27 / 76
CARDIOVASCULAR JOURNAL OF AFRICA • Volume 28, No 2, March/April 2017
AFRICA
93
factor for developing clinical hypertension relative to those
with normotension. It is characterised by blood pressure levels
slightly higher than normal. Research in this area has suggested
that exploration and modification of risk factors could help
ameliorate this trend.
1
Obesity is a disorder characterised by extensive fat
accumulation, and the body fat is distributed in such a way
that health and wellbeing are affected.
2
The condition is an
established risk factor for hypertension, hypercholesterolaemia,
insulin resistance and diabetes.
3,4
The close association of obesity
with blood pressure has long been recognised in both genders
and even in diverse racial/ethnic groups.
5
Anthropometry is the most basic method for assessing body
composition. It describes body mass, size, shape and level of
fatness.
6
It is an easy, economical and effective method that is
used in the initial screening of obesity, hypertension and other
metabolic disorders.
7
Research efforts have developed many anthropometric indices
to specifically describe obesity and fat distribution in humans.
These include body mass index (BMI), ponderal index (PI), waist
circumference (WC), hip circumference (HC), waist–hip ratio
(WHR), waist–height ratio (WHtR), body adiposity index (BAI)
and conicity index (CI).
BMI is promulgated by the WHO as the most useful
epidemiological measure of obesity, but its usefulness suffers
from its inability to account for body fat distribution.
8
BMI and
PI are widely used to describe total or general obesity, while WC,
WHR, WHtR and CI describe more visceral fat; abdominal or
centralised obesity.
9,10
Anthropometric indicators of abdominal
obesity can provide estimates of the visceral adipose tissue, which
in turn is associated with metabolic changes, hyperinsulinaemia,
glucose intolerance, hypertriglyceridaemia and hypertension.
11
The direct association between hypertension and
anthropometric indices of obesity have been studied in
many countries and ethnic groups,
7
but results from different
studies show that the best anthropometric index in predicting
hypertension and other components of the metabolic syndrome
remains inconclusive and controversial.
7,12,13
Some studies
reported that the best single indicator of the risk of hypertension
in Japanese and Cuban populations was BMI.
3,14
Other studies
suggested WC was a better predictor in Greek,
15
Taiwanese
(women) and some Japanese men.
7
WHtR has also been suggested as the best predictor of
hypertension for elderly men in Barbados,
14
Taiwanese men,
16
and Korean men,
17
whereas other studies demonstrated that
WHR was the best predictor for Argentinian men and women,
18
and indigenous Australian men and women.
19
Lee and Kim,
7
and Fuchs
et al
.
13
suggested that combination of two or more
indices could improve the predictive power of an index. Different
studies have therefore posited that the predictive power of an
anthropometric index may be population-dependent and vary
across ethnic groups, age and gender.
20,21
In Nigeria, studies that assessed the performance of
anthropometric indices in predicting risk of some metabolic
disorders such as hypertension are lacking. Okereke
et al
.
22
evaluated the anthropometric indices for the diagnosis of obesity
in pregnant women in Nigeria, and Okafor
et al
.
23
compared
the performance of WC and WHR. However, no study has
comprehensively assessed the performance of anthropometric
indices of obesity in predicting hypertension and prehypertension.
Based on this premise, our study intended to compare the
performance of eight anthropometric measures of obesity: BMI,
PI, WC, HC, WHR, WHtR, BAI and CI as indicators of risk
for hypertension and prehypertension, and the effect (on the
performance) of combining two or more of the best-performing
indices.
Methods
The study was a cross-sectional, church-based survey carried
out in three major cities in Anambra state. The inclusion criteria
were: age range 17–79 years, being a resident in the study areas
Awka, Nnewi and Onitsha cities of Anambra state, being selected
by the random sampling procedure explained below, providing
informed consent/willingness to participate, and complying with
the instructions of the study, for example, avoidance of alcohol,
coffee, drugs and exercise at least 30 minutes before examination.
Information on the prevalence of hypertension in the adult
and paediatric population of Anambra state was lacking, but
based on data in the literature, the prevalence in Nigeria ranges
from 8–30%.
24
Using the ‘stat calc’ function of Epi INFO
(version 7) software, it was determined that a sample size of
900 was adequate to detect the prevalence of hypertension of
10–30% with 3% precision and 95% confidence.
A total of 1 000 participants (we lost 88 to follow up on the
day of testing and administration of the questionnaire) were
randomly selected from 30 primary sampling units. A stratified
random sampling technique was employed. In brief, the three
major cities, Awka, Nnewi and Onitsha were selected for this
study. The cities were stratified by location (rural versus urban
areas) to ensure good representation. Since these populations are
predominantly Christian, the survey was made church based. The
churches constituted the primary units from which individuals or
participants were randomly sampled; 10 churches from each city.
Firstly, pre-visits to the three cities provided us with a list
(dataset) of known churches in the communities within the cities.
For each city, six and four churches were randomly selected from
the urban and rural areas, respectively, using the ‘sample, count’
command of Stata statistical package. A total of 30 churches
were selected from a total of 224 churches (Awka 71, Onitsha 90,
Nnewi 63). The urban areas were more populated and therefore
were sampled more. The selected churches were visited. A list
of members who showed willingness to participate was made
after explaining the objectives and nature of the study to the
congregation, and 11 participants were randomly selected using
the ‘sample, count’ command of Stata.
Information on demography and lifestyle was obtained using
a well-structured and validated questionnaire. Anthropometric
data, which included weight, height, and waist and hip
circumferences were obtained by well-trained personnel. Weight
was measured to the nearest 0.5 kg using a weighing scale with
the participant removing his/her footwear. Height was measured
to the nearest 0.5 cm using a local stadiometer fixed to a wall.
The waist circumference was measured at the level of the iliac
crests,
25
using a flexible tape and passing it along the umbilical
level of the unclothed abdomen. The hip circumference was
measured around the widest portion of the buttocks, with the
tape parallel to the floor.
Blood pressure was taken from the non-dominant arm after
15 minutes of rest, using appropriate cuff size and Accoson