The theoretical reliability of PCR‐based fish viral diagnostic methods is critically affected when they are applied to fish populations with low prevalence and virus loads

• Published in: Journal of applied microbiology 2018-04, Vol.124 (4), p.977-989
• Main Authors: Dopazo, C.P., Moreno, P., Olveira, J.G., Borrego, J.J.
• Format: Article
• Language: English
• Subjects: Animal tissues; Diagnosis; diagnostic; Diagnostic systems; Extreme values; Fish; Fish populations; Gangrene; Hemorrhagic septicemia; Hybridization; Infectious pancreatic necrosis; Likelihood ratio; Monitoring; Pancreas; Parameter sensitivity; Parameters; PCR; Polymerase chain reaction; Population studies; Populations; reliability; Reliability analysis; Reproducibility; Surveillance; virus; Viruses
• ISSN: 1364-5072; 1365-2672
• DOI: 10.1111/jam.13586

Aims The reliability of polymerase chain reaction (PCR) techniques is an important issue in viral diagnosis, and it is even crucial when they must be applied for detection of viruses in asymptomatic carriers. The problems will arise when the aim is to study wild fish populations, where the viral loads and prevalence values are extremely low. We have evaluated several PCR procedures employed by two laboratories for monitoring fish captured in several oceanographic campaigns performed in the Gulf of Cádiz. Methods and Results To evaluate the reliability of different diagnostics test used, we have re‐analysed fish samples that had been previously subjected to diagnosis for a surveillance of viruses performed in 2010–2011 in wild fish populations. The following parameters were employed: the clinical sensitivity (Ss), the clinical specificity (Sp), the predictive positive value, the predictive negative value, and the positive and negative likelihood ratio (LR+ and LR−). For viral nervous necrosis virus, a RT‐PCR procedure supplemented by nested PCR showed the highest values (100%) for all the parameters. For viral haemorrhagic septicaemia virus, the highest values were provided by RT‐PCR supplemented by dot‐blot hybridization. In the case of infectious pancreatic necrosis virus, none of the procedures yielded 100% for any parameter. Conclusions The results obtained for viral prevalence indicate: (i) that the conservation of the samples at −80°C did not affect to the capacity of detection of the virus in the tissues, and (ii) that the reproducibility of the diagnosis can be affected by factors including the staff experience and/or the materials employed. Finally, the use of a combination of procedures in advised to ensure the maximum reliability of the diagnosis when it is applied to asymptomatic fish populations. Significance and Impact of the Study This paper describes a strategy of combining diagnostic tests for the surveillance and monitoring of wild fish populations to reduce underestimation of the prevalence of viruses this type of populations.