The role of cardiovascular magnetic resonance in stratifying paravalvular leak severity after transcatheter aortic valve replacement: an observational outcome study

BACKGROUND: Significant paravalvular leak (PVL) after transcatheter aortic valve replacement (TAVR) confers a worse prognosis. Symptoms related to significant PVL may be difficult to differentiate from those related to other causes of heart failure. Cardiovascular magnetic resonance (CMR) directly quantifies valvular regurgitation, but has not been extensively studied in symptomatic post-TAVR patients. METHODS: CMR was compared to qualitative (QE) and semi-quantitative echocardiography (SQE) for classifying PVL and prognostic value at one year post-imaging in 23 symptomatic post-TAVR patients. The primary outcome was a composite of all-cause death, heart failure hospitalization, and intractable symptoms necessitating repeat invasive therapy; the secondary outcome was a composite of all-cause death and heart failure hospitalization. The difference in event-free survival according to greater than mild PVL versus mild or less PVL by QE, SQE, and CMR were evaluated by Kaplan-Meier survival analysis. RESULTS: Compared to QE, CMR reclassified PVL severity in 48% of patients, with most patients (31%) reclassified to at least one grade higher. Compared to SQE, CMR reclassified PVL severity in 57% of patients, all being reclassified to at least one grade lower; SQE overestimated PVL severity (mean grade 2.5 versus 1.7, p = 0.001). The primary and secondary outcomes occurred in 48% and 35% of patients, respectively. Greater than mild PVL by CMR was associated with reduced event-free survival for the primary outcome (p < 0.0001), however greater than mild PVL by QE and SQE were not (p = 0.83 and p = 0.068). Greater than mild PVL by CMR was associated with reduced event-free survival for the secondary outcome, as well (p = 0.012). CONCLUSION: In symptomatic post-TAVR patients, CMR commonly reclassifies PVL grade compared with QE and SQE. CMR provides superior prognostic value compared to QE and SQE, as patients with greater than mild PVL by CMR (RF > 20%) had a higher incidence of adverse events.