Detection of early vascular changes indicated by lowered coronary flow reserve (CFR) would allow early treatment and prevention of atherosclerosis. The purpose of this study was to test whether it is possible to reproducibly measure CFR with transthoracic Doppler echocardiography (TTE) in healthy volunteers. We measured CFR using dipyridamole infusion in ten healthy male volunteers with two methods: TTE and positron emission tomography (PET) with oxygen-15-labelled water (group A). However, CFR was assessed twice with TTE in eight healthy male volunteers (group B) to study the reproducibility of this method. We compared CFRs obtained using TTE flow measurements in the left anterior descending coronary artery (LAD) and PET flow measurements in the corresponding myocardial area. Coronary flow in LAD could be measured in all subjects using TTE. By TTE, an average CFR based on peak diastolic flow velocity (PDV) was 2·72 ± 1·16, mean diastolic flow velocity (MDV) 2·56 ± 1·06 and velocity time integral (VTI) 1·87 ± 0·49. The results were reproducible in two repeated TTE studies (coefficient of variation in MDV 6·1 ± 4·3%, n=8). By PET, CFR was 2·52 ± 0·84. CFR assessed by TTE correlated closely with that measured by PET (MDV r=0·942, P<0·001; PDV r=0·912, P<0·002 and VTI r=0·888, P<0·006) and intraclass correlation was 0·929 (MDV) and tolerance limits for differences of CFRs was −0·78 to 0·72. We show that CFR measured by TTE has an excellent correlation with CFR measured by PET. We also found that TTE measurements of CFR were highly reproducible.