All patients undergoing transcatheter aortic valve implantation (TAVI) in our centre between November 2012 and August 2013, without CMR-contraindication were included. PAR severity was assessed 5 days after TAVI using: transthoracic echocardiography (TTE) and CMR [regurgitant volume (RV), regurgitant fraction (RF)]. Aortic regurgitation (AR) index was obtained during TAVI. Thirty of 51 patients who underwent TAVI were included (COREVALVE, n = 10; or EDWARDS SAPIEN XT, n = 20). At TTE, PAR was mild in 22, moderate in 3, and severe in 5 patients. Reliable phase-contrast images were acquired at the sino-tubular junction for SAPIEN and at the tubular portion of the ascending aorta for COREVALVE. The reproducibility of CMR was high (coefficient of correlation = 0.99 for intra- and inter-operator variability). At CMR, RV, and RF were significantly (P < 0.0005) correlated with AR severity at TTE, with mean RF values at 9.2 ± 7.6% in mild, 20.3 ± 4.2% in moderate, and 46.8 ± 10.8% in severe PAR. A cut-off value of RF < 14% at CMR accurately discriminated mild from moderate/severe (sensitivity: 100%, specificity: 82%). The mean AR index was 29.4 ± 6 for mild and 13.8 ± 5 for moderate/severe PAR. Three patients had a RF > 14% and a low AR index <25 despite a mild PAR at TTE, suggesting an underestimation at TTE.

CMR is a reproducible, accurate, and reliable method to assess PAR severity. CMR may allow correcting an underestimation at TTE when AR index is doubtful.