Do high-sensitivity cardiac troponin i clinical performance data in package inserts reflect realistic clinical expectations?
Christenson RH, Duh S, Apple FS, Nowak RM, McCord J, De Filippi CR. Do high-sensitivity cardiac troponin i clinical performance data in package inserts reflect realistic clinical expectations? Clin Chem. 2018;64:S26-S27.
Background: Cardiac troponin (cTn) is a cornerstone for diagnosis and management of myocardial infarction (MI). High-sensitivity troponin (hs-Tn) provides earlier MI rule-in/rule-out. The 2015 European Society of Cardiology guidelines proposed hs- Tn algorithms for NSTEMI-management. However, estimates of hs-cTn performancemay vary based on the anchor-time used for analysis. Typically cTn data have been organized relative to the 'first study sample' (1stSS) collection time, including inmanufacturers' package inserts. Alternatively data can be organized based on time of presentation (TOP). We investigated diagnostic performance of the ADVIA Centaur® hs-TnI (TNIH) up to 3.5 hours using TOP or 1stSS anchor-time.Method: Samples from >2,300 'all-comer' suspected MI patients were collected at 29 IRB-approved sites; 310 (13%) patients were adjudicated MI positive. The TNIH assay was validated as hs-TnI: total CV was 2.9% at the female 99th percentile (37 ng/L); using the AACC Universal Sample bank, 58.9% and 85.8% of values from healthy women and men, respectively, exceeded LoD (1.6ng/L).Results: Median delay between local standard-of-care first blood draw and 1stSS was 49 min. Table-Section A shows TNIH sex-specific performance based on TOP analysis (n=4,502 observations). Table-Section B displays sex-specific performance using 1stSS (n=6,346 observations). At 3.5-hours, TOP sensitivity was 95.8% & 89.7% and Negative Predictive Value (NPV) was 99.5% & 98.1% for women and men, respectively. For 1stSS at 3.5-hours, sensitivity was 95.0% & 89.2% and NPV was 99.3% & 97.6% for women and men, respectively. Although at 3.5-hours, TOP (n=198) had more adjudicated MIs than 1stSS (n=114) (p<0.001), sensitivity/NPV was not significantly different than anchoring at 1stSS (p-value=0.44).Conclusion: Reporting performance relative to TOP or the 1stSS does not yield different values for sensitivity/NPV or other MI diagnostic parameters at 3.5 hours. We advocate reporting data anchored to clinical presentation time to facilitate harmonizing with clinical guidelines.