SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: This study investigated ethnic and sex differences in the distribution of fat during childhood and adolescence.

Design and Methods: A cross-sectional sample (n = 382), aged 5–18 years, included African American males (n = 84), White males (n = 96), African American females (n = 118), and White females (n = 84). Measures for total body fat (TBF) mass and abdominal adipose tissue (total volume and L4-L5 cross-sectional area) for both subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) depots were assessed by dual-energy X-ray absorptiometry and magnetic resonance image, respectively. Analyses of covariance (ANCOVAs) were used to determine ethnic and sex differences in TBF (adjusted for age) and ethnic and sex differences in SAT and VAT (adjusted for both age and TBF).

Results: Age-adjusted TBF was greater in African Americans (P = 0.017) and females (P < 0.0001) compared with Whites and males, respectively. In age- and TBF-adjusted ANCOVAs, no differences were found in the SAT. The VAT volume was, however, greater in Whites (P < 0.0001) and males (P < 0.0001) compared with African Americans and females, respectively. Similar patterns were observed in SAT and VAT area at L4-L5.

Conclusions: The demonstrated ethnic and sex differences are important confounders in the prevalence of obesity and in the assignment of disease risk in children and adolescents.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Based on an assessment by body mass index (BMI), the prevalence of obesity is higher among African American youth compared with non-Hispanic White youth (24.3% vs. 14.0%) and is higher among males than females (18.6% vs. 15.0%) [1]. Yet direct measures of total body fat (TBF) and fat distribution reveal distinct ethnic and sex patterns, which portend different health implications. For instance, visceral adipose tissue (VAT) also known as intra-abdominal fat is the padding around all internal vital organs such as stomach, kidney, heart, and pancreas. VAT is particularly associated with cardiovascular disease risk factors and metabolic disorders in both youth [2] and adults [3].

Ethnic patterns for total body fat (TBF) and depot-specific adiposity among youth differ across studies. Some studies find that African Americans have more TBF than Whites [2, 4], whereas others show no difference between ethnic groups [5, 6]. African American youth also typically have more abdominal subcutaneous adipose tissue (SAT) than White youth [2], but the reverse may be true when adjustments are made for TBF [6]. In contrast, White youth tend to have more VAT than African American youth [7], while others find no difference [2, 4]. Clearly, further research with regard to the ethnic differences in TBF, SAT, and VAT is warranted.

Sex differences in depot-specific adiposity in childhood have not been clearly delineated, and results are not consistent across studies. Female youth tend to have more TBF when compared with male youth [6, 10, 11], though some studies find similar amounts of TBF [4, 5]. Females also have been shown to have more abdominal SAT throughout childhood and adolescence [6, 10, 12], but some find no sex difference [2, 4, 5, 13]. Males typically have more VAT than females, yet the age at which this divergence begins differs [2, 5, 13]. Failure to control for age, ethnicity, and TBF, and small samples with limited range of age or adiposity, may account for these contradictory findings.

This study addressed limitations of prior studies by examining sex and ethnic differences in TBF, SAT, and VAT across the whole spectrum of the pediatric age range, taking TBF into account. We hypothesized that: [1] African Americans and females have more TBF than Whites and males, respectively; [2] Whites and females have more TBF-adjusted SAT than African Americans and males, respectively; and [3] Whites and males have more TBF-adjusted VAT than African Americans and females, respectively.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Sample

Participants ranging between 5 and 18 years of age were involved in a cross-sectional study of factors related to abdominal adiposity. Of the 423 participants who enrolled in the study, participants were excluded from the present analysis if self-reported ethnicity was not White or African American (n = 12) and if magnetic resonance image (MRI) or dual-energy X-ray absorptiometry (DXA) scans were not completed due to exceeding weight limits of the equipment, motion artifacts, or participant refusal (n = 28) or were an outlier on one of the primary analysis variables (> ±3 SD from the mean for age and sex) (n = 1). There were 382 participants retained for the present analysis (mean age = 12.3 ± 3.5 years), including African American males (n = 84), White males (n = 96), African American females (n = 118), and White females (n = 84). Parents or guardians of the participants provided signed informed consent and the children provided written assent. All study procedures were approved by the Pennington Biomedical Research Center institutional review board.

Anthropometric measures

Height and weight were measured in the outpatient clinic using standard procedures. The BMI was calculated (kg/m2). BMI percentiles were computed using the SAS program for the 2000 Centers for Disease Control and Prevention (CDC) Growth Charts for the United States [14].

Body composition

TBF (in kilograms) was estimated by DXA using a Hologic QDR 4500A whole-body scanner (Bedford, MA, USA). The participant lay motionless on a table wearing no metal containing objects, while a detector and a scanner emitting low energy X-rays passed along the body. Each scan lasted between 4 and 6 min and the radiation dose was equal to about 12 h of background radiation. The scans were analyzed using QDR software for Windows V11.2. Two distinct energies were used to determine body mineral and soft tissue content. Variations of the attenuation ratio (determined from known tissue content) established the fat content of the tissue at each pixel, thereby calculating TBF.

Abdominal SAT and VAT were measured from cross-sectional MRI scans obtained from a General Electric (GE) Signa Excite (3.0 Tesla) (GE Medical Systems, Milwaukee, WI, USA) scanner. An eight channel torso-array coil was placed over the participant's chest/abdomen area while the participant lay motionless on the scanner table wearing no metal containing objects. A series of scans from the highest point of the liver to the bottom pole of the right kidney were acquired.

Images were analyzed using the Analyze® software package (CNSoftware, Rochester, MN, USA) on a computer workstation by one trained technician. Analyzed slices were 4.78 cm (28 slices) apart and included between five and eight slices depending on the participant's stature. SAT and VAT areas were manually drawn, and the number of pixels was multiplied by voxel width and voxel height for each slice to compute area measured in square centimeters. Area measured at the L4-L5 intervertebral lumbar space was used for analysis (square centimeters). To estimate total abdominal SAT or VAT volume, the area from each slice was multiplied by the number of slices separating each slice (28 slices), then multiplied by 0.000001 by the voxel depth. The five to eight slice volumes were summed to calculate total volume of SAT or VAT in liters (L) for each participant.

To test reliability for quantifying SAT and VAT on the MRI, the technician re-analyzed a sub-set of 20 images at the L4-L5 slice (blinded). Coefficients of variation averaged 0.99 (SD = 1.00) for SAT area and 6.63 (SD = 6.35) for VAT area. Pearson correlation coefficients between the first and second analysis were high for SAT area (r = 0.999) and for VAT area (r = 0.973).

Covariates

Age was computed from birth and observation dates. For the SAT and VAT analyses, TBF measured by DXA was included in the models as a covariate.

Statistical analysis

Analyses of covariance (ANCOVAs) controlling for age were used to examine ethnicity and sex effects on TBF, and ANCOVAs controlling for age and TBF were used to examine ethnicity and sex effects on SAT and VAT volumes and L4-L5 areas. An ethnicity-by-sex interaction term was included in all models. The analyses were repeated after stratification by age group (5- to 9-year-olds, 10- to 14-year-olds, and 15- to 18-year-olds). Statistical analyses were performed using SAS® statistical package V9.3 (SAS Institute Inc., Cary, NC, USA).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Characteristics of the sample are reported in Table 1 stratified by ethnicity and sex. For the total sample, the average BMI percentile was 72.6 ± 28.0. Fat mass averaged 16.8 ± 11.8 kg, SAT volume averaged 4.3 ± 3.9 L, SAT area averaged 193 ± 167 cm2, VAT volume averaged 0.1 ± 0.1 L, and VAT area at L4-L5 averaged 9 ± 7 cm2.

Table 1. Descriptive characteristics of the sample
 MalesFemales
 WhiteAfrican AmericanWhiteAfrican American
N968484118
Age, (years)12.3 ± 3.512.0 ± 3.412.4 ± 3.412.3 ± 3.6
BMI, (kg/m2)21.6 ± 5.622.9 ± 6.621.9 ± 5.225.1 ± 7.8
BMI percentile, (%)67.8 ± 29.573.0 ± 27.970.0 ± 26.078.1 ± 27.7
Total body fat, (kg)13.7 ± 10.214.5 ± 11.016.9 ± 9.420.8 ± 14.1
VAT
Volume (l)0.2 ± 0.20.1 ± 0.10.2 ± 0.10.1 ± 0.1
L4-L5 area, (cm2)9 ± 97 ± 69 ± 78 ± 7
SAT
Volume, (l)3.2 ± 3.13.6 ± 3.44.2 ± 3.25.6 ± 4.8
L4-L5 area, (cm2)148 ± 140162 ± 158193 ± 135251 ± 195

Body fat

In a 2 (ethnicity) × 2 (sex) ANCOVA with age as a covariate and TBF as the dependent variable, there was a significant main effect of ethnicity (P = 0.017) and sex (P < 0.0001). As hypothesized, African Americans had more TBF than Whites, and females had more TBF than males (Figure 1). There was no interaction between ethnicity and sex (P = 0.184). In 5- to 9-year-olds, there was a significant main effect of sex (P = 0.022) where females had more TBF than males, but there was no ethnicity effect (P = 0.274). In 10- to 14-year-olds, there was a significant main effect of ethnicity (P = 0.049) where African Americans had higher TBF than Whites, and a significant main effect of sex (P = 0.007) where females had higher TBF than males. In 15- to 18-year-olds, there was no significant effect of ethnicity (P = 0.225) or sex (P = 0.093) on TBF.

image

Figure 1. Sex- and ethnic-specific (a) TBF adjusted for age, (b) abdominal SAT volume adjusted for age and TBF, and (c) VAT volume adjusted for age and TBF. Error bars indicate standard errors.

Download figure to PowerPoint

Subcutaneous adipose tissue

Contrary to our hypothesis, in a 2 (ethnicity) × 2 (sex) ANCOVA with age and TBF as covariates and abdominal SAT volume as the dependent variable, there was no significant effect of ethnicity (P = 0.153) or sex (P = 0.845) (Figure 1). Similarly, there was no significant effect of ethnicity (P = 0.418) or sex (P = 0.899) on L4-L5 SAT area. There were no main effects of ethnicity or sex on SAT volume or SAT area within any age group (all P values > 0.082) (Figure 2).

image

Figure 2. Abdominal SAT and VAT volume adjusted for TBF across the pediatric age range. Sample sizes ranged from 4 to 27 across the age-, sex-, and ethnic-specific groups. Error bars indicate standard errors.

Download figure to PowerPoint

Visceral adipose tissue

A 2 (ethnicity) × 2 (sex) ANCOVA with age and TBF as covariates and total VAT volume as the dependent variable demonstrated a significant main effect of ethnicity (P < 0.0001) and sex (P < 0.001) (Figure 1). As hypothesized, Whites had greater VAT volume than African Americans, and males had greater VAT volume than females. A similar pattern was demonstrated for L4-L5 VAT area, where Whites had greater VAT area than African Americans (P < 0.0001), and males had greater VAT area than females (P = 0.0008). There was no interaction between ethnicity and sex for either dependent variable (all P values > 0.46).

In 5- to 9-year-olds, there was a significant ethnicity effect on total VAT volume (P = 0.003), but no sex effect (P = 0.473). For 10- to 14-year-olds, there was a significant ethnicity effect (P < 0.0001) and a significant sex effect (P = 0.018) on total VAT volume. For 15- to 18-year-olds, there was a significant ethnicity effect (P = 0.0002) and sex effect (P = 0.0001) on total VAT volume. A similar pattern of findings was demonstrated for L4-L5 VAT area: in 5- to 9-year-olds, there was a significant ethnicity effect (P = 0.0005) but no sex effect (P = 0.787); in 10- to 14-year-olds, there was a significant ethnicity effect (P < 0.0001) but no sex effect (P = 0.115); and in 15- to 18-year-olds, there was a significant ethnicity effect (P = 0.0312) and sex effect (P < 0.0001). In each case, significant findings reveal that Whites had greater VAT than African Americans and males had greater VAT than females (Figure 2).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

In this pediatric sample representing a range of age and adiposity, ethnic and sex differences in TBF and VAT emerged. Overall, as hypothesized, African Americans and females had more age-adjusted TBF than Whites and males, respectively. Examining TBF- and age-adjusted VAT revealed opposite relationships: Whites and males had more total VAT volume and VAT area at L4-L5 than African Americans and females, respectively, which was also consistent with the hypothesis. Contrary to our hypothesis, when controlling for TBF and age, there was no sex or ethnic difference in abdominal SAT measured as total volume or at area at L4-L5. Controlling for TBF accounted for the lack of sex or ethnic differences in SAT, as models that did not control for TBF demonstrated higher TBF in African Americans and females (data not shown). Interestingly, there was no ethnicity by sex interaction for any analysis, indicating that the main effect of ethnicity did not change based on sex, and the main effect of sex did not change based on ethnicity.

Sex and ethnic differences emerged across age groups. Compared with males, TBF was higher in females aged 5-14 years but there was no sex difference in ages 15-18 years. Males and females did not differ in VAT volume until age 10 years or in VAT area at L4-L5 until age 15 years, at which point males had higher VAT than females. The timing of the sex divergence supports previous findings that males accumulate more VAT than females after age 12 years [12] or age 16 years [15], though one study demonstrated sex differences at younger ages [4]. Hormone secretion [16], sexual maturation [17], or skeletal growth [18] may explain sex differences in TBF and VAT that emerge during adolescence.

In comparing the two ethnic groups by age, TBF only differed in the 10- to 14-year-olds, in which African Americans had more TBF than Whites. African Americans may accumulate more TBF during the pubertal period compared with Whites [19], yet the ethnic differences in TBF disappear in the 15- to 18-year-old group as adolescents reach post-puberty. The lack of ethnic differences in TBF for the 15- to 18-year-old group may be due to a limited sample size or may mark a transient phase during which puberty is masking the ethnic differences. Ethnic differences in TBF re-emerge in adulthood: a study of 2037 adults showed that African American women had more TBF than White women whereas White men had more TBF than African American men [20].

Compared with African Americans, Whites had more VAT volume and VAT area at L4-L5 in every age group, whereas SAT did not differ by sex or ethnicity in any age group. The consistency of ethnic effects on VAT and SAT across age groups indicates trends that persist throughout childhood and adolescence. These trends continue into adulthood, as White adults have more VAT than African American adults [21].

A key strength of this study is the diversity in age, adiposity, ethnicity, and sex of the sample. Controlling for TBF in the SAT and VAT analyses exhibited ethnic and sex differences beyond those attributed to total fat. Other strengths were the use of MRI and DXA to directly measure depot-specific adiposity and TBF, and the reporting of both total volume of SAT and VAT as well as area measured at the L4-L5 area.

Limitations of the study warrant discussion. This was a cross-sectional study, so the results do not imply causation. Moreover, differences between age groups cannot be interpreted as reflecting normal developmental growth patterns as participants in each age group were distinct. A longitudinal cohort design would provide more extensive understanding of the ethnic and sex patterns of fat distribution over time. Pelvic morphology may differ by ethnic group. However, it is not known if anatomical differences in the lower vertebral column influence the amount of adiposity deposited at the L4-L5 intervertebral space in children. Demerath et al. [22] reported that ethnic and sex differences in total VAT volume may be diminished when relying on only a single slice at the L4-L5 level among adults. The present findings did not differ whether adiposity was measured as total volume or at L4-L5.

In conclusion, TBF was higher in African Americans and females, and while the SAT did not differ among ethnic or sex groups, the VAT was higher in Whites and males. These ethnic and sex differences are important confounders in the prevalence of obesity and in the assignment of disease risk in children and adolescents.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

We acknowledge the efforts of Emily Mire for data management; Amber Dragg and the clinical staff for data collection; and Julia St Amant and the Imaging Core for analysis of MRI and DXA data.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References