CARDIOVASCULAR JOURNAL OF AFRICA • Vol 23, No 1, February 2012
AFRICA
9
kg/m
2
), the CVD risk was significantly lower (
p
=
0.007) than in
obese females, while in hyperglycaemic states, observed differ-
ences were not significant (Table 3). In the best-subsets linear
regression analysis, the significant predictors of CVD were
sibling history of diabetes, and TG, LDL-C and HbA
1c
levels (
p
<
0.001) (Table 4). These variables accounted for 46.3% of the
variation in the calculated CVD risk.
Discussion
In this study, the CVD risk was significantly higher in subjects
with diabetes. It is well documented that individuals with type
2 DM have increased CVD risk compared to those without
diabetes, and the 30-year risk among subjects with diabetes has
been shown to proportionally increase with BMI.
26,27
However,
no significant differences were observed between obese and
normal-weight diabetics in this study.
In the Framingham cohort, the lifetime risk in obese diabetic
subjects was 78.8 and 86.9% in women and men, respectively.
27
Furthermore, approximately half of those with diabetes were
unaware that they had the disease. Chronic hyperglycaemia,
even in the absence of symptoms, is associated with an increased
risk of diabetic micro-angiopathy and CVD.
7
Undiagnosed
hyperglycaemia has previously been demonstrated in rural and
urban South Africans
2,3
and a recent study found the presence
of undiagnosed hyperglycaemia in patients with coronary artery
disease.
28
While significant differences were observed between the
estimated CVD risk in non-diabetic (excluding IFG and IGT)
normal-weight and obese subjects, particularly in females, high
risk scores were still evident in the normal-weight, normogly-
caemic individuals. The present study builds on a previous study
that found the 10-year CVD risk profile in a similar population
to be less than 10% in subjects under 35 years of age.
29
Ten-year risk estimates have been criticised because they
underestimate the risks, allowing for continued progression of
sub-clinical atherosclerosis. Indeed this is evident in our study
in which we present evidence of a high risk score (
>
20%) in
young individuals and in those subjects with normoglycaemia
(Figs 2, 3). Similarly, younger individuals of the Framingham
cohort with very low 10-year CHD risk were shown to still have
a substantial lifetime risk of CHD.
30
Overall, these data confirm
that irrespective of glycaemic or weight status, an evidence-
based tool is crucial for the identification of high-risk subjects.
Given the findings of this study and the estimated increases in
the incidence of both diabetes and hypertension in South Africa,
this trend may continue to worsen if current trajectories do not
change. Hypertension, a condition that was rarely observed in
non-Western populations in the 1940s, has emerged as the most
common cause of heart failure in Africa.
31,32
This has been attrib-
uted to the adoption of Western lifestyle and diet, with a parallel
increase in obesity, diabetes and hypertension.
32
Evidence from
systematic studies suggests that CVD risks that were almost
unprecedented in non-Caucasian South Africans are now appar-
ent in both rural and urban adult populations.
33-37
The intra-class correlation between the lipid- and BMI-based
equations showed that either method could be used for the esti-
mation of CVD risk. In developing countries, the use of methods
that do not require blood or invasive testing or those that can
use information that is easily obtainable in a primary healthcare
setting has been advocated. Except for confirmation of the
diabetes status, all measurements required to estimate an indi-
vidual’s CVD risk can be obtained in such a setting. Point-of-care
instruments are usually available in these healthcare facilities to
screen for diabetes and to assess one’s blood glucose levels and
could adequately suffice for the purpose of CVD estimation.
Normally, those with high glucose levels or equivocal results
have blood drawn for confirmation. Unfortunately, the subset
linear regression analysis showed that except for one, the other
risk factors that are associated with a high CVD risk also require
blood testing.
Our study observed a strong association between HbA
1c
and the estimated CVD risk. Clinical trials have convincingly
demonstrated that good glycaemic control can reduce the risk
of CVD in patients with both type 1 and type 2 diabetes.
38-40
However, it has recently been indicated that accelerated athero-
sclerosis and CVD in diabetes is likely to be multifactorial and
therefore a glucocentric approach is discouraged in the manage-
ment of diabetes on cardiovascular outcomes.
41,42
On the other
hand, the strong association of triglycerides and LDL-C suggests
that in some asymptomatic individuals, lipid-lowering therapy
may have to be initiated.
The major limitation of this study was the use of a risk
calculator that was developed in a predominantly white middle-
income cohort. The mixed ancestry, sometimes referred to as
coloured, is a South African population group with Khoi, San,
TABLE 3. CVD RISK STRATIFIED BY BMI AND GLYCAEMIC STATE
Females
Male
Non-diabetic
Diabetic
Non-diabetic
Diabetic
BMI
<
25 kg/m
2
20.8 (17.7–23.9)
49.7 (33.5–65.9)
27.6 (22.4–32.9)
55.1 (42.5–67.7)
BMI
≥
25
<
30 kg/m
2
24.8 (20.4–29.1)
65 (57.8–72.2)
43.5 (32.6–54.4)
72 (64.3–79.7)
BMI
≥
30 kg/m
2
31.3 (27.9–34.7)**
57.8 (52.3–63.3)
42.3 (28.3–56.3)
71.9 (60.8–83.0)
Mean (95% CI). Diabetic refers to all diabetic subjects including those diagnosed during the survey; non-diabetic excludes those with IGT or IFG.
**Significant difference between non-diabetic female BMI
<
25 kg/m
2
and non-diabetic female BMI
≥
30 kg/m
2
, (
p
=
0.007).
TABLE 4. BEST-SUBSET LINEAR REGRESSIONANALYSIS.
ADJUSTED R
2
=
0.46
Variable
b*
b
p
Sibling history of DM
0.14
9.16
<
0.0001
TG (mmol/l)
0.34
6.67
<
0.0001
LDL-C (mmol/l)
0.33
7.24
<
0.0001
HbA
1c
(%)
0.34
4.86
<
0.0001
Other variables initially included in the regression analysis were waist
circumference, subscapular and supra-iliac skin thickness, biceps
and triceps circumference, brother, sister, mother or father history of
diabetes and/or hypertension, death of parent from CVD, C-reactive
protein, and HOMA IR.
TG, triglycerides; LDL, low-density lipoprotein cholesterol.
