Troponin elevation is common during sepsis, yet its prognostic significance in patients without known cardiac disease remains uncertain. An analysis presented at the American Heart Association (AHA) 2025 Scientific Sessions evaluated whether modest troponin-T elevations predict in-hospital mortality in septic ICU patients after rigorous adjustment for illness severity.

The study analyzed 1,221 adult first ICU admissions for sepsis from the eICU Collaborative Research Database. Patients with cardiac admission diagnoses were excluded. Missing or implausible laboratory values were addressed using multiple imputation. Logistic regression models included troponin-T, age, sex, lactate, albumin, blood urea nitrogen, creatinine, intubation status, vasopressor use, dialysis requirement, and APACHE-IV predicted mortality. Troponin-T was evaluated both as a continuous variable and using a threshold of >0.04 ng/mL. Non-linear relationships were explored using natural spline modeling, and findings were validated in a 1:1 propensity-matched cohort.

Overall hospital mortality was 11.7%, and troponin-T exceeded 0.04 ng/mL in 38% of patients. After multivariable adjustment, troponin-T >0.04 ng/mL was associated with nearly doubled odds of in-hospital mortality, whereas the continuous model showed only a modest incremental effect. Spline analysis demonstrated a steep rise in mortality risk between 0 and 0.10 ng/mL, with an inflection near 0.05 ng/mL. In the propensity-matched analysis, elevated troponin-T was associated with approximately threefold higher mortality (odds ratio ≈2.8; 95% confidence interval 1.55 to 5.05; p<0.001).

These findings indicate that even modest troponin-T elevations independently predict mortality in sepsis after accounting for global illness severity and organ support. Identification of a low biologic threshold may support bedside risk stratification and inform future studies of septic cardiomyopathy.