Comparative evaluation of three ELISA techniques and an indirect immunofluorescence assay for the serological diagnosis of Epstein-Barr virus infection

Background: The reference method for detecting specific Epstein-Barr virus (EBV) antibodies is indirect immunofluorescence (IF) with EBV-infected cells. The availability of protein purified from infected cells and more recently of recombinant polypeptides designed to contain immunodominant epitopes, has enabled the development of commercial enzyme-linked immunosorbent assays (ELISA) for the specific serodiagnosis of EBV infection.

Objective: Evaluation of ELISA-based EBV serodiagnosis in comparison with indirect immunofluorescence.

Study design: We have first compared three commercial ELISA test systems with our in house indirect immunofluorescence assay for classifying correctly a set of serum samples into clinical categories (acute infection, past infection, interfering non-EBV infection, persistent infection). Additionally a prospective analysis with the best performing ELISA test (Enzygnost) was then carried out by running the ELISA test in parallel with the indirect immunofluorescence assay on 324 consecutive clinical samples sent to our laboratory for EBV serodiagnosis.

Results: For the serodiagnosis of past EBV infection and acute EBV infection all three commercial ELISAs performed well in comparison with indirect immunofluorescence. When testing samples positive for cytomegalovirus (CMV), Toxoplasma or herpes simplex IgM, interference in the IgM tests was observed with the three ELISAs. In some instances we could demonstrate that the positive IgM results were due to EBV reactivation. The observed discrepancies between ELISA and IF for the serodiagnosis of chronic EBV infection or EBV reactivation, point to the difficulty for the serodiagnosis of persistent EBV infection on single serum samples. According to our prospective study the EBV IgG determination was accurate. A positive IgM result was not always indicative of an acute infection. Positive IgM results due to EBV reactivation were observed. A positive EBV nuclear antigen (EBNA) IgG result in those samples precluded acute infection.

Conclusions: 90–95% of samples could be classified correctly into clinical categories by a two parameter ELISA system detecting IgG and IgM against a standardized mixture of EBV antigens, allowing standardization and automation of EBV-specific serology. The absence of EBNA IgG was useful as a second line confirmatory assay for acute EBV infection.