Risk stratification based on troponin I as a predictor of mortality in isolated valve surgeries
R Aragao, J A S Barreto Filho, V E E Rosa, C A S Aragao, A S S A Lopes, J R C Fernandes, L A Hajjar, F Tarasoutchi, R O SampaioAbstract
Background
Cardiac troponin I (cTnI) measured during the postoperative period is a strong predictor of adverse outcomes after coronary artery bypass grafting (CABG). The magnitude of postoperative cTnI release is typically higher in patients undergoing valve surgery compared with CABG. However, only a limited number of studies have evaluated the prognostic value of cTnI after valvular cardiac surgery. This study aimed to assess the independent prognostic value of troponin I levels measured at different time points during the first 72 hours after isolated valvular heart surgery and to determine cTnI cut-off values associated with in-hospital mortality according to valve surgery type.
Methods
All patients who underwent isolated valvular heart surgery at a single tertiary center between 2011 and 2020 were included. Troponin levels were categorized according to postoperative time intervals: ≤12 h, 12–24 h, 24–48 h, and 48–72 h. Logistic regression models were used to evaluate the association between cTnI levels and in-hospital mortality after adjustment for clinical and operative variables. ROC curves were constructed to assess the discriminative performance of cTnI for in-hospital mortality, and optimal cut-off values were determined using Youden’s index.
Results
A total of 4,100 patients were included. Median age was 57 years (IQR 46–66), 54% were female, and 32% had undergone prior cardiac surgery. The predominant etiology was rheumatic heart disease (44%). Median logistic EuroSCORE ranged from 0.9 to 5.1, and 12% of procedures were urgent or emergent. LVEF was 62 (IQR 57-66), with 329 patients (8,1%) presenting LVEF ≤40%. In-hospital mortality was 11.4% (n=470). Troponin levels at all measured time points (≤12 h, 12–24 h, 24–48 h, and 48–72 h) were significantly associated with in-hospital mortality (p<0.001). The highest discriminative performance was observed in the 24–48 h interval (AUC 0.722). Across all surgery types, non-survivors consistently exhibited higher cTnI levels. Optimal cTnI cut-off values for predicting in-hospital mortality were: aortic valve surgery — 132× the upper reference limit (URL) at 24–48 h and 68× URL at 48–72 h; mitral valve surgery — 209× URL at 24–48 h and 100× URL at 48–72 h; multiple valve surgery — 263× URL at 24–48 h and 130× URL at 48–72 h. In the multivariable model with imputed data, cTnI levels measured after 12 h remained independently associated with in-hospital mortality, with the strongest association observed for measurements obtained at 48–72 h (OR 1.91, 95% CI 1.81–2.03; p<0.001).
Conclusion
Postoperative cTnI levels measured between 24 and 72 hours after isolated valvular heart surgery were independently associated with in-hospital mortality. The observed cut-off values were substantially higher than currently recommended thresholds for clinically relevant periprocedural myocardial injury, particularly in multiple valve procedures.