Impact of Implementation of the High-Sensitivity Cardiac Troponin T Assay in a University Hospital Setting

To the Editor: The performance of the high-sensitivity cardiac troponin T assay (hs-cTnT)1 has been evaluated in a multicenter study (1). Effective July 2009, we replaced the fourth-generation troponin T assay (cTnT) with the hs-cTnT assay in clinical practice. This study audits the impact of this implementation. The hs-cTnT, implemented on the cobas e 411 platform (Roche Diagnostics), fully replaced the cTnT performed on the Elecsys 2010 analyzer [cutoff, 30 pg/mL—based on actual assay performance (10% CV concentration)]. We obtained a detection limit of 5 pg/mL, a 99th percentile of 15 pg/mL, limited comparability with the cTnT at concentrations <100 pg/mL (on average, a 30-pg/mL cTnT concentration yielded a value of approximately 65 pg/mL with the hs-cTnT) and mean CVs of 9.1% for the cTnT (at 39 pg/mL) and 8.5% for the hs-cTnT (at 17 pg/mL). We retrieved hs-cTnT results for the first 3 months after implementation (July 16 to October 15, 2009) and cTnT results for the same period 1 year previously. Results were dichotomized as positive or negative with respect to cutoffs. Among troponin-positive patients with at least 2 results during their examination, we divided marker-release curves on the basis of typical or atypical kinetics. We defined as “typical” an increasing or decreasing pattern showing a troponin change between 2 consecutive samples exceeding +46% for increasing troponin results and −32% for decreasing results. Otherwise, the troponin pattern was considered “atypical.” For definition of these percentage changes, we referred to the short-term biological variation for troponin I (2). We are aware, however, that the 2 cardiac troponins may have different biological kinetics in blood, so their biological variation may be different. In the evaluated period, 2287 hs-cTnT tests were performed during 1371 examinations of 1137 patients. Correspondingly, 2170 cTnT tests were performed during 1409 examinations of 1205 patients. After hs-cTnT implementation, a 5.4% increase in the hospital-wide test volume was recorded, despite a slight decrease in the number of admitted patients and examinations. The mean (SD) number of troponin tests per examination was 1.54 (1.0) before and 1.67 (1.1) after hs-cTnT implementation (P < 0.0001), with a single test ordered in 67.5% and 60.2% of examinations, respectively. The distribution of troponin orders and positive-test rates in different wards is shown in Table 1. A positive result was found in 31.7% of cTnT tests and in 58.7% of hs-cTnT tests (relative difference, +85%), corresponding to 25.3% and 51.6% positive examinations, respectively (P < 0.0001). Of all the hs-cTnT positive results, 64% fell in the 16–65 pg/mL interval, previously negative with the cTnT. In the emergency department after hs-cTnT implementation, the number of hospitalized patients with positive troponin results increased from 158 to 292 (+85%), but the rate of admission in intensive care and non–intensive care departments was unchanged (P = 0.108). In the same periods, 16 cTnT-positive patients (8.5%) and 109 hs-cTnT–positive patients (26.6%) were discharged. Of these discharged patients, 1 cTnT-positive patient and 13 hs-cTnT–positive patients were readmitted to the emergency department in the subsequent 2 months (P = 0.804, between the 2 assays). All wards accounting for <10 tests were gathered into this group. All wards accounting for <10 tests were gathered into this group. We audited 458 cTnT and 546 hs-cTnT curves, of which 39.1% and 69.0%, respectively, had at least 1 positive result (P < 0.0001). The difference in the percentage of positive curves displaying a typical marker release was not significant (17.2% for the cTnT vs 20.5% for the hs-cTnT, P = 0.32). A higher absolute number of typical positive curves was observed after hs-cTnT implementation (from 79 to 112). This increased ability to detect events involving acute marker release was fully explained by the number of typically positive curves in which the hs-cTnT result never exceeded 65 pg/mL (n = 38). The replacement of the cTnT with the hs-cTnT markedly increased the rate of positive tests. A similar outcome was previously described for a contemporary sensitive troponin I assay (3). What is unique in our experience is the magnitude of the increase in positive results after hs-cTnT introduction, which was based on implementing this assay in a routine protocol. The number of examinations with positive results increased from approximately 25% to >50%. Although the hs-cTnT could appear confounding as a test less specific for the diagnosis of myocardial infarction (4), we were unable to demonstrate differences in the percentage of curves with a typical marker release when we compared the cTnT and the hs-cTnT. An interpretative approach based on the demonstration of a pathophysiology-defined release of troponin in the blood may allow the same specificity performance to be achieved when using different generations of troponin T assays, thus supporting the use of serial testing for clinical decisions (5). high-sensitivity cardiac troponin T assay fourth-generation troponin T assay.