Abstract Background Exercise stress testing (EST) was the traditional tool to detect myocardial ischaemia, but its use has declined due to its perceived high false positive rate. However, EST was historically validated against the reference standard of obstructive coronary artery disease (CAD). It is now well recognised that myocardial ischaemia can occur in the absence of CAD and this cohort, termed angina with non-obstructive coronary arteries (ANOCA), comprises nearly 50% of all patients presenting with stable angina. Purpose Our study assesses the accuracy of EST in detecting a substrate for myocardial ischaemia, with the reference standard being comprehensive coronary physiology assessment in patients with ANOCA. Methods Patients with typical angina and non-obstructive coronary arteries underwent physiological characterisation of the coronary circulation using a dual sensor-tipped intracoronary guidewire in the left anterior descending artery. The epicardial vessel was assessed by Fractional Flow Reserve (FFR) and the microcirculation by both endothelium-independent function (coronary flow reserve, CFR, using intravenous adenosine) and endothelium-dependent function (acetylcholine flow reserve, AChFR, using intracoronary acetylcholine infusion). Those with FFR≤0.80 were excluded per protocol. Coronary microvascular dysfunction was defined a priori as impaired CFR (<2.5) or impaired AChFR (≤1.5). All patients underwent a treadmill EST, using a standard Bruce protocol, with ischaemic electrocardiographic (ECG) changes defined as the appearance of 0.1mV ST-segment depression 80 milliseconds from the J-point. Patients, physiologists, and researchers were all blinded to the coronary physiology data. All data are presented as mean±SD or median (IQR). Results Ninety-one patients [67% female, 62±9 years, CCS 3 (2 to 3)] were recruited into this study. EST was performed 27 (15 to 139) days after angiography. Twenty-seven patients (30%) developed ischaemic ECG changes during their EST (Ischaemic group), whereas 64 patients (70%) did not (Control group). Both groups were phenotypically similar for demographics, risk factors, exercise time and FFR (Figure 1). 100% of the Ischaemic group and 69% of the control group had CMD. Conversely, 38% of patients with CMD developed ischaemic ECG changes on exercise but none (0%) of those with normal microvascular function did (p=0.001) (Figure 2A). The false positive rates with progressively granular reference standards are depicted in Figure 2B; there were no false positive ESTs using a comprehensive physiological reference standard. Angina reported during EST was less specific for CMD than induced ECG changes (70% vs 100%, p<0.001). Conclusions In patients with ANOCA a positive EST is always indicative of CMD. The potential utility of EST as a rule-in test for CMD makes it an attractive non-invasive widely available test that could be re-integrated into clinical pathways for assessing new onset angina.Patient characteristicsEST versus coronary physiology