Abstract Background Dilated cardiomyopathy (DCM) has a poor prognosis. Aerobic exercise capacity (peak VO2) is an independent predictor of mortality but the central mechanisms contributing to exercise intolerance in DCM are unknown. Purpose To characterise myocardial perfusion reserve (MPR) and determine if cardiovascular magnetic resonance (CMR) measures of structure, function and microvascular function are associated with aerobic exercise capacity in DCM. Methods Single centre, prospective, case-control comparison of adults with DCM and matched controls. Adenosine-stress perfusion CMR to assess cardiac structure, function and quantitative perfusion, and cardiopulmonary exercise testing (CPET) to determine peak VO2, was performed. Controls were propensity matched based on age, sex, body mass index and diabetes status, and between group comparison was adjusted for other key clinical characteristics (CMR: ethnicity and systolic blood pressure; CPET: ethnicity, systolic blood pressure, smoking history and lung disease). Comparison of perfusion in segments with and without late gadolinium enhancement was performed using mixed effects linear regression taking into account individual patient segmental perfusion. Pre-specified multivariable linear regression, including key clinical and cardiac variables, was undertaken in patients with DCM to identify independent associations with percentage predicted peak VO2, which incorporates age, sex, height and weight. Results Sixty-six patients with DCM (median age 61 years, 71% male) were matched to 66 controls (median age 59 years, 71% male). The DCM group had greater left ventricular (LV) end-diastolic and end-systolic volumes, lower systolic and diastolic function, and had more focal and diffuse fibrosis compared to controls (Table 1). There was lower rest (0.58±0.14 versus 0.65±0.17 mL/min/g; P=0.033) and stress (1.53±0.49 versus 2.01±0.60 mL/g/min; P<0.001) myocardial blood flow, and lower MPR (2.69±0.84 versus 3.15±0.84 mL/g/min; P=0.002) in the DCM group. In DCM patients, mixed effects linear regression demonstrated lower rest (adjusted mean 0.54(95% CI 0.50-0.57) versus 0.58(0.55-0.62) mL/g/min; P<0.001) and stress MBF (1.38(1.25-1.51) versus 1.55(1.43-1.67) mL/g/min; P<0.001) in segments with LGE compared to segments without LGE, but there was no difference in MPR (2.71(2.47-2.96) versus 2.76(2.55-2.97) mL/g/min; P=0.496). DCM patients had markedly lower peak VO2 (19.8±5.5 versus 25.2±7.3 mL/kg/min; P<0.001). Multivariable linear regression demonstrated that LV ejection fraction, extracellular volume fraction and MPR, but not LV mass, were independently associated with percentage predicted peak VO2 in DCM (R squared=0.531, P<0.001; Table 2). Conclusions DCM patients have lower rest and stress myocardial blood flow and lower MPR compared to controls. In DCM, MPR, LV ejection fraction and fibrosis are independently associated with aerobic exercise capacity.