Background: Patients with atrial fibrillation (AF) often present with similar symptoms and risk factors as patients with a coronary artery disease (CAD). Accordingly, differentiation between troponin elevation as a consequence of symptomatic AF only versus concomitant myocardial ischemia due to relevant CAD is challenging. Moreover, clinical criteria to identify AF-patients profiting of a coronary angiography (CA) due to attendant CAD remain vague. With high sensitive cardiac Troponin T (hs-cTnT) known to be superior to cardiac Troponin I (cTnI) in CAD-pretesting, we compared both assays, various clinical, echocardiographic and laboratory parameters on their sensitivity to predict significant concomitant CAD in hospitalized symptomatic AF-patients.


Methods: This retrospective observatory monocentric study was performed on AF-patients admitted to the emergency department of the University Hospital Bonn between 2015 and 2019, where the conventional cTnI assay was replaced by hs-cTnT in August 2017. All patients with symptomatic AF undergoing CA during their hospitalization were included. Primary endpoint was the need for revascularization (PCI or bypass surgery). Out of 192 patients enclosed, 98 patients were screened with cTnI while 94 were screened with hs-cTnT. The upper troponin limit was defined as 99th percentile upper reference limit (0.05 µg/l for cTnI and 14 ng/l for hs-cTnT). Additionally, the cut off groups of the ESC NSTE-ACS guideline 2015 ("Rule In", "Other", "Rule Out" for hs-cTnT Elecsys) were included in the analysis. Troponin levels, clinical, echocardiographic and laboratory parameters were evaluated in logistic regression models which were compared using receiver operating curve (ROC).

Results: 47 out of 192 (24%) AF-patients were diagnosed with CAD requiring revascularization. 43 AF-patients were treated with PCI. 145 patients (77%) showed no relevant stenosis. The sensitivity for the presence of CAD with the need for intervention was 65.2% [95%CI,42.7-83.6%] for cTnI and 100% [95%CI,88.3-100%] for hs-cTnT. The median hs-cTnT concentrations were significantly higher in the “intervention-group” than in the “non-intervention-group” (29.9ng/l [95%CI,26.5-54.8ng/l] vs. 16.4ng/l [95%CI,12.9-23.7ng/l],p<0.001). Furthermore, significant differences were found in age (p=0.009), history of CAD (p=0.004), systolic blood pressure (p=0.037) and mitral valve insufficiency (p=0.044). Higher age, previously diagnosed CAD and impaired renal function (creatinine clearance <60ml/min) were independent risk factors of significant stenosis in AF-patients. Calculated regression model including age, history of CAD and hs-cTnT integrating higher “rule out” cut off level of 23.05ng/l - median of all patients with hs-cTnT assay - showed the best pretesting performance with an AUC of 0.83, p=0.002. Poor performance was detected for cTnI (AUC of 0.63, p=0.159).

Conclusions: This study demonstrates a superior pretesting performance of hs-cTnT assay for significant CAD in need of PCI in AF-patients compared to the conventional cTnI assay. Higher age, previous history of CAD, impaired renal function and hs-cTnT assay with almost doubled “rule out” cut off level of 23ng/l (instead of 12ng/l) showed highest pretest probability for relevant CAD in patients hospitalized due to AF.




Fig.1 ROC Analysis of hs-cTnT (top) and cTnI assays (buttom).