Adverse effects of aortic backward waves in a group of African Ancestry
Sibiya, Moekanyi Jeffrey
Although brachial blood pressure (BP) is a well-recognized risk factor for predicting cardiovascular events, aspects of aortic BP may enhance risk prediction. Pulse pressure (PP) is amplified from the aorta to peripheral arteries and variations in differences between brachial and aortic PP (PP amplification) are determined by factors that influence either the aortic forward (Pf) or backward (Pb)(reflected) pressure waves. Although aortic Pb may be more important than Pf in mediating cardiovascular risk, the best approach to assessing backward wave function (augmentation pressures [Pa] and index [AIx] or wave separation analysis); the relative impact of aortic Pb versus Pf on cardiovascular damage; and whether the ability of aortic-to-brachial PP amplification (PPamp) to add to risk prediction reflects backward or forward wave effects, is uncertain. In the present thesis I therefore first assessed in 808 community participants whether gender influences relations between Pa or AIx and left ventricular mass (LVM), a well-accepted end-organ measure. Aortic haemodynamics were determined using radial applanation tonometry and SphygmoCor software and LVM from echocardiography. In men, both AIx derived from Pa/central aortic PP (Pa/PPc) (p<0.01) and AIx derived from the second peak/first peak (P2/P1) of the aortic pulse wave (p<0.0005) were associated with LVM. In contrast, in women neither AIx derived from Pa/PPc (p=0.08) nor AIx derived from P2/P1 (p=0.17) were associated with LVM. Both the strength of the correlations (p<0.001 and p<0.0005) and the slope of the AIx-LVM relationships (p=0.001 and p<0.0005) were greater in men as compared to women. Therefore, in the present study I show that AIx is independently associated with LVMI in men, but not in women. I subsequently evaluated whether in women, measures of aortic systolic pressure augmentation (Pa or AIx) underestimate the effects of reflected waves on cardiovascular risk or whether Pb plays little role in cardiovascular risk prediction. In the same community sample I therefore evaluated sex-specific contributions of reflected (Pb and the reflection index [RI]) versus augmented (Pa and AIx) pressure wave indices to iii variations in PPc (n=1185, 65.0% women), and LVM (n=793, 64.9% women). Aortic Pb and Pf were determined using wave separation analysis. In both women and in men, independent of confounders, RI and Pb contributed more than Pf, whilst Pa and AIx contributed less than incident wave pressure (Pi) to variations in PPc (p<0.0001 for comparison of partial r values). In both men and in women Pb contributed more than Pf (p<0.05) to variations in LVM. Although in men Pa (partial r=0.33, p<0.0001) contributed to a similar extent as Pi ((partial r=0.34, p<0.0001) to variations in LVMI, in women Pa (partial r=0.05, p=0.36) failed to contribute to LVM, whilst Pi was significantly associated with LVM (partial r=0.30, p<0.0001). Similar results were obtained with AIx as opposed to Pa in the regression models. Therefore, in both women and in men, Pb is more closely associated with PPc and LVM than Pf, but indices of aortic pressure augmentation markedly underestimate these effects, particularly in women. As the relative impact of aortic Pb as compared to Pf on cardiovascular damage independent of brachial BP is uncertain, in 1174 participants from a community sample I subsequently assessed the relative impact of Pb and Pf on variations in LVM (n=786), aortic pulse wave velocity (PWV)(n=1019), carotid intima-media thickness (IMT)(n=578), transmitral early-to-late LV diastolic velocity (E/A)(n=779) and estimated glomerular filtration rate (eGFR)(n=1174). Independent of mean arterial pressure and confounders, PPc and both Pb and Pf were associated with end-organ measures or damage (p<0.05 to <0.0001). With adjustments for brachial PP and confounders, Pb remained directly associated with LVM (partial r=0.10, p<0.01), PWV (partial r=0.28, p<0.0001), and IMT (partial r=0.28, p<0.0001), and inversely associated with E/A (partial r=-0.31, p<0.0001) and eGFR (partial r=-0.14, p<0.0001). Similar relations were noted with the presence of end-organ damage (p<0.05 to <0.0001). In contrast, with adjustments for brachial PP and confounders, Pf no longer retained direct relations with LVM, PWV, and IMT or inverse relations with E/A and eGFR. Adjustments for Pb, but not Pf diminished brachial PP-independent relationships between PPc and end-organ measures. Thus, although both Pf and Pb contribute to end-organ measures and damage, independent of brachial iv BP, the impact of aortic BP is accounted for largely by Pb. PPamp is independently associated with cardiovascular outcomes. However, the aortic functional change most likely to account for this effect is uncertain. In 706 community participants I subsequently aimed to identify the aortic functional change that accounts for relations between PPamp and LVM. In multivariate models with the inclusion of brachial PP, 1/PPamp (partial r=0.12, p<0.005), Pb (partial r=0,09, p<0.05), and aortic PWV (partial r=0.09, p<0.05) were independently associated with LVMI. Similarly, in multivariate models with the inclusion of brachial PP, 1/PPamp (p<0.005), Pb (p<0.01), and aortic PWV (p<0.01) were independently associated with LV hypertrophy (LVH). With adjustments for Pb, the brachial PP-independent relationships between 1/PPamp and LVMI or LVH were abolished (p>0.08 for both). However, adjustments for PWV failed to modify brachial PP-independent relations between 1/PPamp and LVMI or LVH. Hence, independent relations between PPamp and LVM or LVH are largely accounted for by Pb. In conclusion, in the present thesis I show that the use of augmented pressures underestimates the impact of reflected pressure wave effects on end-organs, particularly in women; that brachial BP-independent relations between aortic BP and end organs is determined largely by Pb and that relations between PPamp and end organ measures is largely accounted for by Pb. These findings add to our understanding of the adverse effects of aortic functional changes on the cardiovascular system and suggest cost-effective approaches to add to risk prediction.
A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand, for the degree of Doctor of Philosophy. Johannesburg, South Africa September 2017.