23 A Rising High Sensitivity Troponin Trend in Medical Intensive Care Unit Patients Predicts Unfavorable Outcomes

Background Cardiac troponins (cTns) are sensitive and specific biomarkers for cardiomyocyte injury, and their temporal quantification is a critical diagnostic consideration for acute myocardial infarction (AMI). The enhanced analytical sensitivity of modern cTn assays has revealed that myocardial injury is common in other critical illnesses. Recent studies show a single elevated cTn measurement in critically ill, non-cardiac patients in the medical intensive care unit (MICU) is a predictor of poor outcomes. However, the prognostic significance of its temporal trend in non-cardiac patients is less established. This study aimed to determine if temporal changes of cTns provide prognostic information in a MICU population. Methods This retrospective observational cohort study included 106 MICU patients. Plasma samples were retrieved from the clinical laboratory for daily measurements of high sensitivity cTnT (Elecsys TnT Gen 5 Assay, Roche Diagnostics) for the length of MICU admission. Serial cTnT measurements were used to classify patients into five mutually exclusive groups: Nondynamic, Rising, Falling, RiseFall, and FallRise, based on a predefined 20% change across MICU admission. The primary endpoints were unfavorable outcomes (in-hospital death, discharge to hospice, or transition to comfort care) and all-cause 3-month mortality. Statistical analyses included unadjusted and multivariable logistic and Cox proportional hazards models (α = 0.05) with the Nondynamic group as the reference. Results There were no significant differences in age, sex, or medical comorbidities across the five groups (p > 0.05). Measurements of cTnT across each day (ΔcTnT) and MICU admission (overall ΔcTnT) correlated with the troponin trend groupings; for example, the median overall ΔcTnT for the Rising and Falling groups was 14 and -24, respectively, while these values approached 0 for the other three groups (p < 0.001). Using logistic regression, the Rising troponin trend was significantly associated with increased odds of an unfavorable outcome compared to the Nondynamic group (OR 7.33; 95% CI 1.92–32.39; p = 0.005). The Rising troponin trend remained an independent predictor of both unfavorable outcomes (p < 0.001) and 3-month mortality (p < 0.001) when adjusting for age, sex, initial cTnT measurements, and comorbidities. Kaplan-Meier analysis demonstrated significantly lower survival probability for patients in the Rising group compared to other trend patterns (p = 0.0011). Conclusion The temporal trend of high-sensitivity cTnT, particularly a rising pattern, is an independent predictor of unfavorable in-hospital outcomes and 3-month mortality in critically ill patients.