Diagnostic accuracy of two DNA‐based molecular assays for detection of porcine circovirus 3 in swine population using Bayesian latent class analysis. (1st December 2019)

• Record Type: Journal Article
• Authors: Franzo, G.; Segalés, J.; Klaumann, F.; Legnardi, M.; Mweu, M. M.; Mahmmod, Y. S.
• Is Part Of: Letters in applied microbiology. Volume 69:Number 6(2019)
• Page Start: 417
• Page End: 423
• Publication Date: 2019-12-01
• Subjects: diagnostic test evaluation -- direct PCR -- latent class modelling -- porcine circovirus 3 -- quantitative PCR -- sensitivity and specificity; Microbiology -- Periodicals; 660.62
• DOI: 10.1111/lam.13226
• Languages: English
• ISSNs: 0266-8254

Abstract

Abstract: Molecular‐based tools sometimes are the only laboratory techniques available to detect a recently discovered agent, and their validation without the existence of previously described 'gold standard' methods poses a challenge for the diagnosticians. A good example within this scenario is the recently described porcine circovirus 3 (PCV‐3) in the swine population worldwide, from which only few PCR methods have been described. Therefore, the primary objective of this study was to estimate the diagnostic accuracy of a direct PCR (dPCR) and a real‐time qPCR (qPCR) for detection of PCV‐3 in Italian swine population. Bayesian latent class analysis approach was used to rigorously assess their features and applicability in routine diagnostic activity. Data on dPCR and qPCR were available from 116 domestic pigs, which were randomly selected from 55 farms located at different regions in Northern Italy. The sensitivity (Se) estimates of dPCR (94%; posterior credible interval (PCI%) 84–100) and qPCR (96%; PCI% 90–100) were high and similar. The estimated specificity (Sp) of both dPCR and qPCR assays was around 97%. dPCR and qPCR assays showed a high and comparable Se and Sp estimates for the detection of PCV‐3 in Italian swine population. Significance and Impact of the Study: The continuous discovery of new pathogens poses a challenge in the development and evaluation of adequate diagnostic tools. In fact, since molecular‐based tools sometimes are the only available laboratory techniques, it is typically difficult to evaluate their diagnostic performances in the absence of a gold standard. The present study assesses this issue, demonstrating the excellent performances of two PCR‐based assays for porcine circovirus 3 (PCV‐3) detection using a Bayesian latent class analysis approach. Therefore, the molecular tests evaluated under this study constitute reliable tools for the routine diagnosis and surveillance programs of PCV‐3 circulating in swine populations.