Abstract
Context:
Atherogenic dyslipidemia is frequently observed in persons with a greater amount of visceral adipose tissue (VAT). However, it is still uncertain whether VAT is independently associated with the future development of atherogenic dyslipidemia.
Objectives:
The aim of this study was to determine whether baseline and changes in VAT and subcutaneous adipose tissue (SAT) are associated with future development of atherogenic dyslipidemia independent of baseline lipid levels and standard anthropometric indices.
Design and Setting:
Community-based prospective cohort study with 5 years of follow-up.
Participants:
A total of 452 Japanese Americans (240 men, 212 women), aged 34–75 years were assessed at baseline and after 5 years of follow-up.
Main Outcome Measures:
Abdominal fat areas were measured by computed tomography. Atherogenic dyslipidemia was defined as one or more abnormalities in high-density lipoprotein (HDL) cholesterol, triglycerides, or non-HDL cholesterol levels.
Results:
Baseline VAT and change in VAT over 5 years were independently associated with log-transformed HDL cholesterol, log-transformed triglyceride, and non-HDL cholesterol after 5 years (standardized β = −0.126, 0.277, and 0.066 for baseline VAT, respectively, and −0.095, 0.223, and 0.090 for change in VAT, respectively). However, baseline and change in SAT were not associated with any future atherogenic lipid level. In multivariate logistic regression analysis, incremental change in VAT (odds ratio [95% confidence interval], 1.73 [1.20–2.48]; P = .003), triglycerides (4.01 [1.72–9.33]; P = .001), HDL cholesterol (0.32 [0.18–0.58]; P < .001), and non-HDL cholesterol (7.58 [4.43–12.95]; P < .001) were significantly associated with the future development of atherogenic dyslipidemia independent of age, sex, diastolic blood pressure, homeostasis model assessment insulin resistance, body mass index (BMI), change in BMI, SAT, and baseline atherogenic lipid levels.
Conclusion:
Baseline and change in VAT were independent predictors for future development of atherogenic dyslipidemia. However, BMI, waist circumference, and SAT were not associated with future development of atherogenic dyslipidemia.
Obesity is a state of excessive body fat accumulation and is associated with diverse cardiovascular disease risk factors including hypertension, type 2 diabetes, and atherogenic dyslipidemia (1). In a clinical setting, obesity is frequently defined by body mass index (BMI) of ≥30 kg/m2 and is a useful indicator of general adiposity. However, the relationship between the risk of metabolic disease does not depend solely on overall adiposity as might be measured by BMI but further varies (2) according to different regional fat accumulation. It is well known that visceral adipose tissue (VAT) is more closely associated with obesity-related metabolic diseases and cardiovascular disease than subcutaneous adipose tissue (SAT) (3, 4).
In our previous research, greater VAT was associated with higher risk of impaired glucose tolerance, independent of insulin resistance, insulin secretion, and other adipose depots (5). In addition, baseline and change in VAT over time were associated with the development of type 2 diabetes after adjusting for risk factors for diabetes, BMI, and changes in weight and SAT over 10 years of follow-up (6). Furthermore, greater VAT was associated with the development of hypertension, independent of other covariates and baseline systolic blood pressure (7). Similarly, in the Dallas Heart Study, excess visceral fat mass, but not BMI or SAT, was independently associated with incident prediabetes, type 2 diabetes, and hypertension during a median 7 years of follow-up (8, 9).
To date, numerous studies have shown cross-sectional associations between abdominal fat accumulation and atherogenic dyslipidemic phenotype characterized by increased triglycerides and/or decreased high-density lipoprotein (HDL) cholesterol, and the results showed that although both SAT and VAT are associated with atherogenic dyslipidemia, VAT remains more strongly and consistently associated with an atherogenic dyslipidemic profile (4, 10). However, it is still uncertain whether the accumulation of fat in specific locations of the abdomen is independently associated with the risk for future development of atherogenic dyslipidemia because results from longitudinal research on this question have not been published. Therefore, the aim of this study was to determine whether baseline and changes in VAT and SAT are associated with future development of atherogenic dyslipidemia and whether such associations are independent of baseline lipid levels and standard anthropometric indices including BMI and waist circumference.
Subjects and Methods
Study subjects
The study population consisted of men and women enrolled in the Japanese American Community Diabetes Study, a cohort of second-generation (Nisei) and third-generation (Sansei) Japanese Americans of 100% Japanese ancestry. A detailed description of the selection and recruitment of the study subjects has been published previously (11). In brief, study participants were selected as volunteers from a community-wide comprehensive mailing list and telephone directory that included nearly 95% of the Japanese American population in King County, Washington. Among the total of 658 subjects in the original cohort, 84 subjects who did not complete follow-up examinations 5–6 years after the baseline examination were excluded from this study. Additionally, 122 subjects were excluded for one of the following reasons: 1) computed tomography (CT) data to measure abdominal fat were not available at baseline and/or at follow-up (n = 34); 2) lipid data were not available at baseline and/or at follow-up (n = 2); 3) taking lipid-lowering medications at baseline and/or at follow-up (n = 64); or 4) treatment involving oral hypoglycemic agents and/or insulin therapy at baseline (n = 62). Thus, a total of 452 subjects (240 men, 212 women) aged 34–75 years (mean age, 52.6 y) were enrolled for the analysis. The study received approval from the University of Washington Human Subjects Division, and written informed consent was obtained from all subjects.
Clinical and laboratory examination
All evaluations were performed at the General Clinical Research Center, University of Washington. At baseline, a complete physical examination was performed, and personal medical history and lifestyle factors that possibly affect lipid levels, including cigarette smoking, alcohol consumption, and physical activity, were determined using a standardized questionnaire. Smoking was classified into three groups (current smoker, past smoker, and never smoked). Moderate alcohol intake was defined as consuming more than 6 g of ethanol per day (12). The Paffenbarger physical activity index questionnaire was used to determine physical activity level (usual kilocalories spent weekly) (13), and regular physical activity was defined as more than moderate intensity physical activity.
BMI was calculated as weight in kilograms divided by the square of the height in meters. Waist circumference was measured at the level of the umbilicus. Blood pressure was measured with a mercury sphygmomanometer read to the nearest 2 mm Hg with the subjects in a recumbent position. Systolic blood pressure was determined by the first perception of sound, and diastolic blood pressure was determined at the disappearance of sounds (fifth-phase Korotkoff). Average blood pressure was calculated from the second and third of three consecutive measurements.
Biochemical measurements were performed as reported previously (14). All blood samples were obtained after an overnight fast of 10 hours. Plasma glucose was measured by the hexokinase method using an autoanalyzer (Department of Laboratory Medicine, University of Washington, Seattle, Washington). Plasma insulin was measured by RIA (Immunoassay Core, Diabetes Research Center, University of Washington). To estimate insulin sensitivity, the homeostasis model assessment insulin resistance (HOMA-IR) index based on fasting glucose and insulin concentrations was used (15). Lipid and lipoprotein measurements were performed according to modified procedures of the Lipid Research Clinics (Northwest Lipid Research Laboratory, University of Washington). Single 10-mm slice CT scans were performed at the level of the umbilicus to measure cross-sectional fat areas (square centimeters) of SAT and VAT (16). The attenuation range for identification of fat was −250 to −50 Hounsfield Units.
Definitions
Atherogenic dyslipidemia comprises a triad of elevated triglycerides, decreased HDL cholesterol, and elevated small, dense low-density lipoprotein (LDL) particles (17). A previous study suggested that the non-HDL cholesterol level is highly correlated with small, dense LDL cholesterol (r = 0.582; P = .001) (18). Therefore, in this study, atherogenic dyslipidemia was defined as one or more of the following abnormalities in fasting lipid profiles with some modification of the National Cholesterol Education Program Adult Treatment Panel III criteria, and the cutoff for moderate cardiovascular risk group was adopted (17): 1) HDL cholesterol <40 mg/dL (1.03 mmol/L) in men and <50 mg/dL (1.29 mmol/L) in women; 2) triglycerides ≥150 mg/dL (1.69 mmol/L); or 3) non-HDL cholesterol ≥160 mg/dL (4.14 mmol/L) (17). Non-HDL cholesterol was calculated by total cholesterol minus HDL cholesterol. Of the total of 452 subjects at baseline, 183 had none of the aforementioned lipid abnormalities, namely “subjects without atherogenic dyslipidemia.”
A 75-g oral glucose tolerance test was performed to determine glucose tolerance status. Type 2 diabetes was defined by the presence of one of the following: 1) fasting glucose level ≥7.0 mmol/L; 2) treatment involving oral hypoglycemic agents or insulin therapy; or 3) 2-hour post glucose load ≥11.1 mmol/L (19). Hypertension was defined as having a systolic blood pressure ≥140 mm Hg, having a diastolic blood pressure ≥90 mm Hg, or taking antihypertensive medications. The presence of cardiovascular disease was diagnosed by a clinical history of one of the following: 1) coronary artery disease (acute myocardial infarction, angina, coronary artery bypass graft, or coronary angioplasty); 2) cerebrovascular disease (transient ischemic attack, carotid endarterectomy, atherosclerotic stroke, or nonatherosclerotic stroke); 3) peripheral artery occlusive disease (claudication or bypass surgery in lower extremities); or 4) abdominal, thoracic, or other types of aortic aneurysm.
Statistical analyses
Data are expressed as means ± SD for continuous measures or as proportions for categorical variables, except for skewed continuous variables, which are presented as the median (interquartile range, 25–75%). Differences between groups were tested by Student t test or Mann-Whitney U test for continuous variables, and the χ2 test or Fisher's exact test for categorical variables. Multiple stepwise linear regression analysis was used to determine independent associations between abdominal fat depots and standard anthropometric indices and their changes during follow-up and each of atherogenic lipid levels after 5 years. A variance inflation factor (VIF) >5.0 was used as an indicator of multicollinearity. Multiple logistic regression analysis with backward selection was used to determine whether abdominal fat depots and standard anthropometric indices and their changes during follow-up were independently associated with the development of atherogenic dyslipidemia in subjects without atherogenic dyslipidemia at baseline. Odds ratios (ORs) with 95% confidence intervals (CIs) were calculated for independent variables included in logistic models, with a 1-SD increment used for OR calculations for continuous measurements. The presence of interaction was assessed in multivariate models by testing the significance of first-order interaction terms. Presence of nonlinearity was assessed by the method of fractional polynomials (20). All statistical analyses were performed using PASW version 18.0 (SPSS Inc) and STATA version 12 (StataCorp). A P value of <.05 was considered significant.
Results
Table 1 shows the baseline characteristics of the study participants. The mean age of all subjects was 52.6 years, and the mean BMI was 24.2 kg/m2. Mean HDL cholesterol and triglyceride levels were within their respective normal ranges; however, the mean non-HDL cholesterol level was slightly increased. Subjects without atherogenic dyslipidemia at baseline were younger, less obese, more insulin sensitive, and had more favorable lipid profiles compared with subjects with atherogenic dyslipidemia at baseline. In terms of body fat composition, subjects without atherogenic dyslipidemia showed lesser amounts of SAT and VAT compared with subjects with atherogenic dyslipidemia at baseline.
Table 1.
Baseline Characteristics
| Total | Atherogenic Dyslipidemia Absent at Baseline | Atherogenic Dyslipidemia Present at Baseline | P | |
|---|---|---|---|---|
| n | 452 | 183 | 269 | |
| Age, y | 52.6 (11.9) | 48.1 (11.4) | 55.7 (11.3) | <.001 |
| Female, % | 212 (46.9) | 111 (60.7) | 101 (37.5) | <.001 |
| BMI, kg/m2 | 24.2 (3.2) | 23.1 (3.1) | 25.0 (3.1) | <.001 |
| Waist circumference, cm | 82.3 (10.4)a | 77.8 (10.2) | 85.5 (9.3) | <.001 |
| Alcohol consumption, g/d | 0.0 (0.0–5.0) | 0.0 (0.0–4.2) | 0.0 (0.0–5.5) | .35 |
| Moderate to heavy alcohol consumption, % | 101 (22.3) | 38 (20.8) | 63 (23.4) | .51 |
| Current smoking, % | 59 (13.1) | 23 (12.6) | 36 (13.4) | .80 |
| Physical activity, kcal/wk | 2247.0 (1421.1–3381.6) | 2462.0 (1564.3–3726.5) | 2100.0 (1382.3–3269.3) | .026 |
| Moderate to vigorous physical activity, % | 108 (23.9) | 47 (25.7) | 61 (22.7) | .46 |
| Diabetes, % | 71 (15.7) | 13 (7.1) | 58 (21.6) | <.001 |
| Hypertension, % | 163 (36.1) | 48 (26.2) | 115 (42.8) | <.001 |
| Cardiovascular disease, % | 14 (3.1) | 3 (1.6) | 11 (4.1) | .17 |
| Systolic blood pressure, mm Hg | 129.3 (18.0) | 124.0 (16.3) | 132.9 (18.3) | <.001 |
| Diastolic blood pressure, mm Hg | 76.8 (9.5) | 73.7 (8.9) | 78.9 (9.3) | <.001 |
| Fasting plasma glucose, mmol/L | 5.40 (1.09) | 5.04 (0.61) | 5.64 (1.27) | <.001 |
| 2-h plasma glucose, mmol/L | 8.02 (3.10) | 7.07 (1.86) | 8.67 (3.57) | <.001 |
| Fasting plasma insulin, pmol/L | 83.3 (62.5–111.1) | 69.5 (55.6–97.2) | 90.3 (69.5–125.0) | <.001 |
| HOMA-IR | 2.70 (1.98–3.90) | 2.28 (1.78–3.01) | 3.15 (2.21–4.55) | <.001 |
| Total cholesterol, mmol/L | 5.77 (1.02) | 5.09 (0.59) | 6.24 (0.99) | <.001 |
| LDL cholesterol, mmol/L | 3.58 (0.94) | 2.98 (0.50) | 3.99 (0.95) | <.001 |
| Triglycerides, mmol/L | 1.24 (0.86–1.76) | 0.88 (0.69–1.16) | 1.60 (1.18–2.42) | <.001 |
| HDL cholesterol, mmol/L | 1.51 (0.43) | 1.74 (0.41) | 1.35 (0.37) | <.001 |
| Non-HDL cholesterol, mmol/L | 4.26 (1.09) | 3.35 (0.52) | 4.89 (0.92) | <.001 |
| SAT, cm2 | 147.7 (105.9–193.7) | 130.2 (88.6–183.4) | 158.0 (116.8–206.8) | <.001 |
| VAT, cm2 | 79.3 (41.3–115.9) | 46.8 (25.0–78.9) | 95.9 (68.7–131.0) | <.001 |
Data are expressed as means ± SD, median (interquartile range, 25–75%), or proportions.
n = 444.
In all study subjects, of the clinical and biochemical variables, not surprisingly each baseline lipid level showed the strongest associations among all variables examined with future HDL cholesterol, triglyceride, or non-HDL cholesterol levels as judged by the magnitude of the standardized regression coefficient (Table 2). In addition, blood pressure, fasting and 2-hour glucose, and indices reflecting insulin resistance were inversely associated with HDL cholesterol level and positively associated with triglyceride and non-HDL cholesterol levels after 5 years. However, lifestyle factors possibly affecting lipid levels including alcohol consumption, current smoking, and physical activity did not show any associations with future atherogenic lipid levels. Regarding baseline and changes in standard anthropometric indices, baseline BMI and waist circumference were inversely associated with HDL cholesterol level and positively associated with triglyceride and non-HDL cholesterol levels after 5 years; however, changes in BMI and waist circumference were not associated with future atherogenic lipid levels except for an inverse association between change in waist circumference and HDL cholesterol level after 5 years. Although both baseline SAT and VAT were associated with future atherogenic lipid levels, VAT was a stronger predictor than SAT for each future atherogenic lipid level (Table 2). In addition, change in VAT was significantly associated with future log-transformed HDL cholesterol and triglyceride levels (standardized β = −0.128 and 0.131, respectively); however, change in SAT was not associated with any future atherogenic lipid levels (Table 2).
Table 2.
Univariate Linear Regression Analysis of the Prediction of Atherogenic Lipid Levels at 5-Year Follow-Up
| Log HDL Cholesterol | Log Triglyceride | Non-HDL Cholesterol | |
|---|---|---|---|
| Age | 0.103* | 0.073 | 0.249*** |
| Female sex | 0.372*** | −0.135** | −0.160** |
| BMI | −0.398*** | 0.262*** | 0.263*** |
| Waist circumference | −0.450*** | 0.273*** | 0.276*** |
| Alcohol consumption | 0.089 | 0.072 | −0.015 |
| Current smoking | −0.073 | 0.087 | 0.050 |
| Physical activity | 0.013 | −0.041 | 0.003 |
| Systolic blood pressure | −0.143** | 0.204*** | 0.189*** |
| Diastolic blood pressure | −0.202*** | 0.216*** | 0.194*** |
| Fasting plasma glucose | −0.158** | 0.126** | 0.151** |
| 2-h plasma glucose | −0.114* | 0.138** | 0.135** |
| Fasting plasma insulin | −0.294*** | 0.228*** | 0.131** |
| HOMA-IR | −0.276*** | 0.206*** | 0.139** |
| Total cholesterol | 0.036 | 0.213*** | 0.725*** |
| LDL cholesterol | −0.078 | 0.110* | 0.717*** |
| Triglycerides | −0.389*** | 0.441*** | 0.262*** |
| HDL cholesterol | 0.809*** | −0.397*** | −0.288*** |
| Non-HDL cholesterol | −0.288*** | 0.358*** | 0.796*** |
| SAT | −0.110* | 0.124** | 0.174*** |
| VAT | −0.386*** | 0.325*** | 0.331*** |
| Δ BMI | −0.083 | 0.060 | 0.008 |
| Δ Waist circumference | −0.139** | 0.086 | 0.081 |
| Δ SAT | −0.022 | −0.014 | −0.054 |
| Δ VAT | −0.128** | 0.131** | 0.069 |
Data are expressed as standardized β.
P < .05;
P < .01; and
P < .001.
Multiple stepwise linear regression models were fit to determine which of the following variables independently predict future atherogenic lipid levels: age; sex; lifestyle factors; diastolic blood pressure; fasting glucose and insulin levels; baseline atherogenic lipid levels, including HDL cholesterol, triglyceride, and non-HDL cholesterol; baseline and changes in BMI and waist circumference; and baseline and changes in SAT and VAT. In each model, because waist circumference caused multicollinearity with other variables (VIF >5.0 in all models), we excluded it in subsequent models. Besides baseline lipid levels, of the CT-measured abdominal fat depots, only baseline and change in VAT were consistently and independently associated with log-transformed HDL cholesterol, log-transformed triglyceride, and non-HDL cholesterol after 5 years (Table 3). Baseline and change in SAT were not statistically significant independent predictors for future atherogenic lipid levels in these multivariable models (Table 3). In terms of standard anthropometric indices, only change in BMI was associated with future HDL cholesterol and non-HDL cholesterol levels, but not with triglyceride level after 5 years (Table 3).
Table 3.
Multiple Stepwise Linear Regression Analysis of the Prediction of Atherogenic Lipid Levels at 5-Year Follow-Up
| Log HDL Cholesterol | Log Triglyceride | Non-HDL Cholesterol | |
|---|---|---|---|
| Age | 0.104** | ||
| Current smoking | 0.090* | ||
| Triglycerides | 0.365*** | ||
| HDL cholesterol | 0.769*** | ||
| Non-HDL cholesterol | 0.790*** | ||
| VAT | −0.126*** | 0.277*** | 0.066* |
| Δ BMI | −0.122*** | 0.096** | |
| Δ VAT | −0.095** | 0.223*** | 0.090** |
Data are expressed as standardized β. Age, sex, lifestyle factors, diastolic blood pressure, fasting glucose and insulin levels, corresponding baseline atherogenic lipid level, baseline and changes in BMI, waist circumference, SAT, and VAT were included in each model. Blank cells indicate insignificance in each model.
P < .05;
P < .01; and
P < .001.
We next examined whether baseline and changes in standard anthropometric indices and abdominal fat depots were associated with future development of atherogenic dyslipidemia in subjects without atherogenic dyslipidemia at baseline, that is, normal values for all three atherogenic lipid levels at baseline. In univariate logistic regression analysis, systolic and diastolic blood pressure, fasting insulin, HOMA-IR, and all lipid measurements showed positive associations with the development of atherogenic dyslipidemia, whereas baseline HDL cholesterol was inversely associated with this outcome (Table 4). Regarding standard anthropometric indices and abdominal fat depots, BMI, waist circumference, and SAT at baseline, but not VAT, as well as changes in BMI and in VAT were associated with future development of atherogenic dyslipidemia in subjects without atherogenic dyslipidemia at baseline (Table 4). In addition, we divided the subjects without atherogenic dyslipidemia at baseline into VAT change quartiles and determined the change in each of the three lipid measurements of interest by VAT change quartile. Consistent graded trends for all three lipid levels that mirrored the results of the univariate linear regression analyses were seen (Table 2). As the change in VAT increased, change in HDL cholesterol decreased, whereas changes in triglyceride and non-HDL cholesterol increased (Table 5). All continuous variables were found to be linear predictors of atherogenic dyslipidemia using the method of fractional polynomials, with the exception of HDL cholesterol, where (HDL cholesterol)−2 fit the model best.
Table 4.
Univariate Logistic Regression Predicting the Development of Atherogenic Dyslipidemia in Subjects Without This Condition at Baseline
| OR per 1 SD Increment (95% CI) | P | |
|---|---|---|
| Age | 1.16 (0.85–1.58) | .36 |
| Female | 0.75 (0.48–1.71) | .90 |
| BMI | 1.44 (1.05–1.97) | .024 |
| Waist circumference | 1.41 (1.03–1.94) | .033 |
| Moderate to heavy alcohol consumption | 0.54 (0.23–1.27) | .16 |
| Current smoking | 1.91 (0.78–4.66) | .16 |
| Moderate to vigorous physical activity | 0.95 (0.46–1.96) | .89 |
| Diabetes | 1.46 (0.46–4.67) | .53 |
| Hypertension | 1.53 (0.77–3.07) | .23 |
| Cardiovascular disease | 1.14 (0.10–12.80) | .92 |
| Systolic blood pressure | 1.45 (1.06–1.98) | .022 |
| Diastolic blood pressure | 1.51 (1.09–2.10) | .015 |
| Fasting plasma glucose | 1.33 (0.97–1.82) | .08 |
| 2-h plasma glucose | 1.19 (0.87–1.62) | .28 |
| Fasting plasma insulin | 1.41 (1.04–1.93) | .029 |
| HOMA-IR | 1.48 (1.06–2.00) | .019 |
| Total cholesterol | 1.62 (1.15–2.28) | .006 |
| LDL cholesterol | 2.29 (1.56–3.37) | <.001 |
| Triglycerides | 2.04 (1.45–2.87) | <.001 |
| HDL cholesterola | 0.23 (0.15–0.33) | <.001 |
| Non-HDL cholesterol | 2.97 (1.91–4.61) | <.001 |
| SAT | 1.48 (1.07–2.03) | .017 |
| VAT | 1.25 (0.92–1.70) | .15 |
| Δ BMI | 1.49 (1.07–2.08) | .019 |
| Δ Waist circumference | 1.23 (0.89–1.69) | .21 |
| Δ SAT | 1.19 (0.87–1.63) | .27 |
| Δ VAT | 1.47 (1.07–2.04) | .019 |
The best fitting model of HDL cholesterol was nonlinear (HDL cholesterol)−2, and the OR shown is for the comparison of 1.29 to 1.03 mmol/L.
Table 5.
Change in Atherogenic Lipid Levels According to the Quartile of VAT Change During 5 Years of Follow-Up, With the 4th Quartile Being the Greatest VAT Area Change
| 1st Quartile | 2nd Quartile | 3rd Quartile | 4th Quartile | P | |
|---|---|---|---|---|---|
| Δ HDL cholesterol | 0.0 (−5.0, 5.5) | −4.0 (−10.0, 1.0) | −5.0 (−9.0, 1.0) | −6.0 (−11.0, 0.0) | <.001 |
| Δ Triglyceride | −11.0 (−41.0, 27.5) | 5.0 (−20.5, 39.5) | 5.0 (−22.5, 35.5) | 20.0 (−19.5, 76.0) | .001 |
| Δ Non-HDL cholesterol | −11.0 (−30.0, 6.0) | −9.0 (−23.0, 8.0) | −4.0 (−20.5, 11.5) | 3.0 (−11.0, 15.0) | <.001 |
Data are expressed as median (interquartile range).
To determine which variables independently predicted future development of atherogenic dyslipidemia, a backwards selection logistic regression model was fit that included age and sex as well as variables found significantly related to the atherogenic dyslipidemia outcome in univariate analysis from Table 4 (diastolic blood pressure, HOMA-IR, BMI, SAT, change in BMI, change in VAT, HDL cholesterol, triglyceride, and non-HDL cholesterol levels). Waist circumference was excluded from this model due to its causing multicollinearity with other variables (VIF >5.0). Diastolic blood pressure and HOMA-IR are highly correlated with systolic blood pressure and fasting insulin level, respectively, but only the first two were chosen for inclusion in the backwards selection model due to their smaller P values in univariate analysis. In the final model, triglyceride, HDL cholesterol, non-HDL cholesterol, and change in VAT were associated with the risk of future development of atherogenic dyslipidemia after 5 years (Table 6). However, BMI, change in BMI, and SAT were not independently associated with future development of atherogenic dyslipidemia and were therefore not retained in the final model (Table 6). No significant interactions were observed between sex and each of the independent variables shown in the model in Table 6 in predicting the development of atherogenic dyslipidemia. The area under the receiver operating characteristic curve for the multivariable model II in Table 6 predicting atherogenic dyslipidemia was 0.91.
Table 6.
Multivariate Logistic Regression Predicting the Development of Atherogenic Dyslipidemia in Subjects Without This Condition at Baseline
| OR per 1− SD Increment (95% CI) | P | |
|---|---|---|
| Model I | ||
| Diastolic blood pressure | 1.41 (1.002–2.00) | .049 |
| SAT | 1.41 (1.10–1.97) | .041 |
| Δ VAT | 1.45 (1.04–2.03) | .029 |
| Model II | ||
| Triglycerides | 1.68 (1.12–2.52) | .012 |
| HDL cholesterola | 0.32 (0.18–0.58) | <.001 |
| Non-HDL cholesterol | 2.58 (1.63–4.08) | <.001 |
| Δ VAT | 1.60 (1.11–2.31) | .013 |
Model I: adjusted for age, sex, diastolic blood pressure, HOMA-IR, BMI, SAT, Δ BMI, and Δ VAT. Model II: adjusted for Model I + baseline triglyceride, HDL cholesterol, and non-HDL cholesterol.
The best fitting model of HDL cholesterol was nonlinear (HDL cholesterol)−2, and the OR shown is for the comparison of 1.29 to 1.03 mmol/L.
We further analyzed whether baseline or change in VAT predicted the regression of atherogenic dyslipidemia in subjects with this condition at baseline (n = 269). In univariate logistic regression analysis, baseline VAT (OR, 0.68; 95% CI, 0.47–0.99; P = .045) and change in VAT (OR, 0.63; 95% CI, 0.44–0.90; P = .010) were inversely associated with the regression of atherogenic dyslipidemia, defined by none of the abnormal features of atherogenic dyslipidemia after 5 years of follow-up in subjects with one or more features of atherogenic dyslipidemia at baseline.
Discussion
In the current prospective study performed in 452 Japanese American men and women, baseline and change in VAT were associated with atherogenic lipid levels after 5 years independently of baseline atherogenic lipid levels and other general and regional adiposity variables, including baseline BMI, waist circumference, and SAT measured at baseline and their changes over time. Change in BMI was associated with future HDL cholesterol and non-HDL cholesterol levels; however, no associations were noted between baseline BMI and any of future atherogenic lipid levels. On the other hand, baseline and changes in waist circumference and SAT were not associated with any future atherogenic lipid levels. Confined to subjects without baseline atherogenic dyslipidemia, only change in VAT was an independent predictor for future development of atherogenic dyslipidemia among all anthropometric indices and CT-measured abdominal fat depots, whether measured as changes over time or at baseline.
It has been suggested that fat in the VAT depot is more strongly associated with insulin resistance than in the SAT depot (21–24). The reason for this is still unclear, but VAT is highly innervated by β-adrenergic receptors and exhibits greater lipolytic activity than SAT (23), and thus, accelerated mobilization of free fatty acids into the portal circulation could promote insulin resistance in the liver. In addition, VAT might lead to insulin resistance and metabolic abnormalities through a release of inflammatory adipokines (24). In a cross-sectional analysis performed with 382 Caucasian and African American subjects with type 2 diabetes, larger VAT amount was associated with an atherogenic lipoprotein pattern characterized by higher VLDL and LDL particle number, larger VLDL particles, and smaller LDL and HDL particles independent of BMI and SAT. However, neither BMI nor SAT was independently related to lipoprotein parameters (25). In addition, among obese participants in the Dallas Heart Study, VAT was associated with HOMA-IR, lower adiponectin levels, smaller LDL and HDL particle size, larger VLDL size, and higher LDL and VLDL particle number, whereas SAT was not associated with dyslipidemia (26). In addition, there are some interventional studies showing the effect of reduction of VAT on dyslipidemia. In a study performed with 107 nondiabetic, viscerally obese men, a 1-year healthy eating/physical activity lifestyle modification program significantly decreased the volume of VAT measured by CT together with a modest reduction in fasting apolipoprotein B levels (27). In addition, a placebo-controlled trial of 1 year of treatment with rimonabant, a selective cannabinoid type 1 receptor antagonist for weight loss, in obese subjects with atherogenic dyslipidemia significantly reduced body weight and visceral fat and was accompanied by increases in HDL cholesterol level and LDL and HDL particle sizes and a reduction in triglyceride and apolipoprotein B levels (28).
To date, only one other study (the Hitachi Health Study) has examined prospectively the relationship between baseline abdominal fat depots and change in lipid levels, and the results partly agreed with our observation that change in VAT over a 3-year period was associated with changes in triglyceride and HDL cholesterol independently of change in body weight and waist circumference (29). However, to the best of our knowledge, ours is the first study that has prospectively investigated the associations between baseline as well as changes in standard anthropometric indices and areas of abdominal fat depots with future risk of atherogenic dyslipidemia. In addition, there were several differences between our study and the Hitachi Healthy Study. First, the Hitachi Healthy Study subjects had several important limitations, including being limited to men only and having a low 42% follow-up rate for the 3-year repeat CT imaging. Second, our study subjects were followed for a longer period of time, which enabled us to determine whether a more temporally sustained association existed between VAT and future atherogenic lipid levels. Third, we assessed as outcomes not only individual lipid levels but also the presence of atherogenic dyslipidemia. Fourth, in addition to adjustment for anthropometric indices and other covariates, our study adjusted for baseline atherogenic lipid levels because these were the most powerful predictors for future lipid levels and furthermore were strongly associated with baseline fat depots. This adjustment, which was not performed in the Hitachi Health Study, is important to remove potential confounding bias in estimating the associations between fat depots and future change in lipid levels. Lastly, in addition to elevated triglyceride and decreased HDL cholesterol levels, our definition for atherogenic dyslipidemia included elevated non-HDL cholesterol level, an indirect measure of small, dense LDL particles.
Some previous studies have shown SAT to be as strongly or more strongly associated with insulin resistance and metabolic risk than VAT (30, 31). However, in our univariate linear regression analyses, baseline SAT was weakly associated with future atherogenic lipid levels compared to baseline VAT, and moreover, baseline and change in SAT were not independent predictors for future atherogenic dyslipidemia. Therefore, it appears that the role of SAT on future metabolic risk is less pronounced than VAT at least for atherogenic dyslipidemia. To examine the ability of simple anthropometric measurements in predicting our outcomes of interest, we fit models to determine whether simple anthropometric indices alone could predict future atherogenic lipid levels. In these models, baseline BMI and waist circumference predicted future triglyceride level but not future HDL cholesterol or non-HDL cholesterol levels among men and women combined, but among men and women separately no association was seen between baseline BMI and waist circumference and future triglyceride, HDL cholesterol, or non-HDL cholesterol levels (data not shown). Therefore, simple anthropometric indices including BMI and waist circumference have limited value in predicting future atherogenic lipid levels in this Japanese American population. Collectively, together with our previous studies with the same cohort (5–7), VAT appears to be a unifying independent predictor for a diverse spectrum of future metabolic risk including type 2 diabetes, hypertension, and atherogenic dyslipidemia.
This study has some limitations. First, differences in body composition by ethnicity have been reported, with Chinese and South Asians having a relatively greater amount of VAT for a given total body fat and waist circumference compared with Europeans (32), whereas blacks have less VAT compared with whites for a given BMI (33). Therefore, our results, which were obtained in an American population of 100% Japanese ancestry, may not be generalizable to other ethnic groups. In addition, because atherogenic dyslipidemia is closely associated with type 2 diabetes mellitus, the high prevalence of lipid abnormalities in this population (59.5%; 269 of 452) may be explained by the higher risk of diabetes observed in Japanese Americans (34). Second, recent studies have suggested that SAT is further separable into two distinct subdepots (the superficial SAT and the deep SAT) and that they have distinct histological and physiological features and furthermore display different associations with cardiometabolic variables (35). However, we did not measure regional SAT in this study. Third, we did not measure other atherogenic lipid levels that might be more strongly associated with insulin resistance and changes in body composition, including apolipoprotein B and small, dense LDL.
In conclusion, the measurement of abdominal fat depots by CT conveys additional information over simple measures of standard anthropometric indices including BMI and waist circumference in predicting future atherogenic lipid levels. In particular, baseline VAT and change in VAT over time are associated with future atherogenic lipid levels independent of multiple risk factors for lipid abnormalities, baseline atherogenic lipid levels, and standard anthropometric indices over 5 years of follow-up. On the other hand, SAT and its change did not confer any risk for the development of future atherogenic dyslipidemia.
Acknowledgments
We dedicate this manuscript to the memory of Marguerite J. McNeely who for many years played an important role as investigator in the Japanese American Community Diabetes Study. Her critical contributions will be missed. We are grateful to the King County Japanese American community for support and cooperation.
This work was supported by facilities and services provided by the Diabetes Research Center (DK-17047), the Clinical Nutrition Research Unit (DK-35816), and the General Clinical Research Center (RR-00037) at the University of Washington. The VA Puget Sound Health Care System provided support for the involvement of E.J.B. and S.E.K. in this research.
Disclosure Summary: The authors have nothing to disclose.
Footnotes
- BMI
- body mass index
- CI
- confidence interval
- CT
- computed tomography
- HDL
- high-density lipoprotein
- HOMA-IR
- homeostasis model assessment insulin resistance
- LDL
- low-density lipoprotein
- OR
- odds ratio
- SAT
- subcutaneous adipose tissue
- VAT
- visceral adipose tissue
- VIF
- variance inflation factor.
References
- 1. Mokdad AH, Ford ES, Bowman BA, et al. Prevalence of obesity, diabetes, and obesity-related health risk factors, 2001. JAMA. 2003;289:76–79. [DOI] [PubMed] [Google Scholar]
- 2. Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999–2004). Arch Intern Med. 2008;168:1617–1624. [DOI] [PubMed] [Google Scholar]
- 3. Després JP, Moorjani S, Lupien PJ, Tremblay A, Nadeau A, Bouchard C. Regional distribution of body fat, plasma lipoproteins, and cardiovascular disease. Arteriosclerosis. 1990;10:497–511. [DOI] [PubMed] [Google Scholar]
- 4. 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]
- 5. Hayashi T, Boyko EJ, Leonetti DL, et al. Visceral adiposity and the risk of impaired glucose tolerance: a prospective study among Japanese Americans. Diabetes Care. 2003;26:650–655. [DOI] [PubMed] [Google Scholar]
- 6. Wander PL, Boyko EJ, Leonetti DL, McNeely MJ, Kahn SE, Fujimoto WY. Change in visceral adiposity independently predicts a greater risk of developing type 2 diabetes over 10 years in Japanese Americans. Diabetes Care. 2013;36:289–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Hayashi T, Boyko EJ, Leonetti DL, et al. Visceral adiposity is an independent predictor of incident hypertension in Japanese Americans. Ann Intern Med. 2004;140:992–1000. [DOI] [PubMed] [Google Scholar]
- 8. Neeland IJ, Turer AT, Ayers CR, et al. Dysfunctional adiposity and the risk of prediabetes and type 2 diabetes in obese adults. JAMA. 2012;308:1150–1159. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Chandra A, Neeland IJ, Berry JD, et al. The relationship of body mass and fat distribution with incident hypertension: observations from the Dallas Heart Study. J Am Coll Cardiol. 2014;64:997–1002. [DOI] [PubMed] [Google Scholar]
- 10. Oka R, Miura K, Sakurai M, et al. Impacts of visceral adipose tissue and subcutaneous adipose tissue on metabolic risk factors in middle-aged Japanese. Obesity (Silver Spring). 2010;18:153–160. [DOI] [PubMed] [Google Scholar]
- 11. Fujimoto WY, Leonetti DL, Kinyoun JL, Shuman WP, Stolov WC, Wahl PW. Prevalence of complications among second-generation Japanese-American men with diabetes, impaired glucose tolerance, or normal glucose tolerance. Diabetes. 1987;36:730–739. [DOI] [PubMed] [Google Scholar]
- 12. Koppes LL, Dekker JM, Hendriks HF, Bouter LM, Heine RJ. Moderate alcohol consumption lowers the risk of type 2 diabetes: a meta-analysis of prospective observational studies. Diabetes Care. 2005;28:719–725. [DOI] [PubMed] [Google Scholar]
- 13. Paffenbarger RS, Jr, Wing AL, Hyde RT. Physical activity as an index of heart attack risk in college alumni. Am J Epidemiol. 1978;108:161–175. [DOI] [PubMed] [Google Scholar]
- 14. Bergstrom RW, Leonetti DL, Newell-Morris LL, Shuman WP, Wahl PW, Fujimoto WY. Association of plasma triglyceride and C-peptide with coronary heart disease in Japanese-American men with a high prevalence of glucose intolerance. Diabetologia. 1990;33:489–496. [DOI] [PubMed] [Google Scholar]
- 15. Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and β-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28:412–419. [DOI] [PubMed] [Google Scholar]
- 16. Shuman WP, Morris LL, Leonetti DL, et al. Abnormal body fat distribution detected by computed tomography in diabetic men. Invest Radiol. 1986;21:483–487. [DOI] [PubMed] [Google Scholar]
- 17. National Cholesterol Education Program (NCEP) Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in Adults (Adult Treatment Panel III). 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) final report. Circulation. 2002;106:3143–3421. [PubMed] [Google Scholar]
- 18. Gazi I, Lourida ES, Filippatos T, Tsimihodimos V, Elisaf M, Tselepis AD. Lipoprotein-associated phospholipase A2 activity is a marker of small, dense LDL particles in human plasma. Clin Chem. 2005;51:2264–2273. [DOI] [PubMed] [Google Scholar]
- 19. Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus. Diabetes Care. 1997;20:1183–1197. [DOI] [PubMed] [Google Scholar]
- 20. Royston P, Ambler G, Sauerbrei W. The use of fractional polynomials to model continuous risk variables in epidemiology. Int J Epidemiol. 1999;28:964–974. [DOI] [PubMed] [Google Scholar]
- 21. Després JP. Body fat distribution and risk of cardiovascular disease: an update. Circulation. 2012;126:1301–1313. [DOI] [PubMed] [Google Scholar]
- 22. Bergman RN, Kim SP, Catalano KJ, et al. Why visceral fat is bad: mechanisms of the metabolic syndrome. Obesity (Silver Spring). 2006;14(suppl 1):16S–19S. [DOI] [PubMed] [Google Scholar]
- 23. Frayn KN, Karpe F, Fielding BA, Macdonald IA, Coppack SW. Integrative physiology of human adipose tissue. Int J Obes Relat Metab Disord. 2003;27:875–888. [DOI] [PubMed] [Google Scholar]
- 24. Fontana L, Eagon JC, Trujillo ME, Scherer PE, Klein S. Visceral fat adipokine secretion is associated with systemic inflammation in obese humans. Diabetes. 2007;56:1010–1013. [DOI] [PubMed] [Google Scholar]
- 25. Sam S, Haffner S, Davidson MH, et al. Relationship of abdominal visceral and subcutaneous adipose tissue with lipoprotein particle number and size in type 2 diabetes. Diabetes. 2008;57:2022–2027. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Neeland IJ, Ayers CR, Rohatgi AK, 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Pelletier-Beaumont E, Arsenault BJ, Alméras N, et al. Normalization of visceral adiposity is required to normalize plasma apolipoprotein B levels in response to a healthy eating/physical activity lifestyle modification program in viscerally obese men. Atherosclerosis. 2012;221:577–582. [DOI] [PubMed] [Google Scholar]
- 28. Després JP, Ross R, Boka G, Alméras N, Lemieux I, ADAGIO-Lipids Investigators. Effect of rimonabant on the high-triglyceride/ low-HDL-cholesterol dyslipidemia, intraabdominal adiposity, and liver fat: the ADAGIO-Lipids trial. Arterioscler Thromb Vasc Biol. 2009;29:416–423. [DOI] [PubMed] [Google Scholar]
- 29. Matsushita Y, Nakagawa T, Yamamoto S, et al. Effect of longitudinal changes in visceral fat area and other anthropometric indices to the changes in metabolic risk factors in Japanese men: the Hitachi Health Study. Diabetes Care. 2012;35:1139–1143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Abate N, Garg A, Peshock RM, Stray-Gundersen J, Grundy SM. Relationships of generalized and regional adiposity to insulin sensitivity in men. J Clin Invest. 1995;96:88–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Goodpaster BH, Thaete FL, Simoneau JA, Kelley DE. Subcutaneous abdominal fat and thigh muscle composition predict insulin sensitivity independently of visceral fat. Diabetes. 1997;46:1579–1585. [DOI] [PubMed] [Google Scholar]
- 32. Lear SA, Humphries KH, Kohli S, Chockalingam A, Frohlich JJ, Birmingham CL. Visceral adipose tissue accumulation differs according to ethnic background: results of the Multicultural Community Health Assessment Trial (M-CHAT). Am J Clin Nutr. 2007;86:353–359. [DOI] [PubMed] [Google Scholar]
- 33. Tittelbach TJ, Berman DM, Nicklas BJ, Ryan AS, Goldberg AP. Racial differences in adipocyte size and relationship to the metabolic syndrome in obese women. Obes Res. 2004;12:990–998. [DOI] [PubMed] [Google Scholar]
- 34. Fujimoto WY, Boyko EJ, Hayashi T, et al. Risk factors for type 2 diabetes: lessons learned from Japanese Americans in Seattle. J Diabetes Investig. 2012;3:212–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35. Golan R, Shelef I, Rudich A, et al. Abdominal superficial subcutaneous fat: a putative distinct protective fat subdepot in type 2 diabetes. Diabetes Care. 2012;35:640–647. [DOI] [PMC free article] [PubMed] [Google Scholar]