Racial Differences in Insulin Resistance and Mid-Thigh Fat Deposition in Postmenopausal Women

Authors

  • Alice S. Ryan Ph.D.,

    Corresponding author
    1. Department of Medicine, Division of Gerontology, University of Maryland School of Medicine and the Baltimore Geriatric Research, Education, and Clinical Center, Veterans Affairs Maryland Health Care System, Baltimore, Maryland
    Search for more papers by this author
  • Barbara J. Nicklas,

    1. Department of Medicine, Division of Gerontology, University of Maryland School of Medicine and the Baltimore Geriatric Research, Education, and Clinical Center, Veterans Affairs Maryland Health Care System, Baltimore, Maryland
    Search for more papers by this author
  • Dora M. Berman

    1. Department of Medicine, Division of Gerontology, University of Maryland School of Medicine and the Baltimore Geriatric Research, Education, and Clinical Center, Veterans Affairs Maryland Health Care System, Baltimore, Maryland
    Search for more papers by this author

Division of Gerontology, BT/18/GR, Baltimore Veterans Affairs Medical Center, Baltimore, MD 21201. E-mail: aryan@grecc.umaryland.edu

Abstract

Objective: To determine whether racial differences in insulin resistance between African American (AA) and white women exist in postmenopausal women and whether they are related to physical fitness and/or obesity.

Research Methods and Procedures: We studied 35 obese AA (n = 9) and white (n = 26) women of comparable maximal oxygen consumption, obesity, and age. Total body fat was measured by DXA. Abdominal and mid-thigh low-density lean tissue (a marker of intramuscular fat) were determined with computed tomography. Glucose utilization (M) was measured during the last 30 minutes of a 3-hour hyperinsulinemic-euglycemic clamp. Insulin sensitivity was estimated from the relationship of M to the concentration of insulin during the last 30 minutes of the clamp.

Results: The percentage of fat and total body fat mass were similar between AA and white women, whereas fat-free mass was higher in African American women. Visceral adipose tissue was not different between groups, but subcutaneous abdominal fat was 17% higher in the AA than in the white women. AA women had an 18% greater mid-thigh muscle area (p < 0.01) and a 34% greater mid-thigh low-density lean tissue area than the white women. Fasting glucose concentrations were not different, but fasting insulin concentrations were 29% higher in AA women. Glucose utilization was 60% lower in the AA women because of a lower non-oxidative glucose disposal. Insulin sensitivity was 46% lower in the AA women.

Discussion: AA postmenopausal women have more mid-thigh intramuscular fat, lower glucose utilization, and are less insulin sensitive than white women despite comparable fitness and relative body fat levels.

Introduction

The prevalence of obesity continues to rise in the United States, with estimates that 32% of the population is overweight (1). Total and abdominal adiposity increase the risk for coronary artery disease and mortality among women (2,3). Increases in total body fat, visceral fat, and fat deposition in the muscle occur with age and are associated with insulin resistance, dyslipidemia, and risk for type 2 diabetes and coronary artery disease (3,4,5,6,7). These associations have been examined predominately in white populations. Yet, the incidence of obesity is 49% in African American women compared with 33% in white women (8). African American women also have a higher incidence of diabetes than white women (9,10). Physical activity (11,12) and fitness, as measured by Vo2max (13), are significantly lower in African American compared with white obese women. It is unknown whether there are racial differences in fat deposition in the muscle.

Despite greater adiposity, premenopausal African American women have less visceral fat than white women (14,15). Studies in premenopausal women indicate that African American women are more insulin resistant than white women (15,16,17), even after adjustment for visceral and subcutaneous fat (15). Because menopause is a period of increased fat deposition in the total body and abdominal region (18) and because there exist strong associations between obesity and insulin resistance (19,20), we sought to make racial comparisons of insulin sensitivity in postmenopausal women. Furthermore, it is unknown whether racial differences in total or regional adiposity (visceral or intramuscular fat) or Vo2max, contribute to differences in glucose metabolism between African American and white postmenopausal women.

We hypothesized that postmenopausal African American women would have greater mid-thigh low-density lean tissue (a marker of intramuscular fat) and a lower glucose uptake during hyperinsulinemia than white women of similar age and relative body fat. Thus, the goals of this study were to determine whether the previously reported increased insulin resistance in African American premenopausal women compared with white women exists in postmenopausal women and whether it is related to the degree of physical fitness, obesity, and/or regional fat deposition.

Research Methods and Procedures

Subjects

Thirty-five women (9 African American and 26 white) who were healthy, overweight/obese [body mass index (BMI) > 25 kg/m2], and between the ages of 50 and 70 years participated in the study. The women were postmenopausal, i.e., had not menstruated for at least 1 year and/or had plasma follicle-stimulating hormone levels >30 mIU/mL. Race was assigned according to the subject's characterization of herself as either African American or white, and all women were born in the United States. The women were initially recruited to participate in weight loss and exercise studies, and these women responded and qualified for participation. Only women who were weight stable (<2.0 kg weight change in the past year) and sedentary (<20 minutes of aerobic exercise 2 times/wk for the previous 6 months) were recruited. Subjects were screened with a medical history questionnaire, a physical examination, fasting blood profile, and a graded exercise treadmill test in an attempt to exclude those with coronary artery disease. All subjects underwent a 75-g, 2-hour oral glucose tolerance test (21), with blood samples drawn at baseline and at 120 minutes for measurement of plasma glucose levels. All subjects were nonsmokers and had no evidence of diabetes, hyperlipidemia, cancer, liver, renal or hematologic disease, or other medical disorders. Nine women were on medication for hypertension (six African American and three white). All methods and procedures for the study were approved by the Institutional Review Board of the University of Maryland. Each participant provided written informed consent to participate in the study.

Maximal Oxygen Uptake

Vo2max was measured with a continuous treadmill test protocol as previously described (22). Briefly, speed was kept constant while the grade was increased from 0% to 4% at 2 minutes and then increased 2% every minute after 3 minutes until the woman was unable to continue. Validation for attainment of Vo2max included meeting two of the following three criteria: (1) a plateau in oxygen uptake with an increased work load as evidenced by a difference in oxygen uptake of <2 mL/kg per minute, (2) a respiratory exchange ratio >1.10, and (3) a maximal heart rate within 10 beats/min of the age-predicted maximal value.

Body Composition

Anthropometry.

Height (in centimeters) and weight (in kilograms) were measured to calculate BMI (kg/m2). Waist circumference, measured at the narrowest point superior to the hip, was divided by the circumference of the hip as measured at its greatest gluteal protuberance to obtain waist-to-hip ratio (WHR) (23). Thigh circumference was measured at the level of the mid-thigh.

DXA.

Fat mass, lean-tissue mass, and bone mineral content were determined by DXA (model DPX-L; LUNAR Radiation Corp., Madison, WI). Fat-free mass is reported as lean tissue plus bone mineral content. All DXA scans were analyzed with the LUNAR DPX-L (version 1.3z) program for body composition analyses. Regions of interest, including arms, legs, and trunk, were analyzed with the extended analysis of the LUNAR software.

Computed Tomography.

To quantify visceral and abdominal subcutaneous fat areas, a computed tomography (CT) scan of the abdomen was performed with a PQ 6000 scanner (Marconi Medical Systems, Cleveland, OH). A single 5-mm scan was taken at the L4–L5 region while the subject was supine, with her arms stretched above her head. A fat tissue-highlighting technique was used to quantitate the relative proportions of visceral adipose tissue and subcutaneous adipose tissue areas. Sagittal diameter was determined on the images at the level of the umbilicus and the fourth lumbar intervertebral disk. CT data are expressed as cross-sectional area of tissue (in square centimeters) with the Hounsfield units (HU) for adipose tissue as −190 to −30. A second scan performed at the level of the mid-thigh was used to quantify muscle area (HU, 30 to 80), total fat area of the thigh (HU, −190 to −30), and low-density lean tissue (HU, 0 to 29) of both the right and left legs, as previously described (6).

Metabolic Testing

To control nutrient intake, all subjects were provided with a eucaloric diet by a registered dietitian for 2 days before the clamp procedure. The composition of this diet was 50% to 55% carbohydrate, 15% to 20% protein, ≤30% fat, and 300 to 400 mg of cholesterol per day and a polyunsaturated-to-saturated fat ratio of 0.6 to 0.8. The diet was composed of at least 150 g of carbohydrate per day for the 2 days before testing (24). The number of calories given to each woman was estimated from the 7-day food records, and estimates of energy expenditure were based on the Harris–Benedict equation (25) and activity factor described by Rodwell (26). All testing was performed in the morning after a 12-hour overnight fast. All subjects were weight-stable (<1 kg) for at least 2 weeks before metabolic testing.

Hyperinsulinemic-Euglycemic Clamps

Peripheral tissue sensitivity to exogenous insulin was measured with the hyperinsulinemic-euglycemic clamp technique (27). Briefly, an intravenous catheter was inserted by percutaneous venipuncture for the infusion of glucose and insulin. A second catheter was inserted in a retrograde fashion into a dorsal hand or wrist vein, and the hand was enclosed in a grounded, insulated chamber, which was warmed to 70 °C to “arterialize” (28) the blood obtained for all samples. Difficulty in obtaining venous access occurred in one African American and two white women, so the test was not performed in these women. For the assessment of basal glucose and insulin levels, three arterialized blood samples were drawn at 10-minute intervals. Blood samples were obtained every 5 and 10 minutes thereafter for the determination of plasma glucose and insulin levels. A 10-minute priming and continuous infusion of insulin (240 pmol/m2 per minute; Humulin; Eli Lilly, Indianapolis, IN) was performed for 180 minutes. This resulted in a square wave of hyperinsulinemia at a level of 483 ± 14 pM among all women. A 20% glucose solution was used, which was measured as 18%.

The mean plasma glucose level during the 10 to 180 minutes of the euglycemic clamp was computed for each individual study and expressed as a percentage of the desired goal. The plasma glucose levels during each clamp period averaged 5.37 ± 0.19 and 5.34 ± 0.08 mM, which was 98.1 ± 0.3% and 97.6 ± 0.2% of the desired goal in African American and white women with a coefficient of variation of 4.4 ± 0.4% (SD) and 4.8 ± 0.3%, respectively. Plasma insulin concentrations during the 150 to 180 minutes of the hyperinsulinemic-euglycemic clamps were higher in African American than in white women (566 ± 38 vs. 470 ± 11 pM; p < 0.05).

Indirect Calorimetry.

To quantitate carbohydrate oxidation, continuous indirect calorimetry was performed before the start of the glucose infusion and during the last 30 minutes of the insulin infusion by the open circuit dilution technique with a DeltaTrac cart (SensorMedics, Yorba Linda, CA). Rates of glucose oxidation were calculated from measurements of carbon dioxide production and oxygen consumption with established equations (29), with correction for protein oxidation based on 24-hour urinary urea nitrogen (30). Nonoxidative glucose metabolism is the difference between total glucose uptake and glucose oxidation.

Analysis of Blood Samples.

Blood samples were collected in heparinized syringes and placed in prechilled test tubes containing 1.5 mg EDTA/mL of blood. The blood samples were centrifuged at 4 °C, and the plasma was stored at −70 °C until it was analyzed. Plasma glucose was measured with the glucose oxidase method (Beckman Instruments, Fullerton, CA). Immunoreactive insulin was determined by an insulin-specific double-antibody system, as previously described (31), using human insulin standards and tracer. The antiserum was raised against highly purified human insulin and does not cross-react with human proinsulin (<0.1%; Linco, St. Louis, MO). The lower limit of detection of this assay in our laboratory is 12 pM. Intra- and interassay coefficients of variation of pooled control sera averaged 5% and 9%, respectively.

Statistical Analyses

There were no significant differences between the right and left mid-thighs for muscle area, total fat area, and low-density lean tissue area. Therefore, the values of the right leg were used in the statistical analyses for mid-thigh muscle area, fat area, and low-density lean tissue area. For the hyperinsulinemic-euglycemic clamps, the mean concentration of glucose and insulin was calculated for each sample time-point. The trapezoidal rule was used to calculate the integrated response over 30-minute intervals from 30 to 180 minutes for each subject. The integrated response was divided by its time interval to compute mean concentrations. Glucose utilization (M) for 30-minute intervals was calculated from the amount of glucose infused after correction for glucose equivalent space (glucose space correction). Insulin sensitivity was expressed as M/I, which represents the amount of glucose metabolized per unit of plasma insulin (I), and was calculated by dividing the glucose used by the insulin concentration during the last 30 minutes of the clamp for each subject. The metabolic clearance rate of insulin was calculated as described by Elahi et al. (32). Differences between African American and white women were analyzed with unpaired Student's t tests. Relationships between variables were determined with linear regression analyses and calculation of Pearson correlation coefficients. Statistical significance was set at p < 0.05 for all tests. All data were analyzed by SPSS statistical software (SPSS, Chicago, IL). All values are expressed as means ± SE.

Results

Physical Characteristics

The average age, BMI, and Vo2max (liters per minute or milliliters per kilogram per minute) did not differ between African American and white women, but African American women tended to be heavier than the white women (p < 0.05; Table 1). Percentage of body fat, total body fat mass, and arm, leg, and trunk fat mass also did not differ by race (Table 2). However, total body fat-free mass (p < 0.05), arm lean-tissue mass, and leg lean-tissue mass (p < 0.01) were significantly higher in the African American women than in the white women (Table 2).

Table 1.  Subject characteristics of white and African American women
 White (n = 26)African American (n = 9)
  • Vo2max, maximal oxygen consumption. Values are means ± SEM.

  • *

    Significantly different between white and African American women, p < 0.05.

Age (years)60 ± 155 ± 2
Body weight (kg)84.0 ± 2.294.2 ± 4.7*
Height (cm)162.4 ± 1.5163.8 ± 2.0
Body mass index (kg/m2)32.0 ± 0.935.0 ± 1.1
Waist circumference (cm)94.1 ± 2.1101.3 ± 2.1*
Hip circumference (cm)115.1 ± 1.8121.7 ± 4.1
Thigh circumference (cm)60.6 ± 1.164.9 ± 1.7
Waist-to-hip ratio0.82 ± 0.010.84 ± 0.02
Vo2max (liter/min)1.69 ± 0.051.80 ± 0.09
Vo2max (mL/kg per minute)20.2 ± 0.619.3 ± 0.7
Table 2.  Total and regional body composition by DXA in white and African American women
 WhiteAfrican American
  • Values are means ± SEM. Significantly different between white and African American women:

  • *

    p < 0.05

  • p < 0.01.

Total body  
 Body fat (%)46.1 ± 1.146.5 ± 1.5
 Fat mass (kg)38.7 ± 1.843.5 ± 3.3
 Lean mass (kg)42.0 ± 0.946.5 ± 1.6*
 Fat-free mass (kg)44.4 ± 1.049.3 ± 1.7*
Arm  
 Fat mass (kg)4.5 ± 0.35.3 ± 0.4
 Lean-tissue mass (kg)4.5 ± 0.15.5 ± 0.2
Leg  
 Fat mass (kg)14.4 ± 0.717.0 ± 2.3
 Lean-tissue mass (kg)13.8 ± 0.316.5 ± 0.7
Trunk  
 Fat mass (kg)18.3 ± 0.919.4 ± 0.8
 Lean-tissue mass (kg)21.1 ± 0.521.8 ± 0.9

Although the visceral fat area was 15% higher in the white women, this racial difference did not reach statistical significance (Table 3). The subcutaneous abdominal fat area was 17% higher (p < 0.05) and sagittal diameter was 10% greater in the African American women (p < 0.01). The ratio of visceral fat to subcutaneous fat area tended to be greater in the white women than the African American women (0.40 ± 0.04 vs. 0.28 ± 0.03; p = 0.08). African American women had an 18% greater mid-thigh muscle area (p = 0.01) and a 34% greater mid-thigh low-density lean tissue area (p < 0.05) than the white women (Table 3). Mid-thigh subcutaneous fat was not different between African American and white women.

Table 3.  Abdominal and mid-thigh body composition by computed tomography in white and African American women
 WhiteAfrican American
  • Values are means ± SEM. Significantly different between white and African American women:

  • *

    p < 0.01

  • p < 0.05.

Abdominal  
 Visceral fat (cm2)165.6 ± 13.0140.9 ± 14.5
 Subcutaneous fat (cm2)443.1 ± 21.3533.4 ± 34.2
 Sagittal diameter (mm)25.3 ± 0.528.2 ± 0.6*
Mid-thigh  
 Fat (cm2)193.4 ± 10.8213.9 ± 29.5
 Muscle (cm2)72.9 ± 3.088.9 ± 5.9*
 Low-density lean tissue (cm2)21.1 ± 2.131.8 ± 5.2

Insulin Action

Two African American women (22%) and seven white women (24%) had impaired glucose tolerance based on the oral glucose tolerance screening (21). Fasting plasma glucose concentrations were not different between races, but African American women had a 29% higher fasting plasma insulin concentration than white women (p < 0.01; Table 4). Glucose utilization, expressed in μmol/kgFFM per minute, during the last 30 minutes (150 to 180 minutes) of the clamp was 60% lower in the African American than in the white women (p < 0.01; Figure 1) because of a 98% lower nonoxidative glucose disposal (p < 0.01). Basal and insulin-stimulated glucose oxidation rates did not differ between African American and white women. M/I, an index of insulin sensitivity, was 46% lower in the African American than in the white women (0.066 ± 0.010 vs. 0.122 ± 0.008 μmol/kgFFM/min/pM; p < 0.01). The metabolic clearance rate of insulin was slower in African American women than in white women (432 ± 29 vs. 512 ± 11 mL/m2 per minute; p < 0.05).

Table 4.  Glucose metabolism in white and African American women
 WhiteAfrican American
  • Values are means ± SEM.

  • Significantly different between white and African American women:

  • *

    p < 0.05

  • p < 0.01.

Basal  
 Plasma glucose (mM)5.3 ± 0.15.4 ± 0.2
 Plasma insulin (pM)60 ± 585 ± 13*
 Carbohydrate oxidation (μmol/kgFFM per minute)11.0 ± 1.67.0 ± 1.4
150 to 180 minutes of clamp: (μmol/kgFFM per minute)  
 Glucose utilization (M)56.9 ± 3.235.6 ± 3.8
 Nonoxidative glucose disposal33.4 ± 2.916.8 ± 1.9
 Oxidative glucose disposal23.0 ± 1.617.3 ± 2.6
Figure 1.

Oxidative and nonoxidative glucose disposal during 150 to 180 minutes of the hyperinsulinemic-euglycemic clamp in African American (n = 9) and white (n = 26) women. *Glucose utilization in African American vs. white women, p < 0.05.

Relationships

Regression analyses were used to determine if race was a significant predictor of M or M/I, independent of total or regional body composition or fitness. Glucose utilization (μmol/kgFFM per minute) during the last 30 minutes of the hyperinsulinemic-euglycemic period in the total group was negatively correlated with waist circumference (r = −0.57; p < 0.001), WHR ratio (r = −0.55; p < 0.001), sagittal diameter (r = −0.49; p < 0.01), and mid-thigh muscle area (r = −0.49; p < 0.01), but there were no significant relationships between M and fat mass, visceral fat, subcutaneous abdominal fat, or mid-thigh low-density lean tissue. In addition, glucose utilization was inversely correlated with fasting plasma insulin concentrations (r = −0.68; p < 0.001). M/I was negatively correlated with waist circumference (r = −0.56; p < 0.01), WHR (r = −0.63; p < 0.001), VAT (r = −0.33; p = 0.05), sagittal diameter (r = −0.52; p < 0.01), mid-thigh muscle area (r = −0.51; p < 0.01), and mid-thigh low-density lean tissue (r = −0.33; p = 0.05; Figure 2). In a multiple regression model that included race, the ratio of VAT to subcutaneous abdominal fat, and mid-thigh low-density lean tissue, race was the best independent predictor of M (partial r = 0.58; p < 0.01) and the best independent predictor of M/I (partial r = 0.65; p < 0.01).

Figure 2.

Relationship between mid-thigh low-density lean tissue and insulin sensitivity, M/I (r = −0.33, p = 0.05), in African American and white women. •, African American women; ○, white women.

Discussion

Physical inactivity and obesity lead to a deterioration in insulin sensitivity. The physical inactivity, loss of muscle mass, and obesity associated with aging predispose some older women to develop insulin resistance and type 2 diabetes. Results of this study demonstrate that African American women, when compared with similar overweight, sedentary white women, have lower glucose use and insulin sensitivity. These racial differences are evident despite the fact that there are no racial difference in visceral fat area. African American women also have greater muscle mass and greater mid-thigh low-density lean tissue compared with their white counterparts. Thus, African American postmenopausal women have greater intramuscular fat and are more insulin resistant than white women despite comparable fitness and relative body fat levels.

Although the prevalence of type 2 diabetes is twice as high in African American women compared with whites (33,34), the reasons for this disparity are not fully understood or entirely explained by differences in obesity (35). In the Insulin Resistance Atherosclerosis Study, African American men and women were significantly more insulin resistant than whites (36). When young African American and white women were matched for BMI and WHR, African American women had a 36% lower insulin sensitivity index, measured by the Minimal Model, than did white women (16). Furthermore, premenopausal African American women are more insulin resistant than white women after adjustment for differences in visceral and subcutaneous fat (15) or truncal subcutaneous fat (17) are made. Our data support these studies in premenopausal women and show that postmenopausal African American women are more insulin resistant than white women. Moreover, the magnitude of this racial difference in insulin-stimulated glucose utilization (∼60%) and in insulin sensitivity (∼46%) is higher than that reported in premenopausal women. In addition, the difference in glucose use and insulin sensitivity in postmenopausal women occurs despite a similar total amount of body fat, fitness level, and visceral fat.

Surprisingly, it should be noted that the plasma insulin levels during the glucose clamp were statistically higher in the African American women than in the white women. Therefore, we calculated the clearance rate of insulin, which proved to be statistically lower in the African American women. Thus, the higher plasma insulin levels during the hyperinsulinemic-euglycemic clamps in the African American women can be attributed to a lower clearance rate. The mechanisms for the lower clearance rate in the African Americans needs to be investigated.

Increased body fat is associated with reduced rates of insulin-mediated glucose disposal (19,20). Both abdominal subcutaneous fat and visceral adiposity are markers of insulin resistance (4,37). In the current study, insulin sensitivity (M/I), but not glucose utilization (M), was inversely related to visceral obesity. Associations between glucose use and insulin sensitivity with other measures of abdominal adiposity, including waist circumference, sagittal diameter, and WHR, were significant, indicating that abdominal adiposity and fat distribution are important in determining insulin action in postmenopausal women.

In the current study, African American women had a 34% greater mid-thigh low-density lean tissue compared with white women. We (6) and others (4,38,39) have reported that muscle, with lower than normal attenuation (measured by CT scan), increases with age, reduced physical conditioning, and increased body fatness. The reduced attenuation of muscle is believed to reflect increased fat content in and around muscle fibers (40). This tissue is also linked to insulin resistance and cardiovascular risk (4,6,39). Goodpaster et al. (4) reported that mid-thigh muscle attenuation was the strongest single correlate of insulin resistance in obese young men and women. As illustrated in Figure 2, we also found that insulin sensitivity (M/I) was inversely related to mid-thigh low-density lean tissue.

Our findings that glucose use and insulin sensitivity are lower in African American women independent of fitness and visceral adiposity indicates that there are likely to be racial differences in skeletal muscle metabolism. However, the cellular mechanisms that cause African American women to have a lower insulin sensitivity than white women have not been explored. The level of GLUT-4 transporters in skeletal muscle and alterations in key enzymatic processes that regulate glucose uptake and storage as well as interactions between glucose and esterified free fatty acid metabolism (41) could possibly differ between the African American and white women. Alterations in the early steps of the insulin signaling pathway are important components of many insulin-resistant states (42) and may differ between the races. In addition, racial differences in lipid content of skeletal muscle may contribute to racial differences in insulin sensitivity. At the tissue level, intramuscular triglyceride content is increased in obesity (43) and is related to a reduction in glucose storage in muscle in rats (44). Thus, it is possible that the increased fat deposition in the muscle in the African American women decreases insulin sensitivity in this group of women as illustrated by the association between intramuscular fat and insulin sensitivity.

We also examined regional differences in adiposity between African American and white women. The prevalence of overweight is one to one and one-half times greater in African American women than in white women (8). Premenopausal African American women have more upper body fat and a greater fat deposition on the trunk relative to the extremities than premenopausal white women (45). In contrast, we found no difference between African American and white women in fat mass of the trunk, arms, and legs. However, as expected, both African American and white women had greater fat deposition in the trunk than either the arm or leg extremities.

The question of whether the greater trunk adiposity in African American and white women is of visceral or subcutaneous origin has been investigated. Several studies report that premenopausal African American women have less visceral fat than white women (14,15,16). This observation holds when overweight or severely obese women are matched by BMI, circumferences, and WHR (14,16). Obese premenopausal African American women also have a lower visceral to subcutaneous fat ratio than obese white women after controlling for total body-fat mass (15). In contrast to visceral fat, subcutaneous abdominal fat is higher in African American than in white women even when adjusted for total body fatness (46). We also found higher subcutaneous abdominal fat in the African American women than in white postmenopausal women. In this study, we did not detect a significant difference in visceral fat between overweight/obese postmenopausal African American and white women, which confirms a recent study (47) that reported no ethnic differences in visceral fat in perimenopausal women. However, in a previous study of African American and white postmenopausal women matched for fat and lean mass, we found that the white women had significantly higher (22%) visceral fat area but that there was no difference in subcutaneous fat between the women (13). Our results also confirm that the ratio of visceral to subcutaneous abdominal fat is lower in African American compared with white women, suggesting a disparity in the relative amounts of each type of abdominal fat between African American and white women.

We also examined differences in body composition at the level of the mid-thigh. The African American women had greater mid-thigh muscle area and mid-thigh low-density lean tissue area but similar mid-thigh subcutaneous fat than the white women. African American women have higher total body skeletal muscle mass (48) and higher limb lean mass by DXA than white women (49). We are not aware of any studies examining racial differences in mid-thigh low-density lean tissue. Because weight loss and walking reduce mid-thigh low-density lean tissue in overweight postmenopausal women (50), it is possible that this tissue can be altered in African American women. Further studies are needed to determine whether increased intramuscular fat in African American women constitutes a modifiable metabolic risk factor for insulin resistance and cardiovascular disease in this population.

It should be recognized that there are limitations in our study. The small sample size of the African American women indicates that our results are preliminary and should be confirmed in a larger population. In addition, there are limitations in measuring muscle fat with a single-slice CT scan because we have not measured muscle or fat volume and have not directly assessed the fat content in the muscle by biochemical analysis. However, others (40) have reported that skeletal muscle attenuation by CT is highly associated with muscle fiber lipid content and triglyceride content in obese humans.

In summary, our results show that greater intramuscular fat and insulin resistance in obese postmenopausal African American women, compared with similar sedentary white women, are not due to greater total or visceral fat but may be due to greater intramuscular fat. Further studies are needed to determine the cellular mechanisms for the differences in glucose utilization and insulin sensitivity between African American and white postmenopausal women. Also, longitudinal studies are needed to determine whether the greater degree of insulin resistance is associated with the increased development of diabetes in African Americans.

Acknowledgments

This study was supported by NIH grants K01-AG00747 (to A.S.R.), R29-A614066 (to B.J.N.), and K01-AG00685 (to D.M.B.) and by the Geriatrics Research, Education, and Clinical Center, U.S. Department of Veterans Affairs, Baltimore, MD. We thank the participants in this study. We also thank Dr. Andrew Goldberg and Dr. Dariush Elahi for their insightful comments, Adeola Dosunmu and Dana Jones for their assistance in the laboratory, and the nurses and dietitians in the Baltimore Veterans Affairs Medical Center Geriatrics Research, Education, and Clinical Center for technical assistance.

Ancillary