Comparative evaluation of the agar gel immunodiffusion test and two commercial ELISA kits for the serodiagnosis of equine infectious anemia.

Selected sets of serum samples of horses were tested blindly in a comparative investigation for antibodies against Equine Infectious Anemia (EIA) virus. Three commercial kits were used, a well-established agar-gel immuno-diffusion kit which our laboratory has been using routinely for 14 years on one hand, a competitive ELISA kit (CELISA) and a non-competitive ELISA kit on the other hand. The American EIA Reference Laboratory in Ames cotested 56 serum samples with the same 3 products, with highest-level correlation, thereby ascertaining full dependability of our own results. Five EIA experts supplied us critically weak or doubtfully reacting serum samples of experimentally infected horses together with their own test results, by necessity limited to the then available AGID in most instances. A high degree of correlation was found between our and their AGID results. In our own laboratory good correlation was found between the AGID test and one lot of the CELISA product. Time of seroconversion was coincident in some experimentally infected horses, partly AGID, partly CELISA proved more sensitive. Another lot of the CELISA product deteriorated completely long before the warranted validity, an unpleasant finding experienced by many other laboratories alike. The non-competitive ELISA product showed unacceptable inter-lot differences, oscillation between positive and negative results on consecutive samples of one and the same horse, never reacted with the weak positive International Reference Serum, and one lot deteriorated well beyond its expiration date. We discuss our results with the background: high sensitivity versus false-positive horses and advocate to maintain at their present sensitivity levels the AGID and the CELISA tests and not to push them further, as would be technically possible.