Evaluation of six serological ELISA kits available in Italy as screening tests for equine infectious anaemia surveillance.

Background: ELISAs are known to have a higher diagnostic sensitivity than the agar gel immunodiffusion (AGID) when employed for serological diagnosis of equine infectious anaemia (EIA). For this purpose, an "in-house" and five commercial ELISAs available in Italy were assessed by the National Reference Centre for EIA for their analytic specificity (Sp); precocity, defined as capability of detecting first antibodies produced during a new infection; precision based on repeatability and reproducibility, estimated from the coefficient of variation (CV); accuracy, estimated from multiple K and relative Sp and sensitivity (Se). Two serum panels, positive for non-equine retroviruses and the most frequent equine viruses, were employed to measure analytic Sp. ELISA precocity was also compared to that of one "in-house" and three commercial AGID kits, employing a panel of sera, collected weekly from horses infected with modified EIA viruses. Precision and accuracy were defined using results of a panel containing positive and negative sera examined in an inter-laboratory trial with the participation of the ten Official Laboratories. Furthermore, a questionnaire was used to assess the appropriateness of each kit for routine use.
Results: Analytic Sp was 100%, while the 75th percentile of CVs for positive sera varied from 0.4% to 12.73% for repeatability and from 1.6% to 44.87% for reproducibility. Although CV of the negative serum was constantly high, its outcome was unaltered. Relative Se ranged from 98.2% to 100%, relative Sp was constantly 100% and multiple K ranged from 0.95 to 1. Precocity differed among the assays: three kits detected 4.8% and 42.9% positive samples on 21 days post infection (dpi), all assays detected positive samples on 28 dpi, between 47.6% and 95.2%. Precocity of ELISAs was superior to that of the AGIDs except for two assays. In view of the feedback obtained from the questionnaires, all kits were considered appropriate for routine use.
Conclusion: All ELISAs having high Se and precocity are preferable as a screening test in EIA surveillance programmes to the AGID tests examined. These two tests can be incorporated in a serial diagnostic pathway to improve the efficacy of a surveillance plan.