Postoperative ICU Biomarkers and Early Mortality After Surgical Aortic Valve Replacement in Severe Aortic Stenosis With Reduced Left Ventricular Ejection Fraction: Limited Discrimination of High-Sensitivity Troponin T

Early postoperative risk stratification after surgical aortic valve replacement remains challenging in patients with severe aortic stenosis and reduced left ventricular ejection fraction. Although postoperative troponin elevation is expected after open-heart surgery, the discriminatory value of ICU admission high-sensitivity cardiac troponin T (hs-cTnT) in this specific myocardial phenotype remains uncertain. We evaluated whether ICU admission hs-cTnT provides useful early discrimination for 30-day all-cause mortality after surgical aortic valve replacement in patients with severe aortic stenosis and reduced left ventricular ejection fraction and compared its performance with other postoperative ICU biomarkers. We conducted a single-center retrospective cohort study of 184 adults with severe aortic stenosis and left ventricular ejection fraction ≤40% who underwent surgical aortic valve replacement between 2020 and 2024 and had complete 30-day follow-up. ICU admission hs-cTnT was available in 183 patients. Biomarkers included ICU admission hs-cTnT, ICU admission lactate, peak ICU C-reactive protein, and ICU admission B-type natriuretic peptide. Discrimination was assessed using receiver operating characteristic analysis, DeLong comparisons, repeated stratified 5-fold cross-validation, and bootstrap optimism correction. Multivariable models were interpreted as exploratory because of the limited number of mortality events. Thirty-day mortality occurred in 29 of 184 patients (15.8%). ICU admission hs-cTnT showed poor discrimination for 30-day mortality in the hs-cTnT analytic cohort and remained non-discriminatory after repeated cross-validation. In the common four-biomarker cohort, lactate outperformed hs-cTnT by DeLong testing, whereas the four-biomarker model did not significantly improve discrimination beyond lactate alone. Bootstrap optimism correction attenuated the apparent performance of multimarker models, supporting cautious interpretation. Exploratory analyses also suggested that mortality did not increase monotonically across hs-cTnT quartiles. In patients with severe aortic stenosis and reduced left ventricular ejection fraction undergoing surgical aortic valve replacement, ICU admission hs-cTnT did not provide reliable early mortality discrimination. Lactate remained the strongest early individual biomarker, whereas multimarker models showed exploratory discrimination but require external validation. These findings support a phenotype-specific interpretation of postoperative hs-cTnT rather than reliance on universal postoperative troponin assumptions.