Abstract Objectives The laboratory sought to monitor and eliminate reported patient cardiac troponin outlier results using the Abbott ARCHITECT STAT Contemporary Troponin-I test. Methods The laboratory implemented an immediate repeat test for all patient samples yielding an Abbott ARCHITECT STAT contemporary Troponin I concentration greater than the combined male and female 99th percentile of 28 ng/L. The lower troponin concentration of the paired testing was reported. Samples with reference range troponin concentration were reported without repeat testing. Blood collected in Becton Dickinson Vacutainer PST Gel and Lithium Heparin tubes was centrifuged using a HemoCue StatSpin Express 4 centrifuge 4,000 g × 5 minutes and immediately tested on the primary tube. Troponin outliers were defined as paired samples showing a percentage change based on the low value of the pair exceeding the significant change threshold of 4.7x%CVanalytical for brackets of troponin concentrations (P0.0005), representing a significant change value for serial tests of 1.41xZxCVa%, where Z is 3.3. Critical troponin outliers were sample pairs with one result exceeding and one result within the 99th percentile reference range. Turnaround time was defined as time of receipt to time of verification. Results The outlier rate, based on 4,575 troponin tests performed on 3,797 patient samples, was 0.48% (N = 22) of troponin determinations, 0.58% of test samples, and critical outliers 0.13% (N = 6) of troponin determinations, 0.16% of test samples. The significant change threshold at 28 ng/L was 37%. The worst outlier pair was 5,978 ng/L, true value 3,184 ng/L; critical outlier pair 146 ng/L, true value of 14 ng/L. Twenty-one percent of test samples were tested in duplicate; 777 samples had troponin concentration >28 ng/L. Median and mean turnaround time were 32 minutes; 90% of orders completed in 54 minutes. Conclusion The laboratory detected and prevented reporting erroneously high outlier troponin results in 0.58% and critical outlier troponin results in 0.16% of test samples using an immediate repeat testing protocol.
Read full abstract