1Universitäres Herz- und Gefäßzentrum Hamburg Klinik für Kardiologie Hamburg, Deutschland; 2Cardio-CARE AG Medizincampus Davos Davos, Schweiz; 3Universitätsklinikum Hamburg-Eppendorf Klinik für Diagnostische und Interventionelle Radiologie und Nuklearmedizin Hamburg, Deutschland; 4Kardiologische Praxis Orchideenstieg Hamburg, Deutschland
Cardiovascular magnetic resonance (CMR) imaging enables objective quantification of aortic and mitral valve regurgitation. Robust data on the comparability of these parameters with echocardiographic estimates of aortic and mitral valve regurgitation are lacking.
Purpose
We compared aortic (AR) and mitral valve regurgitation (MR) measurements between CMR and echocardiography in unselected middle-aged probands undergoing CMR and echocardiography in the Hamburg City Health Study (HCHS) population.
Methods
The HCHS is a prospective, long-term, population-based cohort study including individuals between 45-74 years conducted at the University Medical Center Hamburg-Eppendorf (UKE), Hamburg, Germany. This analysis refers to the first 10.000 participants of the HCHS, who were included between February 2016 and November 2018. The participants underwent CMR imaging at 3 Tesla magnetic resonance imaging (MRI) scanner (MAGNETOM Skyra, Siemens Healthcare, Erlangen, Germany) using a 2D phase contrast VENC sequence for transaortic flow measurements and a standard cine CMR sequence for LV volumetry. Aortic valve regurgitation fraction (CMR-AR) was directly obtained from transaortic flow measurements (Figure 1). Mitral valve regurgitation fraction (CMR-MR) was calculated as the difference between LV stroke volume by cine CMR volumetry and transaortic flow by VENC CMR (Figure 2). Participants with a history of valvular surgery were excluded from this analysis.
Results
The final study population for this analysis consisted of 1,836 participants (761 (41.4%) females, median age 66.0 [IQR 58.0, 71.0] years). CMR-AR was 1.0 [IQR 0.2, 2.8] % in participants without any AR by echocardiography (n=1,244), 1.8 [IQR 0.9, 4.0] % in participants with mild AR (n=195), 3.0 [IQR 1.8, 8.0] % in participants with moderate AR (n=46) and 25.2 [IQR 20.6, 33.2] % in participants with severe AR (n=4). There was a significant difference in CMR-AR between participants without any AR and participants with any AR (p<0.001, AUC 0.633) as well as between participants without any AR and participants with moderate or severe AR (p<0.001, AUC 0.764). CMR-MR was 11.7 [IQR 5.3, 20.0] % in participants without any (n=1,015), 13.6 [IQR 5.9, 21.5] % in participants with mild MR (n=443), 14.2 [IQR 8.1, 22.9] % in participants with moderate MR (n=33) and 49.3 [IQR 34.1, 55.7] % in participants with severe MR (n=6). There was a significant difference in CMR-MR between participants without any MR and participants with any MR (p=0.017, AUC 0.540) as well as between participants without any MR and participants with moderate or severe MR (p=0.007, AUC 0.636).
Conclusion
Direct aortic flow measurements by VENC CMR provide a reliable assessment of AR in agreement with echocardiography. In contrast, the recommended estimation of MR by CMR is affected by a systematic overestimation of regurgitation fraction by 11.7 %, which needs to be considered for clinical application.