Abstract
Background:
Alterations in the cellular characteristics of subcutaneous adipose tissue (SAT) may reduce its ability to expand in times of caloric excess, increasing the propensity to store excess calories viscerally (visceral adipose tissue [VAT]). We hypothesized (1) that increased SAT density, an indirect marker of fat quality, would be associated with an increased VAT/SAT ratio and increased cardiovascular disease (CVD) risk and (2) that these associations would be independent of the absolute volume of SAT.
Methods:
We investigated the association of SAT density with the VAT/SAT ratio and CVD risk in 3212 participants (48% women, mean age, 50.7 years) from the Framingham Heart Study. Adipose tissue depot density and volume were quantified by computed tomography; traditional CVD risk factors were quantified.
Results:
Higher SAT density was correlated with a higher VAT/SAT ratio in men (r = 0.17; P < .0001) but not in women (r = 0.04; P ≥ .05). More adverse levels of CVD risk factors were observed in the high SAT density/high VAT/SAT ratio group than in the referent group (low density/low ratio). For example, women had an increased risk of diabetes (odds ratio [OR], 6.7; 95% confidence interval [CI], 2.6–17.6; P = .0001) and hypertension (OR, 1.6; 95% CI, 1.1–2.4; P = .009). Additional adjustment for SAT volume generally strengthened these associations (diabetes OR, 10.8; 95% CI, 4.1–29.0; hypertension OR, 2.5; 95% CI, 1.7–3.7; all P < .0001). These trends were similar but generally weaker in men.
Conclusion:
High fat density, an indirect marker of fat quality, is associated with the propensity to store fat viscerally vs subcutaneously and is jointly characterized by an increased burden of CVD risk factors.
Body mass index (BMI) and waist circumference have been used to assess obesity, but these measurements contain little information about the anatomic location of stored excess fat. Regional fat depots may confer differential cardiovascular disease (CVD) risk, with visceral adipose tissue (VAT) volume more closely associated with CVD risk factors than subcutaneous adipose tissue (SAT) volume (1). A large body of work has shown that VAT is associated with increased insulin resistance, type 2 diabetes, metabolic syndrome, and the atherogenic lipoprotein profile (2–7).
In addition to fat volume, previous studies have focused on aspects of fat quality, including cellular characteristics biopsied invasively (8–14). Fat quality may be an important metric to characterize adipose tissue with respect to cardiovascular and metabolic risk but can be difficult to capture in free-living populations. We have previously shown that fat quality can be indirectly estimated using the density of adipose tissue as measured radiographically on computed topography (CT) with Hounsfield units (HU) (15). Small experimental studies in the setting of brown adipose tissue suggest that fat attenuation on CT may reflect extracellular matrix remodeling with high HU associated with increased metabolic activity and low HU associated with more lipid-dense adipocytes (16, 17). Highlighting the complexities of this metric, SAT fibrosis may be reflected by a relatively high density of fat tissue and thus estimated by the higher HU, although this has not been demonstrated in experimental studies (15). We hypothesized that when SAT becomes fibrotic, the adipocytes cannot expand sufficiently in times of calorie excess, forcing excess calories to be stored in VAT, which may be metabolically maladaptive. This shift in fat stores of the body from subcutaneous to visceral depots may be reflected by a high VAT volume/SAT volume ratio. The VAT/SAT ratio has been associated with CVD and metabolic risk, beyond BMI or VAT alone (2, 3, 5, 18, 19). Thus, the goals of our study were (1) to examine whether increases in SAT density were associated with a propensity to store fat viscerally vs subcutaneously and (2) to test whether high fat density and a high VAT/SAT ratio are jointly associated with an increased burden of CVD risk factors.
Subjects and Methods
Study sample
All participants for this study were drawn from the Framingham Heart Study. The Original cohort and their spouses were enrolled in 1948, their children were enrolled as the Offspring cohort in 1971, and finally the Third Generation cohort was enrolled in 2002 (20–22). Data for this current project were collected from the Offspring cohort and the Third Generation cohort enrolled in a multidetector computed tomography (MDCT) substudy (23). The MDCT subcohort contains primarily individuals residing in the US New England area. Participants were eligible if they were >40 years of age for women, were >35 years of age for men, had a body weight of <160 kg, and were not pregnant. Between 2002 and 2005, a total of 3539 participants (1422 from the Offspring cohort and 2117 from the Third Generation cohort) underwent CT scans of the abdomen. Of these 3539 participants, 3394 had acceptable quality CT data. Participants with missing outcomes (VAT and SAT volumes and HU) or missing covariates were excluded, resulting in a final sample size of 3212 (48% women). Informed consent was required for each participant, and this study met the terms of the Declaration of Helsinki and was authorized by the institutional review board of the Boston University Medical Center.
CT measurements
Participants underwent supine 8-slice MDCT scans of the abdomen for SAT, VAT, and other adipose tissue depots resulting in 25 contiguous 5-mm slices. SAT and VAT volumes were manually outlined by drawing the line of the abdominal muscular wall separating these 2 depots. Fat depots were designated for the area with an HU measurement between −195 and −45 HU. This technique demonstrates highly reproducible interreader and intrareader correlations (0.997 for SAT and 0.992 for VAT) (24). Further work evaluated the variability within each depot using 1 cm2 traces from 12 anatomically distinct regions in 3 slices per individual from 25 randomly selected participants. The average SDs for SAT and VAT were 7.6 and 5.5 HU, respectively, suggesting only modest variability of HU measurements within each depot (15). The CT scans were performed an average of 1.7 years after the clinical risk factor data were obtained.
CVD risk factor assessment
CVD risk factor data were obtained from the seventh examination of the Offspring cohort and the Third Generation cohort. Weight was assessed on a Detecto scale during the examination, and waist circumference was measured with a standard tape measure at the level of the umbilicus to the closest ¼ inch. BMI was defined as weight (kilograms) divided by square of height (meters squared). Participant blood pressure was measured at rest twice with a mercury column sphygmomanometer, and hypertension was defined as a systolic pressure of ≥140 mm Hg or a diastolic pressure of ≥90 mm Hg or the routine use of hypertensive medication. Serum metabolic risk factors, including plasma glucose, total cholesterol, high-density lipoprotein cholesterol and triglycerides, were assessed based on participants' fasting blood samples. Diabetes was defined as fasting blood glucose of ≥126 mg/dL or treatment with insulin or a hypoglycemic agent. Metabolic syndrome was classified based on the adapted Adult Treatment Panel III criteria (25).
Covariates
Questions administered by examining physicians were used to assess smoking and alcohol intake status. Current smokers were defined as individuals who had smoked at least 1 cigarette per day for the last year. Alcohol intake was dichotomized based on the criteria for men (>14 drinks/wk) and women (>7 drinks/wk). Physical activity score was determined from a questionnaire that assessed daily activity levels and average number of hours slept. Women were classified as to whether they received hormone replacement therapy.
Statistical analysis
Fat density and the VAT volume/SAT volume ratio were selected as the primary exposures in this analysis. SAT HU was dichotomized at the median value (cutoff value of −103.5 in women and −100.3 in men). The VAT/SAT ratio outcome was dichotomized at the median value (cutoff value of 0.3946 in women and 0.838 in men). SAT HU, an indirect marker of tissue density, was renamed for our analysis as either high or low density, and the VAT/SAT ratio was simplified to high or low ratio.
Four groups were created from our dichotomized primary exposure and outcome: (1) high density/high ratio (ie, high SAT density with a high propensity to store fat viscerally), (2) high density/low ratio (high SAT density with a low propensity to store fat viscerally), (3) low density/high ratio (low SAT density with a high propensity to store fat viscerally), and (4) low density/low ratio (low SAT density with a low propensity to store fat viscerally). We hypothesized a priori that the high density/high ratio group would have the most adverse CVD risk factor profile. For comparative clinical analyses between groups, we designated the low density/low ratio group as our referent group. Because of the known sexual dimorphism in fat distribution, all models were stratified by sex.
The VAT/SAT ratio was log transformed as necessary to produce a more normal distribution. We first computed age-adjusted correlations between the log of the VAT/SAT ratio and CVD risk factors. We tested for an interaction between the VAT/SAT ratio and SAT HU values. Next, multivariable logistic regression was used to test the association between SAT density (high vs low) and the VAT/SAT ratio (high vs low). Finally, multivariable-adjusted logistic and linear regression models were used to test the association of each group compared with the referent group and CVD risk outcomes. Model 1 adjusted for age, sex, alcohol use, smoking status, physical activity, and hormone replacement therapy (women). Model 2 additionally adjusted for SAT volume (continuous) because we were interested in whether associations with high SAT density and a high VAT/SAT ratio were independent of SAT volume
Results
Study sample characteristics
The clinical characteristics of the 3212 study participants are presented in Table 1, stratified by the four groups. In women, mean BMI ranged from 24.7 kg/m2 in the high density/low ratio group to 28.4 kg/m2 in the low density/high ratio group. A similar pattern was seen for the mean absolute SAT volume with a range of 2472 to 3908 cm3 in the high density/low ratio and low density/low ratio groups, respectively. In men, BMI and SAT followed the same pattern. In general, participants with high density/high ratio or low density/high ratio had the most adverse levels of CVD risk factors.
Table 1.
Clinical Characteristics of Study Participants, Subdivided Into Four Groups: High Density/High Ratio, High Density/Low Ratio, Low Density/High Ratio, and Low Density/Low Ratio
| High Density/High Ratio | High Density/Low Ratio | Low Density/High Ratio | Low Density/Low Ratio | |
|---|---|---|---|---|
| Women, no. | 363 | 397 | 400 | 366 |
| Age, y | 55.8 (11.1) | 47.3 (7.5) | 55.2 (9.5) | 49.3 (7.6) |
| Smoking, % | 15.4 (56) | 9.3 (37) | 13.8 (55) | 9.8 (36) |
| Moderate to heavy alcohol use, % | 16.5 (60) | 14.9 (37) | 17.5 (70) | 10.4 (38) |
| Physical activity score, % | 37.8 (6.1) | 37.0 (6.1) | 36.2 (5.5) | 36.0 (5.4) |
| Total cholesterol, mg/dL | 198.7 (36.9) | 185.0 (32.7) | 208.2 (39.0) | 199.0 (32.2) |
| HDL cholesterol, mg/dL | 58.6 (16.6) | 67.4 (17.6) | 58.1 (16.4) | 60.5 (14.9) |
| Low HDL, % | 31.4 (114) | 15.4 (61) | 32.3 (129) | 24.0 (88) |
| Triglycerides, mg/dL | 130.7 (84.2) | 80.0 (42.2) | 135.8 (71.9) | 107.5 (55.1) |
| High triglycerides, % | 37.7 (137) | 8.8 (35) | 39.0 (156) | 21.9 (80) |
| Fasting glucose, mg/dL | 100.8 (24.8) | 90.5 (15.7) | 98.5 (16.2) | 93.7 (12.0) |
| High glucose, % | 22.3 (81) | 6.0 (24) | 28.5 (114) | 16.7 (61) |
| Diabetes, % | 13.8 (50) | 2.8 (11) | 4.3 (17) | 1.4 (5) |
| Hypertension, % | 37.3 (135) | 10.9 (43) | 38.5 (154) | 19.2 (70) |
| Cardiovascular disease, % | 7.7 (28) | 0.8 (3) | 5.5 (22) | 1.1 (4) |
| Metabolic syndrome, % | 38.0 (138) | 7.6 (30) | 42.4 (169) | 20.9 (76) |
| Hypertensive treatment, % | 28.1 (102) | 7.1 (28) | 27.5 (110) | 11.7 (43) |
| Lipid treatment, % | 16.5 (60) | 3.5 (14) | 13.0 (52) | 8.5 (31) |
| Systolic blood pressure, mm Hg | 123.5 (19.1) | 112.6 (13.9) | 125.6 (18.7) | 119.1 (15.3) |
| Diastolic blood pressure, mm Hg | 73.3 (10.1) | 70.9 (8.7) | 75.9 (8.9) | 74.3 (8.1) |
| VAT/SAT ratio | 0.6 (0.3) | 0.3 (0.1) | 0.6 (0.1) | 0.3 (0.1) |
| BMI, kg/m2 | 27.1 (6.9) | 24.7 (5.6) | 28.4 (4.7) | 28.2 (5.1) |
| Waist circumference, cm | 93.5 (18.1) | 84.9 (14.3) | 98.3 (12.2) | 95.7 (13.4) |
| SAT, cm3 | 2706 (1604) | 2472 (1406) | 3515 (1153) | 3908 (1438) |
| VAT, cm3 | 1612 (1020) | 706 (461) | 1922 (678) | 1192 (485) |
| Subcutaneous HU, cm3 | −99.0 (5.3) | −98.7 (4.9) | −105.5 (1.5) | −105.8 (1.7) |
| Visceral HU, cm3 | −92.3 (4.4) | −89.0 (3.2) | −96.1 (3.3) | −92.5 (3.2) |
| Men, n | 497 | 350 | 346 | 493 |
| Age, y | 53.1 (11.8) | 47.7 (9.7) | 51.5 (10.3) | 45.9 (8.5) |
| Smoking, % | 13.1 (65) | 16.9 (59) | 13.0 (45) | 11.4 (56) |
| Moderate to heavy alcohol use, % | 19.9 (99) | 15.1 (53) | 18.5 (64) | 11.0 (54) |
| Physical activity score, % | 38.9 (8.0) | 39.3 (9.0) | 37.9 (7.8) | 37.4 (8.1) |
| Total cholesterol, mg/dL | 193.3 (34.5) | 189.9 (32.7) | 200.2 (33.0) | 196.0 (35.0) |
| HDL cholesterol, mg/dL | 44.7 (12.9) | 50.4 (13.2) | 44.2 (11.0) | 45.1 (11.4) |
| Low HDL, % | 40.1 (199) | 17.7 (62) | 37.0 (128) | 32.7 (161) |
| Triglycerides, mg/dL | 155.0 (124.9) | 107.2 (74.3) | 161.7 (113.4) | 140.6 (99.6) |
| High triglycerides, % | 49.7 (247) | 24.9 (87) | 55.8 (193) | 42.5 (209) |
| Fasting glucose, mg/dL | 105.6 (33.2) | 100.9 (26.7) | 101.0 (11.4) | 100.1 (13.7) |
| High glucose, % | 36.2 (180) | 25.4 (89) | 47.0 (162) | 39.6 (195) |
| Diabetes, % | 12.1 (60) | 7.1 (25) | 3.8 (13) | 5.1 (25) |
| Hypertension, % | 37.8 (188) | 18.0 (63) | 35.3 (122) | 31.2 (154) |
| Cardiovascular disease, % | 11.3 (56) | 5.4 (19) | 9.8 (34) | 4.5 (22) |
| Metabolic syndrome, % | 38.4 (191) | 18.1 (63) | 49.4 (171) | 42.4 (208) |
| Hypertensive treatment, % | 26.2 (130) | 11.1 (39) | 20.2 (70) | 16.4 (81) |
| Lipid treatment, % | 20.3 (101) | 11.4 (40) | 22.0 (76) | 16.0 (79) |
| Systolic blood pressure, mm Hg | 125.0 (15.9) | 119.5 (13.0) | 125.7 (14.4) | 122.7 (13.8) |
| Diastolic blood pressure, mm Hg | 77.3 (9.7) | 75.7 (8.6) | 79.3 (8.2) | 79.7 (9.1) |
| VAT/SAT ratio | 1.2 (0.4) | 0.6 (0.1) | 1.1 (0.3) | 0.6 (0.1) |
| BMI, kg/m2 | 27.5 (4.4) | 27.0 (4.9) | 29.0 (3.6) | 30.0 (4.5) |
| Waist circumference, cm | 98.0 (11.2) | 96.1 (12.7) | 103.0 (8.9) | 105.4 (11.5) |
| SAT, cm3 | 1974 (817) | 2250 (1211) | 2713 (767) | 3527 (1252) |
| VAT, cm3 | 2405 (1074) | 1400 (775) | 2976 (852) | 2108 (733) |
| Subcutaneous HU, cm3 | −96.6 (4.0) | −96.3 (4.4) | −102.3 (1.5) | −103.0 (1.7) |
| Visceral HU, cm3 | −94.8 (4.1) | −90.7 (3.8) | −98.8 (2.6) | −96.4 (3.3) |
Abbreviation: HDL, high-density lipoprotein. For continuous data, the SDs are provided in parentheses. For dichotomous data, the numerator is provided in parentheses. In women, high SAT HU was defined at a cutoff value greater than −103.6 and a high VAT/SAT ratio was defined at a cutoff value of 0.395. In men, the high SAT HU cutoff value was −100.3 and the high VAT/SAT ratio cutoff value was 0.839.
As the VAT/SAT ratio increased, so did CVD risk factors in both women and men with an r value ranging from 0.01 to 0.49 in women and from 0.07 to 0.38 in men (Table 2).
Table 2.
Age-Adjusted Correlations Between Log-Transformed VAT/SAT Ratio and Selected CVD Risk Factors for Women and Men
| Women (n = 1526) | Men (n = 1686) | |
|---|---|---|
| Age | 0.42a | 0.37a |
| BMI | 0.06b | −0.14a |
| Waist circumference | 0.10a | −0.16a |
| Systolic blood pressure | 0.12a | 0.08c |
| Diastolic blood pressure | 0.11a | 0.08c |
| High-density lipoprotein cholesterol | −0.24a | −0.16a |
| Triglycerides (log) | 0.32a | 0.22a |
| Physical activity score | 0.01 | 0.07d |
| Subcutaneous HU | 0.04 | 0.17a |
| Visceral HU | −0.49a | −0.38a |
| SAT | −0.11a | −0.44a |
| VAT | 0.53a | 0.42a |
P < .0001.
P < .05.
P < .001.
P < .01.
Association between SAT density and the VAT/SAT ratio
After adjustment for age, there was an association in men between subcutaneous HU and the VAT/SAT ratio (r = 0.17, P < .0001). In women, however, this correlation was not significant (r = 0.04, P ≥ .05). We tested the interaction between the VAT/SAT ratio and SAT HU (Supplemental Table 1). Nearly all of the P values were highly significant, indicating that the effect of increasing HU on the outcomes is not the same across levels of the VAT/SAT ratio (or that the association of higher VAT/SAT ratio on the endpoints is not the same across levels of SAT HU).
We next used multivariable logistic regression to assess a high VAT/SAT ratio from dichotomous SAT density. In men, high SAT density was associated with a high VAT/SAT ratio (odds ratio [OR], 1.83; 95% confidence interval [CI], 1.49–2.24; P < .0001). Results were similar after adjustment for BMI (OR, 1.70; 95% CI, 1.37–2.09; P < .0001). In contrast, in women, we found that high SAT density was not associated with a high VAT/SAT ratio (OR, 0.89; 95% CI, 0.72–1.11; P = .31), even after adjustment for BMI (OR, 0.96; 95% CI, 0.77–1.21; P = .74).
Clinical risk factor ORs by SAT density and VAT/SAT ratio grouping
As hypothesized, for both women and men, the high density/high ratio group had the most adverse CVD risk factors compared with the referent group. For example, in women, the high density/high ratio group had an increased odds of hypertension (OR, 1.6; 95% CI, 1.1–2.4; P = .009) (Table 3) and diabetes (OR, 6.7; 95% CI, 2.6–17.6; P = .0001) (Table 3) and fasting glucose (4.8 mg/mL; 95% CI, 2.2–7.4; P = .0004) (Table 4). When we further adjusted our models for absolute SAT volume, results were generally stronger. For example, the high density/high ratio group had increased ORs for hypertension (OR, 2.5; 95% CI, 1.7–3.7; P < .0001) (Table 3) and diabetes (OR, 10.8; 95% CI, 4.1–29.0; P < .0001) (Table 3) and fasting glucose (8.9 mg; 95% CI, 6.2–11.5; P < .0001) (Table 4). In men, similar patterns were observed, although the results were less striking. However, with additional adjustment for SAT volume, results were generally stronger (hypertension OR, 1.5; 95% CI, 1.0–2.1; P = .03; diabetes OR, 4.8; 95% CI, 2.6–9.2; P < .0001; fasting glucose, 10.0 mg; 95% CI, 6.5–13.5; P < .0001).
Table 3.
Logistic Regression for Selected CVD Risk Factors Compared With Those in the Referent Group (Low Density/Low Ratio)
| Dependent Variable | High Density/High Ratio |
High Density/Low Ratio |
Low Density/High Ratio |
|||
|---|---|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | OR (95% CI) | P | |
| Women | ||||||
| Low HDL | ||||||
| MV | 1.6 (1.1 to 2.2) | .01 | 0.6 (0.4 to 0.9) | .006 | 1.7 (1.2 to 2.3) | .003 |
| MV + SAT | 2.4 (1.7 to 3.6) | <.0001 | 0.9 (0.6 to 1.4) | .70 | 2.0 (1.4 to 2.8) | <.0001 |
| High triglycerides | ||||||
| MV | 1.5 (1.1 to 2.1) | .02 | 0.4 (0.2 to 0.6) | <.0001 | 1.7 (1.2 to 2.3) | .003 |
| MV + SAT | 2.2 (1.5 to 3.5) | <.0001 | 0.5 (0.3 to 0.9) | .009 | 2.0 (1.4 to 2.8) | .0001 |
| High glucose | ||||||
| MV | 1.2 (0.8 to 1.8) | .40 | 0.3 (0.2 to 0.6) | <.0001 | 1.6 (1.1 to 2.4) | .008 |
| MV + SAT | 1.9 (1.3 to 2.9) | .003 | 0.6 (0.3 to 1.0) | .04 | 2.1 (1.4 to 3.0) | .0002 |
| Diabetes | ||||||
| MV | 6.7 (2.6 to 17.6) | .0001 | 2.4 (0.8–6.9) | .12 | 2.0 (0.7 to 5.6) | .18 |
| MV + SAT | 10.8 (4.1 to 29.0) | <.0001 | 4.2 (1.4–12.5) | .01 | 2.6 (0.9 to 7.2) | .08 |
| Hypertension | ||||||
| MV | 1.6 (1.1 to 2.4) | .009 | 0.6 (0.4 to 0.9) | .01 | 1.8 (1.3 to 2.5) | .001 |
| MV + SAT | 2.5 (1.7 to 3.7) | <.001 | 0.9 (0.6 to 1.4) | .63 | 2.1 (1.5 to 3.1) | <.0001 |
| CVD | ||||||
| MV | 3.9 (1.3 to 11.6) | .02 | 0.8 (0.2 to 3.8) | .82 | 2.9 (1.0 to 8.8) | .06 |
| MV + SAT | 3.4 (1.1 to 10.5) | .03 | 0.7 (0.2 to 3.4) | .68 | 2.8 (0.9 to 8.4) | .07 |
| Metabolic syndrome | ||||||
| MV | 1.7 (1.2 to 2.4) | .005 | 0.3 (0.2 to 0.5) | <.0001 | 2.1 (1.5 to 3.0) | <.0001 |
| MV + SAT | 4.8 (3.1 to 7.4) | <.0001 | 0.8 (0.4 to 1.3) | .30 | 3.8 (2.6 to 5.6) | <.0001 |
| Men | ||||||
| Low HDL | ||||||
| MV | 1.5 (1.1 to 2.0) | .005 | 0.4 (0.3 to 0.6) | <.0001 | 1.3 (0.9 to 1.7) | .11 |
| MV + SAT | 2.5 (1.8 to 3.5) | <.0001 | 0.6 (0.4 to 0.9) | .02 | 1.7 (1.2 to 2.4) | .001 |
| High triglycerides | ||||||
| MV | 1.2 (0.9 to 1.5) | .26 | 0.4 (0.3 to 0.6) | <.0001 | 1.5 (1.2 to 2.0) | .003 |
| MV + SAT | 1.9 (1.4 to 2.7) | <.0001 | .6 (0.4 to 0.9) | .005 | 2.0 (1.5 to 2.8) | <.0001 |
| High glucose | ||||||
| MV | 0.7 (0.5 to 0.9) | .003 | 0.5 (0.3 to 0.6) | <.0001 | 1.1 (0.8 to 1.5) | .47 |
| MV + SAT | 1.0 (0.7 to 1.4) | .98 | 0.6 (0.5 to 0.9) | .009 | 1.4 (1.0 to 1.9) | .03 |
| Diabetes | ||||||
| MV | 1.6 (0.9 to 2.6) | .09 | 1.2 (0.7 to 2.2) | .47 | 0.5 (0.2 to 1.0) | .048 |
| MV + SAT | 4.8 (2.6 to 9.2) | <.0001 | 2.6 (1.4 to 4.9) | .004 | 1.0 (0.5 to 2.0) | .91 |
| Hypertension | ||||||
| MV | 0.8 (0.6 to 1.0) | .08 | 0.4 (0.3 to 0.5) | <.0001 | 0.8 (0.6 to 1.1) | .13 |
| MV + SAT | 1.5 (1.0 to 2.1) | .03 | 0.6 (0.4 to 0.9) | .01 | 1.1 (0.8 to 1.6) | .48 |
| CVD | ||||||
| MV | 1.2 (0.7 to 2.0) | .62 | 0.9 (0.5 to 1.7) | .76 | 1.3 (0.7 to 2.3) | .41 |
| MV + SAT | 1.8 (0.9 to 3.5) | .07 | 1.2 (0.6 to 2.5) | .57 | 1.7 (0.9 to 3.1) | .11 |
| Metabolic syndrome | ||||||
| MV | 0.6 (0.4 to 0.8) | .0002 | 0.3 (0.2 to 0.4) | <.0001 | 1.0 (0.8 to 1.4) | .86 |
| MV + SAT | 2.5 (1.7 to 3.6) | <.0001 | 0.6 (0.4 to 0.9) | .02 | 2.3 (1.7 to 3.2) | <.0001 |
Estimates give differences in odds of the dependent variable in each group compared with the referent group (low SAT HU/low ratio). The multivariable (MV) model adjusts for age, current smoking status, alcohol use, physical activity, and hormone replacement therapy (models in women only). The MV + SAT model also adjusts for the absolute volume of SAT.
Table 4.
Multivariable Linear Regression for Selected CVD Risk Factors Compared to the Referent Group (Low Density/Low Ratio)
| Dependent Variable | High Density/High Ratio |
High Density/Low Ratio |
Low Density/High Ratio |
|||
|---|---|---|---|---|---|---|
| OR (95% CI) | P | OR (95% CI) | P | OR (95% CI) | P | |
| Women | ||||||
| Total cholesterol, mg/dL | ||||||
| MV | −7.0 (−12.2 to −1.8) | .008 | −12.4 (−17.3 to −7.5) | <.0001 | 3.2 (−1.8 to 8.2) | .21 |
| MV + SAT | −4.8 (−10.2 to 0.6) | .08 | −10.0 (−15.2 to −4.8) | .0002 | 4.0 (−1.1 to 9.1) | .12 |
| HDL cholesterol, mg/dL | ||||||
| MV | −3.2 (−5.6 to −0.8) | .01 | 6.3 (4.1 to 8.6) | <.0001 | −3.7 (−6.0 to −1.3) | .002 |
| MV + SAT | −6.7 (−9.1 to −4.3) | <.0001 | 2.5 (0.1 to 4.8) | .04 | −5.0 (−7.2 to −2.7) | <.0001 |
| Log triglycerides, mg/dL | ||||||
| MV | 0.1 (−0.0 to 0.1) | .06 | −0.3 (−0.3 to −0.2) | <.0001 | 0.2 (0.1 to 0.2) | <.0001 |
| MV + SAT | 0.2 (0.1 to 0.3) | <.0001 | −0.1 (−0.2 to −0.1) | .0003 | 0.2 (0.1 to 0.3) | <.0001 |
| Fasting glucose, mg/dL | ||||||
| MV | 4.8 (2.2 to 7.4) | .0004 | −2.5 (−5.0 to 0.0) | .051 | 2.8 (0.3 to 5.4) | .03 |
| MV + SAT | 8.9 (6.2 to 11.5) | <.0001 | 2.1 (−0.5 to 4.6) | .11 | 4.3 (1.9 to 6.8) | .0006 |
| Systolic blood pressure, mm Hg | ||||||
| MV | 0.4 (−2.0 to 2.8) | .75 | −5.3 (−7.5 to −3.0) | <.0001 | 2.8 (0.5 to 5.1) | .02 |
| MV + SAT | 3.8 (1.3 to 6.2) | .002 | −1.5 (−3.8 to 0.8) | .21 | 4.0 (1.8 to 6.3) | .0005 |
| Diastolic blood pressure, mm Hg | ||||||
| MV | −0.4 (−1.7 to 1.0) | .57 | −3.5 (−4.8 to −2.3) | <.0001 | 2.0 (0.7 to 3.3) | .003 |
| MV + SAT | 1.4 (0.0 to 2.8) | .04 | −1.5 (−2.9 to −0.2) | .02 | 2.6 (1.4 to 3.9) | <.0001 |
| VAT/SAT ratio | ||||||
| MV | 0.28 (0.26 to 0.31) | <.0001 | −0.01 (−0.03 to 0.01) | .26 | 0.23 (0.20 to 0.25) | <.0001 |
| MV + SAT | 0.47 (0.44 to 0.51) | <.0001 | −0.08 (−0.11 to −0.04) | .006 | 0.41 (0.38 to 0.45) | <.0001 |
| BMI, kg/m2 | ||||||
| MV | −1.1 (−1.9 to −0.2) | .01 | −3.3 (−4.1 to −2.5) | <.0001 | 0.1 (−0.8 to 0.9) | .88 |
| MV + SAT | 3.3 (2.9 to 3.7) | <.0001 | 1.6 (1.2 to 2.0) | <.0001 | 1.7 (1.3 to 2.1) | <.0001 |
| Waist circumference | ||||||
| MV | −3.5 (−5.7 to −1.3) | .002 | −9.8 (−11.8 to −7.7) | <.0001 | 1.1 (−1.0 to 3.2) | .30 |
| MV + SAT | 7.7 (6.6 to 8.7) | <.0001 | 2.7 (1.7 to 3.7) | <.0001 | 5.4 (4.4 to 6.4) | <.0001 |
| SAT, cm3 | ||||||
| MV | −1231.0 (−1441.0 to −1021.0) | <.0001 | −1368.4 (−1567 to −1169.9) | <.0001 | −457.4 (−660.6 to −254.2) | <.0001 |
| MV + SAT | NA | NA | NA | NA | NA | NA |
| VAT, cm3 | ||||||
| MV | 341.6 (239.5 to 443.6) | <.0001 | −436.0 (−532.5 to −339.5) | <.0001 | 643.2 (544.4 to 741.9) | <.0001 |
| MV + SAT | 819.5 (755.5 to 883.5) | <.0001 | 95.3 (33.8 to 156.7) | .002 | 820.8 (761.0 to 880.6) | <.0001 |
| Subcutaneous HU, cm3 | ||||||
| MV | 6.9 (6.4 to 7.5) | <.0001 | 7.0 (6.5 to 7.5) | <.0001 | 0.5 (−0.1 to 1.0) | .10 |
| MV + SAT | 5.8 (5.2 to 6.3) | <.0001 | 5.7 (5.2 to 6.3) | <.0001 | 0.0 (−0.5 to 0.6) | .90 |
| Visceral HU, cm3 | ||||||
| MV | −0.1 (−0.6 to 0.5) | .80 | 3.4 (2.9 to 3.9) | <.0001 | −3.8 (−4.3 to −3.3) | <.0001 |
| MV + SAT | −1.6 (−2.1 to −1.2) | <.0001 | 1.7 (1.2 to 2.1) | <.0001 | −4.4 (−4.8 to −3.9) | <.0001 |
| Men | ||||||
| Total cholesterol, mg/dL | ||||||
| MV | −1.9 (−6.2 to 2.5) | .40 | −6.5 (−11.1 to −1.8) | .006 | 4.8 (0.1 to 9.5) | .05 |
| MV + SAT | −2.4 (−7.6 to 2.7) | .35 | −6.9 (−12.0 to −1.8) | .008 | 4.5 (−0.5 to 9.4) | .08 |
| HDL cholesterol, mg/dL | ||||||
| MV | −0.9 (−2.4 to 0.6) | .24 | 5.0 (3.4 to 6.7) | <.0001 | −1.4 (−3.1 to 0.2) | .10 |
| MV + SAT | −4.8 (−6.5 to −3.0) | <.0001 | 2.1 (0.3 to 3.8) | .02 | −3.5 (−5.2 to −1.8) | <.0001 |
| Log triglycerides, mg/dL | ||||||
| MV | 0.1 (−0.0 to 0.1) | .12 | −0.3 (−0.3 to −0.2) | <.0001 | 0.1 (0.1 to 0.2) | .0004 |
| MV + SAT | 0.2 (0.1 to 0.3) | <.0001 | −0.1 (−0.2 to −0.0) | .002 | 0.2 (0.2 to 0.3) | <.0001 |
| Fasting glucose, mg/dL | ||||||
| MV | 3.0 (−0.1 to 6.0) | .06 | 0.1 (−3.2 to 3.3) | .97 | −1.0 (−4.3 to 2.3) | .54 |
| MV + SAT | 10.0 (6.5 to 13.5) | <.0001 | 5.5 (2.0 to 9.0) | .002 | 2.8 (−0.6 to 6.2) | .10 |
| Systolic blood pressure, mm Hg | ||||||
| MV | −1.2 (−3.0 to 0.6) | .18 | −4.2 (−6.1 to −2.3) | <.001 | 0.2 (−1.7 to 2.1) | .83 |
| MV + SAT | 2.3 (0.3 to 4.4) | .03 | −1.5 (−3.5 to 0.5) | .15 | 2.1 (0.2 to 4.1) | .03 |
| Diastolic blood pressure, mm Hg | ||||||
| MV | −1.6 (−2.7 to 0.4) | .008 | −3.8 (−5.0 to −2.5) | <.0001 | 0.1 (−1.1 to 1.4) | .82 |
| MV + SAT | 1.0 (−0.4 to 2.3) | .16 | −1.8 (−3.1 to −0.5) | .007 | 1.5 (0.2 to 2.8) | .02 |
| VAT/SAT ratio | ||||||
| MV | 0.58 (0.54 to 0.61) | <.0001 | 0.00 (−0.03 to 0.04) | .92 | 0.47 (0.43 to 0.50) | <.0001 |
| MV + SAT | 0.47 (0.44 to 0.51) | <.0001 | −0.08 (−0.11 to −0.04) | <.0001 | 0.41 (0.38 to 0.45) | <.0001 |
| BMI, kg/m2 | ||||||
| MV | −3.0 (−3.6 to −2.4) | <.0001 | −3.1 (−3.7 to −2.5) | <.0001 | −1.3 (−2.0 to −0.7) | <.0001 |
| MV + SAT | 3.0 (2.7 to 3.4) | <.0001 | 1.5 (1.2 to 1.9) | <.0001 | 1.9 (1.6 to 2.3) | <.0001 |
| Waist circumference | ||||||
| MV | −9.5 (−10.9 to −8.1) | <.0001 | −9.8 (−11.3 to −8.3) | <.0001 | −4.1 (−5.6 to −2.6) | <.0001 |
| MV + SAT | 6.2 (5.4 to 6.9) | <.0001 | 2.4 (1.7 to 3.2) | <.0001 | 4.5 (3.7 to 5.2) | <.0001 |
| SAT, cm3 | ||||||
| MV | −1681.9 (−1814.7 to - 1549.1) | <.0001 | −1302.7 (−1443.7 to −1161.7) | <.001 | −919.7 (−1063.1 to −776.3) | <.0001 |
| MV + SAT | NA | NA | NA | NA | NA | NA |
| VAT, cm3 | ||||||
| MV | 63.1 (−43.2 to 169.4) | .25 | −769.7 (−882.5 to −656.8) | <.0001 | 685.5 (570.7 to 800.3) | <.0001 |
| MV + SAT | 1019.3 (931.8 to 1106.8) | <.0001 | −29.0 (−116.0 to 57.9) | .51 | 1208.4 (1123.8 to 1293.0) | <.0001 |
| Subcutaneous HU, cm3 | ||||||
| MV | 6.4 (6.0 to 6.8) | <.0001 | 6.7 (6.2 to 7.1) | <.0001 | 0.7 (0.3 to 1.2) | .001 |
| MV + SAT | 4.5 (4.0 to 4.9) | <.0001 | 5.2 (4.7 to 5.6) | <.0001 | −0.3 (−0.7 to 0.1) | .17 |
| Visceral HU, cm3 | ||||||
| MV | 1.2 (0.7 to 1.6) | <.0001 | 5.5 (5.0 to 6.0) | <.0001 | −2.7 (−3.2 to −2.2) | <.0001 |
| MV + SAT | −1.1 (−1.6 to −0.7) | <.0001 | 3.7 (3.2 to 4.2) | <.0001 | −4.02 (−4.5 to −3.5) | <.0001 |
Estimates give differences in odds of the dependent variable in each group compared to the referent group (low SAT HU/low ratio). The multivariable (MV) model was adjusted for age, current smoking status, alcohol use, physical activity, and hormone replacement therapy (models in women only). The MV + SAT model also adjusts for the absolute volume of SAT. NA, not applicable.
The low density/high ratio group had a similar but less adverse profile than the high density/high ratio group. For example, in women, compared with the referent group, the low density/high ratio group had increased hypertension (OR, 1.8; 95% CI, 1.3–2.5; P = .001) and high triglycerides (OR, 1.7; 95% CI, 1.2–2.3; P = .003). After SAT adjustment, findings were generally stronger.
The high density/low ratio group demonstrated more variable results with both multivariable and SAT-adjusted multivariable models. This group generally had a less adverse CVD risk profile than the referent group in both sexes. When SAT-adjusted, this group still had the least adverse CVD risk profile compared with that for the referent group, but the difference was attenuated. There was a notable exception of greatly increased risk of diabetes with this group compared with that in the referent group after SAT adjustment (women OR, 4.2; 95% CI, 1.4–12.5; P = .01; men OR, 2.6; 95% CI, 1.4–4.9; P = .004).
Finally, when multivariable models were additionally adjusted for absolute VAT volume in addition to absolute SAT volume, most findings were attenuated. Notably, there was an exception with diabetes in women in the high density/high ratio group compared with those in the referent group (OR, 3.5; 95% CI, 1.1–10.4; P = .03) and men (OR, 2.7; 95% CI, 1.3–5.5; P = .008; data not shown).
Discussion
Principal findings
Our principal findings are 3-fold. First, high SAT density was associated with a high VAT/SAT ratio in men but not in women. Second, the high density/high ratio group was associated with the most adverse CVD risk factor levels in women and men. Finally, these results were generally strengthened after adjustment for absolute SAT volume. Taken together, these findings suggest that increased SAT density along with a high VAT/SAT ratio is associated with a heavy burden of CVD risk factors.
In the context of the current literature
In a state of positive caloric balance, excess free fatty acids are stored in adipose tissue, preferentially in the subcutaneous adipose depot (26). Remodeling of adipose tissue occurs during times of caloric excess via hypertrophy of adipocytes and hyperplasia of preadipocytes into mature adipocytes, as well as changes in the number of stromal and immune cells involved with adipose tissue function (27–29). Adipocyte tissue remodeling is a constant process that may become pathologic in an obese state, with features such as hypoxia, reduced angiogenesis, and invasion of immune cells with a subsequent proinflammatory response and overproduction of extracellular matrix (9–13, 30). When adipose tissue reaches maximum expansion potential, excess free fatty acids overflow to ectopic sites (31). Alternatively, the maximum expansion potential may be reduced, leading to lower effective fat storage and an increased propensity for fat spillover to ectopic sites.
The reduction in maximum SAT expansion potential may be due to at least 5 interconnected mechanisms seen in pathologic obesity: (1) reduced angiogenesis, (2) inflammation, (3) fibrosis, (4) fibrosis-driven physical remodeling, and (5) adipocyte-based or metabolic possibility involving cellular lipid trafficking. Excess caloric intake drives the rapid expansion of adipose tissue. Adipose tissue can become hypoxic if the vasculature cannot keep pace with tissue growth, leading to increased levels of hypoxia-inducible factor 1α (HIF1α) (32–34). HIF1α mediates induction of macrophage proinflammatory cytokines as well as increased expression of fibrotic response genes (32, 34). Mice lacking a form of collagen highly enriched in adipose tissue IV exhibit uninhibited adipocyte expansion, suggesting the potential importance of fibrosis in reducing SAT expansion potential (35, 36). Mechanical compression of human adipose tissue in three-dimensional culture, a model for the physical role of fibrosis on adjacent adipocytes, further induces proinflammatory and fibrotic gene expression of adipocytes (37). Finally, lower attenuation may be a marker of adipose tissue that is lipid dense (16), which may determine adipocyte size (38) and serve as an overflow mechanism for excess free fatty acids. Thus, increasing hypoxia, inflammation, fibrosis and fibrosis-driven remodeling, and lipid trafficking may drive a reduction in SAT storage capacity to adequately match storage needs, although the physiologic mechanisms are not completely known.
In the present study, we used high SAT density as a marker of dysfunctional adipose tissue. There are several potential hypotheses to explain our observations. We hypothesized that as SAT becomes increasingly fibrotic, adipocytes cannot expand in times of caloric excess, causing fat spillover into the ectopic VAT depot. It is now well established that an excess of VAT in obese and nonobese individuals is associated with increased cardiovascular and metabolic abnormalities (39–43). A small number of studies have reported an association between the VAT/SAT ratio and CVD risk factors, with an increasing VAT/SAT ratio associated with increasing CVD risk (2, 3, 5, 18, 19). We now expand the literature by demonstrating an association between SAT density and the VAT/SAT ratio and that, jointly, those with high density and a high ratio carry a high burden of CVD risk factors. In particular, our results for diabetes are notable, even after we statistically adjusted for variation across the four categories in SAT and VAT, and may be consistent with a mechanism of increased adipose tissue fibrosis leading to diabetes. In mouse models, overexpression of HIF1α initiated adipose tissue fibrosis and insulin resistance (33). Moreover, mice lacking collagen VI, the major collagen form seen in adipose tissue fibrosis, exhibit increased insulin sensitivity, suggesting the importance of fibrosis in determining insulin resistance (35, 36). Furthermore, the VAT/SAT ratio was positively associated with fasting glucose in a group of adults and with hepatic insulin resistance in a group of men with type 2 diabetes (whereas VAT or SAT alone was not) (2, 3). Thus, this work improves the framework with which to interpret abdominal CT scan data as related to CVD risk factors. However, it is important to note that these potential mechanisms are speculative, given the lack of definitive knowledge of the histologic correlates of SAT attenuation.
Implications for future work
Our findings suggest that the cellular characteristics of adipose tissue, indirectly measured using fat density, may be an important mechanism in association with relative fat deposition and concomitant metabolic risk. Whereas our data suggest an association between SAT density and the VAT/SAT ratio, the mechanism linking these 2 measurements remains uncertain. Furthermore, it is not clear why individuals with high density and a high VAT/SAT ratio have the most adverse CVD risk factor profile. These questions warrant further investigation, particularly in studies that correlate adipose tissue imaging with invasive biopsy of adipose tissue and subsequent histologic and biochemical work to elucidate the underlying mechanisms of adipose tissue remodeling.
Strengths and limitations
The strengths of our study include a highly specific and reproducible CT-derived assessment of fat volume and quantity. In addition, we made use of the large, well-defined Framingham Heart Study cohort that undergoes regular examinations and provides high-quality biochemical and clinical data. Some limitations of this study include the cross-sectional design, which limits inferences of temporality. Moreover, the observational nature of the data limits causality to be concluded from our results. Our cohort is primarily non-Hispanic white, which may limit the generalizability of our findings. Finally, the underlying biological mechanism of fat density is not well defined; therefore, further laboratory studies are needed.
In conclusion, increased SAT density is associated with an increased VAT/SAT ratio that is a marker of a higher propensity to store excess fat viscerally. High SAT density and a high VAT/SAT ratio are jointly characterized by an increased burden of CVD risk factors.
Acknowledgments
This work was supported by in part using resources and data from the Framingham Heart Study of the National Heart, Lung and Blood Institute of the National Institutes of Health and Boston University School of Medicine (Grant N01-HC-25195).
Disclosure Summary: Dr Alison Pedley is a Merck employee who owns Merck stock. The other authors have nothing to disclose.
Footnotes
- BMI
- body mass index
- CI
- confidence interval
- CT
- computed topography
- CVD
- cardiovascular disease
- HIFα
- hypoxia-inducible factor 1α
- HU
- Hounsfield units
- MDCT
- multidetector computed tomography
- OR
- odds ratio
- SAT
- subcutaneous adipose tissue
- VAT
- visceral adipose tissue.
References
- 1. Fox CS, Massaro JM, Hoffmann U, et al. Abdominal visceral and subcutaneous adipose tissue compartments: association with metabolic risk factors in the Framingham Heart Study. Circulation. 2007;116:39–48. [DOI] [PubMed] [Google Scholar]
- 2. Miyazaki Y, DeFronzo RA. Visceral fat dominant distribution in male type 2 diabetic patients is closely related to hepatic insulin resistance, irrespective of body type. Cardiovasc Diabetol. 2009;8:44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Gastaldelli A, Sironi AM, Ciociaro D, et al. Visceral fat and beta cell function in non-diabetic humans. Diabetologia. 2005;48:2090–2096. [DOI] [PubMed] [Google Scholar]
- 4. Goodpaster BH, Krishnaswami S, Harris TB, et al. Obesity, regional body fat distribution, and the metabolic syndrome in older men and women. Arch Intern Med. 2005;165:777–783. [DOI] [PubMed] [Google Scholar]
- 5. Kim S, Cho B, Lee H, et al. Distribution of abdominal visceral and subcutaneous adipose tissue and metabolic syndrome in a Korean population. Diabetes Care. 2011;34:504–506. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Nieves DJ, Cnop M, Retzlaff B, et al. The atherogenic lipoprotein profile associated with obesity and insulin resistance is largely attributable to intra-abdominal fat. Diabetes. 2003;52:172–179. [DOI] [PubMed] [Google Scholar]
- 7. Preis SR, Massaro JM, Robins SJ, et al. Abdominal subcutaneous and visceral adipose tissue and insulin resistance in the Framingham Heart Study. Obesity (Silver Spring). 2010;18:2191–2198. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Weyer C, Foley JE, Bogardus C, Tataranni PA, Pratley RE. Enlarged subcutaneous abdominal adipocyte size, but not obesity itself, predicts type II diabetes independent of insulin resistance. Diabetologia. 2000;43:1498–1506. [DOI] [PubMed] [Google Scholar]
- 9. Weisberg SP, McCann D, Desai M, Rosenbaum M, Leibel RL, Ferrante AW., Jr Obesity is associated with macrophage accumulation in adipose tissue. J Clin Invest. 2003;112:1796–1808. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Pasarica M, Sereda OR, Redman LM, et al. Reduced adipose tissue oxygenation in human obesity: evidence for rarefaction, macrophage chemotaxis, and inflammation without an angiogenic response. Diabetes. 2009;58:718–725. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Apovian CM, Bigornia S, Mott M, et al. Adipose macrophage infiltration is associated with insulin resistance and vascular endothelial dysfunction in obese subjects. Arterioscler Thromb Vasc Biol. 2008;28:1654–1659. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Farb MG, Ganley-Leal L, Mott M, et al. Arteriolar function in visceral adipose tissue is impaired in human obesity. Arterioscler Thromb Vasc Biol. 2012;32:467–473. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Gealekman O, Guseva N, Hartigan C, et al. Depot-specific differences and insufficient subcutaneous adipose tissue angiogenesis in human obesity. Circulation. 2011;123:186–194. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Divoux A, Tordjman J, Lacasa D, et al. Fibrosis in human adipose tissue: composition, distribution, and link with lipid metabolism and fat mass loss. Diabetes. 2010;59:2817–2825. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Rosenquist KJ, Pedley A, Massaro JM, et al. Visceral and subcutaneous fat quality and cardiometabolic risk. JACC Cardiovasc Imaging. 2013;6:762–771. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Baba S, Jacene HA, Engles JM, Honda H, Wahl RL. CT Hounsfield units of brown adipose tissue increase with activation: preclinical and clinical studies. J Nucl Med. 2010;51:246–250. [DOI] [PubMed] [Google Scholar]
- 17. Hu HH, Chung SA, Nayak KS, Jackson HA, Gilsanz V. Differential computed tomographic attenuation of metabolically active and inactive adipose tissues: preliminary findings. J Comput Assist Tomogr. 2011;35:65–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Kaess BM, Pedley A, Massaro JM, Murabito J, Hoffmann U, Fox CS. The ratio of visceral to subcutaneous fat, a metric of body fat distribution, is a unique correlate of cardiometabolic risk. Diabetologia. 2012;55:2622–2630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19. He H, Ni Y, Chen J, et al. Sex difference in cardiometabolic risk profile and adiponectin expression in subjects with visceral fat obesity. Transl Res. 2010;155:71–77. [DOI] [PubMed] [Google Scholar]
- 20. Dawber TR, Kannel WB, Lyell LP. An approach to longitudinal studies in a community: the Framingham Study. Ann NY Acad Sci. 1963;107:539–556. [DOI] [PubMed] [Google Scholar]
- 21. Kannel WB, Feinleib M, McNamara PM, Garrison RJ, Castelli WP. An investigation of coronary heart disease in families. The Framingham offspring study. Am J Epidemiol. 1979;110:281–290. [DOI] [PubMed] [Google Scholar]
- 22. Splansky GL, Corey D, Yang Q, et al. The Third Generation Cohort of the National Heart, Lung, and Blood Institute's Framingham Heart Study: design, recruitment, and initial examination. Am J Epidemiol. 2007;165:1328–1335. [DOI] [PubMed] [Google Scholar]
- 23. Porter SA, Massaro JM, Hoffmann U, Vasan RS, O'Donnel CJ, Fox CS. Abdominal subcutaneous adipose tissue: a protective fat depot? Diabetes Care. 2009;32:1068–1075. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Maurovich-Horvat P, Massaro J, Fox CS, Moselewski F, O'Donnell CJ, Hoffmann U. Comparison of anthropometric, area- and volume-based assessment of abdominal subcutaneous and visceral adipose tissue volumes using multi-detector computed tomography. Int J Obes (Lond). 2007;31:500–506. [DOI] [PubMed] [Google Scholar]
- 25. Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults. Executive summary of the third report of the National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). JAMA. 2001;285:2486–2497. [DOI] [PubMed] [Google Scholar]
- 26. Després JP, Lemieux I. Abdominal obesity and metabolic syndrome. Nature. 2006;444:881–887. [DOI] [PubMed] [Google Scholar]
- 27. Sun K, Kusminski CM, Scherer PE. Adipose tissue remodeling and obesity. J Clin Invest. 2011;121:2094–2101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. Pessin JE, Kwon H. How does high-fat diet induce adipose tissue fibrosis? J Investig Med. 2012;60:1147–1150. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Suganami T, Tanaka M, Ogawa Y. Adipose tissue inflammation and ectopic lipid accumulation. Endocr J. 2012;59:849–857. [DOI] [PubMed] [Google Scholar]
- 30. Hotamisligil GS, Shargill NS, Spiegelman BM. Adipose expression of tumor necrosis factor-α: direct role in obesity-linked insulin resistance. Science. 1993;259:87–91. [DOI] [PubMed] [Google Scholar]
- 31. Gray SL, Vidal-Puig AJ. Adipose tissue expandability in the maintenance of metabolic homeostasis. Nutr Rev. 2007;65(6 Pt 2):S7–S12. [DOI] [PubMed] [Google Scholar]
- 32. Sun K, Tordjman J, Clément K, Scherer PE. Fibrosis and adipose tissue dysfunction. Cell Metab. 2013;18:470–477. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Halberg N, Khan T, Trujillo ME, et al. Hypoxia-inducible factor 1α induces fibrosis and insulin resistance in white adipose tissue. Mol Cell Biol. 2009;29:4467–4483. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34. Divoux A, Clément K. Architecture and the extracellular matrix: the still unappreciated components of the adipose tissue. Obes Rev. 2011;12:e494–e503. [DOI] [PubMed] [Google Scholar]
- 35. Khan T, Muise ES, Iyengar P, et al. Metabolic dysregulation and adipose tissue fibrosis: role of collagen VI. Mol Cell Biol. 2009;29:1575–1591. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Sun K, Park J, Gupta OT, et al. Endotrophin triggers adipose tissue fibrosis and metabolic dysfunction. Nat Commun. 2014;5:3485. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Pellegrinelli V, Heuvingh J, du Roure O, et al. Human adipocyte function is impacted by mechanical cues. J Pathol. 2014; 233:183–195. [DOI] [PubMed] [Google Scholar]
- 38. Wronska A, Kmiec Z. Structural and biochemical characteristics of various white adipose tissue depots. Acta Physiol (Oxf). 2012;205:194–208. [DOI] [PubMed] [Google Scholar]
- 39. Britton KA, Massaro JM, Murabito JM, Kreger BE, Hoffmann U, Fox CS. Body fat distribution, incident cardiovascular disease, cancer, and all-cause mortality. J Am Coll Cardiol. 2013;62:921–925. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40. Bastien M, Poirier P, Lemieux I, Després JP. Overview of epidemiology and contribution of obesity to cardiovascular disease. Prog Cardiovasc Dis. 2014;56:369–381. [DOI] [PubMed] [Google Scholar]
- 41. Liu J, Fox CS, Hickson DA, et al. Impact of abdominal visceral and subcutaneous adipose tissue on cardiometabolic risk factors: the Jackson Heart Study. J Clin Endocrinol Metab. 2010;95:5419–5426. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42. Shiina Y, Homma Y. Relationships between the visceral fat area on CT and coronary risk factor markers. Intern Med. 2013;52:1775–1780. [DOI] [PubMed] [Google Scholar]
- 43. Nazare JA, Smith JD, Borel AL, et al. Ethnic influences on the relations between abdominal subcutaneous and visceral adiposity, liver fat, and cardiometabolic risk profile: the International Study of Prediction of Intra-Abdominal Adiposity and Its Relationship With Cardiometabolic Risk/Intra-Abdominal Adiposity. Am J Clin Nutr. 2012;96:714–726. [DOI] [PubMed] [Google Scholar]