Skip to main content
SLU publication database (SLUpub)

Conference abstract2009


Öberg, Josefine; Lilliehöök, Inger; Wattle, Ove; Bröjer, Johan; Karlsson, Åsa


Equine serum insulin has previously been analyzed with methods developed for human serum insulin. The performance and results of these human RIA and ELISA methods vary considerably because of different cross-reactivity of the anti-human antibodies and matrix problems. A species optimized, quantitative method (Mercodia Equine Insulin ELISA) for measurement of equine serum insulin has currently been produced by Mercodia, Uppsala, Sweden. It is a direct sandwich ELISA in which two monoclonal antibodies are directed against two separate antigenic determinants on the equine insulin molecule. The detection limit is 0.01µg/L and the detection range is 0.02-1.50 µg/L. Our study evaluated the precision and linearity of the Mercodia Equine Insulin ELISA and compared it with two methods for human insulin (Mercodia Human Insulin ELISA and Coat-A-Count Insulin RIA kit from DPC, Siemens Diagnostics, Los Angeles, CA, US). Biological relevance of the assay was evaluated by measuring insulin before and after feeding. Forty healthy horses were sampled before and 90 minutes after feeding. These 80 serum samples were analyzed with all three methods. Pearson´s correlation values were calculated. Precision for the equine ELISA was evaluated by determining the intra- and inter-assay coefficient of variation (CV). Sera from one horse with a mean concentration of 0.116 µg/L was analyzed in duplicate on fifteen different assay runs. Another sample (mean 0.017µg/L) was analyzed in duplicate four times on one assay the same day. One sample with high insulin value (0.998µg/L) was diluted with physiologic saline in five steps. The results were compared with mathematically predicted values and recovery was determined. Results from the equine ELISA had high correlation both with the human ELISA and the RIA, r² = 0.97 and 0.97 respectively. Inter-assay CV was 10.7% and intra-assay CV was 4.6%. Recovery was 92-122%. Insulin concentration was significantly greater after feeding compared with the pre-feeding sample in 40 horses (Sign test p<0.0001). The median insulin level using the equine ELISA before feeding was 0.07 µg/L compared with 0.29µg/L after feeding. In comparison, the insulin results from human RIA and ELISA increased from median 4.8 and 7.8mU/L to 21.1 and 26.1mU/L respectively. The Mercodia Equine Insulin ELISA had good precision and linearity, and showed high correlation with both the human ELISA and the Coat-A-Count RIA. This commercial assay’s advantages are that it is developed for equine samples and results are quantitative (µg/L) instead of using activity units based on human insulin

Published in