Quantifying wave reflection in children: Invasive vs non-invasive central augmentation index and reflection magnitude and their association with left ventricular mass

Objective: The aims of this study in children were to 1) evaluate two brachial oscillometric devices for estimating central augmentation index (AIx) and reflection magnitude (RM), and 2) test whether AIx or RM are associated with left ventricular mass index (LVMI). Methods: Intra-aortic (IA) AIx was calculated from high-fidelity pressure measured with a Verrata wire (Philips Volcano) in 60 children (9.2 ± 4.7 years) with unobstructed aorta undergoing clinically-indicated catheterisation. AIx was also obtained from SphygmoCor XCEL (SC, AtCor) and/or Mobil-o-Graph (MB, IEM) brachial oscillometric devices. RM(IA) was calculated via wave separation using a representative normalised flow waveform obtained from MRI in a separate group of normal adolescents, RM(SC) via the triangulation method, and RM(MB) provided by the proprietary software. LVMI was estimated via echocardiography. Results: Invasive vs non-invasive AIx and RM are compared in the Table. AIx(IA) correlated weakly with AIx(SC) (R = 0.27, P = 0.04) but not AIx(MB) (P = 0.4). Neither RM(SC) nor RM(MB) correlated with RM(IA) (P = 0.13 and P = 0.96 respectively). RM(IA) was moderately correlated with AIx(IA) (R = 0.69, P < 0.001) and weakly correlated with AIx(SC) (R = 0.36, P = 0.007) but not AIx(MB) (P = 0.7). In a multivariable regression, height (P < 0.001) and RM (IA) (P = 0.04) were independently and positively associated with LVMI (adjusted R2 = 0.24), whereas there were no associations of any AIx or non-invasively estimated RM with LVMI. Conclusion: Central AIx and RM were poorly estimated by SC and MB in children. Unlike RM(IA), none of the non-invasive indices of wave reflection correlated with LVMI, likely due to inadequate estimation of the central pressure waveform shape in this age group.