CARDIOVASCULAR JOURNAL OF AFRICA • Vol 23, No 9, October 2012
AFRICA
513
publications in only the English language.
The eligibility criteria for inclusion were: randomised
control, epidemiological, case-controlled, cohort studies, and
surveys with a mean body mass index (BMI) of
≥
28
kg/m
2
involving participants aged 17 years and older. The rationale
for a lower BMI cut-off point was to allow the consideration
of co-morbidities. Moreover, at that level of body fat and age,
individuals of African descent have been shown to have a lower
BMI compared to that of Caucasians.
27
Studies also had to be
conducted in an African country to be included, and had to have
a comparative aspect (urban vs rural, African-based population
vs Westernised counterpart, gender, etc).
Titles and abstracts were screened and potentially relevant
articles were retrieved. Reference lists of review articles were
searched manually and a few eligible articles were retrieved.
Full texts of potential articles were retrieved and examined for
inclusion by two reviewers.
Data from each eligible study were extracted based on
a standard protocol format recommended by the Cochrane
collaboration.
28,29
The data-extraction form was adapted for this
review after piloting a few studies and making alterations where
necessary, to ensure standardisation prior to final data extraction.
Data were extracted by one reviewer and independently reviewed
by a second reviewer. Uncertainties and discrepancies were
rectified and resolved by discussion with the two reviewers.
Results
The initial search generated 720 titles of potentially relevant
articles. Further scanning of titles and abstracts yielded 58
potentially relevant articles for which full texts were obtained. A
total of 10 articles met the inclusion criteria and were included
in the review. Fig. 1 outlines a summary of the selection process
with reasons for the exclusions.
All the included studies had comparative subgroups. Of the
10
studies included (Table 1), two were observational,
30,31
three
cross-sectional;
32-34
three case–case controlled studies,
35-37
and two
were surveys.
2,38
The age of the participants ranged from 17 to 74 years.
Six of the included studies were conducted on both males
and females,
2,10,30,33,34,36
while four were exclusively on female
populations.
31,32,35,37
Sample sizes varied widely, ranging from 98
to 4 731 participants.
33,37
Two studies reported results from rural and urban
populations, while two others reported findings from rural
and urban populations and compared them with their Western
counterparts.
30,34
The study by Rush
et al
.
31
was the only one
conducted within an urban population, comparing BMI and
percentage body fat differences between women from five ethnic
groups.
Schutte was the first author of three of the studies included.
32,35,37
Participants from Schutte’s studies, known as POWIRS (Profiles
of Obese Women with Insulin Resistance Syndrome), were all
recruited from affluent parts of South Africa. Two of Schutte and
co-workers’ studies compared inflammatory and cardiovascular
risk markers between two ethnic groups (African women vs
Caucasians).
32,35
The third by Schutte
et al
.
considered only
African women, exploring the association of leptin and BMI
between overweight/obese hypertensive (HT) and normotensive
(
NT) participants.
37
Ibhazehiebo’s was the only study conducted within an
institution of higher learning (medical students in a private
university in Nigeria).
36
Ibhazehiebo compared blood pressure
changes between obese and non-obese (control group)
participants following graded exercise.
36
Asfaw’s on the other
hand was the only study that compared the effects of obesity on
four doctor-diagnosed chronic diseases, reporting results from
both Senegal and South Africa.
2
Comparison of demographic variables and BMI
across groups/gender (crude prevalence of obesity
by gender and locality)
Table 2 shows the four studies that reported on demographic
variables. The prevalence of obesity varied extensively between
and within studies. In all the regions studied, the difference
in the prevalence of obesity between males and females was
significant, with women as much as three times more likely to
Fig. 1. Summary of literature search.
Total number of references
retrieved from database and
manual search (
n
=
720)
Duplicates
excluded
(
n
=
23)
Titles/abstract
screening (
n
=
697)
Full text retrieved for
evaluation (
n
=
58)
Potential articles
eligible for inclusion
(
n
=
23)
Total studies included
for review (
n
=
10)
Reasons for exclusion:
•
•
Not pertaining to Africa or
obesity (
n
=
562)
•
•
Not a comparative study
(
n
=
38)
•
•
Age not appropriate
(
n
=
27)
•
•
Reviews (
n
=
7)
•
•
BMI not appropriate
(
n
=
5)
Reason for exclusion:
•
•
Not a comparative study
(
n
=
14)
•
•
BMI not reported (
n
=
5)
•
•
No obese subgroup
(
n
=
4)
•
•
Age not appropriate
(
n
=
4)
•
•
Weight not reported
(
n
=
3)
•
•
Review (
n
=
4)
•
•
Study not in Africa (
n
=
1)
Reasons for exclusion:
•
•
Age not appropriate
(
n
=
5)
•
•
No obese subgroup or no
weight reported (
n
=
7)
•
•
Study not in Africa (
n
=
1)
