Cardiovascular Journal of Africa: Vol 27 No 3 (May/June 2016)

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 27, No 3, May/June 2016

178

AFRICA

for metabolic disorders; not currently pregnant, lactating or

postmenopausal; and being of SA ancestry.

This study was approved by the Human Research Ethics

Committee of the Faculty of Health Science, University of Cape

Town. Procedures and risks were explained to participants, all

of whom gave written informed consent, prior to participation.

The testing procedures and biochemical analyses have been

described previously,

but are described briefly below. A demo-

graphic questionnaire

was administered and included measures

of socio-economic status, including housing density, family

history of T2D and behavioural/lifestyle factors. Contraceptive

use was recorded, and women were categorised as using

hormonal contraception (oral and injection) or not. Smoking

was recorded and women were categorised as current smokers or

not. Alcohol consumption in grams/day was also recorded using

dietary recall.

Physical activity energy expenditure was characterised using

the global physical activity questionnaire (GPAQ).

Moderate-

to vigorous-intensity physical activity (MVPA) was calculated as

minutes of physical activity per week.

Anthropometric measurements of participants were taken,

including height, weight in light clothing, waist circumference

(at the level of umbilicus) and hip circumference (at the largest

gluteal area). Body composition (FM and fat-free mass) was

measured using DXA (Discovery-W, Software version 4.40;

Hologic). Fat mass index (FMI) was calculated as total body fat

(kg)/height (m

). DXA-derived measures of body fat distribution

included trunk, arm and leg FM.

Trunk FM included the region between the neck (line

below the bottom of the jaw) and waist cut-offs (line above the

iliac crest), with the lateral boundaries positioned to achieve

separation of the upper arm and trunk at the glenoid fossa,

and the inclusion of vertical lines on either side of the spine

were positioned to exclude the spine. The arms included the

region below the line through the glenoid fossa. Vertical lines

extending downward from the waist cut-off were positioned to

separate thigh from hands, and oblique lines were positioned to

pass through the femoral neck and join the central vertical line

between the legs, in order to isolate the legs.

A CT scan (Toshiba

X-press Helical Scanner; Toshiba Tokyo, Japan) at the level of

the L4–L5 vertebrae was used to determine VAT and SAT areas.

After an overnight fast (10–12 hours), a blood sample was

drawn from the antecubital vein for the subsequent determination

of plasma glucose, serum insulin, high-density lipoprotein

cholesterol (HDL-C), low-density lipoprotein cholesterol (LDL-

C), total cholesterol (TC) and triglyceride (TG) concentrations.

Plasma glucose concentrations were determined using the glucose

oxidase method (YSI 2300 Stat Plus; YSI Life Sciences, Yellow

Springs, OH). Serum insulin concentrations were determined by

immunochemiluminometric assays using the ADVIA Centaur

(Bayer Diagnostics, Tarrytown, NJ). Blood lipids were measured

using the Roche modular autoanalyser (Roche Diagnostics

GmbH, Mannheim, Germany). LDL-C was calculated using the

Friedewald formula.

HOMA-IR was calculated from fasting

glucose and insulin levels (glucose (mmol/l)

insulin (mU/l)/22.5).

Statistical analysis

Results were analysed using Statistica version 10 (Statsoft Inc,

Tulsa, Oklahoma, USA). Results are presented as median and

interquartile range (IQR). All skewed variables were normalised

by log transformation where required. Ethnic differences in body

composition, IR and lipid levels were determined using one-way

ANCOVA, adjusting for age. Pearson’s chi-squared was used to

determine differences in categorical variables between the black

and white women. Partial correlations were used to determine

the associations between the various body fat distribution

variables and cardiometabolic outcomes in the black and white

women, adjusting for age and FMI. FMI was chosen as the

covariate because it takes into account both the total body

fat and the height of an individual (which differs significantly

between black and white women).

Multiple regression analysis was used to determine the

independent associations between body fat distribution and IR

and serum lipid levels, adjusting for age and FMI. In addition,

the effect of ethnicity on these relationships was tested by

including ethnicity

body fat distribution interaction term in

the model. Backward stepwise regression was used to determine

the model that accounted for most of the variance for each

cardiometabolic outcome. In each model, trunk FM and leg FM

were included in the model, with age, FMI, contraception use,

MVPA, alcohol consumption and smoking. The analyses were

then repeated including VAT and SAT in the model (due to the

smaller sample size).

Results

The black women were younger than the white women [median

(IQR): 22 (22–33) vs 32 (24–39] years,

0.01] and consequently,

all subsequent analyses were adjusted for age. Black women had

higher levels of MVPA compared to white women [335 (90–855)

vs 240 (120–480) min/week,

0.01], and fewer black women

smoked (10.1 vs 17%,

0.04), whereas alcohol consumption

did not differ between the groups [0 (0–2.8) vs 61 (0.5–14.7),

0.95]. There was no significant difference in the proportion of

women who used contraceptives (32.0 vs 31.1%,

0.74), but

more black women used injectable contraceptives (25.7 vs 5.1%,

0.01), while more white women used oral contraceptives (26.0

vs 6.3%,

0.01).

Ethnic differences in body composition and fat distribution

are summarised in Table 1. Black women were significantly

shorter, heavier, had a higher body mass index (BMI) and

greater FM (absolute and %) compared to white women. Black

women had greater absolute measures of trunk, leg and arm

FM compared to white women. However, as a percentage of

total FM, black women had less trunk FM and more leg FM.

Accordingly, the trunk FM/leg FM ratio was greater in white

than black women. As a percentage of total body FM, there was

no significant difference in arm FM between black and white

women. Black women had less abdominal VAT and more SAT

and a lower VAT/SAT ratio compared to white women.

Cardiometabolic risk factors for black and white women

are summarised in Table 2. Although there were no ethnic

differences in fasting plasma glucose concentrations, black

women had higher fasting insulin concentrations and HOMA-

IR than white women. However, after adjusting for differences

in age and FMI, glucose concentrations were significantly lower

in the black compared to the white women, but the differences

in fasting serum insulin concentrations and HOMA-IR were no

longer significant. Black women had lower TC, TG, HDL-C