35 / 64
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 31, No 6, November/December 2020
AFRICA
315
Japanese are ≥ 90 cm for women and ≥ 85 cm for men.
9
To date,
cut-off points for SSA countries have not been established,
with isolated studies from several countries reporting different
values.
8,13-17
Hence, the IDF has since recommended the use of
the European cut-off points to determine AO for SSA countries,
and by extension the MetS, until more specific data are available.
9
The MetS has been reported to be on the rise in SSA
countries, including Botswana. A couple of studies have reported
a MetS prevalence of 27 and 34% among Batswana adults
13,15
in different populations. For the MetS to be prevented, it is
necessary to ascertain the prevalence rates and also characterise
the population most affected. It is vital that the diagnostic tools
used are accurate and population specific. Therefore, the primary
aim of this study was to determine population-specific optimal
cut-off points for AO among Batswana adults. A secondary aim
was to evaluate how the prevalence of the MetS was affected
when the newly determined cut-off values were used to diagnose
AO compared with the IDF-recommended values.
Methods
A cross-sectional study employing a complex, multi-stage,
cluster-sampling method was used to recruit 1 000 participants in
the city of Gaborone and the surrounding villages of Tlokweng
and Mogoditshane. The target population included apparently
healthy male and female citizens aged 25 to 65 years residing in
Gaborone, Tlokweng and Mogoditshane. Data were collected at
shopping malls where people of heterogeneous characteristics
converged. There were a total of 37 malls, classified as either
super mall or satellite mall, spread across Gaborone, Tlokweng
and Mogoditshane. A total of seven malls were randomly
selected for the study.
A sampling frame of all shopping malls in Gaborone and
surrounding villages of Tlokweng and Mogoditshane was
compiled and samples were randomly selected from the list.
In the selected malls, shoppers perceived to be eligible were
systematically recruited as they passed by the testing area. The
recruitment was alternated with a stratifying variable, gender, to
ensure a balance between male and female participants. Once
eligibility had been established through the use of a national
identity card, and consent to participate had been sought, the
participants were enrolled into the study. The opportunity was
also used to share the results with the participants, to help them
appreciate their health parameters and the implications.
Ethical approval was obtained from the institutional review
boards of both the University of Botswana (ref no: IRB0005239)
and Stellenbosch University (ref no: HREC N13/04/052), and
human research office of the Ministry of Health, Botswana.
Data were collected from volunteers who agreed to participate in
the study after an introduction and briefing on the study intent.
After informed consent was obtained, a total of 803 volunteers
were recruited, interviewed and assessed.
Data regarding demographic, anthropometric, biochemical
and behavioural factors were collected using the standardised
methods stipulated in the World Health Organisation (WHO)
STEP-wise approach to chronic disease risk-factor surveillance
instrument (version 2.0). All procedures as specified by the
Health Professions Council of Botswana were strictly adhered
to according to good clinical practices.
Demographic data included questions related to gender, age,
level of education, marital and work status, number of people
above 18 years in the household and average yearly earnings.
Medical history included questions related to history of elevated
BP, diabetes, total cholesterol (TC) and cardiovascular diseases.
Resting BP, height and weight, and waist and hip circumferences
were recorded. Biochemical assessments related to BG and blood
lipids (TC, TG and HDL-C) were determined. BP measurements
were taken using an automated BP monitor (OMRON Intelli
Sense M3W) after a five-minute rest following the interview.
Waist and hip circumferences were measured using a SECA
measuring tape (201 cm) according to the International Society
for the Advancement of Kinanthropometry (ISAK) standard
guidelines.
18
Height was measured using a portable SECA
stadiometer, using the free-standing method with the head
placed in the Frankfort plane, and weight was measured with the
SECA Alpha digital scale (model 770). Capillary BG, TG and
TC were assessed with a finger-prick test with the Care Sense N
BG monitoring system, while blood lipids were measured using
the cardio check PA system [Polymer Technology Systems (Pts),
IN, USA].
Statistical analysis
Statistical analysis was conducted using the Statistical Package
for Social Sciences (SPSS version 22). Most of the variables
analysed for this output were quantitative, so their means ±
standard deviations are reported. All variables tested were
contrasted by gender using the Student’s
t
-tests. Contrasts
for proportions among sub-populations were performed using
homogeneity chi-squared tests. A significant difference was
accepted if
p
-values were less than 0.05 for both tests.
The MedCalc software was used to generate the receiver
operating characteristics (ROC) curves for continuous variables.
The area under the curve (AUC) was measured to summarise
the ability of WC to detect participants with at least two
risk factors for the MetS, as defined by the IDF.
9
The closer
the AUC is to one, the higher the ability of the indicator to
discriminate among subjects. An AUC value of 0.5 indicated
no discriminatory power of the tested indicator. The optimal
cut-off values were calculated by plotting the true-positive
rate (sensitivity) against the false-positive rate (1-specificity):
based on the Yoden Index Parametric method under a normal
distribution approach. The optimal cut-off point was calculated
by maximising sensitivity and specificity across various cut-off
point scenarios.
Results
The participants were mainly residents of Gaborone city and
two surrounding villages of Tlokweng and Mogoditshane, who
make up about 10% of the Batswana. A total of 800 respondents
participated in the study drawn from seven shopping malls,
yielding a response rate of 80%. From the 800 participants,
664 were selected from shopping malls in Gaborone, and the
remaining 136 from Tlokweng and Mogoditshane. The final
results were based on the 756 participants (363 men and 393
women) who had complete WC data.
The men and women were of similar age (36.1 ± 8.9
years) (Table 1). Based on anthropometric and biochemical
measurements, no gender differences were noted forWC, diastolic