14 / 74
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 29, No 6, November/December 2018
346
AFRICA
Hypertension Society guidelines.
2
The readings were recorded
in the patients’ files.
•
ECG: 12-lead digital electrocardiogram, Shenzhen Biocare
Electronics Ltd (model E.C.G-1200). A resting 12-lead ECG
was done using the technique recommended by Noble and
colleagues.
22
The ECG was interpreted by the researcher with
LVH assessed using the Romhilt–Estes five-point score. This
has been reported to yield a specificity of 99%.
23
Participants with problematic alcohol use or smoking were
counselled and referred for assistance. To compensate for time
lost due to participating in the study, all participants were
attended to by a dedicated doctor and arrangements were
made with the pharmacy to immediately dispense medications
ahead of the queue. Data were captured on Microsoft Excel
spreadsheets daily and cross-checked with the second author.
A pilot study was conducted using 30 patients at a nearby
CHC in the same sub-district to assess the feasibility of the study.
The results of the pilot study are not included in the main study
but informed minor adjustments to some questions for ease of
participants’ understanding, for example, that a drink of alcohol
should be expressed in ml and not in oz, and that three possible
responses should be allowed for the question on assessment of
hypercholesterolaemia.
Ethics clearance was obtained from the Human Research
and Ethics Committee of the University of the Witwatersrand
(number M10929). Permission was obtained from the Sedibeng
District Health Services management. To ensure anonymity, the
questionnaires were coded using the corresponding file number
and we did not collect personal identifiable data. Patients who
were found to have a problem with alcohol use or smoking and
with worrying ECG findings were referred for further assistance.
Statistical analysis
Captured data were imported into STATA statistical analysis
software, version 10. A statistician assisted with analysis.
Descriptive statistics were performed to describe participants’
sociodemographic and clinical characteristics. Chi-squared and
t
-tests were used to compare groups, and variables that showed
significant associations on bivariate analysis were inputted into
multivariate analysis. A
p
-value < 0.05 was considered statistically
significant. Main outcome measures included: proportions of
participants with each CV risk factor (tobacco use, alcohol
use, physical inactivity, diabetes, hypercholesterolaemia, family
history of hypercholesterolaemia and fatal CV event) and the
socio-demographic correlates of each CV risk.
Results
There were 328 participants and their characteristics are shown
in Table 1. The mean age of participants was 57.7 years and most
participants were black (86.0%), female (79%) and pensioners
(43.6%). The mean systolic BP was 139/84 mmHg, with 60.7%
(199) having their BP controlled to targets.
In addition to hypertension, the 328 participants reported a
total of 1 232 cumulative CV risk factors; an average of 3.7 CV
risk factors per participant. Table 2 shows that the prevalence
of CV risk factors was as follows: abdominal obesity (80.8%),
physical inactivity (73.2%), diabetes (30.2%), alcohol use (28.0%)
and smoking (11.9%).
Most participants (60.4%,
n
=
198) had normal tracings on
ECG with only 5.2% (
n
=
17) showing LVH. Abnormalities other
than LVH were found in 34.4% (
n
=
113) of participants and
included: sinus bradycardia (52.2%), left-axis deviation (14.2%),
premature ventricular contractions (7.1%), right bundle branch
block (4.4%), T-wave changes (4.4%) and left bundle branch
block (2.6%).
On tests of associations between participants’ characteristics
and CV risk factors (Tables 3–5), age was significantly associated
with current alcohol use (
p
=
0.04), exposure to second-hand
smoke (
p
=
0.00) and physical inactivity (
p
=
0.00). Gender was
significantly associated with being diabetic (
p
=
0.03), physically
inactive (
p
=
0.02), current alcohol use (
p
=
0.00), obesity (
p
=
0.00), snuff use (
p
=
0.00) and cigarette smoking (
p
=
0.00). Race
was significantly associated with cigarette smoking (
p
=
0.00),
snuff use (
p
=
0.01), hypercholesterolemia (
p
=
0.01) and family
history of fatal CV event (
p
=
0.02 for females and
p
=
0.00 for
males). Marital status was associated with cigarette smoking (
p
=
0.03) and family history of fatal CV event (
p
=
0.02). Educational
level was significantly associated with snuff use (
p
=
0.03) and
family history of hypercholesterolaemia (
p
=
0.00). Lastly,
employment status was significantly associated with physical
inactivity (
p
=
0.00).
Table 6 shows the sociodemographic correlates of each CV
risk factor in multivariate regression analysis. Compared to
those aged 20–39 years, older patients were significantly more
likely to report being physically inactive but less likely to report
alcohol use.
Compared to women, men were more likely to report alcohol
use, cigarette smoking, being physically inactive and having
Table 1. Participants’ characteristics
Variable
% (
n
)
Age, years
Gender
Female
79 (260)
Male
21 (68)
Marital status
Divorced
6.4* (21)
Living together
3* (10)
Married
51.8* (170)
Not married
12.8* (42)
Widowed
25.9* (85)
Ethnic group
Asian
0.3 (1)
Black
86.0 (282)
Coloured
0.9 (3)
White
12.8 (42)
Employment status
Employed
30.8 (101)
Pensioner
43.6 (143)
Unemployed
25.6 (84)
Educational level
None
10.7 (35)
Primary
33.5 (110)
Secondary
53.7 (176)
Tertiary
2.1 (7)
Mean age, years (SD)
57.7 (10.8)
Mean weight: study population
85.4
*The total percentage with decimals was slightly less than 100% (98.9%), but
rounded to the nearest integer, it became 100%.