CARDIOVASCULAR JOURNAL OF AFRICA • Vol 24, No 6, July 2013
AFRICA
239
practitioners, 13 were cardiologists, seven endocrinologists,
and 14 internists/specialist physicians. Investigators were drawn
from both the public and private sectors. Public sector sites were
located in tertiary level referral hospitals.
Subjects 18 years or older who had been receiving lipid-
lowering drugs (LLDs) for at least three months (without dose
adjustments for at least six weeks) were eligible. Consecutive
patients who came for their regular scheduled visit to the doctor/
clinic were invited to participate in the survey. Patients who
agreed to participate provided informed written consent.
CEPHEUS was a single-visit non-interventional study. Each
patient’s record form documented patient demographics, current
LLD treatment, smoking status, known diabetes mellitus (DM),
family history of premature vascular disease, known arterial
hypertension (HT) and cardiovascular medical history. Physical
examination by the investigator was limited to measurement
of height, weight, waist circumference and blood pressure.
A fasting blood sample was drawn to evaluate the serum
lipid profile [total cholesterol, LDL-C, high-density lipoprotein
cholesterol (HDL-C), triglycerides and apolipoprotein (apo)
AI and apo B], fasting blood glucose (FG) and glycosylated
haemoglobin (HbA
1c
) levels.
The primary endpoint was the percentage of patients who
achieved the LDL-C goals according to either the NCEP ATP
III/2004 updated NCEP ATP III guidelines or the fourth JETF/
South African guidelines, which were current in South Africa
at the time the CEPHEUS SA study was conducted. Secondary
endpoints included achievement of LDL-C goals in patients with
and without features of the metabolic syndrome, and primary
versus secondary prevention.
The parent study (CEPHEUS-Europe) has been registered
with the US National Institutes of Health (ClinicalTrials.gov),
number NCT00542867. The CEPHEUS study was sponsored
by AstraZeneca. The sponsor oversaw data collection and
monitored study sites. The authors had full access to the
study database and all analyses reported here were performed
independently of the sponsor.
Statistical analysis
We subdivided the cohort by gender and ethnicity for the
purposes of this analysis. The four major ethnic groups in South
Africa were black Africans, Caucasians, Asians (including
patients of Indian descent) and patients with mixed ancestry. The
risk category was determined for each patient and we calculated
a dichotomous variable for each patient indicating whether their
LDL-C had reached the guideline mandated target level.
We generated descriptive statistics for all clinical and
laboratory parameters, following subdivision by gender only,
and then following subdivision by both ethnicity and gender. We
analysed the effect of ethnicity and gender on goal attainment
using logistic regression with the logit function in a model that
incorporated ethnicity and gender simultaneously.
We calculated odds ratios and 95% confidence intervals for
the probability of not attaining LDL-C goal. The probability
of not attaining LDL-C goal was referenced against Caucasian
ethnicity and male gender, for which the odds ratio was set
as 1. All
p
-values are two-sided and we regarded
p
<
0.05
as statistically significant. All analyses were performed with
Statistica [StatSoft Inc (2011), STATISTICA (data analysis
software system) version 10,
.
Results
A total of 3 001 patients consented to participate in the survey.
Full data sets were available from 2 996 patients and form the
basis of this report. About two-thirds of patients were recruited
from the private healthcare sector, with the remaining one-third
coming from public sector institutions.
Demographic, anthropometric and clinical data are shown in
Table 1. Of the total group, 47.1% had known DM but 2.4% of
patients who did not give a history of DM had FG levels that
would qualify for the diagnosis of DM. Glycaemic control in
patients who gave a history of DM was generally poor, with a
mean HbA
1c
level of 8.33%.
The prevalence of a history of HT in this study was 71.6%.
The mean systolic blood pressure in the entire study cohort was
133.2 mmHg, with a diastolic pressure of 80.2 mmHg. In those
subjects with a history of HT (Table 2), the mean systolic blood
pressure was 136.1 mmHg, with a diastolic pressure of 81.3
mmHg. African-ancestry males had the highest systolic blood
pressure and females of mixed ancestry the highest diastolic
blood pressure but the inter-ethnic differences were small.
More Caucasian patients were receiving LLDs for primary
prevention compared to those of African ancestry, few of whom
were on treatment for primary prevention. The majority were
receiving treatment for the CVD risk equivalent of DM.
The percentage of African patients who were on LLDs for
TABLE 1. BASELINE CHARACTERISTICS
Characteristics
Entire study
Caucasian
African
Mixed ancestry
Asian
Male Female Male Female Male Female Male Female Male Female
Number
1572 1424
818
567
168
342
222
259
364
256
Mean age (years)
59.2
59.6
60.9
62.0
57.4
57.4
58.1
59.0
56.8
58.0
Current smoker (%)
18.6
10.8
15.9
14.3
15.6
2.6
25.1
19.7
21.4
5.1
Family history of vascular disease (%)
27.0
29.5
29.0
36.3
4.1
11.1
25.2
30.9
39.2
37.5
Mean body mass index (kg/m
2
)
29.2
30.8
30.0
29.4
29.3
34.2
29.0
31.2
27.6
29.4
Mean waist circumference (cm)
101.0 101.0 105.7 95.6 101.5 102.2 101.7 100.4 99.4
97.6
Known diabetes mellitus (%)
45.8
48.5
34.4
26.8
70.6
74.9
54.1
54.5
38.7
55.1
Known systemic hypertension (%)
68.8
74.6
64.7
64.7
84.6
88.9
76.6
84.2
65.8
67.9
History of coronary heart disease (%)
45.8
23.9
46.3
19.6
19.0
14.3
58.1
38.2
49.4
31.6
History of cerebrovascular disease (%)
5.8
4.8
5.3
4.4
5.9
6.1
6.3
4.2
6.3
4.2
History of peripheral arterial disease (%)
6.2
3.5
7.8
3.2
3.0
2.6
6.8
6.2
3.3
2.7
