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

58 / 102

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 27, No 3, May/June 2016 180 AFRICA
factors, including the following variables in the initial model:
age, FMI, trunk FM and leg FM. Based on the results relating
to the covariates described above, we also included the covariates
contraceptive use and alcohol consumption into the appropriate
models (Table 4). We then repeated the regression analyses
including VAT and SAT in the models.
In black women, FMI and trunk FM accounted for 21% of
the variance in fasting glucose concentrations, whereas in white
women, age and FMI contributed significantly to the model,
accounting for only 12% of the variance. In both black and white
women, trunk FM and leg FM were independently associated
with fasting serum insulin concentrations and HOMA-IR, and
together with age, and in the case of black women, contraceptive
use accounted for 40–45% of the variance in the models. The
addition of VAT and SAT to the models did not contribute
independently or significantly to the models for fasting plasma
glucose and measures of IR in the black and white women.
For the black women, trunk FM and leg FM were
independently associated with TG concentrations, whereas only
FMI and leg FM, as well as contraceptive use were associated
with TG concentrations in the white women. The addition of
VAT and SAT did not contribute significantly to the model in
both black and white women. In the black women, HDL-C
concentrations were independently associated with age, FMI,
leg FM, contraceptive use and alcohol consumption, whereas in
white women, only VAT and contraceptive use contributed to the
model. Notably, the associations between HDL-C concentrations
and contraceptive use were opposite in the black and white
women, showing a negative association in black women and a
positive association in white women.
The model that explained the greatest variance in TC and
LDL-C concentrations in the black women included age, FMI
and abdominal SAT, the latter being negatively associated
with TC and LDL-C concentrations. This contrasts with the
findings for white women, where age, FMI and contraceptive
use accounted for the greatest variance in TC concentrations,
and age and trunk FM contributed to the model for LDL-C
concentrations.
Discussion
The main findings of this study were that compared to white
women, black women had less central and more lower-body
fat, and lower fasting glucose and lipid concentrations, but had
similar levels of IR. Despite these differences, the associations
between body fat distribution and measures of IR, as well as
TG and HDL-C concentrations were similar in black and white
women. The novel finding of this study was that central and
peripheral fat depositions were independently associated with IR
in both the black and white women, and with TG concentrations
in the black women. By contrast, fasting glucose concentrations
were associated with centralisation of body fat in black, but
not white women, whereas TC and LDL-C concentrations were
associated with centralisation of body fat in white, but not black
women.
Black women in this study had more total body fat compared
to their white counterparts. This is in accordance with recent
national SA prevalence data, which reported that black women
had a higher prevalence of obesity than other ethnic groups.5
However, when adjusting for total body fat, black women
had a greater peripheral distribution of fat, characterised by
less central FM and more lower-body FM than their white
counterparts. Furthermore, within the abdominal depot, black
women had less VAT and more SAT compared to white women,
which is commensurate with both SA and American studies.15-19
Less central FM, and to a lesser extent, more peripheral FM
in black women, associated with their lower fasting glucose
concentrations, suggesting that accumulation of central FM
may play a vital role in determining fasting plasma glucose
concentrations, and hence the development of T2D in black
women. By contrast, despite the differences in body fat
distribution, fasting insulin levels and HOMA-IR values were
not significantly different between black and white women.
Numerous studies have shown that compared to white women,
black women have a higher prevalence of IR and T2D for
the same BMI or waist circumference.17,19 These results are
surprising, given that greater central and reduced peripheral FM
were similarly associated with higher fasting insulin and HOMA-
IR values in both black and white women, a finding supported
by similar studies in the USA.15,27 These findings suggest that
Table 4. Multivariate analysis for black and white women, separately
Black women White women
Variable B p p-value Variable B p p-value
Glucose (mmol/l) Glucose (mmol/l)
FMI (kg/m2) –1.11 < 0.01 Age (years) 0.18 0.01
Trunk FM (kg) 1.47 < 0.01 FMI (kg/m2) 0.24 0.00
r = 0.45 r2 = 0.21 < 0.01 r = 0.34 r2 = 0.12 < 0.01
Insulin (mU/l) Insulin (mU/l)
Age (years) –0.33 < 0.01 Age (years) –0.25 0.00
Trunk FM (kg) 1.09 < 0.01 Trunk FM (kg) 1.02 0.00
Leg FM (kg) –0.53 < 0.01 Leg FM (kg) –0.37 0.00
Contraception 0.11 < 0.01
r = 0.64 r2 = 0.40 < 0.01 r = 0.67 r2 = 0.45 < 0.01
HOMA-IR HOMA-IR
Age (years) –0.29 < 0.01 Age (years) –0.21 0.00
Trunk FM (kg) 1.14 < 0.01 Trunk FM (kg) 1.03 0.00
Leg FM (kg) –0.57 < 0.01 Leg FM (kg) –0.38 0.00
Contraception 0.11 0.02
r = 0.65 r2 = 0.42 < 0.01 r = 0.67 r2 = 0.45 < 0.01
TG (mmol/l) TG (mmol/l)
Age (years) 0.12 0.04 FMI (kg/m2) 0.90 0.00
Trunk FM (kg) 0.84 0.00 Leg FM (kg) –0.51 0.00
Leg FM (kg) –0.59 0.00 Contraception 0.26 0.00
r = 0.48 r2 = 0.23 < 0.01 r = 0.49 r2 = 0.24 < 0.01
HDL-C (mmol/l) HDL-C (mmol/l)
Age (years) 0.20 0.07 VAT (cm2) –0.45 0.00
FMI (kg/m2) –1.20 < 0.00 Contraception 0.18 0.01
Leg FM (kg) 0.77 0.01
Contraception –0.23 0.02
Alcohol
consumption
0.23 0.02
r = 0.51 r2 = 0.26 < 0.01 r = 0.51 r2 = 0.26 < 0.01
TC (mmol/l) TC (mmol/l)
Age (years) 0.22 0.0021 Age 0.22 0.00
FMI (kg/m2) 0.38 0.025 FMI (kg/m2) 0.29 0.00
SAT (cm2) –0.36 0.029 Contraception 0.29 0.00
r = 0.30 r2 = 0.10 < 0.01 r = 0.44 r2 = 0.19 < 0.01
LDL-C (mmol/l) LDL-C (mmol/l)
Age (years) 0.19 0.0071 Age 0.16 0.02
FMI (kg/m2) 0.46 0.0073 Trunk FM (kg) 0.38 0.00
SAT (cm2) –0.33 0.049
r = 0.33 r2 = 0.11 < 0.01 r = 0.46 r2 = 0.21 < 0.01