Author + information
- Received February 10, 2014
- Revision received March 11, 2014
- Accepted March 21, 2014
- Published online October 1, 2014.
- Peter Wohlfahrt, MD, PhD∗,†,‡,§,
- Margaret M. Redfield, MD∗,
- Francisco Lopez-Jimenez, MD∗,
- Vojtech Melenovsky, MD, PhD∗,‖,
- Garvan C. Kane, MD, PhD∗,
- Richard J. Rodeheffer, MD∗ and
- Barry A. Borlaug, MD∗∗ ()
- ∗Division of Cardiovascular Diseases, Department of Medicine, Mayo Clinic Rochester, Rochester, Minnesota
- †International Clinical Research Center, St. Anne’s University Hospital, Brno, Czech Republic
- ‡Center for Cardiovascular Prevention of the First Faculty of Medicine, Charles University and Thomayer Hospital, Prague, Czech Republic
- §Department of Preventive Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czech Republic
- ‖Department of Cardiology, Institute for Clinical and Experimental Medicine, Prague, Czech Republic
- ↵∗Reprint requests and correspondence:
Dr. Barry A. Borlaug, Division of Cardiovascular Diseases, Department of Medicine, Mayo Clinic Rochester, Mayo Clinic and Foundation, 200 First Street SW, Rochester, Minnesota 55905.
Data from this article were partially presented at the European Society of Hypertension Working Group on Obesity, Diabetes, and the High Risk Patients Satellite Symposium and Consensus Meeting, June 6 to 7, 2014, Thessaloniki, Greece.
Objectives The aim of this study was to assess the effects of central and general obesity measures on long-term longitudinal changes in ventricular-arterial mechanics.
Background Obesity, female sex, and ventricular-arterial stiffening are associated with the development of heart failure with preserved ejection fraction. Fat distribution and chronic changes in body composition may affect longitudinal changes in LV properties, independent of arterial load.
Methods In 1,402 subjects from a randomly selected, community-based population, comprehensive echo-Doppler echocardiography was performed at two examinations separated by 4 years. From this population, 788 subjects had paired data adequate for determining left ventricular end-systolic elastance (Ees), end-diastolic elastance (Eed), and effective arterial elastance (Ea).
Results Over 4 years, Ea was decreased by 3% in tandem with improved blood pressure control, whereas Ees and Eed were increased by 14% and 8% (all, p < 0.001). Greater weight loss over 4 years was associated with progressively greater decreases in Ea in men and women. After adjustment for Ea change, weight gain was correlated with increases in Eed in both women and men. Central obesity was associated with greater age-related increases in Ees in women but not in men, independent of arterial load, but central obesity did not predict changes in Eed or Ea.
Conclusions In these subjects, weight gain was associated with increases in LV diastolic stiffness, even after adjustment for changes in arterial afterload, whereas weight loss was associated with reductions in arterial stiffness. Age-related LV systolic stiffening was increased in women, but not in men, with central obesity. Strategies for promoting weight loss and reducing central adiposity may be effective in preventing heart failure with preserved ejection fraction, particularly in women.
Obesity has been reported to independently predict incident heart failure (HF) (1–4). In cross-sectional studies, increased waist circumference (WC) was associated with left ventricular (LV) diastolic dysfunction (DD) (5,6). DD and obesity are common in HF with preserved ejection fraction (HFpEF) and may be causally linked (7). Women display increased ventricular and arterial stiffness compared with men and are more likely to develop HFpEF than are men. However, the complex relationships between aging, sex, adiposity, and ventricular mechanics remain poorly understood.
LV stiffness (elastance) at end-systole (Ees) and end-diastole (Eed) has been reported to increase with aging despite reductions in blood pressure (BP) and arterial afterload (effective arterial elastance [Ea]) (8), suggesting that mechanisms other than elevated afterload are involved. Age-related LV stiffening in this study has been correlated with increases in total body mass, but it is unclear how much of this finding was due to weight-related alterations in (Ea), and the impact of fat distribution has not been assessed. In morbidly obese patients, weight loss has been reported to prevent left atrial enlargement (9) and has been correlated with improved myocardial deformation (10), but little is known regarding the effects of fat distribution and long-term weight changes in the community setting and in the absence of medically complicated obesity.
Visceral fat mass is a major risk factor for cardiovascular disease (CVD) (11,12), whereas peripheral fat seems to confer a protective effect (13). We hypothesized that central obesity is more strongly associated with age-related LV stiffening, and that weight gain is associated with diastolic stiffening independent of load. To test these hypotheses, we assessed the associations between central and general obesity measures, weight change, and longitudinal changes in ventricular-arterial coupling over 4 years using sex-specific analyses in a randomly selected, community-based sample of the population.
In 1997, a random sample of individuals residing in Olmsted County, Minnesota, ≥45 years of age was identified by applying a sampling fraction of 7% within 5-year sex- and age-specific strata. A total of 4,203 persons were invited and 2,042 participated in examination 1 (baseline; 1997 to 2000), which consisted of physical examination, echocardiography, and medical-record abstraction. Comparison of data from invited participants to those from invited nonparticipants disclosed no differences in the prevalences of CVD or diabetes or in Charlson index (14). Four years later, all participants were invited to return, and 1,402 participated in examination 2 (2001 to 2004). Diabetes history was based on physician's diagnosis and treatment. Myocardial infarction and hypertension were diagnosed according to criteria from the World Health Organization and the Sixth Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure, respectively. Some clinical and echocardiographic data from this population have been previously reported (6,8,15–17), but the current study findings regarding obesity measures and their influence on changes in ventricular-arterial stiffening over time have not been reported. The institutional review boards of the Mayo Clinic and Olmsted Medical Center approved the study protocol. Participants provided written informed consent for evaluation and medical-record follow-up.
Assessment of LV structure and function
Comprehensive echocardiographic assessment was performed by registered diagnostic cardiac sonographers using standard instruments and techniques as previously described (16,18). At examinations 1 and 2, echocardiography was performed by the same 3 sonographers and reviewed by 2 cardiologists (M.M.R.), who were masked to clinical and exam-1 echocardiography findings. Ventricular dimensions, wall thickness, and chamber volumes were determined in triplicate from 2-dimensional echocardiography. LVEF was determined by the biplane Simpson method. LV stroke volume (SV) was determined from pulse wave Doppler of the LV outflow tract. Cardiac output (CO) was determined as the product of SV and heart rate. Brachial BP was determined by sphygmomanometry. End-systolic pressure (ESP) was determined from the product of 0.9 × systolic BP.
LV diastolic function was assessed by pulse wave Doppler examination of mitral inflow (E, A velocities, E/A ratio) and tissue Doppler examination of septal mitral annular velocities (e′ velocity). The ratio of echo-Doppler and tissue-Doppler early diastolic velocities (E/e′) was used to estimate LV end-diastolic pressure (EDP) as previously validated (11.96 + 0.596 × E/e′) (19). Operating LV end-diastolic elastance (Eed) was estimated by the ratio of LVEDP and end-diastolic volume (20). LV end-systolic elastance (Ees) was determined by the single-beat technique of Chen et al. on the basis of measured SV, EF, BP, and systolic time intervals as previously reported (16,18,21).
Assessment of arterial function and ventricular-arterial interaction
Effective arterial elastance (Ea), a measure of total arterial load that incorporates both mean and pulsatile components, was determined as ESP/SV (17). Systemic vascular resistance (mean BP × 80/CO) and total arterial compliance (SV/pulse pressure) were determined as additional measures of arterial load. The Ea/Ees ratio was used to assess ventricular-arterial coupling (18).
Height and weight were measured with participants standing without shoes or heavy outer garments. Height was determined using a wall-mounted stadiometer. Body weight was measured using a homologated electronic scale. Measurements were carried out by a trained research nurse, with the subjects in the standing position. Body mass index (BMI) was calculated as weight/height2. Body circumference was measured in centimeters using a nonelastic measuring tape, for waist at the top of umbilicus, for the hip at its maximal circumference, and for the neck at the lower half of the neck.
General obesity was defined as BMI >30 kg/m2. We used WC, waist-to-hip ratio, and waist-to-height ratio as measures of central obesity, which can be used for assessing fat mass distribution and visceral fat mass better than can BMI (22). We used neck circumference as an additional measure of central obesity given that it is more closely related to visceral fat mass and metabolic abnormalities than is WC (23,24).
Laboratory analyses were performed at Mayo Clinic Rochester by experienced laboratory personnel. Glucose was measured on the Roche Hitachi 912 Chemistry Analyzer (Roche Diagnostics and Hitachi High-Technologies Corporation, Tokyo, Japan) using the hexokinase reagent from Boehringer Mannheim Corporation (Indianapolis, Indiana). Insulin was determined using a specific two-site immunoenzymatic assay performed on the Access automated immunoassay system (Beckman Coulter, Inc., Chaska, Minnesota), which has a molar cross-reactivity of 0.10% with pro-insulin. The Homeostatic Model Assessment Index 2–Insulin Resistance (HOMA2-IR) index was obtained by the program HOMA Calculator. Lipid analyses were performed using an automated enzymatic method. Adiponectin was measured using the Human Adiponectin Quantikine enzyme-linked immunosorbent assay (R&D Systems Inc., Minneapolis, Minnesota) at the University of Minnesota (Minneapolis, Minnesota). The interassay coefficient of variation was 5.5%.
Data are expressed as mean ± SD, estimated marginal mean (95% confidence interval), or number (percentage). In the figures, estimated marginal mean ± SEM is plotted. Longitudinal changes in categorical variables were assessed with the McNemar test, and continuous variables, by paired the Student t test or the Wilcoxon signed rank test, as appropriate. We used generalized mixed-effects regression models to test the differences in the rates of change of dependent variables by increases in independent variables. In these models, the intercept was treated as a random effect, whereas time, obesity measures, the interaction term between time and obesity measures, and other covariates were treated as fixed effects. The null hypothesis tested was that the change in slope with a 1-unit increase in the obesity measure from exam 1 to 2 is equal to zero. Reported are coefficients with 95% confidence intervals and p values. In mixed-effects models, p values (unadjusted for multiple comparisons) <0.05 were considered significant. For right-skewed data, we used gamma regression. Models were compared by the corrected Akaike information criterion (AIC). When gamma regression of the dependent variable is used, AIC is based on –2 log pseudo-likelihood. Models with smaller AICs show better fit. Calculations were done using SPSS version 19 (IBM SPSS Statistics, IBM Corporation, Armonk, New York). A 2-sided p value <0.05 was considered statistically significant.
Of 1,402 Olmsted County residents participating in examination 2,788 had complete echocardiographic data available for the assessment of Ees and Ea at both examinations. Subjects with adequate-quality data were less obese compared with subjects without adequate data (Online Table 1). The mean age of the subjects at examination 1 was 60 years, and 48% were men (Table 1). Examination 2 was performed 4.0 ± 0.3 years after examination 1. Descriptive statistics of the study sample at exams 1 and 2 were previously reported (8). From examination 1 to 2, measures of both general and central obesity increased slightly but significantly (Table 1).
From examination 1 to 2, systolic, diastolic, and mean BPs were decreased by 4.5 ± 18.4 mm Hg, 4.1 ± 10.3 mm Hg, and 4.3 ± 11.3 mm Hg, respectively, whereas pulse pressure was not changed (0 ± 14 mm Hg; p = 0.4) (Table 1). LV mass was decreased, whereas chamber size and EF remained stable. Over 4 years, Ea was decreased by 3%, whereas Ees and Eed were increased, on average, by 14% and 8%, respectively (Table 1, Figure 1).
Longitudinal changes in arterial elastance
Arterial elastance was higher in women than in men at both examinations (both, p < 0.001), but there was no difference in the slope of the age-related change in Ea between women and men (Figure 1A). None of the measures of central obesity at examination 1 or their changes over time were associated with the change in Ea, even after adjustment for weight loss (Table 2). After adjustment for age and sex, the magnitude of weight change from examination 1 to 2 was associated with the change in Ea in both men and women (Table 2, Figure 2). BMI at examination 1 was negatively associated with Ea change on univariate analysis. However, baseline BMI was strongly associated with weight change (r = –0.78; p < 0.001). After adjustment for weight change, BMI was no longer associated with Ea changes, whereas the association with weight change remained significant. Thus, the negative association between Ea change and BMI was driven by greater weight loss among heavier subjects.
Longitudinal changes in end-diastolic elastance
As previously reported by our group, Eed was higher in women than in men at both examinations (both, p < 0.001), the slope of Eed increase with aging was higher in women than in men (p = 0.02) (Figure 1B), and increasing body weight was directly correlated with increasing Eed on univariate analysis (r: 0.25; p < 0.0001). None of the measures of central obesity at examination 1 or their changes over time were associated with the change in Eed (Table 2). After adjustment for age and Ea change, weight gain remained positively associated with the age-related increase in Eed in both men and women (Figure 3).
Longitudinal changes in end-systolic elastance
Ees was higher in women than in men at both examinations (both, p < 0.001), but there was no difference in the slope of age-related Ees increase between sexes (p = 0.57) (Figure 1C). As previously reported in this population, weight gain was associated with increased Ees on univariate analysis (r = 0.12; p < 0.0001). However, after adjustment for age and changes in Ea, weight change was no longer correlated with change in Ees (Table 2). However, after adjustment for age and changes in Ea, measures of central obesity and neck circumference were positively associated with the age-related increase in Ees. Sex-stratified analysis revealed that the positive correlation between ventricular systolic stiffening and central obesity measures was driven exclusively by female sex and was absent in men (Table 2, Figure 4). As assessed by AIC, neck-to-height ratio showed the strongest association with change in Ees (AIC: –232.4), whereas there was only a small difference between waist-to-height ratio (AIC: –210.7) and waist-to-hip ratio (AIC: –206.4). There were no correlations between changes in central obesity measures and changes in Ees.
Longitudinal changes in ventricular-arterial coupling
The ventricular-arterial coupling (Ea/Ees) ratio decreased in men and women, with no difference in the slope of decrease (p = 0.14) (Figure 1D). Similar to observations with the change in Ees, changes in Ea/Ees were correlated with measures of central obesity in women but not in men (Table 2).
Humoral factors and ventricular-arterial coupling
Plasma levels of high-density lipoprotein cholesterol, total cholesterol, triglycerides, C-reactive protein, adiponectin, and leptin at examination 1 were not associated with Ea and Ees changes (Table 3). Insulin level and insulin resistance (as measured using HOMA2-IR and glucose level at examination 1) were positively associated with increased Ees in women but not in men (Figure 4). Adjustment for measures of central obesity, including neck-to-height ratio and waist-to-height ratio, eliminated the significance of this association, whereas obesity measures remained associated with Ees change. We did not find any relationship between humoral factors and Eed.
This is the first large-scale, population-based study to comprehensively examine correlations between measures of central and general obesity, changes in obesity over time, and longitudinal changes in ventricular and arterial stiffness in humans. We found that weight gain over 4 years was associated with significant increases in LV diastolic stiffness in both men and women, whereas decreases in weight were associated with reductions in arterial load. Measures of central obesity and insulin resistance were associated with exaggerated increases in Ees in women, but not in men, indicating a sex difference in the biology of age-related ventricular systolic stiffening. These findings suggest that strategies to enhance weight loss could mitigate or prevent age-related changes in ventricular structure and function that promote progression to HFpEF. In addition, the presence of central obesity in women may help to identify a group at higher risk for incident HFpEF that might be targeted with preventive treatments.
Obesity might produce hemodynamic alterations that predispose to changes in cardiac morphology and ventricular function (25) and is a potent independent risk factor for HF (1). Cross-sectional studies have shown that BMI and central adiposity are associated with LV dysfunction and arterial stiffness (26). Russo et al. (27) showed that higher BMI was associated with DD, whereas Ammar et al. (6) found that waist-to-hip ratio was a stronger correlate than was BMI. Canepa et al. (28) recently observed that the correlation between central adiposity and DD was independent of general adiposity and was confined to women. However, cross-sectional studies do not provide an accurate picture of true age-related longitudinal changes.
We observed that weight gain was associated with significant increases in LV diastolic stiffness, despite reductions in BP, LV mass, and wall stress. These findings imply that diastolic LV stiffening does not result exclusively from hypertensive LV remodeling but rather is driven by weight gain even after accounting for afterload. These data extend on studies of morbid obesity (29,30) in which surgically induced weight loss was associated with improved DD. Adiposity may impair diastolic function by limiting energetic availability (29) or increasing in myocardial lipid content (31), which is correlated with increased LV diastolic stiffness (32,33). Although some studies have suggested that weight loss may increase mortality (34) or have no effect on CVD events (35), the current data support the hypothesis that weight loss might mitigate or prevent increases in ventricular stiffness with aging.
Weight change was associated with the changes in Ea and Ees, but weight change was no longer predictive of Ees change after adjustment for Ea. These findings suggests that the effects of weight change on Ees might be mediated by the effects of weight change on afterload, but this hypothesis cannot be proven by these data. Increased neck circumference and central obesity was associated with greater increases in Ees in women. Increases in Ees are typical of normal aging but are greater in HFpEF. Elevated Ees has important implications for cardiovascular homeostasis, leading to greater BP lability with load alteration, limited SV reserve during exercise, and increases in the metabolic cost to enhance CO during stress (17,36–38).
The correlation between central obesity and Ees progression was confined to women, suggesting an important sex difference in age-related ventricular stiffening. A similar finding was recently reported in a cross-sectional study in which WC was associated with DD in women but not in men (39). Intriguingly, Manrique et al. (40) observed that while insulin resistance develops in male and female mice exposed to the typical “Western” diet, only female animals developed LV DD, fibrosis, and oxidative stress. These findings, also typical of human HFpEF, provide insight into the greater predilection for women to develop HFpEF compared with men (41) and are consistent with a recently proposed pathogenetic model in HFpEF (7).
Neck circumference, a proxy for upper body fat, was a strong predictor of Ees progression. Increased neck circumference may lead to nocturnal airway obstruction, which has been associated with both insulin resistance (42) and DD (43). Neck circumference is more closely related to visceral adiposity and insulin resistance than is WC (23) and is more closely associated with metabolic disturbances than is subcutaneous fat (12). Insulin level and resistance were not independently associated with Ees after accounting for abdominal adiposity or neck-to-height ratio. It is plausible that increases in neck circumference reflect enhanced ectopic fat accumulation associated with organ-specific fat depots, which have effects primarily on adjacent anatomic organs directly via lipotoxicity, or indirectly via cytokine secretion (44).
The American College of Cardiology/American Heart Association guidelines emphasize the progressive nature of HF from risk (stage A), to asymptomatic structural disease (stage B), to symptoms (stages C and D) (45). Recent studies have reported that interventions guided by risk markers can reduce the risk for LV dysfunction and improve CVD outcomes (46). In this regard, central adiposity may be an important risk marker for stage B HFpEF, allowing for the identification of patients at highest risk for symptomatic (stage C) HFpEF who might benefit from preventive interventions.
Echo-Doppler data inherently have greater variability compared with invasive measurements. Nearly one-half (614 [44%]) of the subjects did not have adequate paired data for elastance measurement, and this group showed greater adiposity, introducing potential bias. In subjects with weight loss, we do not know whether this finding was intentional (“dieting”) or was associated with chronic diseases (cachexia). BMI reflects both fat content and muscle mass (47), and weight change does not necessarily reflect changes in fatness (48), so relationships between absolute fat content and LV stiffness could not be addressed. Measurements of percent body fat as well as visceral and subcutaneous fat might provide additional insight. The study cohort was almost exclusively white, and these results might not apply to other ethnic groups. We did not find any effect of changes in WC or waist-to-hip ratio on ventricular-arterial mechanics, and this finding may have been related to the greater imprecision of tape measurements compared with those of scale measurements. In mixed-effects models, we did not adjust for multiple comparisons, which may have led to false positive associations between obesity measures and parameters of ventricular-arterial coupling. However, we verified our finding by comparing the differences between quartiles of obesity measures. This study did not relate measures of adiposity to clinical outcomes, and future research is required to determine whether the correlations between adiposity and ventricular-arterial mechanics alter outcomes such as HF prevalence or survival.
In these men and women ≥45 years of age, weight loss positively influenced age-related changes in arterial elastance, whereas weight gain increased diastolic LV stiffness, independent of changes in arterial load. Increased neck circumference and central obesity were associated with amplification of the age-related increase in LV Ees in women but not in men. These findings offer further insight into the greater predilection for HFpEF in women and in obese patients, and they suggest that strategies to reduce body fat might be useful for slowing age-related progression of LV stiffness and for reducing the risk for HFpEF, particularly in women.
For a supplemental table, please see the online version of this article.
This research was funded by the National Heart, Lung and Blood Institute (NHLBI-RO1-55502), the European Regional Development Fund (project FNUSA-ICRC [Z.1.05/1.1.00/02.0123]), and the Czech Ministry of Health (NT13434-4/2012). The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- body mass index
- blood pressure
- cardiovascular disease
- diastolic dysfunction
- effective arterial elastance
- end-diastolic elastance
- end-systolic elastance
- ejection fraction
- end-systolic pressure
- heart failure
- heart failure with preserved ejection fraction
- left ventricular
- stroke volume
- waist circumference
- Received February 10, 2014.
- Revision received March 11, 2014.
- Accepted March 21, 2014.
- American College of Cardiology Foundation
- Murphy N.F.,
- MacIntyre K.,
- Stewart S.,
- Hart C.L.,
- Hole D.,
- McMurray J.J.
- Paulus W.J.,
- Tschope C.
- Borlaug B.A.,
- Redfield M.M.,
- Melenovsky V.,
- et al.
- Fox C.S.,
- Massaro J.M.,
- Hoffmann U.,
- et al.
- Redfield M.M.,
- Jacobsen S.J.,
- Borlaug B.A.,
- Rodeheffer R.J.,
- Kass D.A.
- Borlaug B.A.,
- Lam C.S.,
- Roger V.L.,
- Rodeheffer R.J.,
- Redfield M.M.
- Ommen S.R.,
- Nishimura R.A.,
- Appleton C.P.,
- et al.
- Chen C.H.,
- Fetics B.,
- Nevo E.,
- et al.
- Cornier M.A.,
- Despres J.P.,
- Davis N.,
- et al.
- Liang J.,
- Teng F.,
- Li Y.,
- et al.
- Recio-Rodriguez J.I.,
- Gomez-Marcos M.A.,
- Patino-Alonso M.C.,
- Agudo-Conde C.,
- Rodriguez-Sanchez E.,
- Garcia-Ortiz L.
- Russo C.,
- Jin Z.,
- Homma S.,
- et al.
- Ashrafian H.,
- le Roux C.W.,
- Darzi A.,
- Athanasiou T.
- Sarma S.,
- Carrick-Ranson G.,
- Fujimoto N.,
- et al.
- Rosito G.A.,
- Massaro J.M.,
- Hoffmann U.,
- et al.
- Kawaguchi M.,
- Hay I.,
- Fetics B.,
- Kass D.A.
- Chantler P.D.,
- Lakatta E.G.,
- Najjar S.S.
- Borlaug B.A.,
- Redfield M.M.
- Arias M.A.,
- Garcia-Rio F.,
- Alonso-Fernandez A.,
- Mediano O.,
- Martinez I.,
- Villamor J.
- Yancy C.W.,
- Jessup M.,
- Bozkurt B.,
- et al.