Abstract
A pseudovirion-based neutralisation assay (PBNA) has been considered the gold standard for measuring specific antibody responses against human papillomavirus (HPV). However, this assay is labor intensive and therefore very difficult to implement in large-scale studies. Previous studies have evaluated the agreement between virus-like particle (VLP)-based ELISA and PBNA for measuring HPV vaccine-induced antibodies. However, the concordance of these assays to detect antibodies induced by natural infection has not yet been fully elucidated. In this study, the results of an Escherichia coli (E. coli)-expressed VLP-based ELISA were found to be highly concordant with those of a baculovirus-expressed VLP-based ELISA (r = 0.96 and 0.97 for HPV-16 and HPV-18) when detecing HPV vaccine induced antibodies and the concordance was medium (r = 0.68 and 0.68 for HPV-16 and HPV-18) when assessing natural infection induced antibodies. The results of the E. coli expressed VLP-based ELISA correlated well with those of the PBNA when testing 1020 post-vaccination human sera collected at one month after vaccination with the E. coli expressed VLP-based bivalent HPV vaccine (r = 0.83 and 0.81 for HPV-16 and HPV-18). The agreement and correlation were moderate (kappa<0.3 for both HPV types 16 and 18, r = 0.59 and 0.68 for HPV-16 and HPV-18, respectively) when assessing 1600 serum samples from unvaccinated women of age 18-25 years. In conclusion, the VLP-based ELISA is an acceptable surrogate for the neutralizing antibody assay in measuring vaccine responses. However, the use of the VLP-based ELISA in epidemiological studies should be carefully considered.