COMPARISON OF COMMERCIAL ELISA FOR DETECTION OF ANTIBODIES TO THE VIRAL CAPSID ANTIGEN ( VCA ) OF EPSTEIN-BARR VIRUS ( EBV )

Epstein-Barr Yirus (EBY) is the causative agent of infectious mononucleosis (1M). The virus is primarily transmitted via saliva, with seroprevalence of approximately 90% in adults (Rocchi el al., 1977; Yao etal., 1985). Primary EBY infection in infants and young children is usually asymptomatic or at least very mild. In adults, 1M is a selflimiting disease characterised by fever, pharyngitis, generalised lymphadenopathy and splenomegaly (Evans, 1976). Traditionally, the serodiagnosis of EBY infection is based on the detection of heterophile antibodies. However, at least 10% of people with acute 1M fail to produce heterophile antibody with this figure being much higher in children (Evans et al., 1975; Sumaya and Ench, 1985). A more specific method of serodiagnosis of 1M is the detection of EBY -specific antibodies that appear during primary infection. Most serodiagnostic tests are based on the detection of antibodies to the Yiral Capsid Antigen (YCA) (Weidbrauk et al., 1993; Field et al., 1995). YCA antibodies of the IgG and IgM class become detectable early in the acute phase of infection. Anti-YCA IgM levels peak about 3 weeks after the onset of clinical disease, then rapidly decline to become undetectable. In contrast, YCA IgG antibodies persist for life (Youmans et al., 1985; Nikoskelainen et al., 1975). Although the indirect immunofluorscence assay (IF A) is the standard method for detection of antiYCA antibodies, it is not without limitations. IFA is difficult to standardize, time-consuming and not convenient for large scale screening (Wiedbrauk et al., 1993). The development of enzyme-linked immunosorbent assays (ELlSAs) for specific EBY antibodies has provided a sensitive and specific method that is objective and convenient for large scale screening (Hopkins et al., 1982; Luka et al., 1984) and assays for YCA are now available commercially.

In this study we have compared commercially available EUSAs for the detection of EBY yeA IgM and EBY yeA IgG (PanBio, Australia; Gull Laboratories, U.S.A.; Incstar, U.S.A.) using 104 sera presented for routine pathological testing .The EUSAs were performed according to the manufacturers instructions (Table I) with the serum diluent for the Gull and PanBio IgM tests containing goat anti-human IgG to remove competing IgG and rheumatoid factor.A positive result was defined as being positive in all three EUSAs.Discrepant sera were tested using a commercial IFA (MRL, U.S.A.).,with IFA positivity defining a positive result.Similar criteria were used to determine a negative result.
The three yeA IgM EUSAs gave the same results in 9411 04 (90%) of sera.The PanBio and Gull ELiSAs showed the highest agreement (97 %) and correlation (Pearson's r = 0.93, P < 0.000 I) (Table 2, Figure ld).All 3 assays gave high sensitivity when IF A was used to distinguish between the 10 discrepant test results.The Gull yeA IgM ELISA detected allIS IgM positive specimens (sensitivity 100%), while the PanBio and Incstar yeA IgM ELiSAs failed to detect IllS and 2115 specimens respectively.The three assays also showed high specificity: 94% for Incstar and 98 % for the PanBio and Gull ELiSAs (Table 2).
The yeA IgG ELiSAs also showed high agreement, though this was lower than observed with the IgM ELiSAs.For example, the PanBio and Incstar tests showed discrepant results in 17 of 104 sera (overall agreement 84%) (Table 2).Yariation between test values was evident when Pearson's correlation analysis was performed, with the highest correlation between yeA IgG values observed with the PanBio and Gull ELiSAs (r = 0.65) (Figure I a).The three yeA IgG ELISAs gave the same result in 85/ 104 (82 %) of sera, and the ELiSAs showed high sensitivity (>90%) when IFA was used to classify discrepant results (Table 2).However, the PanBio ELISA showed higher specificity (100%) than the Gull and Incstar yeA IgG ELiSAs (73% and 47 %), and the difference between specificity in the PanBio and Incstar ELiSAs was significant (Fisher' s Exact Test, p = 0.0004) (Table 2).
Differences between assay val ues and resul ting sensi ti vity and specificity of the different assays is likely to be a reflection of the assay formats and antigens used by different manufacturers.The PanBio and Gull tests both use native YeA, while the Incstar test uses a synthetic peptide encompassing the major epitopes of the p 18 protein.This protein is one of the major components of the yeA complex (Yan Gransven et at., 1994).The unglycosylated synthetic peptide antigen used in the Incstar ELISAs may be devoid of some epitopes since it would not contain carbohydrate antigens.Furthermore, it represents only part of the yeA complex.This may be the cause of the lower sensitivity observed in the Incstar IgM test, as observed in another study (Weidbrauk et at., 1993).The PanBio and Gull tests showed a higher correlation and this is likely to reflect the use of similar native antigens .However, differences were still observed between these assays which may reflect differences in the native yeA preparations or methods used (e.g.blockers, diluents and plates).This was particularly evident with yeA IgG, where the PanBio ELISA showed perfect specificity (100 %) while the Gull ELISA showed a specificity of only 73 % .Furthermore , the Incstar ELISA showed poor specificity (47%), which may reflect the exposure of cross-reactive epitopes due to incorrect folding or exposure of cryptic non-specific epitopes masked in the nati ve molecule by carbohydrate.Previous studies have reported good sensitivity and specificity for the Gull VCA IgM and VCA IgG ELISAs when Organon Technica and Gull IFA were used as the reference tests (Weidbrauk et al., 1993 ;Gerber et al., 1996).In the same study the Incstar YCA IgM ELISA did not perform as well relative to the Gull IFA (Weidbrauk et (II. , 1993) .The Incstar IgM ELISA showed poor sensitivity (53%) and high specificity (100%),  Incstar IgG (ratio) Incstar IgM (ratio) compared to a sensitivity of 87% and specificity of 94% reported in this study.In this study the commercial IFA (MRL, USA) and ELISAs used were from different manufacturers.The YCA IgG IFA used a native YGA antigen, while the YCA IgM IFA used recombinant YCA expressed in a eukaryotic system.The YCA IgM antigen would be glycosylated, though differently to the native antigen.Consequently, the use of the MRL IFA would eliminate possible bias associated with using IFA and ELISAs from the same manufacturer.This was not the case in a previous study where Gull ELISA and IF A were compared (Gerber et al., 1996).In summary, commercial ELISAs for YCA IgM and IgG provide a simple and reproducible method for the serological diagnosis of infectious mononucleosis.However, the different tests can show variation in performance based on the different methods and reagents (including the antigens) used.

Figure I :
Figure I: Comparison between ELISA va lues in (a) Gull and PanBio VCA IgG; (b) PanBio and Inestar VCA IgG; (c) Gull and Incstar VCA IgG; (d) Gull and PanBio VCA IgM ; (e) PanBio and In csta r VCA IgM; and (0 Gull and Incstar VCA IgM .Assay valu es arc represented on a log scale as a rati o of th e test sa mpl e valu e to th e cutoff se ra va lue.The cuto ff se ra va lue (ratio = 1.0) is sho w n.Pearso n' s r-values, calculated after log tra nsformation of data, are also shown.In all cases, p < 0.0001 when Pearson ' s correlation was performed.

Table 1 .
Comparison Between Commercial EBV VCA ELISA Methods