Hepatitis E virus (HEV) is a RNA non-enveloped virus that comprises four genotypes. The genome of HEV is organized into three Open Reading Frames (ORFs), and the ORF2 is responsible for encoding capsid proteins. HEV can infect a wide range of hosts, and pigs are considered the main reservoir. HEV infection is considered a zoonosis and it is responsible for acute hepatitis in humans, especially in developing countries. The development of a blocking ELISA would be of high value for screening purpose, because there is no need of species specific reagents. The present study was conducted to assess three in-house ELISAs for the detection of HEV infection in 779 sera collected from breeding and fattening farms under field conditions. Two assays were indirect ELISAs, while the third was a blocking ELISA. Two different recombinant antigens were generated from specific sequences of the HEV-ORF2, and a Latent Class approach in a Bayesian framework was used to evaluate the diagnostic accuracy of each ELISA. Because the three ELISAs cannot be thought of as independent, all possible dependence structures were modelled starting from the general case of conditional independence to the most complex situation of three mutually dependent assays. Results showed that none of the three ELISAs was significantly superior to the others in terms of sensitivity (posterior median value ranging from 89% to 94%, all 95% posterior credible intervals (95%PCI) overlapped). In terms of specificity, one of the indirect ELISAs was superior to blocking ELISA (posterior median indirect ELISA: 99%, 95%PCI: 98-100%; blocking ELISA: 90%; 95%PCI: 86-94%). However, this difference could be due to the potential wider spectrum of antibodies that blocking ELISA can actually detect.
Keywords: Blocking ELISA; HEV; Latent Class analysis.
Copyright © 2014 Elsevier B.V. All rights reserved.