Evaluation of a commercially available radioimmunoassay and species-specific ELISAs for measurement of high concentrations of insulin in equine serum

OBJECTIVE: To evaluate a human radioimmunoassay (RIA) and equine and high-range porcine (hrp) species-specific ELISAs for the measurement of high serum insulin concentrations in ponies. SAMPLES: Serum samples from 12 healthy nonobese ponies (7 clinically normal and 5 laminitis prone; 13 to 26 years of age; 11 mares and 1 gelding) before and after glucose, insulin, and dexamethasone administration. PROCEDURES: Intra-and interassay repeatability, freeze-thaw stability, dilutional parallelism, and assay agreement were assessed. RESULTS: Assay detection limits were as follows: RIA, < 389 μU/mL; equine ELISA, < 175 μU/mL; and hrp ELISA, 293 to 8,775 μU/mL. Mean ± SD intra- and interassay repeatability were respectively as follows: RIA, 6.5 ± 5.1 % and 74 ± 3.4%; equine ELISA, 10.6 ± 11.0% and 9.0 ± 4.6%; and hrp ELISA, 19.9 ± 172% and 173 ± 16.6%. Freezing and thawing affected measured concentrations. Dilutional parallelism in the RIA was only evident when insulin-depleted equine serum was used as a diluent (percentage recovery, 95.7 ± 274%); in the ELISAs, dilutional parallelism was observed when a zero calibrator was used. Agreement between RIA and equine ELISA results was good for samples containing concentrations < 175 μU of insulin/mL (bias, -18.5 ± 25.5 μU/mL; higher in RIA). At higher concentrations, assay agreement was poor between RIA and equine ELISA results (bias, -185.3 ± 98.7 μU/mL) and between RIA and hrp ELISA results (bias, 25.3 ± 183.0 μU/mL). CONCLUSIONS AND CLINICAL RELEVANCE: Agreement among results of the 3 assays was variable, and dilutional parallelism was only evident with the RIA when insulin-depleted equine serum was tested. Caution is recommended when evaluating high insulin concentrations measured with the RIA or ELISAs.