Alterations in Ventricular Structure and Function in Obese Adolescents with Nonalcoholic Fatty Liver Disease

Abstract

Objective

To determine the association among nonalcoholic fatty liver disease (NAFLD), metabolic function, and cardiac function in obese adolescents.

Study design

Intrahepatic triglyceride (IHTG) content (magnetic resonance spectroscopy), insulin sensitivity and β-cell function (5-hour oral glucose tolerance test with mathematical modeling), and left ventricular (LV) function (speckle tracking echocardiography) were determined in three groups of age, sex, and Tanner matched adolescents: 1) lean (n=14, BMI=20±2 kg/m²); 2) obese with normal (2.5%) IHTG content (n=15, BMI=35±3 kg/m²); and 3) obese with increased (8.7%) IHTG content (n=15, BMI=37±6 kg/m²).

Results

The disposition index (β-cell function) and insulin sensitivity index (ISI) were ~45% and ~70% lower, respectively, and whole body insulin resistance (HOMA-IR) was ~60% greater, in obese than in lean subjects, and ~30% and ~50% lower and ~150% greater, respectively, in obese subjects with NAFLD than those without NAFLD (P <0.05 for all). LV global longitudinal systolic strain and early diastolic strain rates were significantly decreased in obese than in lean subjects, and in obese subjects with NAFLD than those without NAFLD (P <0.05 for all), and were independently associated with HOMA-IR (β = 0.634). IHTG content was the only significant independent determinant of ISI (β = −0.770), disposition index (β = −0.651), and HOMA-IR (β = 0.738).

Conclusions

These findings demonstrate that the presence of NAFLD in otherwise asymptomatic obese adolescents is an early marker of cardiac dysfunction.

Nonalcoholic fatty liver disease (NAFLD) is a common co-morbidity of obesity. The marked increase in the prevalence rate of obesity in the last 30 years¹ has made NAFLD the most common liver abnormality in both children and adults. The prevalence rate of NAFLD in adults increases with increasing body mass index (BMI), from approximately 15% in non-obese persons, to 65% in persons with class I and II obesity (BMI 30.0–39.9 kg/m²), to 85% in extremely obese people (BMI ≥40 kg/m²).² The prevalence rate of NAFLD is also high among the obese pediatric population, and approximately one-third of obese children, aged 2–19 years old, have NAFLD.³

The presence of NAFLD also has important implications regarding cardiovascular health. The presence of NALFD is associated with metabolic abnormalities that are risk factors for coronary heart disease (CHD), such as multi-organ insulin resistance, dyslipidemia, impaired flow-mediated vasodilation and increased carotid-artery intimal medial thickness,^5–9 and is an important risk factor for CHD.^{10, 11} Data from studies conducted in adult subjects have found that NAFLD is associated with cardiac dysfunction, manifested as myocardial insulin resistance,¹² altered cardiac energy metabolism,¹³ abnormal left ventricular structure and impaired diastolic function.^{14, 15} The duration and severity of these abnormalities are likely important contributors to the increased risk of heart failure and cardiovascular mortality observed in obese people.^{16, 17} Because obesity in adolescence tracks into adulthood,¹⁸ a better understanding of the interrelationship among NAFLD, obesity and cardiac function in adolescents has important physiological and clinical long-term implications.

The purpose of the present cross-sectional study was to test the hypothesis that NAFLD in obese adolescents is associated with early evidence of myocardial structural and functional abnormalities, independent of adiposity and BMI.

METHODS

Three groups of adolescents were studied: (1) lean (n=14); (2) obese (BMI ≥95^th percentile for age and sex¹⁹) with normal intrahepatic triglyceride (IHTG) content (<5.6% liver volume as triglyceride;²⁰ n=15); and (3) obese with increased IHTG (NAFLD) (≥5.6% liver volume as triglyceride; n=15). A comprehensive assessment was made of: (1) body composition, including body fat mass, abdominal fat distribution, and ectopic fat (intrahepatic triglyceride [IHTG], intracardiac triglyceride [ICTG], and skeletal muscle intramuscular triglyceride [IMTG]) by using dual-energy x-ray absorptiometry (DXA), magnetic resonance imaging (MRI) and magnetic resonance spectroscopy (MRS); (2) myocardial structure and function by using integrated backscatter ultrasonography and speckle tracking echocardiography (STE); and (3) insulin sensitivity and β-cell function by using a 5-hour frequently sampled oral glucose tolerance test (FSOGTT).

Subjects were recruited from St. Louis Children’s Hospital outpatient Preventive Cardiology and Gastroenterology clinics. A total of 44 adolescents out of 57 screened subjects were selected based on inclusion and exclusion criteria. There are no specific criteria to define steatosis or NAFLD by using MRS in children and adolescents. Therefore, we used the cut-off value derived from the Dallas Heart Study conducted in adults.²¹ This value is consistent with the standard histological definition of steatosis (≥5% of hepatocytes containing triglyceride) which has been used to diagnose NAFLD in studies conducted in children and adolescents.³ The three groups were matched (one-to-one basis) on age and Tanner stage, and obese adolescents with NAFLD were matched as closely as possible on BMI Z-score with obese adolescents without NAFLD (Table I). All subjects underwent a comprehensive medical evaluation as part of standard of care, which included a detailed history, questionnaires to screen for obstructive sleep apnea,^{22, 23} physical examination, and routine blood tests ( comprehensive metabolic panel, lipid profile, thyroid panel, serology panel, complete blood count, and erythrocyte sedimentation rate). This evaluation helped rule out several other causes of fatty liver, including chronic viral hepatitis, autoimmune hepatitis, hypothyroidism, and genetic defects in lipoprotein metabolism. However, a liver biopsy was not performed in our study subjects so we cannot definitively exclude other potential causes of liver disease, and we did not evaluate surrogate diagnostic serum markers of liver fibrosis, oxidative stress, or apoptosis.²⁴ A 12-lead electrocardiogram and a pregnancy test in female subjects were performed. A 2-hour OGTT was performed in those who had a fasting blood glucose ≥100 mg/dl. Subjects who had diabetes, severe hypertriglyceridemia (>350 mg/dl), liver disease other than NAFLD, positive pregnancy test, poor echocardiographic acoustic window and suspected sleep apnea were excluded. None of the subjects consumed >10 g/day of alcohol, smoked tobacco products, took medications known to cause steatosis or affect metabolism, or participated in regular exercise (defined as >1 hour of structured moderate-to-vigorous physical activity per week for the previous 6 months).

Table 1.

Demographic characteristics and body composition of lean adolescents and obese adolescents with and without nonalcoholic fatty liver disease (NAFLD)

	Lean	Obese		P-values
		Normal IHTG	NAFLD	ANOVA	Obesity	NAFLD
Age (yr)	15 (14, 17)	15 (14, 17)	15 (13, 16)	0.937	0.719	1.000
Tanner stage	4 (4, 5)	5 (4, 5)	4 (4, 5)	0.919	0.683	1.000
Sex (Females/Males)	6/8	9/6	6/9	0.497*	–	–
Race (C/AA/A/NA)	12/2/0/0	8/7/0/0	9/4/1/1	0.255*	–	–
Body mass index (kg/m2)	19.9±1.6	34.5±2.9	37.4±5.9	<0.001	<0.001	0.102
Body mass index Z-score	−0.15±0.52	2.21±0.26	2.41±0.34	<0.001	<0.001	0.174
Body fat (% body weight)	19.5±7.2	39.7±6.1	40.7±5.9	<0.001	<0.001	0.674
Total abdominal fat (cm3)	1444±684	6387±1623	7613±2244	<0.001	<0.001	0.050
Intra-abdominal fat (cm3)	203 (165, 269)	680 (536, 785)	1119 (883, 1919)	<0.001	<0.001	<0.001
Subcutaneous abdominal fat (cm3)	1209±586	5141±1717	6259±1852	<0.001	<0.001	0.050
Intrahepatic triglyceride (%)	1.0 (0.4, 1.9)	2.5 (2.0, 4.0)	8.7 (7.5, 14.5)	<0.001	<0.001	<0.001
Intracardiac triglyceride (%)	0.5 (0.3, 1.1)	0.6 (0.4, 0.8)	0.7 (0.5, 1.5)	0.353	0.303	0.321
Intramuscular triglyceride (%)	0.6 (0.3, 1.0)	1.1 (0.7, 1.3)	1.0 (0.6, 2.3)	0.040	0.013	0.632

Values are mean±SD for normally distributed variables and median (quartiles) for non-normally distributed variables.

P-values are shown for the one-way ANOVA model and the two contrasts tested: Obesity (lean vs. all obese adolescents) and NAFLD (obese adolescents with NAFLD vs. obese adolescents without NAFLD).

^*Sex and race distributions were evaluated by using the χ² test.

Abbreviations: IHTG, intrahepatic triglyceride; C, Caucasian; AA, African American; A, Asian; NA, Native American.

The study was approved by the Human Research Protection Office of Washington University School of Medicine in St. Louis, MO. All subjects agreed to participate in the study after a detailed explanation of the study was provided to them and their parents. Written informed consent was obtained from each subject’s parent(s) and written informed assent was obtained from each subject before being enrollment.

Body composition analyses, cardiac structure and function studies and assessment of metabolic function were performed in the Center for Clinical Imaging Research, the Echocardiography Laboratory, and the Clinical Research Unit of Washington University School of Medicine on separate days, within a 7-day period. Subjects were studied as outpatients after they fasted for 12 hours overnight before each visit.

Body composition analyses

Body fat mass and fat-free mass were determined by using DXA (Hologic QDR 4500, Waltham, MA). Subcutaneous abdominal and intra-abdominal fat volumes were determined by using MRI with a 1.5 T scanner (Siemens, Iselin, NJ). Eight 10-mm thick axial images were obtained beginning at the L₄-L₅ interspace, and analyzed for subcutaneous and intra-abdominal fat content by using Analyze 6.0 software (Mayo Foundation, Biomedical Imaging Resource, Rochester, MN); the volume of fat was calculated for each slice and the values were added. The amount of triglyceride in the liver (IHTG), soleus muscle (IMTG) and the heart (ICTG) was determined by using MRS (1.5 Tesla MAGNETOM Sonata whole body system, Siemens Medical Systems, Erlangen, Germany), as described previously.^{25, 26}

Cardiovascular structure and function

Blood pressure and heart rate

Systolic and diastolic blood pressures were determined by using oscillometry with appropriate cuff size (two-third relative to arm circumference and length), and heart rate was determined by using auscultation. Three measurements were obtained after 5 min of rest in the sitting position and the median was calculated.

Left ventricle (LV) volume and mass and relative posterior wall thickness

A transthoracic complete 2D and Doppler echocardiographic examination (GE Vivid 7™ ultrasonic imaging system with a M3S phased array transducer, General Electric Medical Systems, Waukesha, WI, USA) was used to determine LV end-diastolic volume, end-systolic volume, and mass according to the recommendations of the American Society of Echocardiography.²⁷ Left ventricular mass was indexed to height in the power of 2.7.²⁸ The LV end-diastolic and end-systolic volumes were measured by using the modified Simpson method, and LV ejection fraction was calculated as (LV end-diastolic volume minus LV end-systolic volume) / LV end-diastolic volume.²⁹ Relative wall thickness (RWT) of the LV was calculated as the ratio of 2 × posterior wall thickness at end-diastole / LV end-diastolic diameter.

Global myocardial strain and strain rate

Global ventricular function was evaluated by measuring LV strain and strain rate. Strain is a dimensionless measure of myocardial deformation, which is usually reported as percentage change with negative connotation, with more negative values indicating greater strain. Strain rate, a time derivative of strain, correlates linearly with LV peak elastance, which is a load-independent global measure of ventricular systolic function.³⁰ These quantitative measures may precede and are more robust and sensitive than conventional load-dependent indices of cardiac functional changes such as ejection fraction.³¹ Longitudinal strain in pediatric age is relatively independent of maturational changes and normal values in pediatric subjects have been established, taking into consideration the effects of growth, maturation, and body size.³² Myocardial strain in two dimensions is measured by a novel echocardiographic method that determines myocardial deformation from continuous frame-by-frame tracking of speckles.^{31, 33, 34} Speckle tracking echocardiography (STE) is angle-independent, can track in two and three dimensions, and therefore, can assess the magnitude and timing of regional and global ventricular deformation in different directions. Global LV longitudinal, circumferential and radial strain and strain rate were obtained by using two-dimensional STE and quantified by using the EchoPac™ software package (General Electric Medical Systems, Waukesha, WI).^{31, 35}

Myocardial tissue characterization

Cyclic variation of backscatter was used to assess the structural characteristics of myocardial tissue. The principle for using this procedure is based on the known quantitative relationships between ultrasonic backscatter signals and the inherent properties of myocardial tissue.^36–39 Data from studies have illustrated regional dependence of backscattered ultrasound on intrinsic myofiber organization⁴⁰ and increases in both the measured ultrasonic attenuation and backscatter correlate well with increased collagen concentration.^{36, 41} Cyclic variation of backscatter data was determined in each subject by measuring the mean backscatter value within a region of interest in the posterior wall of the LV in the parasternal long axis view for each image frame recorded over the acquired heart cycles. Cyclic variation of backscatter data was characterized by its magnitude and normalized time delay relative to the systolic interval.^{40, 42}

We have defined myocardial dysfunction as an alteration in LV structure or function that is different from values observed in healthy lean subjects.

Metabolic function

Beta-cell function and insulin sensitivity were evaluated by using a 5-h FSOGTT. An intravenous catheter was placed into an antecubital vein, which was heated by using a thermostatically controlled box, to obtain arterialized venous samples. At time 0, participants ingested 75 g of glucose. Blood samples were collected 10 min before and immediately before (t=0) ingesting the glucose load and then at 10, 20, 30, 60, 90, 120, 150, 180, 240, and 300 min after glucose ingestion to determine plasma glucose, insulin, and C-peptide concentrations. Plasma glucose was measured immediately by using an automated glucose analyzer (Yellow Springs Instruments, Yellow Springs, OH). Plasma insulin and C-peptide were measured by using a chemiluminescent immunoassay (Immulite 1000, Diagnostic Products Corporation, Los Angeles, CA).

Calculations

The total areas under the curve (AUC) for glucose and insulin were calculated by using the trapezoid rule. Whole-body insulin resistance was calculated by using the homeostasis model assessment of insulin resistance (HOMA-IR).⁴³ The insulin sensitivity index (ISI) was estimated from plasma glucose and insulin concentrations obtained during the FSOGTT, by using the oral glucose minimal model.⁴⁴

Βeta-cell function was assessed from plasma glucose and C-peptide concentrations obtained during the FSOGTT, by using the oral minimal model of C-peptide secretion and kinetics⁴⁵ and incorporating published parameters for C-peptide kinetics and volume of distribution.⁴⁶ This model provides a global index of postprandial insulin secretion in response to ingested glucose (overall responsivity, Φ_o). The disposition index (DI) was calculated by multiplying the index of global β-cell responsivity (Φ_o) by insulin sensitivity (ISI) to determine insulin secretion in relationship to the degree of insulin resistance.⁴⁷

Statistical Analyses

All datasets were tested for normality according to the Shapiro-Wilks criteria, and non-normally distributed variables were ranked for analysis. Differences among groups in anthropometric characteristics, body composition, metabolic variables, hemodynamic variables, cardiac structure (LV mass, RWT, and measures of the cyclic variation of backscattered energy) and function (EF, global strain and strain rate) were evaluated by using one-way ANOVA, with two planned contrasts for the comparisons of interest: obesity (lean vs. all obese adolescents) and NAFLD (obese adolescents with NAFLD vs. obese adolescents without NAFLD). Differences among groups in frequencies (sex and race) were evaluated by using the χ² test. Pearson correlation and stepwise multiple linear regression analyses were performed to assess the relationships between variables of interest. A P-value <0.05 was considered significant. All statistical analyses were conducted with SPSS version 19 (IBM SPSS, Chicago, IL).

RESULTS

Body composition

Obese subjects with and without NAFLD were matched on BMI, BMI Z-score and percent body fat (Table I). Intra-abdominal fat volume and IHTG content were greater in obese subjects with NAFLD than those without NAFLD. Although IHTG content was greater in obese subjects with NAFLD than those without NAFLD, IMTG and ICTG were not different between the two groups (Table I). Values for IHTG content correlated positively with percent body fat (r = 0.570, P <0.001) and intra-abdominal fat volume (r = 0.743, P <0.001), and less strongly with IMTG (r = 0.345, P = 0.027) and ICTG (r = 0.363, P = 0.020).

Metabolic function

Fasting plasma glucose concentration and glucose AUC during the FSOGTT were not significantly different among groups, but fasting plasma insulin and C-peptide concentrations and insulin AUC were significantly greater in obese than lean adolescents, and in obese adolescents with NAFLD than those without NAFLD (Table II). The β-cell response to the oral glucose load (Φ_o) was greater, but insulin secretion adjusted for insulin sensitivity (disposition index) was lower, in obese than in lean subjects, and in obese subjects with NAFLD than those without NAFLD (Table II). Whole-body insulin resistance (HOMA-IR) was greater, and insulin sensitivity (ISI) was lower, in obese than in lean subjects, and in obese subjects with NAFLD than those without NAFLD (Table II). Although plasma alanine aminotransferase concentrations were greater in obese than in lean adolescents and tended to be greater in obese adolescents with NAFLD than those without NAFLD, mean values were within the normal range in all groups (Table II).

Table 2.

Metabolic variables in lean adolescents and obese adolescents with and without nonalcoholic fatty liver disease (NAFLD)

	Lean	Obese		P-values
		Normal IHTG	NAFLD	ANOVA	Obesity	NAFLD
Basal plasma concentrations
Glucose (mg/dl)	89 (83, 93)	89 (87, 96)	91 (86, 102)	0.283	0.197	0.367
Insulin (mU/L)	8 (5, 12)	15 (12, 22)	36 (23, 49)	<0.001	<0.001	0.001
C-peptide (pmol/L)	500 (440, 606)	699 (559, 917)	1032 (879, 1352)	<0.001	<0.001	<0.001
Areas-under-the-curve
Glucose (mg/dl·min)	32673 (30109, 34474)	31269 (29618, 33466)	32972 (31255, 35816)	0.245	0.932	0.096
Insulin (mU/L·min)	13038 (10195, 16597)	16537 (12750, 22844)	43598 (25382, 59409)	<0.001	<0.001	<0.001
Insulin secretion indices
Global beta-cell responsivity, Φo	13.1±3.4	15.4±5.6	24.4±7.6	<0.001	0.002	<0.001
Disposition index	0.0064±0.0049	0.0041±0.0028	0.0028±0.0017	<0.001	<0.001	0.034
Insulin sensitivity indices
Insulin sensitivity index (ISI)	0.0006±0.0005	0.0003±0.0002	0.0001±0.0001	<0.001	<0.001	<0.001
HOMA-IR score	1.9 (1.2, 2.6)	3.1 (2.6, 4.8)	7.8 (5.5, 10.8)	<0.001	<0.001	0.001
Plasma lipids
Triglyceride (mg/dl)	64±37	90±53	122±70	0.027	0.024	0.129
HDL-cholesterol (mg/dl)	45±9	43±13	36±6	0.057	0.136	0.077
Total cholesterol (mg/dl)	127±25	136±25	165±42	0.006	0.025	0.017
LDL-cholesterol (mg/dl)	69±20	75±20	105±34	0.001	0.017	0.003
Hepatic enzymes
Alanine aminotransferase (U/L)	15 (12, 18)	17 (13, 17)	24 (17, 39)	0.01	0.01	0.06
Aspartate aminotransferase (U/L)	22 (19, 27)	20 (19, 24)	23 (21, 26)	0.30	0.96	0.12

Values are mean±SD for normally distributed variables and median (quartiles) for non-normally distributed variables. P-values are shown for the one-way ANOVA model and the two contrasts tested: Obesity (lean vs. all obese adolescents) and NAFLD (obese adolescents with NAFLD vs. obese adolescents without NAFLD).

Abbreviations: IHTG, intrahepatic triglyceride; HOMA-IR, homeostasis model assessment of insulin resistance; HDL, high-density lipoprotein; LDL, low-density lipoprotein.

β-cell responsivity (Φ_o) was positively correlated and the disposition index was negatively correlated with IHTG content (Figure 1, A and B; available at www.jpeds.com), percent body fat (r = 0.515 and r = −0.440, respectively; P <0.005), and intra-abdominal adipose tissue volume (r = 0.632 and r = −0.502, respectively; P ≤0.001). The ISI was negatively correlated and HOMA-IR score was positively correlated with IHTG content (Figure 1, C and D), percent body fat (r = −0.629 and r = 0.613, respectively; P <0.001), and intra-abdominal adipose tissue volume (r = −0.659 and r = 0.651, respectively; P <0.001). Values for ICTG did not correlate with any indices of β-cell function or insulin sensitivity. In multiple stepwise regression analysis with age, Tanner stage, percent body fat, and intra-abdominal adipose tissue volume as predictive variables, IHTG content was the only significant independent determinant of ISI (β = −0.770), HOMA-IR (β = 0.738), overall β-cell responsitivity Φ_o (β = 0.682) and disposition index (β = −0.651) (all P <0.001). However, body composition variables are often highly correlated with each other, which makes it difficult to reliably separate one candidate from another.

Figure 1.

(online). Relationship between intrahepatic triglyceride (IHTG) content and indices of β-cell function and insulin sensitivity in lean adolescents and obese adolescents with and without nonalcoholic fatty liver disease (NAFLD). A) Global β-cell responsivity to oral glucose (Φ_o), B) insulin secretion adjusted for insulin sensitivity (disposition index, DI), C) insulin sensitivity index (ISI), and D) homeostasis model assessment of insulin resistance (HOMA-IR). Data are ranked. All relationships are statistically significant (P ≤ 0.001).

Cardiovascular structure and function

Systolic and diastolic blood pressures were greater in obese than lean subjects and in obese subjects with NAFLD than those without NAFLD (Table III). The LV mass and LV mass index were greater in obese than in lean subjects, but were still within the normal range⁴⁸ and were not different between obese subjects with and without NAFLD. Relative posterior wall thickness was not significantly different among groups and was normal for age (<0.41^{48, 49}) in most subjects; two obese adolescents with NAFLD, who had RWT values of 0.43 and 0.44 (Table III). No significant differences among groups were detected in the magnitude of cyclic variation of backscatter, but there was a trend of increasing normalized time delay of cyclic variation of backscatter in obese compared with lean subjects (Table III).

Table 3.

Myocardial structure and function and hemodynamics in lean adolescents and obese adolescents with and without nonalcoholic fatty liver disease (NAFLD)

	Lean	Obese		P-values
		Normal IHTG	NAFLD	ANOVA	Obesity	NAFLD
Systolic blood pressure (mmHg)	109±11	120±10	127±7	<0.001	<0.001	0.046
Diastolic blood pressure (mmHg)	68±5	74±9	79±6	<0.001	<0.001	0.030
Heart rate (bpm)	78±9	82±5	83±4	0.084	0.097	0.646
RWT	0.29±0.06	0.33±0.04	0.35±0.04	0.070	0.070	0.120
LV mass (g)	96±23	115±28	132±28	0.004	0.003	0.102
LV mass index (g/m2.7)	24±4	29±5	31±7	0.002	0.001	0.370
LV ejection fraction (%)	58.7±4.2	57.4±4.0	59.1±3.9	0.480	0.727	0.248
LV longitudinal global strain (%)	−18.2±2.6	−16.6±2.2	−14±3.8	0.002	0.005	0.030
LV longitudinal global strain rate in systole (%/s)	−0.97 (−1.08, −0.95)	−0.90 (−0.97, −0.80)	−0.87 (−1.00, −0.65)	0.013	0.001	0.564
LV longitudinal global strain rate in early diastole (%/s)	1.58±0.32	1.37±0.22	1.14±0.32	0.001	0.001	0.037
LV longitudinal global strain rate in late diastole (%/s)	0.55 (0.44, 0.81)	0.50 (0.37, 0.60)	0.50 (0.43, 0.57)	0.455	0.338	0.419
Magnitude of cyclic variation (dB)	4.3±1.7	4.5±1.5	3.7±2	0.460	0.704	0.242
Normalized time delay of cyclic variation	0.8 (0.5, 1.03)	0.92 (0.64, 1.95)	1.0 (0.62, 2.31)	0.247	0.097	0.252

Values are mean±SD for normally distributed variables and median (quartiles) for non-normally distributed variables.

P-values are shown for the one-way ANOVA model and the two contrasts tested: Obesity (lean vs. all obese adolescents) and NAFLD (obese adolescents with NAFLD vs. obese adolescents without NAFLD).

Abbreviations: IHTG, intrahepatic triglyceride; RWT, relative posterior wall thickness; LV, left ventricular.

The LV global longitudinal strain was significantly decreased in obese than in lean subjects and in obese subjects with NAFLD than those without NAFLD (Figure 2 and Table III), indicating greater systolic dysfunction in the NAFLD group. Compared with lean adolescents, the obese groups had decreased systolic and early diastolic longitudinal strain rates. Early diastolic longitudinal strain rates were significantly decreased in obese subjects with NAFLD than those without NAFLD (Figure 2 and Table III), indicating greater diastolic dysfunction in the NAFLD group. The LV ejection fraction was normal and similar among the three groups.

Figure 2.

Cardiac function in lean adolescents and obese adolescents with and without nonalcoholic fatty liver disease (NAFLD). A) Left ventricular (LV) global strain, B) LV longitudinal global strain rate in systole, and C) LV longitudinal global strain rate in early diastole. Values are means±SD (A, C) or medians with quartiles (B). * Value significantly different from value in obese subjects (P < 0.05), ^† value significantly different from value in obese subjects without NAFLD (P < 0.05).

The LV global longitudinal strain and systolic and early diastolic strain rates correlated positively with IHTG content (r = 0.25–0.40, P <0.05), percent body fat (r = 0.30–0.40, P <0.05), and intra-abdominal adipose tissue volume (r = 0.40–0.60, P <0.05), but not ICTG (all P >0.55). In multiple stepwise regression analysis with age, Tanner stage, BMI, blood pressure, percent body fat, intra-abdominal adipose tissue volume, total cholesterol and LDL-cholesterol, IHTG content, and insulin sensitivity and β-cell function indices as predictive variables, the HOMA-IR score was the only independent variable associated with global longitudinal strain (β = 0.634, P <0.001), global longitudinal systolic strain rate (β = 0.487, P = 0.001), and longitudinal early diastolic strain rate (β = 0.527, P = 0.001). IHTG content did not significantly contribute further to explaining the variance in these measures of myocardial function.

DISCUSSION

The presence of increased IHTG content is an important marker of metabolic dysfunction (multi-organ insulin resistance and dyslipidemia).^{5, 6} Our data demonstrate that obese adolescents with NAFLD have greater abnormalities in cardiac function, manifested by decreased systolic and diastolic myocardial strain and strain rate than obese adolescents who have normal IHTG content. The cardiac functional abnormalities in our obese adolescents with NAFLD were independent of traditional cardiac risk factors (i.e. high systolic and diastolic blood pressures, total and LDL cholesterol, and BMI), and correlated with insulin resistance. Furthermore, insulin sensitivity was decreased in obese adolescents with NAFLD compared with those without NAFLD, as shown previously in obese adults.^{7, 8} In addition, increased IHTG content was associated with increased insulin secretion in response to ingested glucose, which has also been shown previously in adults with NAFLD.^50–52 However, the disposition index, which provides a more comprehensive assessment of β-cell function (i.e. insulin response to glucose in relationship to the degree of insulin resistance) was lower in obese adolescents with NAFLD than in those without NAFLD. Therefore, our findings demonstrate that the presence of NAFLD in obese adolescents is an early marker of myocardial dysfunction, and support the notion that NAFLD is an integral component of the cardiometabolic abnormalities that increase the risk of future diabetes and congestive heart failure in obese children.

The precise mechanism responsible for the link between IHTG content and myocardial dysfunction cannot be determined from our study. Although all our obese subjects were clinically asymptomatic and had a normal ejection fraction, they demonstrated evidence of impaired systolic and diastolic function manifested by decreased LV systolic strain, strain rate, and early diastolic rate. These alterations in myocardial function were worse in obese adolescents with NAFLD than in those with normal IHTG content. The abnormalities in cardiac function observed in our obese subjects likely involve myocardial fiber organization that affects ventricular intrinsic deformation, which we measured with the relatively load-independent parameters of strain and stain rate. These results indicate that conventional measures of cardiac function, which are load dependent, do not provide a sensitive assessment of early stages of cardiac dysfunction. Data from another study in obese children and adolescents, employing strain and strain rate measurements by 3D-wall motion tracking echocardiography, also demonstrated abnormal myocardial deformation changes.⁵³ Increased blood pressure can affect myocardial deformation,⁵⁴ and our obese subjects had higher systolic and diastolic blood pressure values than our lean subjects. Myocardial deformational changes caused by high blood pressure vary according to the LV adaptive response, manifested by abnormalities in LV mass and RWT. The adaptive response with normal geometry has normal deformation and other adaptive responses have abnormal deformation. It is thus unlikely that differences in blood pressure contributed to differences in myocardial function among our study groups, because all our subjects had LV mass below the 95th percentile and all but two obese subjects with NAFLD had RWT values that were within the normal range,^{48, 49} indicating normal LV geometry (normal RWT and LV mass) with no signs of concentric remodeling), concentric hypertrophy (increased RWT, increased LV mass), or eccentric hypertrophy (decreased RWT, increased LV mass).^55–57 In addition, regression analysis found the impairment in ventricular function was independent of blood pressure.

In contrast with data reported from studies conducted in adult subjects,^{20, 58} we found that impaired cardiac function was not accompanied by an increase in ICTG content. Although we cannot exclude a type II error, this result suggests that asymptomatic alterations in cardiac function in our obese adolescents represent an early stage in the cascade of events, presumably induced by alterations in myocardial energy metabolism, which preceded ICTG accumulation. However, cardiac dysfunction was associated with insulin resistance, suggesting that an insulin resistance-induced increase in myocardial fatty acid uptake⁵⁹ causes both an impairment in cardiac function and an accumulation of ICTG. Given the tendency for obesity to track from adolescence into adulthood,¹⁸ persistence of these metabolic disturbances can lead to progressive structural remodeling of the heart as suggested by increasing normalized time delay of cyclic variation of backscatter and obesity-associated cardiomyopathy.⁶⁰

The presence of NAFLD in obese adolescents is associated with cardiac ventricular dysfunction, independent of BMI, conventional cardiovascular risk factors, and ICTG content. In addition, IHTG content itself is associated with β-cell dysfunction and insulin resistance which are likely involved in the pathogenesis of obesity-related myocardial dysfunction. These results demonstrate that NAFLD should be considered part of the constellation of metabolic and myocardial abnormalities in obese adolescents that increase their risk for developing type 2 diabetes and heart failure. Determining the specific criteria needed to optimize the stratification of cardiovascular risk in obese adolescents requires more research, including a careful assessment of NAFLD severity.

Abbreviations and Acronyms

AUC

area under the curve

BMI

body mass index

CHD

coronary heart disease

FSOGTT

frequently sampled oral glucose tolerance test

HOMA-IR

homeostasis model assessment

ICTG

intracardiac triglyceride

IHTG

intrahepatic triglyceride

IMTG

skeletal muscle intramuscular triglyceride

ISI

insulin sensitivity index from minimal modeling

LV

left ventricle

MRI

magnetic resonance imaging

NAFLD

nonalcoholic fatty liver disease

STE

speckle tracking echocardiography