How to estimate aortic characteristic impedance from magnetic resonance and applanation tonometry data?

E. Bollache (1, 2) , N. Kachenoura (1) , I. Bargiotas (1) , A. Giron (1) , A. de Cesare (1) , M. Bensalah (1) , D Lucor (2) , A. Redheuil (1, 3) , E. Mousseaux (4)

1 LIB - Laboratoire d'Imagerie Biomédicale [Paris]
2 DALEMBERT - Institut Jean Le Rond d'Alembert
3 Institut de cardiologie [CHU Pitié-Salpêtrière]
4 PARCC - UMR-S U970 - Paris-Centre de Recherche Cardiovasculaire

Objectives: Compare seven previous methods for the estimation of aortic characteristic impedance, which contributes to left ventricle pulsatile load, from phase-contrast cardiovascular magnetic resonance (CMR) and applanation tonometry data. Methods: We studied 77 healthy (43 ± 16 years) individuals and 16 hypertensive (61 ± 9 years) patients, who consecutively underwent ascending aorta CMR and carotid tonometry, resulting in flow and pressure waveforms, respectively. Characteristic impedance was semi-automatically estimated in time domain from these latter waveforms, using seven methods. The methods were based on the following: methods 1-4, magnitudes at specific times; method 5, early-systolic up-slope; method 6, time-derivatives peak; and method 7, pressure-flow loop early-systolic slope. Results: Aortic characteristic impedance was significantly increased in hypertensive patients when compared to elderly controls (n = 32) with a similar mean age of (59 ± 8 years) when using methods based on 95% of peak flow, up-slopes, and derivatives peaks (P < 0.05). When considering healthy individuals, impedance indices were significantly correlated to central pulse pressure for all methods (P < 0.005). Finally, characteristic impedance was correlated to the frequency-domain reference values (r > 0.65, P < 0.0001), with a slight superiority for the same three methods as above (r > 0.82, P < 0.0001). Conclusions: This is the first study demonstrating phase-contrast CMR and tonometry usefulness in aortic characteristic impedance temporal estimation. Methods based on 95% of peak flow, as well as those based on derivative peaks and up-slopes, which are fast and independent of curve preprocessing, were slightly superior. They can be easily integrated in a clinical workflow and may help to understand the complementarity of this pulsatile index with other CMR aortic geometry and stiffness measures in the setting of left ventricle-aortic coupling.

Domaines: Bio-informatique [q-bio.QM]

E. Bollache, N. Kachenoura, I. Bargiotas, A. Giron, A. de Cesare, et al.. How to estimate aortic characteristic impedance from magnetic resonance and applanation tonometry data?. Journal of Hypertension, 2015, pp.9. ⟨10.1097/HJH.0000000000000448⟩. ⟨hal-01150778⟩