Cardiovascular Journal of Africa: Vol 30 No 6 (November/December 2019)

18 / 66

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 30, No 6, November/December 2019 326 AFRICA
to the greatest variance in triglyceride concentrations (13%).
Additionally, leg %FM also accounted for the greatest variance
in HDL-C (14%) levels. Body composition did not add to
variance in TC and LDL-C concentrations above that of age
and gender.
There were gender-specific differences in the associations
between body fat distribution and measures of cardiometabolic
risk. While fat mass (kg) and abdominal SAT were associated
with two-hour glucose (Fig. 1A, C), HOMA-IR (Fig. 1B, D) and
fasting insulin levels in both men and women, the association
was stronger in men compared to women (all interactions p ≤
0.027). Conversely, the associations between serum triglycerides
and the distribution of body fat were significant in women,
but not in men. Specifically, central adiposity measures (trunk
%FM and android %FM) were positively associated with serum
triglyceride concentrations (Fig. 2A, B, all interactions p ≤
0.014), and peripheral fat mass (leg %FM and gynoid %FM)
were negatively associated with serum triglyceride levels (Fig.
2D, E, both interactions p ≤ 0.022) in women, but not in men.
The association between VAT and triglyceride concentrations
was stronger in women than men (Fig. 2C, interaction p = 0.012).
The associations between body composition and cardio-
metabolic risk factors in pre-and post-menopausal women are
shown in Table 4. For the most part, the association between
body fat distribution and cardiometabolic risk did not differ
between the pre- and post-menopausal women. Significant
interactions were however seen between fat mass (kg), central
fat distribution (%FM, trunk %FM, VAT and SAT area) and
SBP (all interactions p ≤ 0.042) and VAT and DBP (interactions
p = 0.030), such that these were significant in pre-menopausal
women but not in post-menopausal women.
Similarly, fat mass (kg), fat mass (%), VAT and SAT were
associated with TC (all interactions p ≤ 0.002) and LDL-C levels
(all interactions p ≤ 0.007) in pre-menopausal women only. Fat
mass (kg) was associated with fasting insulin in both pre- and
post-menopausal women, but the association was stronger in
pre- than post-menopausal women (interaction p = 0.019), while
fat mass (kg) was associated with triglyceride concentrations in
pre-menopausal women only (interaction p = 0.016). Peripheral
fat (arm %FM and gluteofemoral %FM) was associated with
LDL-C (both interaction p ≤ 0.032) and hsCRP levels (both
interactions p ≤ 0.012) in pre- but not post-menopausal women.
Discussion
This is the first study to investigate the relationship between
body composition and cardiometabolic risk profile in mixed-
ancestry South Africans. The main findings of the study are that
body fat and, in particular central adiposity, were associated
with unfavourable cardiometabolic risk profiles, while lower-
body peripheral fat was associated with favourable risk profiles.
However, the associations between body fat distribution and
cardiovascular risk profile differed by gender and menopausal
status, such that the associations were stronger in men and
pre-menopausal women.
Although the women in our sample had nearly twice as
much body fat mass, and had higher obesity rates than the
men, the prevalence of cardiometabolic risk factors was similar
between genders (apart from two-hour glucose and HDL-C
concentrations being higher in women). This may be explained
by the fact that despite marked differences in total body fat,
VAT area was similar in men and women. Indeed, VAT was the
most consistent and significant correlate of cardiometabolic risk
(insulin resistance, glucose tolerance, triglyceride and HDL-C
concentrations) in this sample. Furthermore, the association
between VAT and cardiometabolic risk did not differ by gender.
Similarly, a recent study among Korean men and women
showed that DXA-derived VAT was the best correlate of diabetes
and pre-diabetes. 42 Likewise, the meta-analysis by Zhang and
co-workers8 supports VAT as the strongest correlate of insulin
resistance, followed by total fat mass. The mechanisms linking
VAT accumulation to metabolic complications include a higher
production of pro-inflammatory cytokines and higher lipolytic
activity compared to SAT, with the consequent increase in
cytokine and free fatty acid delivery to the hepatic portal system
impacting on insulin sensitivity.5 VAT is also proposed to be a
marker of insulin resistance as a consequence of lipotoxicity, in
particular an increase in fat deposition in the liver.5
In contrast to VAT, women had more abdominal SAT than
men.14,15 Notably, the relationship between both total adiposity
and abdominal SAT and insulin resistance was stronger in
men than women. These differences may relate to the fact that
oestrogens regulate insulin sensitivity and that female adipocytes
are more insulin sensitive compared with male adipocytes. 43
Alternatively, the gender-specific relationship between abdominal
SAT and insulin resistance could, in part, be explained by the
fact that men have greater deep SAT (dSAT) and less superficial
SAT (sSAT) than women.44
Nazare et al. showed that of the two SAT layers, dSAT had
a higher association with inflammation and oxidative stress,
suggesting that dSAT is an important determinant of the
Table 4. Associations between body composition and cardiometabolic risk
factors in the pre- and post-menopausal women
Body composition SBP DBP
Fasting
insulin TG TC LDL-C hsCRP
FM (kg)
Pre-meno 0.264a 0.292a 0.629A 0.502A 0.459A 0.394A 0.592A
Post-meno –0.092# 0.028 0.380 B# 0.081# –0.120# –0.079# 0.198
FM (%)
Pre-meno 0.198 0.246a 0.509A 0.357A 0.462A 0.451A 0.528A
Post-meno –0.123# 0.006 0.318B 0.001 –0.073# –0.053# 0.226b
Trunk fat (%FM)
Pre-meno 0.443A 0.507A 0.562A 0.585A 0.199 0.128 0.504A
Post-meno –0.002# 0.118 0.358B 0.525B 0.126 0.004 0.067#
Arm fat (%FM)
Pre-meno 0.209 0.283a 0.156 0.203 0.190 0.260a 0.400A
Post-meno 0.131 0.158 0.215b 0.076 –0.098 –0.076# –0.023#
Gynoid (%FM)
Pre-meno –0.449A –0.531A –0.615A –0.591A –0.242a –0.164 –0.679A
Post-meno –0.158 –0.107 –0.388B –0.450B 0.023 0.137# 0.074#
VAT (cm)2
Pre-meno 0.415A 0.436A 0.737A 0.635A 0.411A 0.339A 0.597A
Post-meno –0.047# 0.037# 0.519B 0.336B –0.086# –0.098# 0.243b
SAT (cm)2
Pre-meno 0.274a 0.334A 0.601A 0.533A 0.430A 0.369A 0.605A
Post-meno –0.117# 0.063 0.387B 0.132 –0.059# –0.040# 0.231b
Values are Spearman’s correlation coefficients. %FM, expressed as percentage of
sub-total fat mass (FM). SBP, systolic blood pressure; DBP, diastolic blood pressure;
TG, triglycerides; TC, total cholesterol; LDL-C, low-density lipoprotein cholesterol;
hsCRP, high-sensitivity C-reactive protein.
ap < 0.05 and Ap < 0.01 for pre-menopausal women; bp < 0.05 and Bp < 0.01 for post-
menopausal women; #p < 0.05 for age × body composition interaction.