Clinical evaluation of the ADVIA Centaur XPT chemiluminescent immunoassay for equine insulin measurement

Abstract

Background: Insulin dysregulation is a key component of equine metabolic syndrome and is commonly assessed using basal insulin concentrations. Different analyzers may yield variable insulin results, limiting comparability. Aims/Objectives: To assess the analytical performance of the ADVIA Centaur XPT chemiluminescent immunoassay (CLIA) for equine insulin measurement and to establish adapted decision thresholds based on comparison with a previously used CLIA method. Methods: Precision, linearity, and dilution recovery were assessed for the ADVIA Centaur XPT. A total of 89 equine serum samples submitted for basal insulin testing were measured using the ADVIA Centaur XPT and Immulite 2000 XPi analyzers. Results were compared using Deming regression, Bland-Altman analysis, and weighted kappa statistics to evaluate agreement in classification. Results: Intra-assay coefficients of variation were 3.56 %, 2.01 %, and 1.92 % for low, medium, and high insulin concentrations, respectively; inter-assay variation was 5.19 %, 5.78 %, and 5.68 %. Deming regression showed a proportional bias, with the ADVIA Centaur XPT consistently measuring lower insulin concentrations compared to the Immulite 2000 XPi. Bland-Altman analysis confirmed this bias across the measurement range. Spearman correlation between the two methods was 0.92 (95 % CI: 0.87–0.94; P < 0.0001), indicating a strong association in rank order. Classification agreement using adapted decision limits yielded a weighted kappa of 0.82, indicating strong agreement. Conclusion: The ADVIA Centaur XPT demonstrated acceptable precision. When insulin concentrations were classified using adapted thresholds, the agreement between the two CLIA analyzers was strong, suggesting that the ADVIA Centaur XPT may be suitable for measurement of equine insulin concentration.