Abstract Body (Do not enter title and authors here): Background
The coronary sinus reducer (CSR) has placebo-controlled evidence of angina benefit from two randomised placebo-controlled trials. The characteristics of patients with most to gain from this therapy remain unknown.

Research Question
What is the relationship between endocardial to epicardial blood flow ratio at stress and the placebo-controlled angina response to the CSR?

Methods
The Coronary Sinus Reducer Objective Impact on Symptoms, MRI Ischaemia and Microvascular Resistance (ORBITA-COSMIC) trial was a randomised, double-blind, placebo-controlled trial of the CSR in patients with angina on the maximum tolerated antianginal medication, myocardial ischaemia, epicardial coronary artery disease and no further options for revascularisation. Patients underwent a quantitative adenosine-stress perfusion cardiac magnetic resonance (CMR) scan and started daily angina reporting on a smartphone application (ORBITA-app) on enrolment to the trial. At the enrolment CMR, myocardial blood flow (MBF) in all 16 myocardial segments was quantified using an automated perfusion quantification sequence, with further stratification into endocardial and epicardial layers. Patients were randomised in the cardiac catheterisation laboratory to CSR or placebo and entered a 6-month period of blinded follow-up prior to a repeat stress CMR and scheduled unblinding.

Results
Enrolment quantified perfusion CMR data were available for 48/51 (94.1%) patients randomised in ORBITA-COSMIC, 22 CSR and 26 placebo (median age 67 (IQR 60 to 73), 42/48 (87.5%) male). The median endocardial to epicardial perfusion ratio (EN:EP) of stress MBF in ischaemic myocardial segments was 0.76 (IQR 0.72 to 0.87). The lower the EN:EP stress MBF, the greater the placebo-controlled improvement with CSR. There was strong evidence that a patient with EN:EP stress MBF in the lower quartile at baseline would have greater placebo-controlled benefit in angina episodes with the CSR than a patient in the upper quartile (EN:EP stress MBF 0.72 versus 0.87, OR 1.26, 95% CrI 1.13 to 1.41, probability of interaction>99.9%).

Conclusion
EN:EP stress MBF is the first biological variable which has been shown to predict placebo-controlled benefit with the CSR. This may have a role in the selection of patients for treatment.