A-375 Verifying analytical performance of i-STAT high sensitivity troponin cartridges in an urban academic medical center emergency department

Abstract

Troponin I is a commonly used laboratory test for detection of myocardial infarction. A point-of-care (POC) option for Troponin I allows for quicker clinical decision making, which is particularly useful in the Emergency Department (ED). However, the Abbott contemporary cartilages currently used in this ED for POC testing are scheduled for discontinuation. This study aimed to verify that the analytical performance of the i-STAT hs-TnI cartridge is acceptable for ED implementation. The validation study was performed in an urban academic medical center using four i-STAT analyzers. Precision was assessed with two levels of quality control (QC) material ran four times daily over five days, for a total of twenty runs per level per analyzer. Analytical measurement range (AMR) was verified using manufacturer-provided materials (2.9-1000 ng/L). Method comparison utilized whole blood samples collected from ED patients with troponins <1000 ng/L, measured on i-STAT in duplicates, then centrifuged with separation of serum for measurement on the Beckman DxI hs-TnI 800 assay in duplicates. Acceptance criteria included coefficient of variation (CV) =10%, AMR slope 0.90–1.10, correlation coefficient (R) >0.95, and allowable total error (ATE) ±30% or 3 ng/L. Sex-specific reference intervals (=13 ng/L female; =28 ng/L male) were evaluated in healthy volunteers, with BNP and core lab hs-TnI also measured to confirm lack of heart disease in these volunteers. Precision goals were met across all four analyzers and QC levels. Low-level QC (19.3-20.8 ng/L) exhibited CVs of 6.5–9.5%, while high-level QC (527.9–550.2 ng/L) exhibited CVs of 6.9–9.5%. AMR verification confirmed that three analyzers met the expected 2.9–1000 ng/L range within ±20% at both ends, with slopes of 0.893–0.928 and acceptable calibration verification values. One analyzer demonstrated a slope outside the desired range but maintained acceptable recovery, allowing overall AMR acceptance. Method comparison using 46 ED samples demonstrated acceptable correlation with the core laboratory assay (slope=0.918, intercept=2.776, R=0.995, mean bias ∼5.6%). Reference interval verification confirmed sex-specific cutoffs, with 19/20 males and 20/22 females falling below manufacturer upper 99th percentile. These three volunteers had core lab hs-TnI within normal limits. Manufacturer-reported interferences were largely minimal, with bilirubin noted as a possible source of ±10% bias in whole blood. No safety or toxicity risks were identified in reagent review. The Abbott i-STAT hs-TnI assay satisfied analytical performance requirements for precision, analytical range, accuracy, and reference intervals when compared to a high-sensitivity laboratory platform. The observed analytical performance supports clinical implementation of the i-STAT hs-TnI for ED use.