Three-dimensional echocardiography of the athlete's heart: a comparison with cardiac magnetic resonance imaging

Three-dimensional echocardiography (3DE) is the most accurate cardiac ultrasound technique to assess cardiac structure. 3DE has shown close correlation with cardiac magnetic resonance imaging (CMR) in various populations. There is limited data on the accuracy of 3DE in athletes and its value in detecting alterations during follow-up. Indexed left and right ventricular end-diastolic volume (LVEDVi, RVEDVi), end-systolic volume, ejection fraction (LVEF, RVEF) and left ventricular mass (LVMi) were assessed by 3DE and CMR in two-hundred and one competitive endurance athletes (79% male) from the Pro@Heart trial. Sixty-four athletes were assessed at 2 year follow-up. Linear regression and Bland-Altman analyses compared 3DE and CMR at baseline and follow-up. Interquartile analysis evaluated the agreement as cardiac volumes and mass increase. 3DE showed strong correlation with CMR (LVEDVi r = 0.91, LVEF r = 0.85, LVMi r = 0.84, RVEDVi r = 0.84, RVEF r = 0.86 p < 0.001). At follow up, the percentage change by 3DE and CMR were similar ( increment LVEDVi r = 0.96 bias - 0.3%, increment LVEF r = 0.94, bias 0.7%, increment LVMi r = 0.94 bias 0.8%, increment RVESVi r = 0.93, bias 1.2%, increment RVEF r = 0.87 bias 0.4%). 3DE underestimated volumes (LVEDVi bias - 18.5 mL/m(2), RVEDVi bias - 25.5 mL/m(2)) and the degree of underestimation increased with larger dimensions (Q1vsQ4 LVEDVi relative bias - 14.5 versus - 17.4%, p = 0.016; Q1vsQ4 RVEDVi relative bias - 17 versus - 21.9%, p = 0.005). Measurements of cardiac volumes, mass and function by 3DE correlate well with CMR and 3DE accurately detects changes over time. 3DE underestimates volumes and the relative bias increases with larger cardiac size.