Cardiovascular Journal of Africa: Vol 28 No 2 (March/April 2017)

CARDIOVASCULAR JOURNAL OF AFRICA • Volume 28, No 2, March/April 2017

AFRICA

Discussion

To our knowledge, this study is the first to comprehensively

compare the performance of a large set of anthropometric

indices as correlates and potential predictors of risk for

hypertension and prehypertension in a typical Nigerian (West

African) population. We analysed the performance of some

anthropometric indices of obesity as potential predictors of

hypertension and prehypertension.

The mean values of the following anthropometric measures,

BMI, WC, HC, CI and BAI were significantly higher in women.

This could have been attributed to the general inactivity of women

in this population. The mean values of all the anthropometric

indices studied were higher in the prehypertensive and highest

in the hypertensive participants relative to the normotensive

participants. This is an indication that participants with a higher

obesity index tend to have high blood pressure values. This

finding is consistent with reports from previous studies.

5,8,26

The correlation analysis showed that all the studied

anthroprometric indices were correlatedwith SBP andDBP. BMI,

WC, PI and WHtR had correlation coefficients greater than 0.25,

while BAI and CI correlated poorly with blood pressure. Our

results also showed BMI, WC, WHtR and PI performed best as

potential predictors of the risk for hypertension on comparing

respective AUCs from ROC curve analysis. The prevalence

ratios for general obesity index were lower than that of central

obesity in both the crude and adjusted models, however these

differences were not large enough to suggest that central obesity

index (WC or WhtR) outperformed general obesity index (BMI)

in this study. There was no significant difference between the

performances of BMI, WC and WHtR in predicting risk for

hypertension. A similar finding was reported previously by Lee

and co-workers.

BAI and HC showed a fair performance in predicting

hypertension and prehypertension risk. CI had a poor predictive

power for hypertension and totally lacked the capacity to

distinguish prehypertensive cases from normotensive cases. The

results of the ROC and correlation analyses were consistent and

showed similar trends.

Anthropometric indices (BMI, WC, WHtR and PI), which

had higher correlation coefficients with blood pressure (SBP and

DBP), had very high AUCs that were statistically significant (

0.05). The reverse was true for poorly correlated anthropometric

indices such as BAI, WHR and CI. CI was the poorest correlate

of hypertension and prehypertension (AUC

0.5,

0.05). BMI,

WHtR and WC emerged the best predictors of hypertension and

prehypertension in this study.

These findings conform with and confirm the findings of

Silva

et al

in Brazillian women and men, Sanchez-Viveros

in Mexican women and men, and Uhernik

et al

in Croatian

men and women. They differ from those of Feldstein

et al

Argentina and Li

et al

in Australia where none of BMI, WC

or WHtR emerged as the best predictors of hypertension or

prehypertension. These results also provide evidence to support

the findings that suggested the superiority of WC and BMI over

BAI.

As mentioned above, epidemiological studies on the predictive

potentials of anthropometric indices for hypertension and

cardiovascular-related diseases are limited in Nigeria. Okafor

al.

reportedWCwasabetterpredictorof obesityandhypertension

than WHR in a population with similar characteristics to

our study population, while Sonuyi and co-workers

reported

normative values of selected anthropometric variables in Lagos,

Nigeria. Both findings were consistent with our results.

The differences in the results of some of the previous studies

mentioned could have been attributed to differences in the

characteristics of the populations. Evidence of racial/ethnic, gender

and age variations in anthropometry is well established.

Sakurai

et al

reported that the percentage body fat in Asians, as measured

by dual-energy X-ray absorptiometry is greater than in African

Americans and whites with a similar BMI. Variations in the level of

leptin (the product of the gene largely responsible for obesity) across

different ethnic groups and races is also well established.

Human

body composition is evidently a result of complex multifactorial

interactions between lifestyle, culture, environmental and genetic

differences,

which vary from place to place and impact differently

on the results of studies in different populations.

Secondly, rigours, technicalities and lack of universally

accepted standards in measuring some anthropometric measures

could account for some of the reported differences in different

studies.

Our study also provided evidence to suggest that the

predictive potential of anthropometric indices may vary with

age. BMI, PI and WHtR performed well in predicting risk for

hypertension and prehypertension in three age categories (≤ 20,

26–40 and

≥

40 years), while BAI was better in one age category

(

≥

40 years). HC and CI were not particularly outstanding in any

of the age categories. This differential performance in different

age categories could also account for the variations in the results

from different studies.

Our predicted cut-off points for some of the anthropometric

predictors of hypertension were somewhat similar to that

proposed by the WHO and other studies

8,10,11,35

in Korean,

Brazilian and Pakistani populations, respectively. However, the

cut-off points for WC and WHtR were higher in our study when

compared to the WHO cut-off value. This could be attributed

to the higher WC and lower height of females in the population.

Africans and Westerners have quite distinct anthropometry

occasioned by differences in culture, environment, genetics,

nutrition as well as economy. Most of the recommended

cut-off points are more representative of Western populations.

The cut-off points for the anthropometric indices in our study

differed markedly in women for prehypertension; the predicted

cut-off points were higher in women and lower in men.

The performance of the anthropometric indices in predicting

both conditions differed by gender in this study. All the

indices studied tended to predict risk for hypertension and

prehypertension better in males than in females. These differences

have also been corroborated by previous independent studies.

3,7,36

There is evidence that fat distribution in men and women differs.

Visceral fat is more dominant in men and subcutaneous fat

in women. This may provide an explanation for the existence

of gender differences in the performance of anthropometric

indices. Visceral fat has a stronger association with metabolic

abnormalities than subcutaneous fat,

and this could also explain

why we found a higher risk for hypertension with regard to

obesity in males than females in this study.

Our study presents evidence that the relationship between

obesityand the twoconditions, hypertensionandprehypertension,

differed in terms of the performance of anthropometric indices.

This is to be expected as prehypertension has been described by

JNC-7 as a new category of hypertension with high risk for the