Precursors of Hypertensive Heart Phenotype Develop in Healthy Adults

Objectives This study used high-resolution 3-dimensional cardiac magnetic resonance to define the anatomical and functional left ventricular (LV) properties associated with increasing systolic blood pressure (SBP) in a drug-naïve cohort. Background LV hypertrophy and remodeling occur in response to hemodynamic stress but little is known about how these phenotypic changes are initiated in the general population. Methods In this study, 1,258 volunteers (54% women, mean age 40.6 ± 12.8 years) without self-reported cardiovascular disease underwent 3-dimensional cardiac magnetic resonance combined with computational modeling. The relationship between SBP and wall thickness (WT), relative WT, end-systolic wall stress (WS), and fractional wall thickening were analyzed using 3-dimensional regression models adjusted for body surface area, sex, race, age, and multiple testing. Significantly associated points in the LV model (p < 0.05) were identified and the relationship with SBP reported as mean β coefficients. Results There was a continuous relationship between SBP and asymmetric concentric hypertrophic adaptation of the septum and anterior wall that was associated with normalization of wall stress. In the lateral wall an increase in wall stress with rising SBP was not balanced by a commensurate hypertrophic relationship. In normotensives, SBP was positively associated with WT (β = 0.09) and relative WT (β = 0.07) in the septal and anterior walls, and this regional hypertrophic relationship was progressively stronger among pre-hypertensives (β = 0.10) and hypertensives (β = 0.30). Conclusions These findings show that the precursors of the hypertensive heart phenotype can be traced to healthy normotensive adults and that an independent and continuous relationship exists between adverse LV remodeling and SBP in a low-risk population. These adaptations show distinct regional variations with concentric hypertrophy of the septum and eccentric hypertrophy of the lateral wall, which challenge conventional classifications of LV remodeling.

established by the time therapy has been initiated (2)(3)(4), and are independent determinants of all-cause mortality (5)(6)(7). It remains uncertain what determines the onset and pattern of left ventricular (LV) remodeling, but the duration and severity of elevated blood pressure (BP), as well as genetic, metabolic, and environmental factors are all likely to be important (8).
Although LV remodeling is known to begin at below-hypertensive levels, there is little data on how untreated healthy subjects adapt to rising BP, at what stage adverse LV changes are initiated and what mechanical factors drive hypertrophy (9,10).
Our knowledge of the natural history of hypertensive heart disease in human populations has been developed from volume and mass measurements derived using conventional 2-dimensional (2D) echocardiography and cardiac magnetic resonance (CMR).
Conventional classifications of hypertrophy and remodeling using LV mass/volume ratio rely on strong geometric assumptions about the anatomic uniformity of phenotypic adaptations (11). Three-dimensional (3D) CMR is a novel technique that has advantages over conventional phenotyping as it enables quantitative whole-heart assessment of cardiac physiology and noninvasive modeling of the predictors of LV morphology, function, and wall stress (WS) (12). In this study, we used 3D-CMR to define the relationship between systolic blood pressure (SBP) and regionspecific LV adaptations in a population of healthy adults not taking antihypertensive medications.

METHODS
STUDY POPULATION. In this study, 1,258 adult volunteers (680 women, age range 18 to 80 years, mean age 40.6 AE 12.8 years) were recruited prospectively via advertisement for the UK Digital Heart Project at Imperial College London. We excluded participants who had known cardiovascular or metabolic disease.
Subjects taking prescription medicines were excluded but simple analgesics, antihistamines, and oral contraceptives were acceptable. Female subjects were excluded if they were pregnant or breastfeeding.
Standard safety contraindications to magnetic resonance imaging were applied including a weight limit of 120 kg. All subjects provided written informed consent for participation in the study, which was approved by a research ethics committee. PARTICIPANT (13).

QUANTIFICATION OF LEFT VENTRICULAR MASS
AND VOLUME. Volumetric analysis of 2D LV cine images was performed using CMRtools (Cardiovascular Imaging Solutions, London, United Kingdom).
Cardiac volumes and mass were indexed to body surface area.

THREE-DIMENSIONAL ASSESSMENT OF VENTRICULAR
STRUCTURE AND FUNCTION. Briefly, the heart was segmented from the 3D images using previous knowledge of cardiac anatomy from a set of manually annotated atlases ( Figure 1) (14). Each segmentation was coregistered to ensure anatomical consistency between subjects. Wall thickness (WT) was calculated at over 40,000 points in the 3D model at end-diastole and was measured as the distance between the endocardial and epicardial surfaces perpendicular to the midwall plane. Relative wall thickness (RWT) was determined using a scale transformation of the myocardial surfaces to correct WT for variations in the size and shape of the LV. Differences in the endocardial and epicardial surfaces was determined relative to an average cardiac shape such that an outward or inward change would add or subtract from the volume on a regional basis. Fractional wall thickening (FWT) was calculated as: (end-diastolic WT À end-systolic WT)/end-diastolic WT Â 100 (12).
The definitions of LV geometry used in conventional assessment of chamber size were adapted for 3D datasets (15).  mm Hg) as shown in Figure 2. A summary of subject High spatial resolution imaging data is used to build a statistical model of how systolic blood pressure (SBP) influences the left ventricular (LV) phenotype. This paradigm allows the adaptations of whole heart structure and function in response to a stimulus to be explored.
de Marvao et al.

Effects of Blood Pressure in Healthy Hearts
N O V E M B E R 2 0 1 5 : 1 2 6 0 -9 characteristics split by these SBP thresholds is shown in Table 1. These data are presented for the whole cohort and separately for men and women in Online Table 1. A summary of the regression models using the conventional CMR data for the whole cohort is shown in Online Table 2. Among SBP subgroups, SBP was positively related to indexed LV mass in normotensives (b ¼ 0.53, p < 0.001) and pre-hypertensives Indexed LV mass and cardiac output were significantly lower in normotensives than in pre-hypertensives and hypertensives.
3D LV GEOMETRY. The pattern of LV hypertrophy and remodeling observed with respect to each of the phenotypic variables is illustrated in Figure 3 (and in each sex in Online Figure 1).
In normotensives, the 3D regression models In hypertensives, there was a strong and asymmetric association between SBP and both WT (b ¼ 0.42, significant area ¼ 0.5%) and RWT (b ¼ 0.30, significant area ¼ 9.6%) that was observed in the basal septal and anterior walls, with negative coefficients in the lateral wall.
Overall, there was an outward increase in the LV epicardial surface, relative to the mean shape, associated with SBP (b ¼ 0.08, significant area ¼ 97%). The majority of the LV endocardial surface also showed an outward expansion in response to Taken together, these data show a predominantly eccentric Effects of Blood Pressure in Healthy Hearts 180 mm Hg (black outline) shows how LV geometry varies between these 2 BP. Arrows indicate the relationship between each coefficient and SBP. Abbreviations as in Figure 1.    (11,(27)(28)(29).
We observed that rising SBP is associated with a normalization of Laplacian WS in the septum where concentric hypertrophy is predominant. However, in the majority of the LV, the increase in WS with rising   Prognostically adverse cardiovascular features, including LV hypertrophy and increased vascular stiffness, have been observed in pre-hypertensives with a high prevalence of obesity and diabetes (35).
Our data show that a "hypertensive pattern" of remodeling is also observed in asymptomatic normotensive adults. This suggests that rising SBP may have a much earlier impact on cardiac structure and function than previously recognized and might point toward a nuanced relationship between homeostasis and cardiovascular risk. It was notable that with rising SBP the increase in septal RWT is progressively more pronounced, and this has been previously recognized as a phenotype of hypertension especially in the elderly (36). The mechanism for increasing hypertrophic sensitivity to rising SBP remains to be determined and may depend on vascular, mechanical, and/or neurohormonal factors that contribute to the acceleration in the regional RWT increase as SBP rises (37)(38)(39) datasets provide an accurate patient-specific model of curvature and WT that does not rely on geometric assumptions; however, we assumed that material properties of the myocardium were uniform as we did not have data on mechanical characteristics or fiber orientations in this cohort. Therefore, the relationship between WS and hypertrophy is subject to unmeasured factors and should be interpreted cautiously.
Although 3D imaging offers advantages for wholeheart phenotyping, the images also have lower signal-to-noise ratio and poorer temporal resolution compared with those seen with 2D imaging (12). Effect sizes in healthy individuals were more modest compared with those in the hypertensive range of SBP.
We did not assess subjects for insulin resistance or subclinical diabetes, which is thought to influence the pattern of LV remodeling, although findings have been inconsistent (43), nor did we account for the potential effects of smoking or alcohol consumption.
We controlled for racial group, but this study did not have sufficient power to explore ethnic differences in LV remodeling (44). Each vertex in the 3D model was considered independently of its neighbors and cluster-based significance thresholding could be more sensitive for detecting extended areas of anatomical association.

CONCLUSIONS
The precursors of the hypertensive heart phenotype can be traced to healthy normotensive adults and a continuous relationship exists between adverse LV remodeling and SBP in a low-risk population. Rising SBP is associated with concentric hypertrophy of the septum and eccentric hypertrophy of the lateral wall.