OUP user menu

Impact of blood pressure and insulin on the relationship between body fat and left ventricular structure

Kristjan Karason, Lars Sjöström, Ingemar Wallentin, Markku Peltonen
DOI: http://dx.doi.org/10.1016/S0195-668X(03)00312-9 1500-1505 First published online: 2 August 2003

Abstract

Aims Obesity leads to hypertension and metabolic disturbances, as well as left ventricular hypertrophy and altered left ventricular geometry. However, the underlying mechanisms behind these relationships are not clear. The aim of this study was to investigate how body composition, blood pressure and metabolic factors relate to left ventricular mass and geometry.

Methods and results We included 60 patients with obesity (BMI 31–52) and 43 non-obese subjects (BMI 18–27). Body weight, blood pressure and metabolic parameters were measured and echocardiography was performed. Body composition was determined by gender-specific anthropometric equations. Multivariate analyses showed that both body fat and lean body mass were independently and positively associated with left ventricular mass, whereas adipose tissue alone was related to relative wall thickness. When blood pressure was added to the model, the associations between body fat and left ventricular mass and geometry were weakened. Further adjustment for insulin levels eliminated the relationship between adipose tissue and relative wall thickness.

Conclusions Both total adipose tissue and lean body mass predict an increase in myocardial mass, while adipose tissue alone is related to a rise in relative wall thickness. The concentric left ventricular geometry associated with body fat accumulation appears to be mediated, at least in part, by blood pressure and insulin levels.

  • Obesity
  • Blood pressure
  • Insulin
  • Hypertrophy
  • Echocardiography

1 Introduction

Obesity is frequently associated with the development of hypertension and metabolic aberrations, as well as disturbances in cardiac structure.1Changes in left ventricular structure associated with body fat accumulation include increments in both chamber size and wall thickness, leading to left ventricular hypertrophy,2a powerful risk factor for cardiovascular morbidity and mortality.3,4In obese people, wall thickness frequently increases to a greater extent than chamber size2,5causing a rise in relative wall thickness, an additional structural measure that has been related to increased cardiovascular risk.6,7

Some of the cardiovascular variation found in obese people is related to the haemodynamic changes that occur with an accumulation of excess body fat.4,8As weight increases, total blood volume and cardiac output rise and cause volume overload, a condition commonly followed by left ventricular dilatation. A parallel increase in wall thickness reduces the subsequent rise in ventricular wall stress and leads to eccentric left ventricular geometry. Moreover, obesity is frequently associated with the development of arterial hypertension,9,10a state of pressure overload that results in concentric wall thickening without concomitant chamber dilatation. With time obesity may thus lead to various degrees of eccentric and concentric left ventricular hypertrophy.

Stimuli other than haemodynamic factors are also likely to be of importance in the development of cardiovascular disturbances in obese people.11Obesity isassociated with a cluster of metabolic and hormonal disturbances12,13and it has been suggested that some of them could be involved in the modulation of left ventricular structure. Indeed, some investigators have reported a correlation between measurements of insulin resistance and left ventricular mass,14,15while others have not been able to confirm this.

As obesity develops, both adipose and lean body mass increase, but little is known about the separate effects of these body compartments on left ventricular structure. The aim of this study was to investigate how body composition, blood pressure and metabolic factors are related to left ventricular mass and geometry.

2 Patients and methods

The total study group consisted of 119 subjects, comprising 61 men and 58 women, with ages ranging from 37–61 years. We recruited 76 patients who were defined as obese (BMI 31–52) and 43 individuals who were classified as non-obese (BMI 18–27). The obese subjects were enrolled from the ongoing Swedish Obese Subjects (SOS) study, a nationwide trial designed to determine whether the mortality and morbidity rates among obese individuals who lose weight by surgical means differ from the rates associated with conventional treatment.16The non-obese subjects were recruited from the SOS-reference study, in which inhabitants of the Municipality of Mölndal were randomly recruited to serve as a reference group for the SOS study. Obese and non-obese subjects were matched with respect to gender, age and height. The present study was approved by the Göteborg University ethical committee and all study subjects gave their informed consent.

Body weight was measured with the subjects in light clothing without shoes and rounded off to the nearest 0.1kg. Height measurements were rounded off to the nearest centimetre. BMI was calculated as body weight divided by height squared. Total adipose tissue volume (litres) was calculated from weight and height measurements using gender-specific equations introduced by Kvist et al. 1.36×weight/height-42.0 for males and 1.61×weight/height—38.3 for females.17Total adipose mass (kg) was derived by multiplying volume measurements with adipose tissue density (0.923kg/ml). Lean tissue mass was obtained by subtracting adipose tissue mass from total body weight.

The equations from Kvist et al. were derived from examinations using the 22 scan computed tomography model. The prediction of adipose tissue volume in primary material was associated with errors of 7% and 9% for men and women respectively and, in a cross-validation study, the corresponding prediction errors were 9% and 11%. This accuracy is high for anthropometric techniques and is comparable with bioelectrical impedance analysis, which is associated with a measurement error of around 5% when compared with methods of dual X-ray energy analysis.18

Blood pressure was measured with a mercury sphygmomanometer in the supine position after 10min of rest using an appropriate cuff. Hypertension was defined as systolic blood pressure higher than 140mmHg or diastolic blood pressure higher than 90mmHg or current antihypertensive medication.

Blood samples were obtained in the morning after 12h of fasting from an antecubital vein. S-cholesterol, s-HDL-cholesterol, s-triglycerides, b-glucose and p-insulin levels were measured. A rise in fasting insulin levels was regarded as a marker of insulin resistance.19

Echocardiography studies were performed on each subject using a commercially available Acuson128XP computed sonography system with 2.5MHz transducers. The echocardiograms were obtained at rest with the subject in the left lateral decubitus position. Two-dimensional guided M-mode measurements of left ventricular end-diastolic dimension (LVED), interventricular septum (IVS) and posterior wall thickness (PW) were performed, as recommended by the American Society of Echocardiography.20Wall thickness was calculated as the sum of IVS and PW thickness, while relative wall thickness (RWT) was calculated as the ratio of wall thickness to left ventricular internal dimension. Left ventricular mass (LVM) was calculated according to the regression equation from Devereux et al.: 0.80×(1.04×(LVED+IVS+PW)3−(LVED)3)+0.6gm.21All the recordings were made by the same physician (Dr Wallentin), who was unaware of other data relating to the subjects. In the obese group, 16 patients (21%) had technically unsatisfactory echocardiograms and were excluded from the study. This resulted in a total final study group of 103 patients. There were no differences between patients for whom echocardiograms were technically satisfactory and those for whom they were not. Double determinations of echocardiographic measurements in obese subjects revealed 10% variability in estimations of left ventricular mass.

Statistical analyses were performed using the Stata 7.0 statistical software package (Stata Corporation, College Station, TX). Descriptive statistical results are given as the mean (standard deviation). The chi-square test was used to compare categorical variables; otherwise the differences between obese and non-obese groups were assessed by an unpaired t-test. After pooling data from obese and non-obese subjects univariate regression analyses were performed to estimate associations between clinical variables and measurements of left ventricular mass and geometry. Multiple regression analyses were then used to determine which of the clinical variables were independently associated with the indices of left ventricular structure. The results of the multivariate analyses are given as partial correlation coefficients. A P-value of less than 0.05 was considered significant.

3 Results

The clinical, anthropometric, metabolic and echocardiographic characteristics of the study groups are shown in Table 1. The two groups did not differ in terms of gender, age or height. The higher body weight in the obese group was mainly explained by an increase in adipose tissue, but obese patients also had more lean body mass than non-obese control subjects. Systolic and diastolic blood pressure were higher in the obese group, as was the prevalence of hypertension and diabetes. Among the obese subjects, 25% were taking antihypertensive medication, while the corresponding figure for the non-obese subjects was just 2%. No difference was found in smoking habits between the two groups. Obese and non-obese subjects did not differ significantly in terms of cholesterol levels, but the obese group had significantly lower HDL-cholesterol. Obese patients also had significantly higher fasting values for triglycerides, glucose and insulin. The left ventricular end-diastolic diameter did not differ significantly between the two groups, but, in comparison to the non-obese subjects, the obese patients had increased absolute and relative wall thickness, as well as an increase in left ventricular mass.

View this table:
Table 1

Clinical, anthropometric, metabolic and echocardiographic characteristics of the study groups

VariableNon-obese (n=43)Obese (n=60)P value
Clinical/anthropometric
Gender (M/F)23/2031/29
Age (yrs)49(7)49(6)0.927
Height (m)1.73(0.09)1.72(0.09)0.402
Weight (kg)70(11)115(14)<0.001
Body mass index (kg/m2)23(2)39(4)<0.001
Lean body mass (kg)53(11)61(15)0.004
Total adipose tissue (kg)17(5)55(13)<0.001
Systolic BPa(mmHg)116(14)142(19)<0.001
Diastolic BPa (mmHg)72(10)88(11)<0.001
Hypertension (%)557<0.001
Diabetes (%)0120.040
Current smoker (%)19201.000
S-Triglycerides (mmol/l)1.2(0.7)2.4(2.1)<0.001
B-Glucose (mmol/l)4.1(0.3)5.4(1.9)<0.001
P-Insulin (mU/l)7(3)20(10)<0.001
Echocardiographic
LVEDb(cm)5.1(0.4)5.3(0.5)0.081
Left ventricular mass (g)185(51)280(72)<0.001
Wall thickness (cm)1.9(0.3)2.5(0.4)<0.001
Relative wall thickness0.38(0.05)0.49(0.09)<0.001
  • a BP=blood pressure.

  • b LVED=left ventricular end-diastolic dimension. Values are expressed as means (standard deviation).

Univariate regression analyses on pooled data from obese and non-obese subjects were performed in order to evaluate relationships between clinical variables and left ventricular measurements (Table 2). Both total adipose tissue and plasma insulin levels were strongly correlated with absolute and relative wall thickness, as well as left ventricular mass, but not with chamber diameter. The correlation coefficients for the association between left ventricular structure and other metabolic variables were lower than those for insulin (data not shown). Lean body mass and blood pressure displayed a significant correlation with all four echocardiographic variables. The correlation coefficients were somewhat higher for systolic blood pressure than diastolic.

View this table:
Table 2

Univariate correlation coefficients (r) of body compartments, systolic blood pressure and plasma-insulin on WT, LVDD, LVM and RWT

LVEDaLVMbWTcRWTd
rP-valuerP-valuerP-valuerP-value
Total adipose tissue (kg)0.140.1640.47<0.0010.55<0.0010.49<0.001
Lean body mass (kg)0.41<0.0010.53<0.0010.45<0.0010.250.012
Systolic BPe(mmHg)0.240.0160.54<0.0010.57<0.0010.45<0.001
Plasma-insulin (mU/l)0.040.6910.48<0.0010.58<0.0010.55<0.001
  • a LVED=left ventricular end-diastolic diameter.

  • b LVM=left ventricular mass.

  • c WT=wall thickness.

  • d RWT=relative wall thickness.

  • e BP=blood pressure.

We evaluated the independent associations of body composition, blood pressure and insulin resistance with left ventricular structure in multivariate regression analysis. After controlling for age, gender and antihypertensive treatment, stepwise adjustments were made for adipose tissue, lean body mass, systolic blood pressure and insulin levels resulting in four different models for each left ventricular measure (Tables 3 and 4).

View this table:
Table 3

Partial correlation coefficients for body compartments, systolic blood pressure and plasma insulin from regression on left ventricular end-diastolic diameter (LVED) and left ventricular mass

LVED (cm)Left ventricular mass (g)
Model12345678
Total adipose tissue (kg)0.27c0.22b0.180.23b0.60d0.38d0.26b0.23b
Lean body mass (kg)−0.03−0.030.020.19a0.22b0.20b
Systolic BP (mmHg)0.040.030.26c0.27c
Plasma insulin (mU/L)−0.22b0.11
Adjusted R Squared19.018.317.320.948.449.852.852.9
  • a P<0.10.

  • b P<0.05.

  • c P<0.01.

  • d P<0.001. All models also included age, gender and antihypertensive treatment as covariates.

View this table:
Table 4

Partial correlation coefficients for body compartments, systolic blood pressure and plasma insulin from regression on absolute and relative wall thickness

Wall thickness (cm)Relative wall thickness
Model12345678
Total adipose tissue (kg)0.63c0.44c0.31b0.26a0.49c0.32b0.22a0.15
Lean body mass (kg)0.140.180.130.090.110.05
Systolic BP (mmHg)0.29b0.31b0.21a0.24a
Plasma insulin (mU/L)0.29b0.33b
Adjusted R squared48.148.652.556.826.726.528.937.2
  • a P<0.05.

  • b P<0.01.

  • c P<0.001. All models also included age, gender and antihypertensive treatment as covariates.

Total adipose tissue was significantly and positively associated with left ventricular end-diastolic diameter, whereas lean body mass and blood pressure where not (Table 3, Model 4). Insulin was negatively related to chamber size when other covariates where accounted for (Table 3, Model 4).

Total adipose tissue, lean body mass and systolic blood pressure were all significantly and positively related to left ventricular mass, but insulin levels were not (Table 3, Model 8). Adjustments for lean body mass and blood pressure reduced the relationship between body fat and the mass of the ventricle (Table 3, Models 5–7). Corrections for lean body mass also eliminated gender differences in ventricular mass (data not shown).

In initial models, adipose tissue was positively associated with absolute and relative wall thickness, whereas lean body mass was not (Table 4, Models 2 and 6). Further adjustment for blood pressure and insulin levels weakened or eliminated the relationship between body fat and measures of wall thickness (Table 4, Models 4 and 8).

4 Discussion

The present study confirms that obesity is related to changes in left ventricular structure consisting of increased left ventricular mass and higher relative wall thickness. Analyses of the influence of body composition revealed that both adipose and lean body masses predicted the mass of the left ventricle (Table 3, Model 6), whereas adipose tissue alone was correlated with concentric left ventricular geometry (Table 4, Model 6). Systolic blood pressure and insulin levels eliminated the association between body fat and relative wall thickness (Table 4, Model 8), suggesting that these factors may mediate the impact of adipose tissue on left ventricular structure.

4.1 Body fat and left ventricular mass

The high prevalence of left ventricular hypertrophy in obese people is of great importance as it may increase their risk of premature morbidity and mortality. However, the underlying mechanisms that link obesity with increased left ventricular mass have not yet been elucidated. In this respect, separate effects of different body compartments on heart structure are of interest. In obesity, both adipose and lean body masses increase and, in the present study, both body compartments contributed to an increase in left ventricular mass independent of age, gender and antihypertensive treatment(Table 3, Model 6).

An increment in left ventricular mass related to a rise in lean body mass should probably be regarded as a normal cardiac adaptation, matching the augmented perfusion needs of expanded fat free tissue. Support for this is found in our multivariate analysis, which shows that adjustments for lean body mass eliminated gender differences in left ventricular mass. Similar findings have been reported by Hense et al.22and lean body mass has been proposed as the optimal normalization of left ventricular mass to body size.23

In contrast, an increase in cardiac mass related to the accumulation of adipose tissue mass is probably a more maladaptive process caused by abnormal haemodynamic, metabolic and hormonal stimuli. Accordingly, the relationship between body fat and ventricular mass in our analyses was reduced when systolic blood pressure was taken into account (Table 3, Model 7). It therefore seems rational to avoid correcting left ventricular mass for measures that are strongly affected by the existing amount of adipose tissue, such as body mass index and body surface area.

4.2 Body fat and relative wall thickness

Recent reports have demonstrated that the relationship between wall thickness and chamber diameter is of importance with respect to cardiovascular risk.6,7A concentric left ventricular pattern characterized by high relative wall thickness is associated with a poorer prognosis than eccentric hypertrophy in which the ratio between wall thickness and chamber diameter is normal. Previously, obesity was thought to cause pure eccentric left ventricular hypertrophy due to volume overload and it was assumed that concentric remodelling would only occur in the presence of hypertension.24More recent studies have, however, shown that relative wall thickness tends to increase along with body fat accumulation irrespective of blood pressurelevels.2,5

Multivariate analysis revealed that relative wall thickness was strongly related to adipose tissue mass but not to lean body mass (Table 4, Model 6). The disparate effects of these body compartments on cardiac geometry further indicate that changes associated with body fat accumulation are more maladaptive than those related to increases in lean body mass. When systolic blood pressure and insulin levels were accounted for, the association between adipose tissue and relative wall thickness ceased to exist (Table 4, Model 8). This suggests that the impact of body fat on wall thickness may be mediated, not only by pressure overload but also via metabolic disturbances. Consequently, thick ventricular walls could be regarded as a consequence of the metabolic syndrome.

Recent studies have shown that adipose tissue is an active endocrine gland, including a local renin angiotensin system.25Apart from expressing angiotensin peptides locally, it has been suggested that adipose tissue may contribute to circulating angiotensin II,26an important regulator of blood pressure and a potent growth factor in myocardial tissue.27This provides another potential mechanism by which excess body fat may directly contribute to concentric geometry.

The strong association between body fat and concentric left ventricular geometry found in the present study contrasts to some degree with earlier reports in which obesity has mainly been related to eccentric hypertrophy. One possible explanation could be that our patients were older and therefore had higher a frequency and longer duration of obesity-related complications compared with previous study populations. It is not unlikely that eccentric left ventricular geometry in an early phase of uncomplicated obesity may turn into a more concentric type along with ageing and the accumulation of secondary haemodynamic, metabolic and hormonal disturbances. Alpert et al.28reported a positive correlation between the duration of obesity and left ventricular mass, demonstrating the importance of obesity duration with respect to left ventricular structure.

4.3 Impact of insulin

Relationships between measures of insulin resistance and left ventricular structure have been studied in different populations with varying results. Significant relationships between insulin resistance and left ventricular mass have been reported in both obese14,15and hypertensive 29,30populations. However, other investigators have only found a weak correlation between theseparameters31or no association at all32–35after adjustments for covariates. In the present study, fastinginsulin levels did not predict the weight of the left ventricle, possibly due to the inverse relationship between insulin levels and chamber diameter (Table 3, Model 4), an important contributor to left ventricular mass. In contrast, insulin levels correlated strongly with left ventricular concentric geometry even after adjustments for other relevant covariates (Table 4, Model 8). These findings are in accordance with a study on a worksite population by Ohya et al.36and a study of elderly men by Sundström et al.37in which insulin resistance was more closely related to thick ventricular walls than to ventricular mass.

Epidemiological studies have shown that high fasting insulin levels are associated with an adverse cardiovascular outcome, independent of other risk factors,38and this could be explained in part by a remodelling effect by insulin on left ventricular structure. It has been proposed that insulin may exercise its influence on cardiac geometry by acting as a growth factor39and trophic effects by insulin on myocardial tissue have been demonstrated in cell cultures40and animalmodels.41Moreover, it has been suggested that hyperinsulinemia stimulates sympathetic nervous systemactivity,42which may in turn affect ventricular structure directly, due to growth-stimulating effects, or indirectly, by contributing to increases in heart rate and blood pressure levels. However, further mechanistic studies are required in order to clarify the links between insulin resistance, compensatory hyperinsulinemia and aberrations in cardiovascular structure.

5 Conclusion

Cardiac changes in obese people include increases in left ventricular mass, as well as altered left ventricular geometry. Obesity is associated with increases in both adipose and lean body mass, but these body compartments appear to have disparate effects on left ventricular structure. Whereas both compartments predict increased myocardial mass, adipose tissue alone is related to a rise in relative wall thickness. The concentric left ventricular geometry associated with body fat accumulation appears to be mediated, at least in part, by increased blood pressure and insulin levels.

Acknowledgments

This study was supported by grants from the Swedish Heart Lung Foundation and the Swedish Medical Research Council (grant nos. 05239 and 10880).

References

View Abstract