Regional Fat Distribution and Blood Pressure Level and Variability: The Dallas Heart Study

Abstract

Our aim was to investigate the associations of regional fat distribution with home and office blood pressure (BP) levels and variability. Participants in the Dallas Heart Study, a multiethnic cohort, underwent five BP measurements on three occasions over 5 months (two in-home and one in-office) and quantification of visceral (VAT), abdominal subcutaneous adipose tissue (SAT), and liver fat by MRI, and lower body subcutaneous fat by dual x-ray absorptiometry. The relation of regional adiposity with short-term (within-visit) and long-term (over all visits) mean BP and average real variability (ARV) was assessed with multivariable linear regression.

2,595 participants mean age 44 years; 54% women, 48% black; and mean BMI 29 kg/m² were included. Mean SBP/DBP was 127/79mmHg and ARV_SBP was 9.8mmHg over 3 visits. In multivariable-adjusted models, higher amount of VAT was associated with higher short-term (both home and office) and long-term mean SBP (β[SE]: 1.9[0.5], 2.7[0.5], and 2.1[0.5], respectively, all P<0.001), and with lower long-term ARV_SBP (β[SE]: -0.5[0.2], P<0.05). In contrast, lower body fat was associated with lower short-term home and long-term mean BP (β[SE]: -0.30[0.13] and -0.24[0.1], respectively, both P<0.05). Neither SAT or liver fat were associated with BP levels or variability.

In conclusion, excess visceral fat was associated with persistently higher short- and long-term mean BP level and with lower long-term BP variability whereas lower body fat was associated with lower short- and long-term mean BP. Persistently elevated BP, coupled with lower variability, may partially explain increased risk for cardiac hypertrophy and failure related to visceral adiposity.

Introduction

Obesity is a well-established risk factor for the development of hypertension.^1-8 Most studies linking obesity with hypertension have used office blood pressure (BP) measured on a single occasion, a “snapshot” of BP that may not fully characterize an individual's BP status.^{9, 10} Alternative BP phenotypes, including out-of-office BP measurements and BP variability, have been implicated in the pathogenesis of vascular damage and stroke independent of office BP.^11-14 However, previous population-based studies have shown an inconsistent relationship between obesity and BP level and variability.^14-19

Many population studies examining the relationship between obesity and BP have used body mass index (BMI) as a measure of adiposity.^1-4 However, obesity is a complex and heterogeneous condition, with varying amounts of ectopic fat deposition (e.g., visceral adipose tissue, VAT) despite similar BMI among individuals.^{20, 21} Therefore, prior studies assessing BMI alone may not accurately characterize the impact of adiposity on BP. How individual variation in regional fat distribution, a key determinant of adverse cardiovascular phenotypes,^{20, 22, 23} contributes to out-of-office BP levels and short- and long-term BP variability remains uncertain since previous studies have focused primarily on the relationship of adipose depots with in-office BP levels in isolation.^{1-4, 6-8}

Therefore, using data from the Dallas Heart Study (DHS), a multiethnic population cohort, we assessed the relationship between regional fat distribution and BP phenotypes including short-term (home and office) BP levels, long-term BP levels (3 visits over 5 months), and short- and long-term BP variability. Given that visceral adiposity may mediate sympathetic nerve activity, impaired (pressure) natriuresis, insulin resistance, and adipokine and aldosterone dysregulation,^{3, 5, 24, 25} we hypothesized that: (1) those with excess visceral adiposity would have sustained short- and long-term BP elevation with higher BP variability, and (2) the association between visceral adiposity and BP would be modified by circulating natriuretic peptides, adipokines (i.e., adiponectin and leptin), insulin resistance, and aldosterone levels.

Methods

The DHS is a multiethnic (blacks, whites, and Hispanics), probability-based population study, recruiting Dallas County residents ages 18 to 65 years. Details of the DHS have been described previously.²⁶ Briefly, between 2000 and 2002, 2,595 participants completed 3 visits during which BP was measured as described below. The 1^st and 2^nd examinations were an in-home survey during which a trained surveyor measured the participant's BP using an oscillometric device; the 3^rd examination was an in-office survey during which BP was measured with the same oscillometric device. The mean (±SD) interval between each DHS visit was 2.5 ± 4.5 months. Figure 1 provides a timeline for the sequence and timing of the BP measurements over the three DHS visits. Participants also underwent laboratory testing, and assessment of body fat distribution by abdominal MRI and DEXA scanning. All participants provided written informed consent, and the University of Texas Southwestern Medical Center institutional review board approved the protocol.

Figure 1. Defining Blood Pressure Phenotypes.

The figures show an example of individual blood pressure (BP) measures across 3 visits (two in-home and one in-office). Five BP measures were taken at each of the 3 visits (black circles, panel A), and their average was defined as mean home and office BP, respectively (red dashed line in panel A and red circle, panel B). Short-term BP variability was defined as average real variability (ARV) and standard deviation (SD) over 5 reading at each of the 3 visits. Short-term home ARV is calculated as (ΔH1+ΔH2+ΔH3+ΔH4)/4 and short-term office ARV as (ΔO1+ΔO2+ΔO3+ΔO4)/4. SDs were calculated from all 5 BP values within each visit, and coefficient of variation (CV) was calculated as SD×100/mean BP. Maximum and minimum BP difference (MMD) was calculated as maximum BP minus minimum BP within each visit.

In panel B, the absolute differences of mean BP between successive visits are shown as Δ1 and Δ2. Red circles represent average of five BP measures taken at each of the 3 visits. For example, Δ1 represents the difference in SBP between the visit 1 BP and visit 2 BP. Long-term ARV is calculated as (Δ1+Δ2)/2. Mean BP and SD were calculated from all 3 BP values from visit 1 to visit 3 for each individual, and CV was calculated as SD×100/mean BP. MMD was calculated as maximum BP minus minimum BP over the entire follow-up.

The mean (±SD) interval between visits was 2.5±4.5 months.

BP measurements and BP phenotype definitions

Trained medical staff took five BP measurements with one minute intervals using the appropriate cuff size after 5 minutes of rest in seated position at both the home and office settings. An automated oscillometric device (Series #52,000, Welch Allyn, Arden, North Carolina) was used.⁸ Staff were trained to choose the appropriate cuff size for the participants and wrap the cuff around the arm with the center of the bladder over the brachial artery. The average of five BP measures taken during the 1^st and 2^nd in-home visit was defined as home BP. The average of five BP measures taken during the in-office visit was defined as office BP. Hypertension was defined as average systolic BP (SBP) ≥140 mm Hg, average diastolic BP (DBP) ≥90 mm Hg, or antihypertensive medication use.

Alternative BP phenotypes evaluated here include mean BP at each visit and over all 3 visits (BP level), within-visit (defined as short-term) home and office BP variability, and visit-to-visit (defined as long-term) BP variability over the 3 visits. For BP variability, we calculated the standard deviation (SD_SBP and SD_DBP), coefficient of variation, the maximum and minimum BP difference, and average real variability (ARV_SBP and ARV_DBP) over 5 readings taken at each visit (i.e., within-visit home or office BP variability) or across 3 visits (i.e., visit-to-visit BP variability) (Figure 1). These measures have been used to describe BP variability in previous studies.^{11-19, 27-30} ARV is calculated as (ΔBP1+ΔBP2+ΔBP3+ΔBP4)/4 for short-term ARV and as (ΔBP1+ΔBP2)/2 for long-term ARV where ΔBP is the absolute difference between successive BP measurements. In contrast with SD, ARV takes the order of the BP measurements into account.^{12, 13}

Variable definitions

Data on smoking, physical activity, medication use, clinical history of cardiovascular disease (CVD), and fasting laboratory values were collected using standardized protocols.²⁶ Physical activity was derived using self-reported frequency and type of leisure-time physical activity and a standard conversion for metabolic equivalence units (METs) and is reported as MET-min/week.³¹ Diabetes was defined by fasting blood glucose≥ 126 mg/dL, nonfasting blood glucose ≥200 mg/dL, or ongoing medical treatment for diabetes. Estimated glomerular filtration rate (eGFR) was calculated by the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) formula.³² Serum aldosterone was measured using a standard assay. Measurement methods of circulating N-terminal pro–B-type natriuretic peptide (NT-proBNP: Roche Diagnostics, Indianapolis, Indiana), BNP (Alere Inc., San Diego, California), adiponectin (Millipore, Billerica, MA, USA), leptin (Linco Research Inc., St. Charles, Missouri) have been described previously.²⁴ The homeostasis model assessment of insulin resistance index (HOMA-IR) was calculated with the following: (fasting insulin [μIU/mL] × fasting glucose [mmol/L]) divided by 22.5.³³ The intraassay and interassay coefficients of these tests were all <12%.

Body fat distribution and imaging measures

BMI was calculated as weight (kg)/height² (m). Retroperitoneal, intraperitoneal, and subcutaneous adipose tissue (SAT) abdominal fat masses were quantified by 1.5-T MRI (Intera, Philips Healthcare, Best, the Netherlands), with a single MRI slice taken at the L2–L3 level using manual contours, as previously validated against cadaveric samples.³⁴ Areas were converted to mass (kg) using previously determined regression equations.³⁵ VAT was defined as the combination of both retroperitoneal and intraperitoneal fat masses to express the total intra-abdominal (visceral) fat mass. DEXA (Delphi W scanner, Hologic, and Discovery software version 12.2) was used to measure lower body subcutaneous fat. Lower body subcutaneous fat was delineated by 2 oblique lines crossing the femoral necks and converging below the pubic symphysis and included gluteal-femoral fat.³⁶ Participants also underwent ¹H-magnetic resonance spectroscopy for hepatic triglyceride quantification, as previously described.³⁷

Statistical analyses

Descriptive statistics are presented as means and SD, and proportions where appropriate. Correlations between BP phenotypes and clinical characteristics were calculated by Pearson correlation method. Cubic smoothing spline curves were drawn to describe the shape of the associations between adiposity variables and BP phenotypes.

Multivariable-adjusted linear regression models were used to assess the association between adiposity variable and BP phenotype (both as continuous variables). We verified the model assumptions of linearity, normality of residuals, homoscedasticity and absence of collinearity.³⁸ Variance inflation factors were calculated to examine for substantial multicollinearity among adiposity-related parameters, and values >5.0 were considered to indicate collinearity. Covariates included demographic variables: age, sex, and race and clinical characteristics: BMI, smoking, physical activity, fasting glucose, total cholesterol, high-density lipoprotein cholesterol, use of antihypertensive medications, prevalent diabetes, and prevalent CVD. These covariates were selected a priori because they have known correlations with BP levels or BP variability.^11-19 Analyses for heterogeneity of effect between the adiposity variable and BP phenotype by sex and race were performed. We also assessed the associations between adiposity-related biomarkers and BP phenotypes and whether the relationship between adiposity variables and BP phenotypes would be attenuated after adjustment for heart rate, eGFR, and adiposity-related biomarkers (i.e., circulating NT-proBNP, BNP, adiponectin, leptin, HOMA-IR, and aldosterone levels), which all represent potential mediators in the adiposity-BP relationship.^{20, 24, 25} We conducted additional sensitivity analyses by: (1) excluding individuals taking antihypertensive medications; and (2) assessing ARV defined by the two in-home visit BPs only. Statistical significance was defined as a P<0.05 on 2-sided tests. All statistical analyses were performed with SAS version 9.4 (SAS Institute, Cary, North Carolina).

Results

Of the 2,595 participants, 54 % were women, 48 % were black, mean age was 44 years, and 20% used antihypertensive medication (Table 1). Long-term mean (±SD) SBP and DBP, SD_SBP, and ARV_SBP were 127±17mm Hg, 79±9 mm Hg, 8±6mm Hg, and 10±7 mm Hg, respectively. The coefficient of variation and the maximum and minimum BP differences were strongly correlated with SD (Pearson's r>0.95; Supplementary Table S1), and therefore we only report BP variability by SD and ARV. The mean BP level at each visit and short-term BP variability are shown in Supplementary Table S2. The mean (±SD) interval between visits was 2.5±4.5 months (Figure 1). Short-term home and office ARV_SBP were only weakly associated with long-term ARV_SBP (r=0.11-0.17; P<0.001). Higher age, male sex, black race, higher BMI, lower physical activity, and diabetes were associated with higher SBP in all settings (i.e., at home, office, and over 3 visits), as well as higher short- and long-term ARV_SBP (P<0.05 for all).

Table 1. Clinical characteristics of study cohort (n=2,595).

Descriptive variable
Age, years	43.9±10.0
Men, %	46
Ethnicity, %
Non-Hispanic white	32.0
African American	48.4
Hispanic	17.4
Other	2.2
Hypertension, %	33
Diabetes mellitus, %	10
Current smoker, %	28
Physical activity, MET-min/week	452±772
Antihypertensive medication, %	20
Total cholesterol, mg/dL	181±39
High-density lipoprotein cholesterol, mg/dL	50±15
Fasting glucose, mg/dL	102±43
Triglyceride, mg/dL	124±105
eGFR, mL/min/1.73 m2	100.4±9.0
Pre-existing CVD, %	7
Long-term BP values, over 3 visits
Mean SBP, mmHg	126.8±16.7
Mean DBP, mmHg	79.0±9.0
SDSBP, mmHg	7.8±5.5
SDDBP, mmHg	4.8±2.8
ARVSBP, mmHg	9.8±7.1
ARVDBP, mmHg	6.0±3.8
Adiposity variables
Body mass index, kg/m2	29.3±6.6
VAT, kg	2.2±1.0
Retroperitoneal fat mass, kg	0.8±0.5
Intraperitoneal fat mass, kg	1.3±0.7
SAT, kg	4.9±3.0
VAT/SAT ratio	0.5±0.3
Hepatic triglyceride, %	5.6±5.7
Lower body subcutaneous fat, kg	9.5±4.6

Data are expressed as the means ± SD or percentage. Participants underwent three examinations on different occasions during 5 months, including the 1^st visit (in-home survey), the 2^nd visit (in-home survey), and the 3^rd visit (in-office survey), over 5 months. VAT was defined as the combination of both retroperitoneal and intraperitoneal fat mass.

SBP indicates systolic blood pressure; DBP, diastolic blood pressure; CVD, cardiovascular disease; eGFR, estimated glomerular filtration rate; SD, standard deviation; ARV, average real variability; VAT, visceral adipose tissue; SAT, subcutaneous adipose tissue.

Figure 2 and Supplementary Figure S1 show the associations of BMI and VAT with each BP phenotype. Higher BMI and VAT were both associated with higher mean SBP in all settings. In contrast, BMI was positively associated with long-term ARV_SBP, whereas the association between VAT and long-term ARV_SBP was inverse.

Figure 2. Body mass index, Visceral Adiposity, and Blood Pressure Phenotype.

The figures show associations of BMI (A) and visceral fat mass (B) with BP phenotype over all 3 visits. Red lines represent mean SBP over visits and blue lines represent long-term ARV_SBP. Solid line is the cubic smoothing spline fit, and dotted line is 95% confidence interval.

Linear regression models examining the associations of adiposity variables with SBP phenotypes are shown in Table 2 and Supplementary Table S3, and with DBP phenotypes in Supplementary Table S3. With adjustments for covariates including BMI, higher VAT was associated with higher mean SBP in all settings, whereas greater lower body subcutaneous fat was associated with lower mean SBP over 3 visits (Table 2). When VAT, SAT, and lower body subcutaneous fat were analyzed jointly, the associations with VAT and lower body subcutaneous fat remained statistically significant (Table 3, left column). When retroperitoneal fat mass and intraperitoneal fat mass were considered jointly, both were independently associated with mean BP in all settings (Table 3, right column). Within VAT, more retroperitoneal but not intraperitoneal fat mass was associated with higher short-term office ARV_SBP, and with lower long-term ARV_SBP (Table 2). Results were generally similar after additional adjustment for heart rate (data not shown), eGFR, circulating NT-proBNP, BNP, adiponectin, leptin, HOMA-IR, and aldosterone levels (Supplementary Table S4). We examined the associations between adiposity-related biomarkers and short- and long-term BP levels (Supplementary Table S5) and with BP variability (Supplementary Table S6) accounting for adiposity. We did not find an association of HOMA-IR with short- and long-term BP levels or variability. Positive associations were observed between NT-proBNP and short- and long-term BP levels and variability and between adiponectin and aldosterone levels with higher long-term BP variability.

Table 2. Multivariable-adjusted linear regression models to examine the association between adiposity variable and short-term and long-term mean BP and BP variability (n=2,595).

Adiposity variable	Short-term home SBP		Short-term office SBP		Long-term SBP over 3 visits
	β (SE)	R2,%	β (SE)	R2,%	β (SE)	R2,%
	Mean SBP
VAT	1.91 (0.52)‡	28	2.66 (0.51)§	28	2.14 (0.46)§	33
Retroperitoneal fat mass	1.82 (0.83)*	28	2.26 (0.81)†	27	1.93 (0.73)†	32
Intraperitoneal fat mass	1.54 (0.60)†	28	2.29 (0.58)§	27	1.78 (0.52)‡	32
SAT	0.41 (0.23)	27	0.44 (0.23)	27	0.38 (0.20)	32
VAT/SAT ratio	3.30 (1.60)*	27	3.39 (1.58)*	27	2.72 (1.41)	32
Hepatic triglyceride	0.02 (0.07)	28	0.08 (0.07)	27	0.03 (0.06)	33
Lower body subcutaneous fat	-0.30 (0.13)*	28	-0.22 (0.13)	27	-0.24 (0.12)*	32
	ARVSBP
VAT	0.08 (0.08)	5	0.09 (0.08)	5	-0.46 (0.23)*	10
Retroperitoneal fat mass	0.08 (0.12)	5	0.27 (0.13)*	5	-0.95 (0.36)†	10
Intraperitoneal fat mass	0.06 (0.09)	5	-0.02 (0.09)	5	-0.11 (0.26)	9
SAT	0.01 (0.03)	5	0.04 (0.04)	5	0.09 (0.10)	9
VAT/SAT ratio	0.13 (0.24)	5	0.04 (0.25)	5	-0.29 (0.70)	9
Hepatic triglyceride	-0.01 (0.01)	5	-0.01 (0.01)	5	-0.02 (0.03)	10
Lower body subcutaneous fat	0.01 (0.02)	5	0.003 (0.02)	5	-0.07 (0.06)	10

Each adiposity-related parameter was analyzed separately. Home SBP was defined as SBP values measured at the 2^nd in-home survey (visit 2 examination), office SBP was defined as SBP values measured at in-office survey (visit 3 examination), and SBP over 3 visits were calculated from visit 1 BP (the 1^st in-home survey), visit 2 BP (the 2^nd in-home survey), and visit 3 BP (the 3^rd in-office survey).

^βrepresents unstandardized regression coefficient, and R² represents a measure for the model prediction. Adjusted βs (95% CIs) associated with a 1kg increment (retroperitoneal fat mass, intraperitoneal fat mass, VAT, SAT, lower body subcutaneous fat), a 1% increment (hepatic triglyceride), and a 1 unit increment (VAT/SAT ratio) are shown. As adjustment factors, all models include demographic variables (age, sex, and race)+clinical characteristics at visit 1 (BMI, smoking, physical activity, fasting glucose, total cholesterol, HDL, use of antihypertensive drugs, prevalent diabetes, and prevalent CVD).

VAT indicates visceral adipose tissue; SAT, subcutaneous adipose tissue. Statistical significance was defined as P <0.05.

^*P<0.05;

^†P<0.01;

^‡P<0.001.

^§P<0.0001.

Table 3. Multivariable-adjusted linear regression models to examine the associations between adiposity variable and short-term and long-term mean SBP (n=2,595).

Adiposity variable	β (SE)	VIF	R2,%		β (SE)	VIF	R2,%
	Short-term mean home SBP
VAT	1.93 (0.52)‡	2.4	28	Retroperitoneal fat	3.97 (0.78)§	1.3	26
SAT	1.15 (0.23)§	3.8		Intraperitoneal fat	3.62 (0.52)§	1.3
Lower body subcutaneous fat	-0.44 (0.14)†	4.1
	Short-term mean office SBP
VAT	2.48 (0.52)§	2.4	27	Retroperitoneal fat	4.02 (0.76)§	1.3	27
SAT	0.87 (0.22)§	3.8		Intraperitoneal fat	3.87 (0.51)§	1.3
Lower body subcutaneous fat	-0.36 (0.14)†	4.1
	Long-term mean SBP
VAT	2.06 (0.46)§	2.4	33	Retroperitoneal fat	3.80 (0.69)§	1.3	31
SAT	0.96 (0.20)§	3.8		Intraperitoneal fat	3.56 (0.46)§	1.3
Lower body subcutaneous fat	-0.36 (0.12)†	4.1

In left column, VAT, SAT, LBT, and total fat-free (lean) mass were analyzed jointly. In right column, retroperitoneal fat mass and intraperitoneal fat mass were analyzed jointly. β means unstandardized regression coefficient, and R² means a measure for the model prediction. VAT was defined as retroperitoneal plus intraperitoneal fat masses. Adjusted βs (95% CIs) associated with a 1kg increment (VAT, SAT, lower body subcutaneous fat, retroperitoneal fat mass, and intraperitoneal fat mass) are shown.

As adjustment factors, we used demographic variables (age, sex, and race), clinical characteristics (smoking, physical activity, fasting glucose, total cholesterol, HDL, use of antihypertensive drugs, prevalent, and prevalent CVD), and total fat free (lean) mass were used.

VAT indicates visceral adipose tissue; SAT, subcutaneous adipose tissue. Statistical significance was defined as P <0.05.

^*P<0.05;

^†P<0.01;

^‡P<0.001.

^§P<0.0001.

When DBP was analyzed in place of SBP and SD was used to represent BP variability instead of ARV, results were similar (Supplementary Table S3). Results were generally similar in sex-stratified analyses (Supplementary Table S7, men and Supplementary Table S8, women) and there was no heterogeneity of effect between adiposity variables and mean SBP level or long-term ARV_SBP by sex or race (all P>0.06). Results were similar when participants taking antihypertensive medications were excluded (n=2,070, Supplementary Table S9). When long-term ARV_SBP was defined using only the two in-home visit SBPs, results were similar (data not shown). Neither BP level nor BP variability in all settings was associated with SAT or liver fat after multivariable adjustment (Table 2).

Discussion

In this multiethnic community-based cohort of younger and middle-aged adults, higher visceral adiposity was associated with persistently elevated mean SBP and DBP with lower long-term BP variability over 5 months, independent of BMI. Conversely, greater lower body subcutaneous fat was associated with lower long-term mean SBP, independent of both BMI and visceral adiposity. Neither short- or long-term mean SBP and DBP nor BP variability was associated with SAT or liver fat after multivariable adjustment. These findings highlight the heterogeneity of obesity sub-phenotypes and demonstrate the divergent relationships between various fat depots and BP level and variability.

Higher short- and long-term BP variability have been shown to be associated with higher risk for CVD morbidity and mortality, and all-cause mortality, independent of mean BP.^{11-13, 28, 29} The associations between obesity (defined by BMI) and BP variability, have been inconsistent.^14-16 In our study, participants with higher BMI had higher long-term ARV_SBP with a modest increase in mean SBP across BMIs, whereas an increase in mean SBP with less variable long-term ARV_SBP was observed with increasing visceral adiposity. Regional fat distribution may therefore be a determinant of individual BP level and variability, independent of BMI, which would partially explain the complex associations of obesity with BP variability. Although visceral adiposity may increase sympathetic tone, its effect on baroreflex sensitivity is controversial. Much of the literature on baroreflex sensitivity associated with visceral obesity has only addressed baroreflex control of heart rate (cardiovagal baroreflex).³⁹ However, the role of visceral obesity on baroreflex control of muscle sympathetic nerve activity (sympathetic baroreflex), which is more important for BP regulation, is more controversial.^{40, 41} While one study showed impaired sympathetic baroreflex,⁴⁰ another showed no impairment in humans with visceral obesity.⁴¹ Thus, high muscle sympathetic nerve activity coupled with relatively intact sympathetic baroreflex could explain high BP without increasing BP variability in individuals with visceral obesity in our study.

When visceral adiposity was analyzed by its components, retroperitoneal but not intraperitoneal fat mass was associated with higher long-term BP level with lower variability. A previous observation from the DHS with a median follow-up of 7 years demonstrated that retroperitoneal fat mass (i.e. fat surrounding the kidneys) was more strongly associated with incident hypertension compared with intraperitoneal fat mass.⁸ The Framingham Heart Study also found that excess retroperitoneal fat around the kidneys was associated with hypertension risk independent of BMI or visceral adiposity.⁴²

Multiple biological mechanisms may underlie the associations we observed between VAT and sustained BP elevation. First, natriuretic peptide deficiency and abnormal adipokine regulation might contribute to sustained BP elevation in those with excess VAT.^{5, 24} However, in our study, the association between excess VAT and higher long-term BP level with less variability was not attenuated after adjustments for circulating natriuretic peptide, adiponectin, and leptin levels. Second, neurohormonal activation, such as the renin-angiotensin-aldosterone system and sympathetic nerve activity,^{3, 5, 20, 21} might contribute to sustained BP elevation. To test this possibility, analyses were performed adjusting for aldosterone or heart rate, with similar results observed. However, aldosterone measurements reflect only one aspect of renin-angiotensin-aldosterone system, and resting heart rate is an indirect marker of sympathetic nerve activity and strongly affected by vagal activity.⁴³ Third, impaired natriuresis may occur due to physical compression of the kidneys by fat (particularly, retroperitoneal fat),⁵ resulting in volume expansion and sodium retention. The potential mechanisms and clinical implications of less long-term ARV_SBP at higher levels of visceral adiposity require further investigation. Positive associations were observed between NT-proBNP and short- and long-term BP levels and variability. The relationship between natriuretic peptides and BP is complex; natriuretic peptides reduce BP by decreasing blood volume and systemic vascular resistance, whereas higher serum concentration of natriuretic peptides reflect pressure-induced cardiac damage.⁴⁴ Given the nature of cross-sectional analysis, causality cannot be determined and is beyond the scope of this study.

The strengths of this study include a large multiethnic epidemiologic cohort with well-phenotyped measures of adiposity. However, several limitations should be noted. First, since this is a cross-sectional study, we are unable to assign causality to our findings. Further studies are needed to determine whether reducing visceral adiposity or improving regional fat distribution can result in BP reduction in younger and middle-aged adults. Second, studies have shown that BP variability assessment will vary according to the number of measurements obtained, with a larger number of measurements allowing more accurate characterization of the BP variability.²⁷ In our study, we were able to assess only 5 BP measurements obtained at each of the 3 study visits for determination of short and long term BP variability. We acknowledge that our study would be better powered to identify additional relationships between fat depots and ARV if further measurements were made. Further study with a greater number of visits and longer term follow-up is warranted. Third, there remains possible residual confounding in the adiposity-BP phenotype relationship, such as individual dietary patterns, environmental factors, or genetics, which we are unable to account for. Fourth, home BP was obtained by surveyors, not by self-measurement. Therefore, the definition of home BP used in our study may not be exactly congruent with other definitions used in the literature.⁴⁵ However, the use of local lay, ethnically congruent field staff members likely minimized the alerting reaction during home BP measurements in the DHS, making the measurement more consistent with a “home” measurement. Furthermore, a recent study from the DHS used the same home BP definition to describe the associations of white coat and masked hypertension with target organ complications and cardiovascular events.⁴⁶ Fifth, we did not adjust for multiple testing, although findings remained consistent after multivariable adjustment; validation of our findings in other epidemiological cohorts is warranted. Lastly, our findings may not be generalizable to individuals older than 65 years or other racial/ethnic groups not included in the DHS (e.g., Asians).

Perspective

In this study of younger and middle-aged adults, higher visceral adiposity was associated with higher BP level with lower variability over multiple visits. The presence of persistently elevated BP, coupled with lower BP variability, may impose a higher cardiac workload and may partly explain the increased risk for cardiac hypertrophy and heart failure seen with excess visceral adiposity.^22,23 Further studies are needed to determine whether modification of adipose tissue distribution may help improve diagnostic and therapeutic strategies for hypertension.

Novelty and Significance.

What's New

Excess visceral fat is associated with persistently higher long-term mean BP level with lower variability, whereas more lower body fat is associated with lower long-term BP level, independent of BMI.

What Is Relevant?

Persistently elevated BP, coupled with lower variability, may partially explain increased risk for cardiac hypertrophy and failure related to visceral adiposity.

Summary

Regional fat distribution is an important determinant of individual BP level and variability independent of BMI.