is considered an emerging zoonotic disease whose control is based on serological testing and the subsequent culling of seropositive animals or the full depopulation of affected flocks. Here we assessed the performance of several serological tests (Rose Bengal Test [RBT], indirect ELISA [i-ELISA], blocking ELISA [b-ELISA], and two competitive ELISAs [c-ELISA]) for diagnosing swine brucellosis caused by B. suis biovar 2. Both frequentistic and Bayesian statistical inference were used. A frequentistic analysis, using sera from known gold standard (GS) populations (i.e., from truly infected or brucellosis free animals), resulted in maximum (100%) diagnostic sensitivity (Se) and specificity (Sp) in the RBT, i-ELISA and b-ELISA tests. However, c-ELISAs resulted in lower diagnostic Se (ranging from 68.5% to 92.6%, according to the different cut-offs selected). A Bayesian analysis of tests yielding the best diagnostic performance with GS sera (RBT, i-ELISA and b-ELISA), but using a large collection of field sera, resulted in similar Se among tests but markedly lower (˜80%) than that resulting from the frequentistic analysis using the GS serum populations. By contrast, the estimated Sp in the Bayesian analysis was only slightly lower than 100%, thus similar to that obtained frequentistically. Our results show that adequate diagnostic tests for brucellosis in swine are available, but also emphasize the need for more extensive validation studies before applying these tests under field conditions.

AB - Swine brucellosis due to Brucella suis is considered an emerging zoonotic disease whose control is based on serological testing and the subsequent culling of seropositive animals or the full depopulation of affected flocks. Here we assessed the performance of several serological tests (Rose Bengal Test [RBT], indirect ELISA [i-ELISA], blocking ELISA [b-ELISA], and two competitive ELISAs [c-ELISA]) for diagnosing swine brucellosis caused by B. suis biovar 2. Both frequentistic and Bayesian statistical inference were used. A frequentistic analysis, using sera from known gold standard (GS) populations (i.e., from truly infected or brucellosis free animals), resulted in maximum (100%) diagnostic sensitivity (Se) and specificity (Sp) in the RBT, i-ELISA and b-ELISA tests. However, c-ELISAs resulted in lower diagnostic Se (ranging from 68.5% to 92.6%, according to the different cut-offs selected). A Bayesian analysis of tests yielding the best diagnostic performance with GS sera (RBT, i-ELISA and b-ELISA), but using a large collection of field sera, resulted in similar Se among tests but markedly lower (˜80%) than that resulting from the frequentistic analysis using the GS serum populations. By contrast, the estimated Sp in the Bayesian analysis was only slightly lower than 100%, thus similar to that obtained frequentistically. Our results show that adequate diagnostic tests for brucellosis in swine are available, but also emphasize the need for more extensive validation studies before applying these tests under field conditions.

KW - fluorescence polarization assay

KW - melitensis infection

KW - swine brucellosis

KW - validation

KW - suis

KW - specificity

KW - sensitivity

KW - dependence

KW - antigens

KW - efficacy

U2 - 10.1016/j.vetimm.2012.02.012

DO - 10.1016/j.vetimm.2012.02.012

M3 - Article

VL - 146

SP - 150

EP - 158

JO - Veterinary Immunology and Immunopathology

JF - Veterinary Immunology and Immunopathology

SN - 0165-2427

ER -