The optimal diagnostic threshold for hs-cTnI was higher in patients with known CAD (27.7 ng/L; AUC 0.906) compared to those without known CAD (14.6 ng/L; AUC 0.942), with a significant interaction for 30-day MI or death (p=0.009).
Cohort (n=681)
No
Does the presence of known coronary artery disease modify the diagnostic accuracy and prognostic impact of high-sensitivity troponin I in patients presenting with suspected acute coronary syndrome?
Known coronary artery disease significantly modifies the diagnostic and prognostic interpretation of high-sensitivity troponin I in acute chest pain, suggesting that universal thresholds may obscure true risk gradients.
p-value: p=0.009
Abstract Background High-sensitivity cardiac troponin (hs-cTn) has transformed the management of acute chest pain, enabling faster rule-out and improved triage of acute coronary syndromes (ACS). However, interpretation remains challenging in patients with established coronary artery disease (CAD), who often exhibit chronically elevated troponin values unrelated to acute ischemia. Current diagnostic thresholds do not account for prior CAD, potentially reducing specificity and prognostic discrimination in high-risk tertiary care populations. Purpose To determine whether the presence of known CAD modifies the diagnostic accuracy, optimal cut-off values, and short-term prognostic impact of high-sensitivity troponin I (hs-cTnI) in patients presenting with suspected ACS. Methods We conducted a pragmatic, prospective, single-center cohort study at a major cardiology hospital in Brazil between March and November 2024. Adults presenting with acute chest pain were included if evaluated under the institutional 0/1-hour or single rule-out hs-cTnI protocol. Exclusion criteria were ST-elevation myocardial infarction, myocarditis, pulmonary embolism, aortic dissection, or non-cardiac pain. hs-cTnI was measured using a validated high-sensitivity immunoassay. Known CAD was defined as prior PCI, CABG, or ≥50% coronary stenosis on angiography or CTA. The primary endpoint was myocardial infarction (MI) or all-cause death within 30 days. Diagnostic performance was assessed by ROC analysis, and hs-cTnI × CAD interaction was tested by multivariable logistic regression. Results Among 681 patients (mean age 64.2 ± 11 years; 63% male), 63.5% had known CAD. ROC analysis revealed distinct hs-cTnI performance according to coronary history. In patients with known CAD, the optimal diagnostic threshold was 27.7 ng/L (AUC 0.906; sensitivity 56.7%; specificity 92.6%; NPV 0.925), reflecting excellent discrimination despite lower sensitivity. In contrast, patients with no known CAD showed a lower optimal cut-off of 14.6 ng/L (AUC 0.942; sensitivity 83.5%; specificity 74.8%; NPV 0.974), maintaining very high negative predictive accuracy. A significant interaction between hs-cTnI and CAD status was observed for MI/death (p = 0.009). In those with known CAD, troponin elevation often represented chronic myocardial injury, whereas in those with no known CAD, even small increases were strongly predictive of acute ischemia and adverse outcomes. Conclusions The same troponin concentration carries different clinical meaning depending on coronary history. Known CAD significantly modifies both diagnostic and prognostic interpretation of hs-cTnI in acute chest pain. Applying a single universal threshold may obscure true risk gradients. These findings support individualized troponin interpretation, with higher rule-in and lower rule-out thresholds stratified by CAD status to enhance accuracy and safety in real-world ACS care.Box-plots of hs-cTnI ROC curves for hs-cTnI by known CAD
Ferreira et al. (Fri,) conducted a cohort in suspected acute coronary syndromes (ACS) (n=681). high-sensitivity troponin I (hs-cTnI) vs. known CAD vs no known CAD was evaluated on myocardial infarction (MI) or all-cause death within 30 days (p=0.009). The optimal diagnostic threshold for hs-cTnI was higher in patients with known CAD (27.7 ng/L; AUC 0.906) compared to those without known CAD (14.6 ng/L; AUC 0.942), with a significant interaction for 30-day MI or death (p=0.009).