Associations of Visceral,Subcutaneous,Epicardial,and Liver Fat with Metabolic Disorders up to 14 Years After Weight Loss Surgery

Abstract

Background:

Bariatric surgery leads to long-term remission and reduced incidence of diabetes, hypertension, and dyslipidemia. Short-term studies suggest reduction in specific fat depots may be more predictive of health improvement than reduced body mass index (BMI). Visceral, subcutaneous, epicardial, and liver fat, measured 11 years after bariatric surgery, were associated with long-term remission and incidence of diabetes, dyslipidemia, and hypertension.

Methods:

Fat depots an average of 11 (maximum 14) years after surgery were quantified by noncontrast computed tomography in subjects who did (N = 261; 86% gastric bypass) or did not (N = 243) have bariatric surgery. Multiple regression related fat depots to disease endpoints with and without adjustment for change in BMI and surgical status.

Results:

Visceral fat was 42% lower, subcutaneous fat 20% lower, epicardial fat 30% lower, and liver-to-spleen density ratio 9% higher at follow-up in the bariatric surgery group compared with the nonsurgery group (all P < 0.01). Higher visceral fat at follow-up exam was significantly associated with reduced remission and increased incidence of diabetes, hypertension, and dyslipidemia. Subcutaneous fat was not associated with disease. The liver-to-spleen ratio was associated with the remission and incidence of hypertriglyceridemia and not with other fat depots. Epicardial fat was related to incidence of elevated low-density lipoprotein cholesterol and low high-density lipoprotein cholesterol.

Conclusions:

Whether or not a patient shows greater long-term diabetes, dyslipidemia, or hypertension remission or incidence after bariatric surgery appears dependent on the amount of fat within specific fat depots measured at follow-up. Furthermore, associations of the three disease endpoints with different fat depots suggest varied fat depot pathology.

Introduction

Obesity is strongly associated with increased risks for premature mortality, driven in large part by cardiovascular disease.^1,2 Obesity, particularly visceral obesity, is also related to increased incidence of prediabetes and type 2 diabetes.³ Even in persons with normal body mass index (BMI), visceral obesity remains an important risk factor associated with decreased survival.⁴ Metabolic syndrome is clearly related to overall obesity, but visceral obesity and fatty liver disease are particularly important features of this syndrome, each contributing to systemic inflammation.^5,6 Furthermore, while the liver plays a key role in both glucose and lipid metabolism, nonalcoholic fatty liver disease (NAFLD) is bidirectionally related to the metabolic syndrome.^7,8 In other words, metabolic syndrome promotes NAFLD and vice versa.

However, the metabolic consequences of fat deposition appear to differ depending on depot location. For example, liver fat may have greater effects on glucose metabolism than visceral fat.⁹ Subcutaneous fat may be protective against metabolic derangements, but this may depend on its location.¹⁰ Whether epicardial fat reflects visceral fat accumulation or has independent effects on the tissue or vasculature of the heart remains unclear.^11,12 Changes in different depots of adipose tissue and liver fat after bariatric surgery, and their relationships to changes in the components of the metabolic syndrome, are of substantial interest as they may give insights into potential mechanisms of benefit following bariatric surgery.

Long-term success in treatment of severe obesity is most reliably obtained using bariatric surgical procedures.^13

–17 The use of such procedures has been associated with improvements in survival and substantial reductions in the incidence of new heart failure and myocardial infarction.^18,19 Reductions in stroke are less evident.¹⁹ After gastric bypass surgery, clinical and biochemical variables related to cardiovascular disease are significantly improved, and persist until at least 12 years after the surgery.¹⁶ Importantly, there is an increased rate of remission and reduced incidence of diabetes, hypertension, and dyslipidemia over the 12-year period after gastric bypass surgery.¹⁶ Most of these beneficial changes were strongly related to the amount of weight lost after gastric bypass surgery. However, the amounts of fat tissue in different adipose beds and the relationships of postsurgical regional fat masses to clinical outcomes are uncertain due to the small numbers of bariatric patients enrolled (most studies with <50 patients) and the limited follow-up time after surgery (<2 years).^{20

–29}

Results from these limited studies showed that visceral adipose tissue and subcutaneous adipose tissue both significantly decreased, but that visceral adipose tissue had a larger decrease than subcutaneous adipose tissue. The larger study of 126 patients who had gastric bypass surgery showed that most metabolic variables measured were associated with changes in visceral adipose tissue, but only high-density lipoprotein cholesterol (HDL-C) was associated with subcutaneous adipose tissue,²² suggesting different physiological features of the two fat depots. None of these studies looked at liver or epicardial fat depots or whether the visceral adipose tissue and subcutaneous adipose tissue changes were maintained long term.

This study identifies which fat depots, measured an average of 11 years after a baseline exam or subsequent bariatric surgery, have the strongest associations with the observed improvement in disease risk. This study also addresses whether or not the specific fat depots are related to a significant amount of disease remission observed an average of 11 years after surgery and whether these associations remain even after adjusting for changes in BMI.

Methods

Study subjects

This study represents an extension of a previously reported long-term prospective study that involved 1156 subjects examined at baseline, with follow-up exams on average of 2, 6, and 12 years.^16,30 In this prospective study, the 1156 subjects included 418 who had gastric bypass surgery within a year after the baseline exam and 746 subjects with severe obesity (BMI ≥35 kg/m²) who did not have surgery within the first year of the baseline exam. All participants had the same baseline and follow-up examinations and all surgeries were performed by three surgeons at Rocky Mountain Associated Physicians. At the 12-year examination, 95% of the entire cohort still alive (1023/1073) had in-person detailed exams (n = 722) or medical record abstraction (n = 301).¹⁶ At the 12-year examination, participants age 45 years and older were for the first time offered a computed tomography (CT) scan, of which 504 participants underwent CT. Data obtained from the 504 subjects with a CT scan were used for the current study. All participants who underwent CT examination were classified into two groups, those who had undergone bariatric surgery since baseline exam (ever had bariatric surgery group; n = 261) compared with those who had not had weight loss surgery (never had bariatric surgery group; n = 243). For the ever had bariatric surgery group, 178 had gastric bypass surgery within the first year following baseline (follow-up average of 11 years, maximum 14 years; the same 3 surgical partners performed surgeries) and 83 had bariatric surgery at least 1 or more years after baseline exam [gastric bypass (n = 46), duodenal switch (n = 4), sleeve gastrectomy (n = 1), gastric band (n = 19), and procedure not reported (n = 13); follow-up average of 7 years, maximum 11 years since surgery; by multiple surgeons]. For the never had bariatric surgery comparison group, follow-up averaged 11 years with maximum 14 years. All subjects provided written informed consent and the study was approved by an Institutional Review Board.

Clinic procedures

At each clinical exam, height (Harpenden anthropometer; Holtain, Crymych, United Kingdom), weight (Scaletronix, Wheaton, IL), waist circumference at the umbilicus (metal Lufkin tape), blood pressures (Dinamap; GE Healthcare, Tampa, FL), plasma lipids, glucose, and glycated hemoglobin were measured as previously described.^16,30 Diabetes was defined as fasting glucose ≥126 mg/dL, glycated hemoglobin ≥6.5%, or taking antidiabetic medications. Hypertension was defined as systolic blood pressure ≥140 mmHg, diastolic blood pressure ≥90 mmHg, or taking antihypertensive medications. The three dyslipidemia endpoints were defined as plasma low-density lipoprotein cholesterol (LDL-C) ≥160 mg/dL, plasma triglyceride ≥200 mg/dL, plasma HDL-C <40 mg/dL, or on lipid medication. Remission or incidence of each disease was determined at the last follow-up exam.

Clinical CT scan data were only obtained at the 12-year follow-up exam at the University of Utah medical center. Noncontrast CT scans were performed using a Siemens Definition scanner. Quantification of the CT scan data was obtained for all subjects in the study in a standardized manner by one trained technician and reviewed by a cardiologist. Prospective, axial electrocardiographic-gated images of the chest using 3 mm slices were obtained from the bifurcation of the carina to the bottom of the heart. The top of the liver and spleen were visible in all scans. An additional 3 mm slice was obtained for measurement of visceral fat at the level of L4–5 (Fig. 1). Using commercially available software, epicardial fat was quantified by manually tracing the border of the visceral pericardium in each slice from the bottom to the top of the heart. Using computer-assisted identification, all pixels of fat density (between −45 and −195 Hounsfield units) between the epicardial border and the pericardium were identified and counted (Fig. 2).³¹ The number of pixels with fat attenuation was multiplied by the slice thickness to obtain the volume of fat in that slice. The volume from each of the slices was summed to yield total epicardial fat volume.

FIG. 1.

Examples of abdominal slices from noncontrast CT scans of representative subjects. On the left are images from a nonsurgery subject who has large amounts of both subcutaneous and visceral fat (top) and a surgery subject who has moderate subcutaneous fat with much less visceral fat (bottom). On the right are examples of fat quantification. The top image shows definition of visceral fat based on HU of −45 to −195 (orange color). The bottom image shows both subcutaneous and visceral fat (pink). The difference between the two is the subcutaneous fat. The area of fat tissue is multiplied by the slice thickness to obtain a volume. CT, computed tomography; HU, Hounsfield units.

FIG. 2.

Examples of epicardial fat in a nonsurgery subject (left) and a subject who underwent bariatric surgery (right). Arrows indicate epicardial and pericardial borders (epicardial fat is between the epicardium and the pericardium). A nonsurgery subject has much larger epicardial fat volume than the surgery subject. Quantification performed as in Fig. 1.

Liver fat was estimated using a ratio of liver and spleen attenuation values (average of measurements in three regions of interest in each organ that were free of any visible artifact). An attenuation ratio <1.0 is often used to define fatty liver.³² Visceral fat and subcutaneous fat were measured in the same way as epicardial fat. The peritoneal circumference and the cutaneous circumference were manually traced and the number of pixels with fat attenuation in the visceral and subcutaneous compartments recorded, summing each specific bed from the visceral images using the same cutoff values as for epicardial fat. In this case, the fat volume within the single slice is reported.

Approximately 21% of patients had edges of the subcutaneous fat compartment that were cut off by the scanner due to body size. We measured the largest size that we could on these scans and the amount that was cut off represents a small fraction of the actual subcutaneous fat. Analyses with and without these subjects were performed.

Statistical analyses

Statistical analyses were performed on the two study groups using general linear models for comparison of means or logistic regression to calculate odds ratios (ORs) for incidence and remission. Prevalent disease at baseline was excluded for incidence calculations and only prevalent disease at baseline was included for remission calculations. Covariates included in most models were gender, baseline age, baseline BMI, and length of follow-up. Three stepwise regression models were used to determine the best subset of variables explaining incidence and remission of diabetes, hypertension, and dyslipidemia. The first stepwise models allowed each of the fat measurements to enter the model as surrogates for surgery status but excluded change in BMI or surgery status. To identify fat depots that were related to incidence or remission of each major disease even after correction for surgery status, the second model forced surgery status into the model before the stepwise procedure. Finally, since the effects of surgery may involve more than just weight or fat loss, change in BMI was first forced into the models to control for weight change, with surgery status and fat depot variables allowed to subsequently enter the model if significant at P < 0.05.

Results

Table 1 shows the mean and standard deviations of the study variables for the ever had and never had bariatric surgery groups. Baseline age and BMI differed between groups (P < 0.001). Length of follow-up was also significantly different between groups because the initial nonsurgery subjects who subsequently had surgery were only followed from the time of their surgery. Therefore, to account for these differences, all subsequent models adjusted for baseline age, baseline BMI, and length of follow-up.

Table 1.

Means and Standard Deviations of Study Variables at Each Time Point for the Two Study Groups Who Underwent Computed Tomography Scans at 12-Year Follow-Up Examination

	Ever had bariatric surgery	Never had bariatric surgery
Baseline	N = 261	N = 243
Gender (M/F)	50/211	55/188
Age (years)	48.7 ± 7.5	52.1 ± 7.3^a
Current smoker (yes/no)	9/252	6/237
BMI (kg/m²)	46.6 ± 7.0	43.6 ± 6.4^a
Waist (cm)	135 ± 17	131 ± 16
Glucose (mg/dL)	104.2 ± 29.9	108.1 ± 31.9
HbA1c (%)	5.9 ± 0.9	6.0 ± 1.0
Systolic blood pressure (mmHg)	130 ± 20	128 ± 18
Diastolic blood pressure (mmHg)	72 ± 12	73 ± 10
LDL-C (mg/dL)	110 ± 27	107 ± 29
HDL-C (mg/dL)	46.8 ± 11.1	47.0 ± 11.9
Triglycerides (mg/dL)	181 ± 85	183 ± 115
ALT (U/L)	27.5 ± 28.4	27.7 ± 14.4
Antihypertensives	40% (104/261)	41% (100/243)
Lipid medications	19% (50/261)	14% (33/243)
Diabetic medications	20% (51/261)	19% (46/243)
CT-based follow-up examination
BMI (kg/m²)	35.5 ± 8.4	43.3 ± 7.6^a
Waist (cm)	115 ± 23	133 ± 18^a
Visceral fat (cm³)	77.5 ± 51.9	135.7 ± 51.3^a
Subcutaneous fat (cm³)	208 ± 78	246 ± 58^a
Epicardial fat (cm³)	89.4 ± 43.8	130.4 ± 52.1^a
Liver density (HU)	55.3 ± 8.4	50.3 ± 11.8^a
Spleen density (HU)	41.7 ± 7.2	42.3 ± 8.5
Liver-to-spleen density ratio^b	1.4 ± 0.3	1.2 ± 0.4^a
Glucose (mg/dL)	92.3 ± 28.3	110.4 ± 45.2^a
HbA1c (%)	5.7 ± 0.9	6.4 ± 1.6^a
Systolic blood pressure (mmHg)	121 ± 18	126 ± 20^a
Diastolic blood pressure (mmHg)	70 ± 10	69 ± 10
LDL-C (mg/dL)	109 ± 30	112 ± 34
HDL-C (mg/dL)	60.5 ± 17.8	47.6 ± 13.9^a
Triglycerides (mg/dL)	107 ± 41	150 ± 77^a
ALT (U/L)	20.6 ± 10.4	23.7 ± 12.4^c
Change between exams
Length of follow-up (years)	10.0 ± 2.8	11.5 ± 0.7^a
BMI (kg/m²)	−11.1 ± 6.8	−0.2 ± 5.5^a

Data represent baseline examination and CT-based follow-up examination for ever had bariatric surgery and never had bariatric surgery groups.

P < 0.001 versus the bariatric surgery group means.

A higher liver-to-spleen ratio represents less liver fat.

P = 0.003 versus the bariatric surgery group means.

All other tests P > 0.05.

BMI, body mass index; CT, computed tomography; HbA1c, glycosylated hemoglobin; HDL-C, high-density lipoprotein cholesterol; HU, Hounsfield units; LDL-C, low-density lipoprotein cholesterol.

All of the different fat depots were significantly different between the two groups (P < 0.001) at the time of follow-up (Table 1). The bariatric surgery group had a higher liver-to-spleen density ratio, while spleen density, the denominator of the liver-to-spleen density ratio, was not significantly different between groups. These findings suggest less liver fat after bariatric surgery. To check for indications of bias between groups, subjects with gastric bypass surgery just after their baseline exam were compared with those who had any type of bariatric surgery during the follow-up period and with the subset of subjects who only had gastric bypass surgery during the follow-up period (Supplementary Table S1). No significant differences were seen among groups, except weak significance for baseline fasting glucose and glycosylated hemoglobin (HbA1c) between those who had surgery at baseline and those who had surgery during follow-up. None of the longitudinal changes in measurements were significantly different between any of the groups.

Table 2 shows the group means and comparisons after statistical adjustment for gender, baseline age, baseline BMI, and length of follow-up. The means of each fat depot remained significantly different between study groups (P < 0.001). Visceral fat volume estimated at the L4/L5 abdominal level was 42% lower (77.8 vs. 135.3 cm³) and subcutaneous fat volume was 20% lower (202.3 vs. 252.1 cm³) in the ever had bariatric surgery group compared with the never had surgery group. Epicardial fat was 30% lower (90.7 vs. 129.0 cm³) and liver-to-spleen density ratio 9% higher (1.35 vs. 1.25) in the bariatric surgery group compared with the nonsurgery group. Adding follow-up smoking and medication use for diabetes, hypertension, or lipids as covariates to the analyses of Table 2 had little effect on the estimates or P values.

Table 2.

Adjusted Means (±Standard Errors) of Fat Depots at Computed Tomography-Based Follow-Up Examination for Subjects Who Were Severely Obese and Had Bariatric Surgery After the Baseline Exam Versus Subjects Who Were Severely Obese and Never Had Bariatric Surgery

Variable	Ever had bariatric surgery (N = 261)	Never had bariatric surgery (N = 243)
BMI (kg/m²)	34.3 ± 0.4^a	44.7 ± 0.4
Visceral fat (cm³)	77.8 ± 3.0^a	135.3 ± 3.1
Subcutaneous fat (cm³)	202 ± 4^a	252 ± 4
Epicardial fat (cm³)	90.7 ± 3.0^a	129.0 ± 3.1
Liver-to-spleen density ratio^b	1.36 ± 0.024^c	1.25 ± 0.026

Fat depot measurements are in volumes obtained at the L4/L5 level of the abdomen.

P < 0.0001 between groups, adjusted for gender, baseline age, baseline BMI, and length of follow-up.

A higher liver-to-spleen ratio represents less liver fat.

P = 0.005 between groups, adjusted for gender, baseline age, baseline BMI, and length of follow-up.

Table 3 shows the partial correlation coefficients separately for the surgical (upper diagonal) and nonsurgical groups (lower diagonal). Visceral fat, subcutaneous fat, and epicardial fat were significantly correlated with follow-up BMI in both groups, but the liver-to-spleen density ratio was not correlated with follow-up BMI in either group. In the nonsurgical group, epicardial fat was significantly correlated with visceral fat, but not with subcutaneous fat. In the surgical group, epicardial fat was correlated with both fat depots, but was more strongly correlated with visceral fat than with subcutaneous fat.

Table 3.

Correlations Between Fat Measurements at Follow-Up Within Study Group

Variables at follow-up	BMI	VFV	SFV	EFV	LDR
BMI (kg/m²)		0.66^a	0.73^a	0.37^a	0.03
Visceral fat (cm³)	0.44^a		0.45^a	0.52^a	0.03
Subcutaneous fat (cm³)	0.37^a	−0.20^a		0.28^a	−0.04
Epicardial fat (cm³)	0.33^a	0.46^a	−0.00		0.04
Liver-to-spleen density ratio	−0.11	−0.02	0.04	−0.01

Ever had bariatric surgery group correlations are in the top right diagonal and the never had bariatric surgery group correlations in the bottom diagonal.

P < 0.05, adjusted for gender, baseline age, baseline BMI, and length of follow-up.

EFV, epicardial fat volume; LDR, liver-to-spleen density ratio; SFV, subcutaneous fat volume; VFV, visceral fat volume.

To estimate the relationship between the follow-up amount of fat in each depot and the remission and incidence of important obesity-associated diseases during the follow-up period, the first analysis combined the surgery and nonsurgery groups and excluded the study group variable in the stepwise regression model (Table 4). Higher visceral fat at follow-up was significantly associated with both decreased remission and increased incidence of each of the five obesity-associated diseases that were considered.

Table 4.

Predictors of Incidence and Remission in the Combined Ever Had and Never Had Bariatric Surgery Groups Without Adjusting for Surgical Group or Body Mass Index Change

Endpoint	Yes/total N at follow-up	Predictor	OR (95% CI)	Standard deviation^a	P
Diabetes
Incidence	55/338	Visceral fat	2.25 (1.64–3.09)	54.7	<0.0001
Incidence	55/338	Baseline BMI	0.69 (0.49–0.98)	6.7	0.04
Remission	96/129	Visceral fat	0.39 (0.24–0.65)	65.0	0.0003
Hypertension
Incidence	65/212	Visceral fat	2.76 (1.94–3.94)	54.4	<0.0001
Remission	196/256	Visceral fat	0.59 (0.41–0.83)	61.7	0.003
Remission	196/256	Age	0.67 (0.49–0.92)	7.5	0.012
Low HDL-C
Incidence	35/334	Visceral fat	1.69 (1.07–2.66)	59.2	0.02
Incidence	35/334	Epicardial fat	1.96 (1.30–2.97)	52.8	0.002
Remission	58/132	Visceral fat	0.35 (0.22–0.56)	58.8	<0.0001
Remission	58/132	Gender (M vs. F)	0.33 (0.15–0.73)		0.006
High LDL-C
Incidence	154/367	Visceral fat	2.50 (1.78–3.52)	58.4	<0.0001
		Epicardial fat	1.59 (1.13–2.22)	53.6	0.007
		Baseline BMI	0.73 (0.56–0.93)	7.1	0.013
Remission	68/101	Visceral fat	0.40 (0.22–0.72)	63.1	0.002
Remission	68/101	Age	0.54 (0.32–0.91)	7.4	0.02
High triglyceride
Incidence	23/310	Liver-to-spleen ratio	0.52 (0.33–0.81)	0.37	0.004
Incidence	23/310	Visceral fat	1.75 (1.13–2.70)	60.9	0.012
Remission	41/158	Visceral fat	0.29 (0.17–0.49)	56.4	<0.0001
		Liver-to-spleen ratio	1.70 (1.12–2.60)	0.37	0.014
		Baseline BMI	1.97 (1.20–3.23)	6.6	0.007
		Age	2.08 (1.28–3.38)	7.7	0.003

Stepwise logistic regression using the combined ever had bariatric surgery and never had bariatric surgery subjects without allowing gastric bypass group or BMI change to enter the model. Variables available for entry were: gender, baseline age, baseline BMI, income level, marital status, education level, visceral fat, subcutaneous fat, epicardial fat, and liver-to-spleen density ratio.

ORs represent 1 standard deviation changes of the predictor variables, which are shown in this column.

CI, confidence interval; OR, odds ratio.

Higher epicardial fat was significantly associated with incidence of low HDL-C [OR = 1.96; 95% confidence interval, CI: (1.30–2.97) per standard deviation change in epicardial fat] and high LDL-C [OR = 1.59; 95% CI (1.13–2.22)]. Both of these associations were independent of visceral fat, as visceral fat also entered the stepwise regression model without removing the significance of epicardial fat. A lower liver-to-spleen density ratio (more relative liver fat) was associated with a higher incidence of high triglyceride levels [OR = 0.52; 95% CI (0.33–0.81)]. Less relative liver fat at follow-up was associated with greater odds of remission of high triglyceride levels [OR = 1.70; 95% CI (1.12–2.60)].

Table 5 includes results of analysis specifically addressing the resulting effects of bariatric surgery without regard to change in BMI by forcing the surgical status variable into the stepwise regression and testing the independent additional contributions of each fat depot. These data show that the amount of visceral fat makes an additional contribution beyond surgical status to incidence but not remission of diabetes and hypertension. The association of visceral fat with incidence of hypertension made the surgery status variable nonsignificant.

Table 5.

Predictors of the Incidence and Remission of Study Endpoints with Surgery Status (Ever Had Bariatric Surgery vs. Never Had Bariatric Surgery) Forced into the Stepwise Regression Model

Endpoint	Yes/total N at follow-up	Predictor	OR (95% CI)	Standard deviation^a	P
Diabetes
Incidence	55/338	Surgery versus nonsurgery	0.06 (0.02–0.19)		<0.0001
		Visceral fat	1.52 (1.05–2.20)	54.7	0.03
		Subcutaneous fat	0.61 (0.40–0.92)	72.1	0.02
Remission	96/129	Surgery versus nonsurgery	18.2 (5.2–64.2)		<0.0001
Hypertension
Incidence	65/212	Surgery versus nonsurgery	0.60 (0.29–1.21)		0.15
Incidence	65/212	Visceral fat	2.45 (1.67–3.62)	54.4	<0.0001
Remission	196/256	Surgery versus nonsurgery	2.88 (1.52–5.47)		0.0012
Remission	196/256	Age	0.65 (0.47–0.88)	7.5	0.006
Low HDL-C
Incidence	35/334	Surgery versus nonsurgery	0.22 (0.08–0.61)		0.003
Incidence	35/334	Epicardial fat	2.29 (1.61–3.27)	52.8	<0.0001
Remission	58/132	Surgery versus nonsurgery	3.80 (1.57–9.20)		0.003
		Visceral fat	0.46 (0.28–0.77)	58.8	0.003
		Gender (M vs. F)	0.30 (0.13–0.70)		0.005
High LDL-C
Incidence	154/367	Surgery versus nonsurgery	0.21 (0.12–0.40)		<0.0001
		Visceral fat	1.82 (1.28–2.60)	58.4	0.0009
		Epicardial fat	1.46 (1.04–2.06)	53.6	0.03
		Gender (M vs. F)	2.04 (1.04–4.00)		0.04
Remission	68/101	Surgery versus nonsurgery	14.8 (3.20–68.6)		0.0006
Remission	68/101	Age	0.52 (0.31–0.88)	7.4	0.01
High triglycerides
Incidence	23/310	Surgery versus nonsurgery	0.38 (0.14–1.04)		0.06
Incidence	23/310	Liver-to-spleen ratio	0.57 (0.36–0.91)	0.37	0.02
Remission	41/158	Surgery versus nonsurgery	7.52 (2.57–22.0)		0.0002
		Visceral fat	0.52 (0.30–0.90)	56.4	0.02
		Liver-to-spleen ratio	1.64 (1.07–2.50)	0.37	0.02
		Age	2.02 (1.23–3.34)	7.7	0.006

Stepwise logistic regression with gastric bypass group (ever had bariatric surgery vs. never had bariatric surgery) forced into the model. Variables available for entry were: gender, baseline age, baseline BMI, income level, marital status, education level, visceral fat, subcutaneous fat, epicardial fat, and liver-to-spleen density ratio.

ORs represent 1 standard deviation changes of the predictor variables, which are shown in this column.

Epicardial fat had a significant association with the incidence of low HDL-C and high LDL-C even after correcting for surgical group. The liver-to-spleen density ratio was more strongly associated with the incidence of high triglycerides than study group, as study group became nonsignificant (P = 0.06). The liver-to-spleen ratio, visceral fat, and study group were significant for remission of high triglycerides.

Finally, to test whether or not the study group effect was limited to change in BMI (i.e., the surgery had no other effects on incidence or remission other than weight change), the regression models were repeated forcing change in BMI into the model and allowing study group to enter if it explained a significant amount of additional variance (Table 6). For hypertension, study group did not enter either the remission or incidence models, with BMI change being highly predictive. For diabetes remission and incidence, study group entered the model and BMI change became a nonsignificant predictor. In each of the remaining cases where study group entered the regression model, the change in BMI became nonsignificant.

Table 6.

Predictors of the Incidence and Remission of Study Endpoints for Combined Ever Had Bariatric Surgery and Never Had Bariatric Surgery Groups with Body Mass Index Change Forced into the Stepwise Model

Endpoint	Yes/total N at follow-up	Predictor	OR (95% CI)	Standard deviation^a	P
Diabetes
Incidence	55/338	BMI change	0.98 (0.53–1.80)	8.3	0.94
		Surgery versus nonsurgery	0.06 (0.02–0.21)		<0.0001
		Visceral fat	1.53 (1.01–2.31)	54.7	0.04
		Subcutaneous fat	0.61 (0.39–0.97)	72.1	0.04
Remission	96/129	BMI change	0.83 (0.45–1.53)	8.2	0.55
Remission	96/129	Surgery versus nonsurgery	14.6 (3.4–61.9)		0.0003
Hypertension
Incidence	65/212	BMI change	1.13 (0.73–1.77)	8.5	0.58
Incidence	65/212	Visceral fat	2.58 (1.67–3.96)	54.4	<0.0001
Remission	196/256	BMI change	0.50 (0.35–0.69)	8.1	<0.0001
Remission	196/256	Age	0.64 (0.47–0.87)	7.5	0.005
Low HDL-C
Incidence	35/334	BMI change	2.76 (1.58–4.80)	8.0	0.0003
		Epicardial fat	2.10 (1.45–3.05)	52.8	<0.0001
		Subcutaneous fat	0.55 (0.32–0.94)	70.4	0.03
Remission	58/132	BMI change	0.76 (0.41–1.39)	8.9	0.37
		Surgery versus nonsurgery	2.96 (1.05–8.33)		0.04
		Visceral fat	0.50 (0.29–0.87)	58.8	0.013
		Gender (M vs. F)	0.27 (0.11–0.65)		0.004
High LDL-C
Incidence	154/367	BMI change	1.05 (0.74–1.50)	8.1	0.77
		Surgery versus nonsurgery	0.23 (0.12–0.44)		<0.0001
		Visceral fat	1.79 (1.23–2.61)	58.4	0.003
		Epicardial fat	1.46 (1.03–2.05)	53.6	0.03
		Gender (M vs. F)	2.09 (1.05–4.16)		0.04
Remission	68/101	BMI change	0.74 (0.37–1.47)	8.7	0.39
		Surgery versus nonsurgery	9.90 (1.68–58.3)		0.011
		Age	0.52 (0.31–0.89)	7.4	0.017
High triglycerides
Incidence	23/310	BMI change	0.92 (0.55–1.54)	8.5	0.63
		Liver-to-spleen ratio	0.51 (0.32–0.81)	0.37	0.002
		Visceral fat	1.82 (1.11–2.97)	60.9	0.02
Remission	41/158	BMI change	1.07 (0.56–2.04)	7.9	0.84
		Surgery versus nonsurgery	8.11 (2.21–29.8)		0.0016
		Visceral fat	0.51 (0.29–0.90)	56.4	0.02
		Liver-to-spleen ratio	1.65 (1.07–2.54)	0.37	0.02
		Age	2.04 (1.23–3.37)	7.7	0.006

Stepwise logistic regression with BMI change (Exam 4 to Exam 1) forced into the model. Variables available for entry were: ever had surgery/never had surgery status, gender, baseline age, baseline BMI, income level, marital status, education level, visceral fat, subcutaneous fat, epicardial fat, and liver-to-spleen density ratio.

ORs for continuous variables are for 1 standard deviation changes of the predictor variables, which are shown in this column.

Plasma ALT is a marker of nonalcoholic liver disease and associated with increased liver fat, although it is insensitive for NAFLD diagnosis.³³ The presence of fatty liver disease defined by a relative liver density <1.0 was 21%. The correlation of ALT with relative liver density was r = −0.28, P < 0.0001 compared with a correlation of ALT with visceral fat (r = 0.17, P = 0.0002), waist circumference (r = 0.14, P = 0.003), and subcutaneous fat (r = −0.03, P = 0.56), after adjusting for sex, age, and baseline BMI. Further adjustment for follow-up BMI only slightly reduced the correlation of ALT with relative liver density (r = −0.26, P < 0.0001), but had larger reductions for visceral fat (r = 0.10, P = 0.04) and waist circumference (r = 0.03, P = 0.49). ALT became strongly inversely correlated with subcutaneous fat (r = −0.17, P = 0.0004). Substituting visceral fat adjustment for follow-up BMI adjustment had little effect on the ALT and relative liver density correlation (r = −0.27, P < 0.0001).

Relative liver density, adjusted for sex, age, baseline BMI, visceral fat, and subcutaneous fat, was not correlated with follow-up fasting glucose (P = 0.31) or HbA1c (P = 0.82), but was correlated with follow-up triglyceride levels (r = −0.13, P = 0.004). Visceral fat, however, had stronger correlations with fasting glucose (r = 0.33, P < 0.0001) and HbA1c (r = 0.40, P < −0.0001), after adjustment for sex, age, baseline BMI, and subcutaneous fat. Visceral fat also correlated with triglyceride levels (r = 0.42, P < 0.0001).

Discussion

In this study of participants with an average baseline BMI of 45 kg/m², the quantity of fat in several major fat depots after an average of 11 years follow-up was lower in subjects who had bariatric surgery than those who did not have surgery. The lower amounts of fat in these depots at follow-up significantly correlated with change in BMI with the exception of liver fat. Higher visceral fat at follow-up clearly was the strongest predictor of less disease remission (0.6 times) and higher disease incidence (2.3 times) for a 1 standard deviation increase in visceral fat volume. Statistical modeling supports the notion that visceral fat is more specific and closer to the underlying pathway of diabetes incidence and remission than BMI per se. Furthermore, the benefits of bariatric surgery for the five endpoints tested were mostly explained by the amount of visceral fat; however, our data suggest that there may be additional benefits of weight loss surgery beyond a simple reduction in visceral fat.

Previous studies have clearly shown that liver fat decreases after bariatric surgery.^34
–36 Most of these studies, as well as other studies not mentioned, generally included small patient numbers with limited follow-up. A larger study of 243 individuals showed that those who lost the most weight following bariatric surgery had the lowest percent of liver fibrosis.³⁵ Liver fat has been shown to be correlated with visceral fat.^37,38 However, in this study of postbariatric surgery subjects, there were no significant correlations of liver fat, as estimated by the liver-to-spleen density ratio, with any of the other fat depots. The relationships of increased liver and visceral fat with triglycerides may directly impact the development of NAFLD induced by increased fatty acid storage in the hepatocytes. However, supporting the lack of a postsurgical correlation found between visceral and liver fat was their statistically independent prediction of hypertriglyceridemia incidence and remission, as both entered and remained in the stepwise prediction models. While visceral fat was associated with all five outcomes in this study, liver fat was only correlated with hypertriglyceridemia.

The metabolic consequences of liver fat may differ in some respects from the consequences of visceral fat. One study showed that when matching subjects so that they had the same amount of visceral fat, higher triglyceride levels in liver led to impaired insulin sensitivity in liver, adipose, and muscle tissues, but when matching on liver fat, subjects with higher visceral fat did not show such impairment.⁹ Perhaps it is not surprising that liver fat and visceral fat have differing effects since fat in the liver likely acts through effects on the hepatocytes, whereas an adipose depot behaves as a relatively independent organ and the adipocytes themselves secrete hormones and cytokines that then affect other organs and body parts.

It is generally believed that subcutaneous fat and visceral fat behave very differently and have divergent effects on metabolism and health. Subcutaneous fat has been associated with higher leptin levels, but not with insulin resistance, dyslipidemia, or atherosclerosis.^39,40 Subcutaneous fat was not associated with glucose tolerance while visceral fat was correlated.¹⁰ However, others have suggested that subcutaneous and visceral adipose tissue have similar correlations with insulin resistance.⁴¹ Our study supports the notion that these fat beds are regulated differently, even many years after bariatric surgery and that each bed has different associations with long-term protection from diabetes, hypertension, and dyslipidemia.

Epicardial fat was significantly related to HDL-C and LDL-C incidence. The relationship of epicardial fat with higher LDL-C and lower HDL-C suggests a possible mechanism for the proposed link between epicardial fat and coronary atherosclerosis.^42,43 This does not exclude additional, more local (paracrine or mechanical) effects of epicardial fat on coronary arteries.⁴⁴ Epicardial fat was correlated with BMI and visceral fat, but not with hepatic or subcutaneous fat. The 30% reduction in epicardial fat at 11 years was comparable to that of another study that showed a 36% reduction in epicardial fat 6 months after a very low-calorie diet,¹¹ particularly since some of our subjects regained some of their weight over the follow-up period. In a 4-year follow-up of 374 subjects, none of whom had bariatric surgery, epicardial fat change was related to changes in weight and waist circumference.⁴⁵ The subset who increased weight >5% showed increases in epicardial fat, whereas those who lost >5% weight had decreases in epicardial fat. Epicardial fat did not seem to be related to hypercholesterolemia in that study.⁴⁵ In a study of 10 patients who were diabetic, obese, and had gastric bypass surgery, pericardial fat had an even larger decrease than epicardial fat 16 weeks after surgery.⁴⁶ A study of 37 subjects who underwent either gastric bypass or sleeve gastrectomy operations showed a reduction in epicardial fat, but associations with blood parameters were not tested.⁴⁷

Subcutaneous fat was not strongly related to any of the disease endpoints, and when it did enter the statistical model, higher subcutaneous fat appeared to be protective even though it was measured abdominally, as has been observed previously.^39,48,49 One theory of the metabolic consequences of obesity is that subcutaneous fat accumulates excess lipid, acting as a “metabolic sink” to prevent the deposition of ectopic fat in liver and muscle.^50,51 However, once the subcutaneous fat depot cannot further expand for fat storage, lipid accumulation in these other depots has adverse consequences. In contrast to subcutaneous fat, which seems to behave more like a passive storage site, there is evidence that visceral fat has a proinflammatory effect mediated by the production and secretion of inflammatory adipocytokines. Subcutaneous fat does not appear to have this same effect. It has been hypothesized that differences between subcutaneous and visceral fat quantities may underlie some of the well-known gender differences in susceptibility to coronary artery disease.

The answer to the question as to whether or not there were additional benefits of bariatric surgery beyond just weight loss appears to be yes, at least from this study. For most of the endpoints, even when change in BMI was forced into the stepwise regression model, the dichotomous surgery group variable also entered the model, making the quantitative change in BMI nonsignificant. This suggests that there were other hormonal, nutritional, or anti-inflammatory factors playing a role in disease incidence or remission of diabetes and dyslipidemia independent of weight loss. The one exception was the incidence and remission of hypertension, which only seemed to be related to follow-up visceral fat or change in BMI, as surgical group was not significant. This also seems to make sense, as the hormones related to gastrointestinal alterations are not as strongly related to blood pressure as they are to lipids and diabetes.

One limitation of this study is that fat depots were not assessed at baseline. To infer that the changes in fat depots would have provided the same information as comparing the follow-up fat depot measurements between the two surgery groups requires an assumption that the baseline fat amounts were the same in the two groups. Because BMI was significantly greater in the surgical group at baseline (although not waist circumference), this may not be a valid assumption. The comparison of the fat depots between groups and with BMI is cross-sectional only. However, the incidence and remission of disease was assessed over time from baseline. The interpretation of these results is that the less fat stored in particular fat depots after 11 years of follow-up after bariatric surgery, the more disease incidence is reduced and remission is increased, and that the relationships between fat depots and disease are depot specific.

Only the dome of the liver was used to estimate the liver-to-spleen ratio. However, liver fat seems to be sufficiently uniformly distributed for this analysis to be meaningful.⁵² As noted in the Methods section, 21% of subjects were too large to have all of the subcutaneous fat included in the scan field. Analyses were rerun after excluding the 21% resulting in only minor numerical differences and no differences in conclusions. The group differences in subcutaneous fat remained consistent and significant, whereas the correlation coefficients that were significant in the whole group changed by only 0.02 or less after exclusion.

We note that bariatric surgeries during the first year following baseline examination were all gastric bypass performed by the same three surgeons (partners) and, therefore, procedures were homogeneous in nature. Surgeries after the first year of baseline examination were performed by multiple surgeons and included varied procedures, resulting in greater variation. A majority of bariatric surgery studies, including the Swedish Obesity Subjects (SOS) study, have combined surgery types for primary analysis. Due to the small number of nongastric bypass surgeries, we compared gastric bypass surgeries versus all other bariatric procedures to assess the impact on the study results.

Finally, those subjects who had bariatric surgery during the follow-up period had shorter follow-up time (7 years) than subjects with surgery at baseline (11 years). Despite the differences in follow-up time, the most pronounced changes occur during the first couple of years postsurgery. Therefore, even with a difference of about 4 years of follow-up, there were no significant differences between the two surgery groups for baseline age or changes in BMI, blood pressure, waist circumference, glucose, HbA1c, ALT, LDL-C, and triglycerides. Only a change in HDL-C was nominally significantly different between the two groups (P = 0.013), which became nonsignificant after multiple comparison adjustment. None of these clinical differences was likewise significant when comparing those who had gastric bypass versus other surgery procedures at follow-up.

In conclusion, we have shown that bariatric surgery compared with nonsurgery in subjects with severe obesity leads to lower fat at follow-up in all major fat depots, but that specific depots are not all related to the same disease endpoints. Liver fat was uncorrelated with other fat depots and was most strongly associated with triglyceride levels, while epicardial fat was most strongly associated with cholesterol levels. Visceral fat was related to multiple diseases, but most strongly with diabetes. Because the fat depots have different associations with specific cardiovascular risk factors, the postbariatric surgery reductions in fat help explain the improvements in multiple systemic pathways that have been associated with bariatric surgery and the subsequent increase in survival. A major strength of this study is the large sample size and the long duration of follow-up. The large sample size also allowed more robust estimation of incidence and remission of diabetes, hypertension, and dyslipidemia.

Footnotes

Author Disclosure Statement

The authors declared the funding listed as the only potential conflict of interest. No other competing financial interests exist.

Funding Information

The study was supported by a grant (DK-55006) from the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health, a U.S. Public Health Service research grant (MO1-RR00064) from the National Center for Research Resources, and Biomedical Research Program funds from Weill Cornell Medicine and Intermountain Healthcare.

Supplementary Material

Supplementary Table S1

References

Flegal

, Graubard

, Williamson

, et al. Cause-specific excess deaths associated with underweight, overweight, and obesity. JAMA, 2007; 298:2028–2037.

Britton

, Massaro

, Murabito

, et al. Body fat distribution, incident cardiovascular disease, cancer, and all-cause mortality. J Am Coll Cardiol, 2013; 62:921–925.

Neeland

, Turer

, Ayers

, et al. Dysfunctional adiposity and the risk of prediabetes and type 2 diabetes in obese adults. JAMA, 2012; 308:1150–1159.

Sahakyan

, Somers

, Rodriguez-Escudero

, et al. Normal-weight central obesity: Implications for total and cardiovascular mortality. Ann Intern Med, 2015; 163:827–835.

Pou

, Massaro

, Hoffmann

, et al. Visceral and subcutaneous adipose tissue volumes are cross-sectionally related to markers of inflammation and oxidative stress: The Framingham Heart Study. Circulation, 2007; 116:1234–1241.

Lim

, Meigs

. Links between ectopic fat and vascular disease in humans. Arterioscler Thromb Vasc Biol, 2014; 34:1820–1826.

Marchesini

, Bugianesi

, Forlani

, et al. Nonalcoholic fatty liver, steatohepatitis, and the metabolic syndrome. Hepatology, 2003; 37:917–923.

Yki-Jarvinen

. Non-alcoholic fatty liver disease as a cause and a consequence of metabolic syndrome. Lancet Diabetes Endocrinol, 2014; 2:901–910.

Fabbrini

, Magkos

, Mohammed

, et al. Intrahepatic fat, not visceral fat, is linked with metabolic complications of obesity. Proc Natl Acad Sci U S A, 2009; 106:15430–15435.

10.

Despres

, Nadeau

, Tremblay

, et al. Role of deep abdominal fat in the association between regional adipose tissue distribution and glucose tolerance in obese women. Diabetes, 1989; 38:304–309.

11.

Iacobellis

, Singh

, Wharton

, et al. Substantial changes in epicardial fat thickness after weight loss in severely obese subjects. Obesity (Silver Spring), 2008; 16:1693–1697.

12.

Mikhalkova

, Holman

, Jiang

, et al. Bariatric surgery-induced cardiac and lipidomic changes in obesity-related heart failure with preserved ejection fraction. Obesity (Silver Spring), 2018; 26:284–290.

13.

Sjostrom

, Lindroos

, Peltonen

, et al. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med, 2004; 351:2683–2693.

14.

Sjostrom

, Narbro

, Sjostrom

, et al. Effects of bariatric surgery on mortality in Swedish obese subjects. N Engl J Med, 2007; 357:741–752.

15.

Adams

, Gress

, Smith

, et al. Long-term mortality after gastric bypass surgery. N Engl J Med, 2007; 357:753–761.

16.

Adams

, Davidson

, Litwin

, et al. Weight and metabolic outcomes 12 years after gastric bypass. N Engl J Med, 2017; 377:1143–1155.

17.

Schauer

, Bhatt

, Kirwan

, et al. Bariatric surgery versus intensive medical therapy for diabetes—5-Year outcomes. N Engl J Med, 2017; 376:641–651.

18.

Sundstrom

, Bruze

, Ottosson

, et al. Weight loss and heart failure: A Nationwide Study of Gastric Bypass Surgery Versus Intensive Lifestyle Treatment. Circulation, 2017; 135:1577–1585.

19.

Fisher

, Johnson

, Haneuse

, et al. Association between bariatric surgery and macrovascular disease outcomes in patients with type 2 diabetes and severe obesity. JAMA, 2018; 320:1570–1582.

20.

Busetto

, Tregnaghi

, Bussolotto

, et al. Visceral fat loss evaluated by total body magnetic resonance imaging in obese women operated with laparascopic adjustable silicone gastric banding. Int J Obes Relat Metab Disord, 2000; 24:60–69.

21.

Johansson

, Roos

, Kullberg

, et al. Lipid mobilization following Roux-en-Y gastric bypass examined by magnetic resonance imaging and spectroscopy. Obes Surg, 2008; 18:1297–1304.

22.

Pontiroli

, Frige

, Paganelli

, et al. In morbid obesity, metabolic abnormalities and adhesion molecules correlate with visceral fat, not with subcutaneous fat: Effect of weight loss through surgery. Obes Surg, 2009; 19:745–750.

23.

Carroll

, Franks

, Smith

, et al. Visceral adipose tissue loss and insulin resistance 6 months after laparoscopic gastric banding surgery: A preliminary study. Obes Surg, 2009; 19:47–55.

24.

Miller

, Carr

, Fernandez

. Regional fat changes following weight reduction from laparoscopic Roux-en-Y gastric bypass surgery. Diabetes Obes Metab, 2011; 13:189–192.

25.

Umemura

, Sasaki

, Nitta

, et al. Effects of changes in adipocyte hormones and visceral adipose tissue and the reduction of obesity-related comorbidities after laparoscopic sleeve gastrectomy in Japanese patients with severe obesity. Endocr J, 2014; 61:381–391.

26.

Bazzocchi

, Ponti

, Cariani

, et al. Visceral fat and body composition changes in a female population after RYGBP: A two-year follow-up by DXA. Obes Surg, 2015; 25:443–451.

27.

Galanakis

, Daskalakis

, Manios

, et al. Computed tomography-based assessment of abdominal adiposity changes and their impact on metabolic alterations following bariatric surgery. World J Surg, 2015; 39:417–423.

28.

Toro-Ramos

, Goodpaster

, Janumala

, et al. Continued loss in visceral and intermuscular adipose tissue in weight-stable women following bariatric surgery. Obesity (Silver Spring), 2015; 23:62–69.

29.

Auclair

, Martin

, Bastien

, et al. Is there a role for visceral adiposity in inducing type 2 diabetes remission in severely obese patients following biliopancreatic diversion with duodenal switch surgery?. Obes Surg, 2016; 26:1717–1727.

30.

Adams

, Pendleton

, Strong

, et al. Health outcomes of gastric bypass patients compared to nonsurgical, nonintervened severely obese. Obesity (Silver Spring), 2010; 18:121–130.

31.

Rosito

, Massaro

, Hoffmann

, et al. Pericardial fat, visceral abdominal fat, cardiovascular disease risk factors, and vascular calcification in a community-based sample: The Framingham Heart Study. Circulation, 2008; 117:605–613.

32.

Zeb

, Li

, Nasir

, et al. Computed tomography scans in the evaluation of fatty liver disease in a population based study: The multi-ethnic study of atherosclerosis. Acad Radiol, 2012; 19:811–818.

33.

Mofrad

, Contos

, Haque

, et al. Clinical and histologic spectrum of nonalcoholic fatty liver disease associated with normal ALT values. Hepatology, 2003; 37:1286–1292.

34.

van Werven

, Schreuder

, Aarts

, et al. Hepatic steatosis in morbidly obese patients undergoing gastric bypass surgery: Assessment with open-system ¹H-MR spectroscopy. AJR Am J Roentgenol, 2011; 196:W736–W742.

35.

Abdennour

, Reggio

, Le Naour

, et al. Association of adipose tissue and liver fibrosis with tissue stiffness in morbid obesity: Links with diabetes and BMI loss after gastric bypass. J Clin Endocrinol Metab, 2014; 99:898–907.

36.

Hedderich

, Hasenberg

, Haneder

, et al. Effects of bariatric surgery on non-alcoholic fatty liver disease: Magnetic resonance imaging is an effective, non-invasive method to evaluate changes in the liver fat fraction. Obes Surg, 2017; 27:1755–1762.

37.

Thomas

, Parkinson

, Frost

, et al. The missing risk: MRI and MRS phenotyping of abdominal adiposity and ectopic fat. Obesity (Silver Spring), 2012; 20:76–87.

38.

, Wong

, Hsu

, et al. The correlation between body fat, visceral fat, and nonalcoholic fatty liver disease. Metab Syndr Relat Disord, 2017; 15:304–311.

39.

Cnop

, Landchild

, Vidal

, et al. The concurrent accumulation of intra-abdominal and subcutaneous fat explains the association between insulin resistance and plasma leptin concentrations: Distinct metabolic effects of two fat compartments. Diabetes, 2002; 51:1005–1015.

40.

Neeland

, Ayers

, Rohatgi

, et al. Associations of visceral and abdominal subcutaneous adipose tissue with markers of cardiac and metabolic risk in obese adults. Obesity (Silver Spring), 2013; 21:E439–E447.

41.

Ross

, Aru

, Freeman

, et al. Abdominal adiposity and insulin resistance in obese men. Am J Physiol Endocrinol Metab, 2002; 282:E657–E663.

42.

Iacobellis

, Ribaudo

, Assael

, et al. Echocardiographic epicardial adipose tissue is related to anthropometric and clinical parameters of metabolic syndrome: A new indicator of cardiovascular risk. J Clin Endocrinol Metab, 2003; 88:5163–5168.

43.

Yerramasu

, Dey

, Venuraju

, et al. Increased volume of epicardial fat is an independent risk factor for accelerated progression of sub-clinical coronary atherosclerosis. Atherosclerosis, 2012; 220:223–230.

44.

Nagy

, Jermendy

, Merkely

, et al. Clinical importance of epicardial adipose tissue. Arch Med Sci, 2017; 13:864–874.

45.

Nakazato

, Rajani

, Cheng

, et al. Weight change modulates epicardial fat burden: A 4-year serial study with non-contrast computed tomography. Atherosclerosis, 2012; 220:139–144.

46.

van Schinkel

, Sleddering

, Lips

, et al. Effects of bariatric surgery on pericardial ectopic fat depositions and cardiovascular function. Clin Endocrinol, 2014; 81:689–695.

47.

Kokkinos

, Alexiadou

, Liaskos

, et al. Improvement in cardiovascular indices after Roux-en-Y gastric bypass or sleeve gastrectomy for morbid obesity. Obes Surg, 2013; 23:31–38.

48.

Adams

, Montague

, Prins

, et al. Activators of peroxisome proliferator-activated receptor gamma have depot-specific effects on human preadipocyte differentiation. J Clin Invest, 1997; 100:3149–3153.

49.

Borel

, Nazare

, Smith

, et al. Visceral, subcutaneous abdominal adiposity and liver fat content distribution in normal glucose tolerance, impaired fasting glucose and/or impaired glucose tolerance. Int J Obes (Lond), 2015; 39:495–501.

50.

Virtue

, Vidal-Puig

. It's not how fat you are, it's what you do with it that counts. PLoS Biol, 2008; 6:e237.

51.

Shulman

. Ectopic fat in insulin resistance, dyslipidemia, and cardiometabolic disease. N Engl J Med, 2014; 371:2237–2238.

52.

Davidson

, Kuk

, Church

, et al. Protocol for measurement of liver fat by computed tomography. J Appl Physiol, 2006; 100:864–868.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.01 MB