Abdominal Adiposity Change in White and Black Midlife Women: the Study of Women’s Health Across the Nation

Abstract

Objective

The principal objective of this investigation was to compare the naturalistic intra-abdominal adipose tissue (IAAT) change among black and white women during midlife.

Methods

A cohort of 222 (56%) white and 171 (44%) black midlife women were investigated in the Fat Patterning study at the Chicago site of the Study of Women’s Health Across the Nation. The subjects’ total body fat was assessed by dual-energy X-ray absorptiometry and IAAT by a planimetric computed tomography (at the level of L₄–L₅) annually over up to 4 years.

Results

Total body fat at initial evaluation was higher in black women (45.1±8.2%) compared to white women (41.3±8.7%, p<0.001), and did not significantly change over the longitudinal follow up. No significant racial differences were found in the mean annualized gain of intra-abdominal adipose tissue (4.4±0.5%) in models adjusted for total body fat, initial IAAT, age, race, time and race interaction, physical activity, depressive symptoms, menopausal status, and menopausal hormone therapy.

Conclusions

During a naturalistic observation, black and white midlife women had similar abdominal fat gain adjusted for differences in baseline adiposity. This data informs future research aimed to prevent intra-abdominal adipose tissue gain during the critical midlife period of rising cardiovascular risk.

INTRODUCTION

Racial disparities in cardiovascular disease and metabolic conditions have fueled the interest to explore adipose tissue distribution among white and black women. Intra-abdominal adipose tissue (IAAT) contributes to cardiovascular and metabolic risk regardless of body weight.¹ Midlife women are at higher risk for IAAT accumulation.^{2, 3}

Cross-sectional studies report that white women have more IAAT than black women for a given total body adiposity.^{4, 5, 6, 7, 8} While it has been shown that both white and black women gain IAAT during midlife,³ there is no clear understanding about the racial differences in the longitudinal change of IAAT.

Clinical trials of weight loss have reported conflicting results on differences in short-term IAAT change between white and black women. Some reports suggest a greater average IAAT loss in white women compared to black women, whereas others demonstrate no difference in the percentage of the IAAT loss among the two races.^{8, 9, 10, 11} There is a need to clarify the racial specifics of the naturalistic long-term IAAT changes among midlife women.

The principal focus of this investigation was to compare the naturalistic long-term IAAT changes among black and white women during midlife.

METHODS

The Study of Women’s Health Across the Nation (SWAN), a community-based 7-center observational study enrolled premenopausal women age 42–52 years as previously reported.¹² The ancillary Fat Patterning study conducted at the Chicago site coincided with annual SWAN follow-up visits 4 through 11 between 2000 and 2008. Additional 138 premenopausal women were recruited, because many SWAN participants had already transitioned through menopause. These women were originally screened for the SWAN study but were too young to participate at that time. Details of the Fat Patterning Study have been described elsewhere.⁵ Enrollees of the Fat Patterning Study were similar in all baseline characteristics to those who did not enroll. Of the 435 women enrolled, 41 were excluded from our analysis due to surgical hysterectomy (due to distortions of intra-abdominal anatomy) or diagnosis of diabetes (due to diabetes and therapy-related effects on IAAT). Further 57 women were excluded as they did not have longitudinal IAAT data. Thus the final cohort consisted of 337 women, 78% of all Fat Patterning Study participants.

IAAT assessments were conducted annually by planimetric computed tomography obtaining a single 10-mm thick abdominal image at the level of L₄–L₅ vertebral interspace, as reported previously.^{5, 13} Scans were read using software developed by the reading center at the University of Colorado Health Sciences Center (RSI, Boulder, CO). This protocol is in line with other longitudinal studies that use planimetric (cm²) rather than volumetric methodology to estimate IAAT volume or weight (cm³ or kg). Planimetric IAAT assessment at the level of L₄–L₅ vertebral interspace correlates well with total IAAT volume, and has been used most frequently in other large studies.^{14, 15} Total body fat was assessed on the same day with whole body dual-energy X-ray absorptiometry (GE-Lunar, Madison, WI), and expressed as a percent of fat in soft tissue. Scans were analyzed with GE-Lunar enCORE software (GE-Lunar, Madison, WI).

A standardized SWAN protocol was used to assess age, race, menstrual bleeding patterns, hormone therapy use, depressive symptoms, and physical activity, as previously described.¹⁶ For the purpose of the present analysis, the women who had no menstrual bleeding in the 12 months prior to assessment were classified as postmenopausal. Hormone therapy was determined as use of estrogen or progesterone alone or in combination. Physical activity was estimated at the initial assessment as a sum score of modified the Kaiser Physical Activity Survey¹⁷ and depressive symptoms were assessed with the Center for Epidemiologic Studies Depression Scale (CES-D).¹⁸

Distributions of the variables were checked, and violations of normality were detected in the outcome of interest, IAAT, which was log-transformed. Time was calculated in years based on assessment completion dates. A series of linear mixed-effects models were conducted with log-transformed IAAT as the outcome and race as the principal predictor, adjusting for time, age at initial assessment, total body fat (time-varying) and IAAT at the initial assessment (model 1); with additional adjustment for time and race interaction (model 2); adding menopausal status, hormone use, physical activity and depressive symptoms (model 3). The annualized percent change in IAAT was calculated multiplying the time estimate by 100.¹⁹

Model fit was evaluated with appropriate diagnostics. All analyses were conducted using PC-SAS^® 9.2 (SAS Institute Inc., Cary, NC).

RESULTS

The race-stratified cohort characteristics at the initial assessment are shown in Table 1. There were no racial differences in age or menopausal stage. Black women were less likely to use hormonal therapy, had lower physical activity and higher depressive symptom scores.

Table 1.

Cohort characteristics at initial evaluation stratified by race (unadjusted variable comparisons).

Variables	Black n=171 (44%)	White n=222 (56%)
Age (years), mean±SD	50.7±3.9	50.4±3.6
Premenopausal stage,1 n (%)	102 (60)	145 (65)
Menopausal hormone therapy, n (%)	8 (5)4	36 (16)4
CES-D2 score, mean±SD	8.1±8.35	6.0±6.75
KPAS3 sum score, mean±SD	7.3±1.64	8.2±1.54
Total body fat (%), mean±SD	45.1± 8.24	41.3±8.74
Initial IAAT (cm2), geometric mean (95% CI)	81.3 (74.9–88.3)	77.4 (71.4–84.0)

¹Premenopausal stage was defined as regular menstruations within the past 3 months.

²CES-D score – Center of Epidemiological Study-Depression scale (scores ≥16 are indicative of depression).

³KPAS – Kaiser Physical Activity Scale.

⁴P<0.001

⁵P<0.01

The average follow up time between the first and the last IAAT assessments was 2.1 ± 0.4 years, with a median of 3 IAAT assessments per participant. The annualized total body fat change −0.32 (95% CI −0.74 – 0.10) was non-significant in both races. As shown in Table 2, there were no differences in annualized IAAT gain estimate between black and white women. The estimated mean annualized percent of IAAT gain was 3.9%, after adjustment for age, race, initial IAAT and time-varying total body fat (model 1). Likewise, the mean annualized IAAT gain was 4.4 % in the models that additionally adjusted for annual IAAT and race interaction (model 2), and also menopausal status, menopausal hormone therapy, depressive symptoms, and physical activity (model 3). Likewise no significant racial differences in the IAAT change were found when the models did not include the adjustment for total body fat (results not shown).

Table 2.

Intra-abdominal adipose tissue (IAAT) gain: the mixed effects model analysis with random intercept and logarithmic transformation of IAAT outcome.

Variables	Model 1		Model 2		Model 31
	Estimate (SE)	p-value	Estimate (SE)	p-value	Estimate (SE)	p-value
Time (years)	0.039 (0.004)	<0.001	0.044(0.005)	<0.001	0.044 (0.005)	<0.001
IAAT at initial assessment (cm2)	0.010 (0.0002)	<0.001	0.008 (0.0002)	<0.001	0.008 (0.0003)	<0.001
Total body fat (%)	0.019 (0.001)	<0.001	0.019 (0.001)	<0.001	0.019 (0.001)	<0.001
Age at initial assessment (years)	0.005 (0.003)	0.068	0.005 (0.003)	0.067	0.007 (0.004)	0.036
Black race	−0.040 (0.022)	0.071	−0.027 (0.024)	0.242	−0.025 (0.025)	0.329
Time and race interaction	-	-	−0.014 (0.009)	0.122	−0.015 (0.009)	0.097

¹Adjusted for menopausal status, menopausal hormone therapy, depressive symptoms, and physical activity.

DISCUSSION

The principal finding was that community-dwelling black and white midlife women had similar rates of naturalistic IAAT gain over up to 4 years of adiposity assessments. The comparisons were conducted with adjustment for total adiposity, as cross-sectional studies suggested racial differences in IAAT after adjustments for total body fat.^{4, 5, 6, 7} Although short-term IAAT changes during weight loss have been studied previously,^{8, 9, 10, 11} our report provides a unique racial comparison of long-term IAAT gain in a naturalistic setting among midlife women.

Black women had significantly higher total body fat than white women, which may explain no significant racial differences in unadjusted IAAT area at the initial assessment. This finding is consistent with previously published baseline IAAT comparisons from the SWAN Fat Patterning Study,^{4, 5, 13} and aligns with other published research suggesting that white women have more IAAT than black women after adjustment for differences in total body fat.^{6, 7, 8}

The principal strength was a naturalistic longitudinal observation of adiposity changes in a community-based, representative sample of midlife women from a well-characterized cohort, with validated methodology applied for assessment of intra-abdominal and total body fat across the representational adiposity spectrum. Our report is also distinct for demonstrating the similar rates of long-term IAAT gain among black and white midlife women, without significant change in total body fat over time (i.e., similar rates of adipose tissue redistribution during midlife in black and white women). Our findings are in line with a similar study,²⁰ which also found no racial differences and larger changes in IAAT than in total body fat. However, their cohort was younger, normal weight or with recent intentional weight loss to achieve normal weight.²⁰

The potential limitation of this study was planimetric rather than volumetric methodology used for IAAT assessments. To date, more studies have used planimetric (cm²) rather than volumetric methodology (cm³ or kg) to estimate IAAT. A planimetric IAAT assessment at the level of the L₄–L₅ vertebral interspace (as a proxy for a volumetric IAAT assessment) has been used commonly in epidemiological studies, where larger sample sizes compensate for the lower precision of planimetric methodology.^{14, 15} Volumetric measurement of IAAT change has been necessary for smaller studies, limited by the need for higher measurement precision.^{14, 15} Our study used the annualized percent of IAAT change as a practical metric that permits comparisons across studies or methodologies for IAAT estimation.^{14, 15} Thus, our findings can be applied to design intervention trials for prevention of IAAT gain during the midlife transition.

CONCLUSION

The mean annualized percent of IAAT change was similar among black and white midlife women after adjustment for total body fat, baseline IAAT, age, menopausal status, hormone therapy, depressive symptoms, and physical activity. Although weight loss interventions are important, a more rational way to reduce cardiometabolic risk during midlife is to prevent menopause-associated IAAT gain in the first place. Since the IAAT gain was similar among black and white women, the same preventive abdominal adiposity targets apply to both races.

What is already known about this subject

• Intra-abdominal adipose tissue is directly associated with cardiometabolic risk.

• At the same levels of total adiposity, intra-abdominal adipose tissue is higher in white than in black women.

What this study adds

• At the same levels of initial adiposity, naturalistic gain of intra-abdominal adipose tissue during midlife is similar among black and white women.