A-007 Comparisons of High-Sensitivity Cardiac Troponin I Assays: Quantitative Differences in Non-Selectivity in Patient Samples

Abstract

High-sensitivity cardiac troponin (hs-cTn) I&T measurements are a cornerstone for diagnosis and risk assessment of heart disease and are essential for optimum patient care. Unfortunately, different assays show different results, and neither a widely-adopted cTnI reference material nor reference method exists. hs-cTnI assay comparisons using a ‘commutable’ reference-material or calibrator might be a viable strategy toward hs-cTnI harmony provided assays have similar high-selectivity for the analyte/measurand. Conversely, ‘differences in nonselectivity’ (DINS) expresses an assay’s inability to distinguish a specific molecular-target from other similar molecules. Typically, DINS is expressed qualitatively; here a quantitative method for assessing DINS is used for comparing hs-cTnI assays. Nine FDA-cleared hs-cTnI assays (see Figure Legend) are quantitatively examined in pairs for a total of 36 unique method-comparisons. Each paired-comparison was conducted using 40 individual lithium-heparinized samples from consented subjects having acute coronary syndrome with release of hs-cTnI levels ranging from 19 to 3600 ng/L (Siemens Atellica TnIH values). Selectivity data were analyzed by a novel statistical tool (Fauskanger et al., Biom J. 2025 Feb;67(1):e70032) that quantifies nonselectivity by computing the amount of heterogeneity between instruments that cannot be explained by measurement error alone. Each of the 36 method comparisons was quantitatively categorized as: Absence DINS; Acceptable DINS; or Excessive DINS to identify hs-cTnI method-comparisons that are influenced by excessive nonselectivity differences. With this method an acceptable DINS is Zeta=3.0 (see Figure). Each of the patient samples was compared in quadruplicate with each method-pair. The total number of measurements for this study was n=40 samples X 4 replicates X2 methods =320 comparisons per method-pair X 36 method-pairs for a grand total of 11,520 measurements. The Figure shows results for all 36 hs-cTnI method comparisons. The estimated DINS for each comparison is represented by a dot, and the vertical bars show the associated 95% confidence limits. Seven of the 36 method comparisons (19.4%) showed results that had an absence of DINS. These were all method comparisons of instruments from the same manufacturer. Another 6 of the 36 method comparisons (16.7%) had DINS rated as acceptable. The remaining 13 method comparisons (36.1%) had DINSthat did not overlap with the acceptable Zeta=3 of the absence/acceptable lower limit and are graded as having excessive DINS. Historically, differences in selectivity between method comparisons were relegated to qualitative speculations about differences in reagent antibodies, difference in capture versus detection reagents and other factors that are difficult to evaluate. Our assessment showed that some pairs of assays, often from the same manufacturer appeared to have similar selectivity (low DINS values). However, our results also demonstrated substantial differences in selectivity for some assay pairs, illustrating the difficulty of comparing results from different assays. Of interest, some manufacturers, e.g. Siemens are able to use different antibody reagents, and detection strategies and still achieve an absence of DINS. Also, method comparisons including Abbott, Beckman-Coulter, Siemens and Pathfast assays are able to produce assays that demonstrate acceptable DINS. Leveraging strategies to lower DINS would improve method agreement and be a substantial step toward harmonization/standardization.