Evaluation of coronary artery disease in asymptomatic persons with type 2 diabetes using a non-invasive acoustic detection system

Tine W. Hansen Hansen1, Ida Rasmussen1, Mathias Sørgaard2, Bernt von Scholten1, Samuel Emil Schmidt3, Peter Rossing1, Klaus Kofoed2
1Steno Diabetes Center Copenhagen, 2Department of Cardiology, The Heart Centre, Rigshospitalet, 3Aalborg university


Background and aims: Coronary artery disease (CAD) is frequent in persons with type 2 diabetes (T2D) and the diagnosis is often delayed due to silent myocardial ischemia. Therefore, screening for CAD in asymptomatic persons with T2D using simple and inexpensive methods is desirable. A noninvasive system to rule-out CAD has been developed. The system uses acoustic detection to isolate and analyze the auditory signature of obstructive CAD and calculates a CAD-score. The system has been validated in suspected symptomatic CAD patients. We evaluated if the system individually or in combination with presence of albuminuria was suited for risk assessment of CAD in asymptomatic persons with T2D and high risk for CAD. Materials and methods: Observational pilot study in persons with T2D without symptoms or any history of cardiovascular disease. The cohort was enriched for presence of albuminuria to increase CAD risk. Information on clinical characteristics was collected, the CAD-score was measured and coronary computed tomography angiography (CCTA) was performed. Results: A total of 108 patients (34% females) were enrolled, median (IQR) age was 65 (58-69) years and 48 (44%) had albuminuria (>30mg/g). A CAD-score and a CCTA result were available in 86 persons; hereof 12 (14%) had a stenosis (≥ 70% in any of the coronary arteries) identified by CCTA. The AUC for the CAD-score was 42.7% versus CCTA. Using a cut-off at 20, the sensitivity and negative predictive value were 75.0% and 81.2%, respectively. In contrast, the specificity and positive predictive value were low (below 20%). Conclusions: The system alone was inefficient as a screening tool in a low CAD prevalent population of asymptomatic persons with T2D, with its current rule-out algorithm directed at symptomatic patients. Obtaining further acoustic recordings and validated clinical data from this specific asymptomatic patient group could potentially expand the rule-out population indicated.