Mancini, Nastasia

Abstract Background and Aims A simple diagnostic method able to reliably exclude left main (LM) coronary artery disease (CAD) or LMCAD-equivalent would expand implementation of an initial non-invasive strategy in patients with chronic coronary syndrome (CCS). This study assessed the diagnostic utility of an approach using clinical and ECG stress testing (EST) variables in excluding LMCAD/LMCAD-equivalent in CCS patients. Methods In a multicentre case-control study, CCS patients undergoing invasive coronary angiography (CAG) after a maximal EST were evaluated. Cases were patients with angiographic ≥ 50% LM stenosis or ≥70% stenosis of both proximal left anterior descending and proximal circumflex arteries, matched with similar patients without them (controls) in a 1:3 ratio. A risk model developed through logistic regression was internally and externally validated. Results Three hundred and thirty-five cases were matched with 797 controls. The model area under the curve (AUC) was .78. Assuming LMCAD prevalence of 5% and a misclassification cost ratio of 1:100 (ratio of cost of performing CAG in a control to cost of not performing CAG in a case), negative predictive value was 98.2%. Thus, CAG could be safely avoided in 41% of patients, missing one LMCAD/LMCAD-equivalent diagnosis for every 58 CAGs safely spared in patients without them. Conclusions Among CCS patients, LMCAD/LMCAD-equivalent can be excluded with high negative predictive value through a model based on clinical and EST parameters, allowing initial non-invasive management of most patients able to exercise. This approach is potentially useful particularly in communities where access to computed tomography coronary angiography is limited.