Diagnostic performance measures of ELISA and quantitative PCR tests for porcine circovirus type 2 exposure using Bayesian latent class analysis

Charles Haley, Bruce Wagner, Sumathy Puvanendiran, Juan Abrahante, Michael P. Murtaugh

Research output: Contribution to journalArticlepeer-review

11 Scopus citations

Abstract

Porcine circovirus 2 (PCV2) is believed to be a necessary but not sufficient underlying cause of porcine circovirus associated disease (PCVAD) in swine (Opriessnig et al., 2007). Since the potential threat of PCVAD is dependent on the prevalence of PCV2 in swine populations, accurate diagnostic tests are important for epidemiologic surveillance. Therefore, we evaluated the diagnostic sensitivity (Se) and specificity (Sp) of a new indirect ELISA and two quantitative PCR tests for PCV2 in a series of latent class models that used Bayesian estimation procedures. A total of 4140 samples from finisher pigs were tested for evidence of PCV2 by the ELISA and a TaqMan (TM) quantitative PCR, 995 by the ELISA and a SYBR Green (SG) dye-binding PCR, 998 by both PCRs and 993 by all three tests. Overall, the median (95% probability interval) ELISA Se and Sp was 0.85 (0.83-0.87) and 0.74 (0.68-0.79), respectively, when all three tests were analyzed together at an ELISA absorbance (optical density or OD) cutoff of ≥0.3. The TM PCR Se and Sp was 0.86 (0.84-0.88) and 0.94 (0.87-0.97), respectively, and the SG PCR Se and Sp was 0.83 (0.81-0.85) and 0.98 (0.94-1.00), respectively when all three tests were analyzed together at an ELISA OD cutoff of ≥0.3. Sensitivity analysis revealed that Sp estimates in general had less stability than Se estimates, but the SG PCRSp was the most stable. Limited conditional dependence between the two PCR tests was detected. We conclude that the ELISA had the highest diagnostic Se at an absorbance cutoff of ≥0.3, while the SG PCR had the highest diagnostic Sp. The prevalence levels for exposure to PCV2 in finishing swine populations across all analyses ranged from 58 to 100%.

Original languageEnglish (US)
Pages (from-to)79-88
Number of pages10
JournalPreventive Veterinary Medicine
Volume101
Issue number1-2
DOIs
StatePublished - Aug 1 2011

Bibliographical note

Copyright:
Copyright 2013 Elsevier B.V., All rights reserved.

Keywords

  • Bayesian
  • Conditional dependence
  • Latent data
  • Sensitivity
  • Specificity
  • Swine

Fingerprint Dive into the research topics of 'Diagnostic performance measures of ELISA and quantitative PCR tests for porcine circovirus type 2 exposure using Bayesian latent class analysis'. Together they form a unique fingerprint.

Cite this