Prognostic relevance of cardiac troponin T levels and their dynamic changes measured with a high-sensitivity assay in acute ischaemic stroke: Analyses from the TRELAS cohort

- Elevated cTnT (>99th percentile) occurs in >50% of acute ischaemic stroke patients when measured in a highly sensitive assay
- Moderate myocardial injury (undetectable with conventional cTnT assays) is associated with early unfavourable functional outcome
- The optimal cut-off for prediction of unfavourable outcome is 16 ng/l.
- Dynamic changes in cTnT are associated with in-hospital death.

Background/objectivesTo assess the prognostic relevance of cardiac troponin T (cTnT) values and their serial changes as measured with a high-sensitivity assay in acute ischaemic stroke.Methods and resultsIn total, 1016 consecutive acute ischaemic stroke patients were registered prospectively within a two-year period. High-sensitivity cTnT was quantitated on admission and the following day. Patients were stratified into three groups: below the 99th percentile of a healthy reference population (< 14 ng/l; n = 409), moderately elevated cTnT (below the median of remaining patients, 14-30 ng/l; n = 299) and high cTnT (> 30 ng/l; n = 308). Outcome measures were unfavourable outcome (modified Rankin Scale ≥ 2 at hospital discharge, n = 604) and in-hospital mortality (n = 36). Peak cTnT levels were significantly associated with unfavourable outcome (adjusted odds ratios versus reference for moderately elevated cTnT 1.7, 95% confidence interval [CI] 1.1-2.7 and for high cTnT 3.1, 95% CI 1.8-5.6). The optimal cut-off for determining unfavourable outcome proved to be 16 ng/l. Adding this cut-off to clinical variables led to a significant improvement of c-statistics (0.851 versus 0.838, p = 0.02). Dynamic changes (δ ≥ 50%) of cTnT were detected in 137 patients (13%). Multiple Cox regression analysis showed an independent association of dynamic changes in cTnT with in-hospital mortality (hazard ratio 2.3, 95% CI 1.1-4.7). The model-fitting (p < 0.001) and c-statistics (0.889 versus 0.863, p = 0.08) improved most after dynamic change in cTnT had been added to a regression model which included clinical variables alone.ConclusionsOur study adds novel findings relevant for interpretation of highly sensitive cTnT assays in acute ischaemic stroke: a) Myocardial injury is detectable in more than half of patients; b) even moderately elevated cTnT is associated with unfavourable outcome (optimal cut-off 16 ng/l); and c) dynamic changes in cTnT indicate an increased risk of in-hospital death.