点击显示 收起
1 From the Divisions of Cardiology (JGL, DA, AV, MR, and TH) and Medical Genetics (TMK), University of Texas, Houston Medical School, Houston, TX
2 Supported by grants no. 5RO1 HL073162-02 from the National Heart, Lung, and Blood Institute, National Institutes of Health (to HT) and M01RR002558 from the General Clinical Research Center, University of Texas, Houston. 3 Address reprint requests to H Taegtmeyer, University of Texas, Houston Medical School, 6431 Fannin, MSB 1.220, Houston, TX 77030. E-mail: heinrich.taegtmeyer{at}uth.tmc.edu.
ABSTRACT
Background: Obesity is an important contributor to many cardiovascular risk factors and has been associated with abnormalities in cardiac contractile function. Causes of impaired contractile function are not fully understood and may include an oversupply of substrates.
Objective: We tested the hypothesis that metabolic dysregulation may adversely influence cardiac function. Specifically, we examined the effects of plasma free fatty acids and insulin sensitivity on left ventricular function in patients with clinically severe obesity.
Design: We measured metabolic and cardiac variables in 64 obese patients [body mass index (BMI; in kg/m2) > 35], including 2-D complete echocardiogram with M-mode and tissue Doppler imaging, anthropometric measurements, and analysis of blood chemistries.
Results: The median (25th and 75th percentile) age and BMI were 46 y (36, 53 y) and 51.5 (42.5, 56.5), respectively. The prevalence of diabetes, hypertension, and insulin resistance were 38%, 53%, and 90%, respectively. Plasma free fatty acid (FFA) concentrations were elevated in the cohort. No association was observed between insulin sensitivity or anthropometric measurements and left ventricular contractile function. However, FFA concentration was independently associated with diastolic function (r = –0.33, P = 0.01), and 40% of the cohort showed age-adjusted diastolic impairment as measured by tissue Doppler imaging.
Conclusion: The negative association between FFA and diastolic function, in the setting of insulin resistance, suggests that excess FFA may exert a lipotoxic effect on the heart.
Key Words: Obesity • free fatty acids • insulin resistance • diastolic function • echocardiography • lipotoxicity
INTRODUCTION
Obesity affects >30% of the adult population in the United States (1) and is associated with multiple comorbid conditions. Many obese persons exhibit the constellation of traits that make up the metabolic syndrome of insulin resistance, hypertension, and dyslipidemia. Although these traits are important risk factors for cardiovascular disease, obesity itself is an independent risk factor for heart failure (2).
Evidence suggests subclinical left ventricular (LV) systolic and diastolic dysfunction in healthy young obese persons when assessed by tissue Doppler imaging (TDI) echocardiography (3). The causes of cardiac contractile dysfunction in obesity are not entirely clear and may involve multiple mechanisms such as changes in hemodynamics (4), obesity-related inflammation (5), and local and systemic metabolic derangements (6, 7). Both human studies (8, 9) and animal models (10) suggest that cardiac dysfunction in obesity may be due to an imbalance between increased substrate uptake and decreased substrate oxidation. This imbalance may lead to alterations in contractile function by the production of reactive oxygen species (8), impaired calcium handling (11), and cellular toxicity from increases in lipid and glucose metabolites (12, 13).
The goal of this study was to evaluate the effects of systemic metabolic substrates on cardiac structure and function in a group of patients with clinically severe obesity. We measured fasting concentrations of glucose, insulin, and free fatty acids (FFAs), and we assessed LV function by TDI, an echocardiographic technique that is relatively load independent (14).
SUBJECTS AND METHODS
Patient selection
We offered participation to consecutive patients, of any race or ethnicity, from the Bariatric Surgery Center at the University of Texas, Houston, who met the candidacy requirements for bariatric surgery as outlined previously (15). In brief, criteria include body mass index (BMI; in kg/m2) > 40 (or 35 with significant obesity-related comorbidities); psychiatric stability; a history of multiple, failed, medically managed weight-loss attempts; and an absence of any genetic or reversible endocrinologic cause for obesity. Exclusion criteria were known coronary artery disease, ischemic cardiomyopathy, severe peripheral vascular disease, current smoking, pregnancy, and age < 18 y. Patients with a significant risk of coronary artery disease, as defined by their Framingham risk score or clinical symptoms, underwent either perfusion imaging or angiography to rule out the presence of coronary artery disease or ischemic cardiomyopathy.
All patients signed an informed consent before enrollment in the study. The study was approved by the Committee for the Protection of Human Subjects at the University of Texas, Houston.
Study protocol
A total of 64 patients were enrolled at the Clinical Research Center of the University of Texas, Houston. Four patients were excluded from the final analysis because their echocardiograms were of insufficient quality. Patients were subjected to an overnight fast and instructed to take their normal medications with water, if needed. Participants filled out questionnaires about medical, social, and family history, as well as quality-of-life and symptom-based questionnaires. On enrollment patients underwent a physical examination, and an electrocardiogram was performed and anthropometric measurements were obtained. Fasting blood samples were drawn and measured at our institution. Participants underwent an echocardiogram within 2 h of their blood draw. Insulin was measured by using a chemiluminescence assay (Immulite, Los Angeles, CA), and plasma FFAs were measured spectrophotometrically (Hitachi 912; Roche, Alameda, CA). Insulin sensitivity was assessed by using the homeostasis model of assessment 2 [HOMA2 (16)]. Insulin resistance was defined as insulin sensitivity < 100%, according to the HOMA2 computer model (17). Diabetes was determined by the patient’s medical history or by a fasting serum glucose concentration of 126 mg/dL, based on the criteria of the American Diabetes Association (18). Dyslipidemia was defined by the National Cholesterol Education Program criteria (19) or by a current regimen of medication for dyslipidemia. Blood pressure was measured at rest. The diagnosis of hypertension was based on the patient’s history or current treatment with antihypertensive agents. In all other patients, a blood pressure of >140/90 mm Hg on 3 separate resting measurements was used to define hypertension.
Echocardiography
Two-dimensional echocardiographic, M-mode, and cardiac Doppler echocardiograms were all performed with a commercially available system (Acuson Sequoia, Malvern, PA). Participants were studied in the left lateral decubitus position, and images were obtained by using standard parasternal and apical acoustic windows to record 10 beats. Myocardial contrast agents were used to improve endocardial resolution.
The echocardiographic measurements of LV internal dimension and interventricular septal and posterior wall thickness were performed according to recommendations of the American Society of Echocardiography (20). When LV M-mode measurements could not be optimally obtained, LV internal dimensions and wall thickness measurements were made by using the leading edge convention as described by the American Society of Echocardiography (21). Measurements from 3 consecutive cardiac cycles were averaged. End-diastolic LV dimensions were used to calculate LV mass by using a previously validated formula (22). End-diastolic and end-systolic LV volumes were calculated by the method of Teichholz et al (23).
The LV ejection fraction (LVEF) was calculated by using the following formula:
RESULTS
Patient characteristics
The baseline characteristics of the cohort are shown in Table 1, where data are presented as the median (25th and 75th percentile) to accurately represent the physiologic range of clinical values. The median age of the patient population was 44 y, and patients had markedly elevated BMI. All patients met the criteria for abdominal obesity as measured by waist circumference. Patients were taking different classes of medications to treat their comorbid conditions, including antihypertensive therapy, oral hypoglycemic medications, and lipid-lowering agents (Table 2). Most of the patients were women and white (Table 1). Patients with insufficient echocardiographic data differed from the rest of the cohort in terms of BMI, but no significant differences were observed for the other measured markers.
View this table:
TABLE 1. Baseline characteristics of subjects1
View this table:
TABLE 2. Medication profile of subjects1
Baseline blood chemistries
The fasting blood chemistries are listed in Table 3. Concentrations of plasma FFAs were greater than previously described normal ranges. Regression models for FFA concentrations showed no significant relations with medications, diabetes, duration of obesity, or age (Table 4). Although insulin and glucose concentrations were both elevated, it was insulin concentrations that explained the decreased insulin sensitivity (ß = –0.44, P < 0.0001). Insulin sensitivity was negatively associated with both weight (r = –0.26, P = 0.04) and waist circumference (r = –0.27, P = 0.03). Adiponectin concentrations were negatively correlated with insulin concentrations (r = –0.41, P = 0.01). Regression analysis did not show a significant interaction of oral hypoglycemic agents with adiponectin concentrations (ß = 0.14, P = 0.35). Measured serum high-sensitivity C-reactive protein (hsCRP) indicated that our cohort was at low risk of coronary artery disease (25), and hsCRP was positively associated with leptin concentrations (r = 0.29, P = 0.02) and BMI (r = 0.32, P = 0.01). Lipid-lowering agents did not have an effect on hsCRP concentrations (ß = 0.33, P = 0.30).
View this table:
TABLE 3. Fasting baseline chemistries1
View this table:
TABLE 4. Linear regression for interaction on plasma free fatty acid concentrations
Left ventricular structure
No correlations were observed between cardiac structure and either systolic or diastolic function. Relative wall thickness was negatively associated with insulin sensitivity (r = 0.49, P < 0.0001). We did not find any correlations between LV structure and anthropometric or other metabolic measurements.
Left ventricular function
Global LV systolic function, measured by LVEF, percentage of fractional shortening, and mitral annular systolic velocity, was preserved in the cohort (Table 5). With the use of previously reported age-adjusted values for Em (26), 40% of the cohort exhibited criteria of impaired diastolic myocardial velocities. The E:A, a classic measurement of diastolic function, was positively correlated with the TDI measures of diastolic function (r = 0.46, P < 0.0001).
View this table:
TABLE 5. Left ventricular structure and function1
Univariate analysis showed that only FFA concentrations and subject age were significantly associated with Em (Table 6). With the use of stepwise regression, FFA concentrations had an independent negative association with Em (P = 0.034; Table 7). Medication had no significant association with diastolic function (Table 8).
View this table:
TABLE 6. Pearson correlation coefficients for early diastolic myocardial velocity1
View this table:
TABLE 7. Stepwise linear regression for early diastolic myocardial velocity1
View this table:
TABLE 8. Regression analysis for the interaction of medications on diastolic function
DISCUSSION
The main finding of our study is an inverse relation between plasma FFA concentrations and LV diastolic function in a group of patients with clinically severe obesity. In addition, age-adjusted FFA concentrations independently predicted Em. We also found that decreasing insulin sensitivity was associated with measures of concentric hypertrophy, which is a poor prognostic factor for heart failure (27). We did not find significant associations between obesity-related inflammatory factors or adipokines and cardiac function.
Fatty acids are the predominant fuel for respiration in the postnatal heart (28). In heart muscle, as in skeletal muscle and liver, there is a balance among the uptake of fatty acids by the cell, transport of fatty acids into mitochondria, and their subsequent oxidation. Metabolic dysregulation, especially when the rate of fatty acid uptake exceeds the rate of fatty acid oxidation (6, 10, 29), leads to a wide range of metabolic disturbances, including fatty acyl-CoA accumulation, insulin resistance (30, 31), and triacylglycerol accumulation (32-34).
In transgenic mice, the overexpression of fat acid transport protein-1, known to import long-chain FFAs into mammalian cells, leads to diastolic dysfunction in a setting of metabolic stability (ie, no hypertension or diabetes) (35). Lipid accumulation is also a feature of the failing human heart (8) and is associated with a broad range of cellular and metabolic derangements that are collectively called lipotoxicity (32, 36). The negative association between FFAs and diastolic function in our cohort suggests that FFAs may exert a lipotoxic effect on the heart of patients with clinically severe obesity.
Although diabetes (and thus insulin resistance) is associated with increases in FFA uptake and utilization (37), diabetes did not show a significant effect on the association between FFAs and diastolic function. However, 90% of the cohort had evidence of insulin resistance, which suggests a dysregulation of glucose uptake and therefore impaired utilization leading to increased flux into the hexosamine pathway, which in turn increases substrate concentrations for O-GlcNAcylation (O-linked ß-N-acetylglucosamine enzymatic glycosylation). O-GlcNAcylation is associated with decreases in sacroplasmic reticulum calcium-APTase 2a (SERCA2a) promoter activity in diabetic cardiac myocytes (38). Recently, Hu et al (12) found improved calcium handling in diabetic hearts that were treated with an adenoviral O-GlcNAcase, which led to improved diastolic function. It is tempting to speculate that, as one possible mechanism for the association between FFAs and diastolic function, derangements in calcium homeostasis, in a state of insulin resistance and excess FFAs, lead to an impaired contractile response in diastole.
It is interesting that no associations were observed between the inflammatory factors tumor necrosis factor- and hsCRP or between the adipokines leptin and adiponectin and cardiac contractile function. All of these factors are associated with obesity and impaired insulin sensitivity (5, 39-41). Furthermore, regression analysis to examine the effect of medications on function and structure showed no significant interaction. Significant associations may be observed in more severe conditions of cardiac contractile dysfunction than the subclinical dysfunction of our cohort.
This study has several limitations. First, this was a descriptive study that involved a select group of obese patients, ie, those healthy enough to undergo bariatric surgery. Thus, the results may not be applicable in all persons with obesity. Second, the study population was heterogeneous with respect to the various comorbid states associated with the subjects’ obesity and the medications taken (eg, the presence of hypertension and the use of lipid-lowering and insulin-sensitizing agents). Although comorbid conditions and medications did not influence the associations between FFAs and diastolic function, the differences in the patient profiles could be the reason that other associations are not detected.
In conclusion, obesity is an important contributor to metabolic dysregulation and is a known risk factor for heart failure. In the current study, we found that patients with clinically severe obesity have elevated concentrations of FFAs in association with worsening diastolic function. The finding of metabolic dysregulation in the setting of subclinical cardiac dysfunction may eventually help define a mechanism for the heart failure that is associated with obesity.
ACKNOWLEDGMENTS
HT and DA contributed to the study concept and design. JGL, AV, and MR contributed to data collection. JGL, TMK, and DA contributed to the data analysis and interpretation of the results. JGL and HT contributed to the writing of the manuscript. All authors reviewed the final manuscript. None of the authors had a financial or personal conflict of interest.
REFERENCES