In-depth phenotyping by cardiovascular magnetic resonance uncovering differences between ethnic groups in hypertensive heart disease

Abstract

Background: Black African/African-Caribbean individuals with hypertension (BH) are at greater risk of heart failure than those of white European ethnicity (WH). The mechanisms underlying this dissimilarity remain poorly understood. Purpose(s): To investigate the influence of ethnicity on left ventricular (LV) remodelling using multi-parametric cardiovascular magnetic resonance (CMR). Method(s): BH (n=44), WH (n=38) and healthy-volunteers (HV; n=25, 5 of black ethnicity) underwent comprehensive CMR. The exam included: i) Arterial Stiffness/Afterload pulse-wave-velocity (PWV), aortic elastance (Ea) and systemic vascular resistance (SVR) by phase-contrast velocityencoding imaging; ii) Ventricular remodelling/Function LV and right ventricular (RV) volumes, mass, ejection fraction (EF), LV peak-filling rate by short-axis cine images; myocardial strains were measured by feature tracking; iii) Left atrial (LA) remodelling/Function volumes and functions by long-axis cine images; iv) Tissue characterisation: extracellular volume by pre/post-contrast T1-mapping and late gadolinium enhancement (LGE) for interstitial and replacement myocardial fibrosis, respectively. Multivariate linear regression models were developed to investigate how LV remodelling associates with ethnicity, arterial afterload, including elastance (Ea) and stiffness [PW], and SVR. Models were adjusted for age, gender, bodymass- index, LV volumes or function and LA volumes. Result(s): Subject characteristics are summarised in the Table. PWV and Ea and SVR were greater in hypertensives, particularly in BH, than HV; this was paralleled by higher LV mass, interventricular septum thickness (IVS), LA volumes but lower LV-EF. These findings were confirmed after adjusting for age. On the Model-1, IVS was associated with Ea (B=0.335, P=0.008) and black ethnicity (B=0.226, P=0.019) but not with SVR or PWV. For each increment of Ea there was a similar increase of IVS in BH and WH (P=0.602 for interaction), however BH had greater IVS than WH at each Ea value (Figure, fully-adjusted Model-1). On Model-2, LV end-diastolic volume was associated with Ea (B=-0.268, P=0.001), SVR (B=-0.319, P=0.019) but not with PWV or ethnicity. However, the inverse relation between LV size and Ea was significantly attenuated in BH (P=0.039 for interaction), (Figure, fullyadjusted Model-2). On model-3, LV-EF was associated with Ea (B=0.223, P=0.009) but not with ethnicity, PWV or SVR. LV-EF reduction for each Ea increment was similar for BH and WH (P=0.597 for interaction). Conclusion(s): BH and WH show a distinctive LV remodelling phenotype. BH had a greater susceptibility to hypertrophy and an attenuated reduction of chamber size in response to arterial afterload. Further research to disentangle the genetic and environmental factors underlying these ethnic group-specific differences is utterly required.