Background: It is unclear whether serum high-sensitive troponin T (hs-TnT) levels at the convalescent stage of ST-elevation myocardial infarction (STEMI) are associated with long-term mortality. Methods: This study enrolled a total of 2944 consecutive STEMI patients who were registered in the Osaka Acute Coronary Insufficiency Study between 2000 and 2009, and whose hs-TnT levels were evaluated at the convalescent stage. Patients were divided into four hs-TnT category groups according to the results of survival classification and regression tree (CART) analysis. The impact of hs-TnT levels on 5-year mortality was evaluated using multivariate Cox regression analysis. Results: Only one patient had hs-TnT level below the detection limit of the assay (<0.003 ng/mL). The median hs-TnT level was 0.025 (quartile 0.011-0.083) ng/mL. During the median follow-up period of 1782 days, 188 patients died. Survival CART analysis revealed that the 1st, 2nd, and 3rd discriminating hs-TnT levels to discern 5-year mortality were 0.028, 0.008, and 1.340 ng/mL, respectively. The adjusted hazard ratios for the medium-low (0.009-0.028 ng/mL), medium-high (0.029-1.340 ng/mL), and high-risk (≥1.341 ng/mL) groups were 3.03 (95% confidence interval 1.18-7.77, p = 0.021), 4.29 (1.63-11.28, p = 0.003), and 8.68 (2.20-34.27, p = 0.002), respectively. Integrated discrimination improvement (IDI) analysis revealed that incorporation of this hs-TnT classification scheme with other clinical variables statistically improved the discriminatory accuracy for 5-year mortality, with a time-dependent IDI of 0.0076 (p = 0.033). Conclusions: hs-TnT levels at the convalescent stage were associated with long-term mortality in STEMI patients. Even subclinical elevation of hs-TnT levels was associated with increased 5-year mortality.
- Classification and regression tree
- High-sensitive troponin T
- Integrated discrimination improvement
- ST-elevation myocardial infarction
ASJC Scopus subject areas
- Cardiology and Cardiovascular Medicine