Nigerian Journal of Paediatrics

Body mass index assessment using three reference standards among school adolescents in Sokoto Metropolis North Western Nigeria

Niger J Paediatr 2018; 45 (2):92 - 98

ORIGINAL

Isezuo KO

CC – BY Body mass index assessment using

Jiya NM

Audu LI

three reference standards among

Ibitoye PK

school adolescents in Sokoto

Sani UM

Ahmad MM

Metropolis, North-Western

Garba BI

Nigeria

Waziri UM

DOI:http://dx.doi.org/10.4314/njp.v45i2.4

Accepted: 30th March 2018

Abstract :

Background: Under

was analysed with SPSS version

nutrition and overweight in chil-

22. The level of agreement be-

Isezuo KO

(

)

dren co-exist in developing coun-

tween the reference standards in

Jiya NM, Ibitoye PK, Sani UM

tries and may persist into adult-

diagnosing nutritional status was

Ahmad MM, Garba BI, Waziri UM

hood. Interpretation of body mass

assessed using Kappa statistics.

Department of Paediatrics,

index (BMI), a useful measure of

Level of significance was put at p<

Usmanu Danfodiyo University

nutritional status in children re-

0.05.

Teaching Hospital, Sokoto, Nigeria

quires age related reference stan-

Results: The prevalence rate of

Email: khadisez@yahoo.com

dards, many of which were devel-

thinness was highest with the

oped from different international

IOTF at 22.6% compared to

Audu LI

sources. It is necessary to deter-

19.4%, and 19.6% by the WHO

Department of Paediatrics,

mine which of these reference

and CDC respectively. The preva-

National Hospital, Abuja, Nigeria

standards is most suitable for use

lence rates of overweight and obe-

in the assessment of BMI distribu-

sity were highest with the WHO

tion among adolescents in Nige-

reference standards at 7.0% and

ria.

1.3% followed by the IOTF (6.1%

Objective:

determine

the

and 0.6%) and lowest with the

prevalence of thinness, over-

CDC (5.8% and 0.3%). Substantial

weight and obesity using the

agreement was observed between

World

Health

Organization

WHO and IOTF (Kappa = 0.77),

(WHO), International Obesity

while the level of agreement was

Task Force (IOTF), and Centre

0.68 for IOTF/CDC and 0.64 for

for Disease Control (CDC) refer-

WHO/CDC. All agreement statis-

ence standards and assess their

tics were significant (p < 0.001).

level of agreement amongst ado-

Conclusion: The WHO and IOTF

lescents in Sokoto metropolis.

reference charts may be more suit-

Methods: 800 students were se-

able for our population.

lected through multi-stage sam-

pling technique from secondary

Key words: body mass index,

schools. BMI was classified ac-

WHO, IOTF, CDC, Sokoto, ado-

cording to three reference stan-

lescents

dards (WHO, CDC, IOTF). Data

Introduction

fordable proxy measure to assess nutritional status in

children but requires age related references often devel-

oped from different international sources. Nutritional

Under nutrition and overweight in children are both

prevalent in developing countries. This nutrition para-

disorders of childhood may persist into adulthood lead-

dox occurs because of social inequalities and rapid epi-

ing to increased morbidity, mortality and reduced life

demiological transition. Under nutrition in children es-

expectancy hence, early detection is vital to ameliorate

associated complications.

pecially if severe can lead to problems such as poor re-

sistance to infection, cognitive problems and also con-

The most widely used references for nutritional status

tributes to morbidity and mortality from other ill-

assessment in children and adolescents are the Centres

nesses.

2,3

Obesity leads to many physical and psychoso-

for Disease Control and Prevention (CDC) reference

cial complications which also persist into adulthood and

which was developed in 2000 from five previous nation-

ally representative surveys of American children, the

reduce the quality of life if not addressed.

2,3

Body mass index (BMI), is a widely accepted and af-

International Obesity Task Force (IOTF) reference de-

veloped in 2005 by experts, who extrapolated the adult

z = Standard normal deviate set at 1.96

BMI cut-off points for overweight (25 kg/m ) and obe-

p = There was no previous study in this North

sity (30 kg/m ) to data sets from six countries, which

western region of Sokoto and environs.

Therefore, a prevalence of 50% (0.5) was used.

were later extended to include cut offs for thinness in

2012, and the WHO criteria developed in 2007 from

q = 1 - p = 1 – 0.5 = 0.5

child’s growth data from the United States National

d = degree of accuracy desired = 0.05

Centre for Health Statistics and data from six countries.

A minimum sample size of 384 was arrived, however,

These reference standards were developed for different

about 800 participants were sampled to increase the

reasons from different reference populations; therefore,

chances of statistically valid results.

they give different prevalence rates of childhood BMI

status. It is necessary to compare their performance to

Sample selection

ensure proper use in populations of children likely to

exhibit anthropometric parameters different from refer-

The study participants were selected through a multi-

ence population. A previous study in Port Harcourt, in

stage sampling process from the secondary schools in

Southern Nigeria compared the three references among

the Metropolis which is made up three LGAs. Two

female adolescents only and this may not be reflective of

schools were selected from each LGA to give six

both genders. Hence, the current study was carried out to

schools. Proportionate numbers of subjects were allo-

compare the various growth reference standards/values

cated to each school based on their population. The allo-

in the assessment of BMI status among adolescents in

cated number to each school was further divided propor-

Sokoto, an urban centre in Northern Nigeria where no

tionately amongst six class levels [Junior Secondary (1-

such study has been done previously.

3) to Senior Secondary(1-3)] in each school. In each

class, the allocated number was drawn by simple ran-

dom sampling.

Objectives

Inclusion and exclusion criteria

To determine the prevalence of thinness, overweight and

obesity using the World Health Organization (WHO),

All the subjects who assented and whose parents gave

International Obesity Task Force (IOTF), Centre for

written informed consent were included. Those excluded

Disease Control (CDC) reference standards amongst

were subjects with history and clinical signs of acute

secondary school students aged 10 to 18 years in Sokoto

and chronic illnesses that could affect cause significant

metropolis and to assess the level of agreement between

weight loss, those with oedema and those on drugs that

these reference standards in determining BMI status.

could cause weight gain like steroids. Also, those with

gross limb abnormalities were excluded as these would

affect the BMI calculated.

Subjects and Methods

Data collection tools

Study location

A structured questionnaire was used to record each sub-

The study was carried out in Sokoto, the capital of

ject’s demographic data, weight and height measure-

Sokoto State of Nigeria. The city lies on latitude 13º3 ′5″

ments. A Camry Electronic Weighing Scale (Model EB

1002. ISO 9001: 2008) and Stature Meter Bioplus

N and longitude 5º15 ′53″E of the Equator. The pro-

jected population of the city in 2015(from 2006 census

(Model number 26M/ 1013522; Bharat Enterprises)

figures) is 558,130at an annual growth rate of 3%. The

were used to obtain the weights and height respectively

three Local Government Areas (LGA) which constitute

after proper standardization. The Body mass index

the metropolis, are Sokoto North, Sokoto South and

charts for CDC, IOTF, and WHO were used.

Wammako. Inhabitants are mainly ethnic Hausa and

Fulani but many other ethnic groups also reside in the

Measurements

State. All the socioeconomic classes are represented in

the population.

Weight was measured with a battery powered electronic

weighing scale. The batteries were changed daily to en-

Study Population

sure consistent results. Subjects were weighed lightly

dressed in school uniform without shoes, stockings,

The study population were students aged 10 to 18 years

caps, veils, sweaters or jackets, and all pockets were

from secondary schools in Sokoto Metropolis.

emptied. Measurements were taken to the nearest 1g.

Height was measured with a Stature Meter, which was

Study design

mounted on a flat wall surface to nearest 0.1cm. After

removal of all footwear and caps, subjects stood erect

The study was descriptive and cross-sectional in design.

with their heels, buttocks, shoulders and occiput against

Sample size determination

the wall, so that the external auditory meatus and lower

Sample size was calculated using the formula

border of the eyes were in the same horizontal plane.

n = z pq/ d

Body mass index was derived from the weight (kg) and

Where n = minimum sample size

height (m) for each subject, using the formula: BMI =

Weight (kg) / Height (m ).

Results

BMI values were classified according to three reference

Socio-demographic characteristics

standards (WHO, CDC, IOTF) as follows:

WHO: BMI for age < 5th percentile as Thinness, > 5th

Of the 800 subjects, 424 (53.0%) were males (M:F ratio-

to < 85th percentile as Normal, between 85th to 97th

1.1:1) and 565 (70.6%) were from public schools. The

percentiles as Overweight and above 97th percentile as

mean age was 14.5 ± 2.0 years (95% CI = 14.3 – 14.6

Obesity.

years). Mid adolescents within the age range of 14 to 16

CDC : BMI for age < 5th percentile as Thinness, > 5th to

years accounted for 48.1%, followed by early adoles-

< 85th percentile as Normal, between 85th to 95th per-

cents aged 10 to 13 years (34.0%), then late adolescents

centiles as Overweight and above 95th percentile as

aged between 17 to 18 years (17.9%). Those from lower

Obesity.

socio-economic class were in the majority; accounting

IOTF : Thinness is defined as BMI of ≤ 18.5Kg/m , Nor-

for 376 (47%).

mal as BMI of >18.5 to < 25kg/m and Overweight as

BMI of ≥25kg/m and Obesity as BMI of ≥ 30kg/m .

2 7

Prevalence of thinness, underweight and obesity accord-

Socio-Economic classification was based on the method

ing to the three reference systems

described by Oyedeji.

Scores were awarded to each

child based on the occupation and educational attain-

In Table 1, BMI classification according to the different

ment of the parents or their caregivers.

references is shown. Thinness was the most prevalent

abnormality in the population according to all the refer-

Ethical approval

ences. The IOTF criteria yielded the highest prevalence

of thinness. The WHO criteria yielded the highest preva-

Ethical approval for the study was obtained from the

lence of overweight (7.0%) and obesity (1.3%) followed

Ethics Committee of Usmanu Danfodiyo University

by the IOTF(6.1%, 0.6%), then the CDC (5.8%, 0.3%)

Teaching Hospital, Sokoto. Approval was obtained from

as seen in Figure 1.

the Ministry of Education of Sokoto State. Principals

and teachers of participating schools gave permission for

Table 1: BMI Classification according to WHO, IOTF

the conduct of the study in their schools. Written con-

and CDC references

sent was obtained from the parents or caregivers of the

Thin

Normal

Overweight

Obese

participating students. Assent was obtained from the

n (%)

participating students.

WHO*

Male

112 (26.4)

282 (66.5)

27 (6.4)

3 (0.7)

Statistical analysis

Female

43 (11.4)

297 (79.0)

29 (7.7)

7 (1.9)

Total

155 (19.4)

579 (72.4)

56 (7.0)

10 (1.3)

IOTF**

Data entry and analysis was done using statistical pack-

Male

125 (29.5)

278 (65.6)

21 (5.0)

0 (0.0)

age for social sciences (SPSS) version 22.0. (IBM Corp.

Female

56 (14.9)

287 (76.3)

28 (7.4)

5 (1.3)

Released 2013. IBM SPSS Statistics for Windows, NY,

Total

181 (22.6)

565 (70.6)

49 (6.1)

5 (0.6)

CDC***

USA). Continuous variables were presented as means

Male

108 (25.5)

299 (70.5)

17 (4.0)

0 (0.0)

and standard deviations. Categorical variables were pre-

Female

49 (13.0)

296 (78.7)

29 (7.7)

2 (0.5)

sented as percentages and distribution of BMI status

Total

157 (19.6)

595 (74.4)

46 (5.8)

2 (0.3)

determined by the different reference standards were

* X = 30.0, df = 3, p = 0.00, ** X = 29.7, df = 3, p = 0.00, *** X =

compared using Chi squared test.

The level of agreement between the reference standards

24.5, df = 3, p = 0.00

in diagnosing nutritional status was assessed using

Kappa statistics. According to Cohen,

kappa coeffi-

cients between 0.1 and 0.20 indicate slight agreement,

between 0.21 and 0.40 are considered as fair, between

0.41 and 0.60 as moderate, between 0.61 and 0.80 as

substantial, and between 0.81 and 1 as almost perfect.

Multivariate logistic regression analysis was carried out

to assess the variables that predict the probability of

being overweight or obese according to the three BMI

references. The dependent variable was BMI status di-

chotomously classified as not overweight (thin and nor-

mal) and overweight (overweight and obese). The inde-

Fig 1: Showing the prevalence of thinness, overweight and

pendent variables entered into the model were those that

obesity

were significantly related to BMI status on bivariate

analysis (Chi-square). For all analysis done, p value less

When analysed by gender, it was seen that thinness was

than 0.05 was considered statistically significant.

higher in males while overweight and obesity was

higher in females across the three references shown in

Figure 2 (p < 0.001). None of the males was obese

according to the IOTF and CDC references

Fig 2: Showing the prevalence of thinness, overweight and

Table 3: BMI Classification (WHO, IOTF and CDC) in rela-

obesity by gender

tion to social stratification

Social

Thin

Normal

Over-

Obese

stratification

n (%)

weight

n (%)

WHO*

Upper

23(14.4)

107 (66.9)

22 (13.0)

8 (5.0)

Middle

19 (17.2)

226 (85.6)

19 (7.2)

0 (0.0)

Lower

113 (30.1)

246 (65.4)

15 (4.0)

2 (0.5)

Total

155 (19.4)

579 (72.4)

56 (7.0)

10 (1.3)

IOTF**

Upper

26 (16.3)

107 (66.9)

24 (15.0)

3 (1.9)

Middle

28 (10.6)

220 (83.3)

16 (6.1)

0 (0.0)

Lower

127 (33.8)

238 (63.3)

9 (2.4)

2 (0.5)

Total

181 (22.6)

565 (70.6)

49 (6.1)

5 (0.6)

In Table 2, all the three references showed that preva-

CDC***

lence of overweight was highest amongst the mid-

Upper

26 (16.3)

111 (69.4)

23 (14.4)

0 (0.0)

adolescents aged 14 to 16 years, while obesity was

Middle

26 (9.8)

224 (84.8)

14 (5.3)

0 (0.0)

higher amongst the late adolescents. The prevalence of

Lower

105 (27.9)

260 (69.1)

9 (2.4)

2 (0.5)

thinness was highest using the IOTF amongst the mid-

Total

157 (19.6)

595 (74.4)

46 (5.8)

0 (0.3)

adolescents, while the WHO and CDC references

* X = 92.0, df = 6, p = 0.00, **X = 84.0, df = 6, p = 0.00,

yielded highest prevalence of thinness in the early ado-

***X = 63.2, df = 6, p = 0.00

lescents and late adolescents respectively

Agreement level between the three reference systems

Table 2: BMI Classification by age according to WHO, IOTF

using Kappa statistics

and CDC references

Thin

Normal

Overweight

Obese

Substantial agreement was observed between all the

n (%)

references, however this was highest between WHO and

WHO*

IOTF (Kappa = 0.77), while IOTF/CDC was 0.68, and

10 – 13

54 (19.9)

196 (72.1)

16 (5.9)

6 (2.2)

WHO/CDC was 0.64. All were significant (p < 0.001).

14 – 16

73 (19.0)

277 (71.9)

35 (9.1)

0 (0.0)

The details are shown in Table 4.

17 – 18

28 (19.6)

106 (74.1)

5 (3.5)

4 (2.8)

Total

155 (19.4)

579 (72.4)

56 (7.0)

10 (1.3)

Table 4: BMI Classification by age according to WHO, IOTF

IOTF**

and CDC references

10 – 13

57 (21.0)

199 (73.2)

14 (5.1)

2 (0.7)

14 – 16

93 (24.2)

263 (68.3)

29 (7.5)

0 (0.0)

Thin

Normal

Over-

Obese

17 – 18

31 (21.7)

103 (72.0)

6 (4.2)

3 (2.1)

weight

Total

181 (22.6)

565 (70.6)

49 (6.1)

5 (0.6)

WHO

CDC***

10 – 13

53 (19.5)

204 (75.0)

15 (5.5)

0 (0.0)

Thin

142

14 – 16

71 (18.4)

288 (74.8)

26 (6.8)

0 (0.0)

Normal

IOTF

533

17 – 18

33 (23.1)

103 (72.0)

5 (3.5)

2 (1.4)

Overweight

Total

157 (19.6)

595 (74.4)

46 (5.8)

0 (0.3)

Obese

* X = 15.1, df = 6, p = 0.02, **X = 11.3, df = 6, p = 0.08,

Level of agreement: kappa statistic: 0.77,p <0.001

***X = 12.4, df = 6, p = 0.05

WHO

Thin

118

In Table 3, the BMI classification according to social

Normal

CDC

531

class is shown. The lower social class accounted for a

Overweight

higher prevalence of thinness regardless of the BMI

Obese

classification used. While the upper social class ac-

Level of agreement: kappa statistic: 0.64,p <0 .001

counted for a higher proportional prevalence of obesity

IOTF

and overweight with all the classifications except for the

Thin

120

CDC classification where obesity was only seen in the

Normal

CDC

526

lower social class. However, all the findings were sig-

Overweight

nificant (p < 0.001).

Obese

Level of agreement: kappa statistic: 0.68,p <0.001

Multivariate logistic regression analysis of factors

associated with overweight and obesity

On multivariate logistic regression analysis, a model

was built with BMI status as the dependent variable and

age, gender and social class (which were significant on

bivariate analysis) as the independent categorical vari-

ables. It is seen that those of upper social class were 5 to

7 times more likely to be overweight or obese compared

to the lower social class after controlling for other fac-

cents in a study by Minghelli, the WHO reference gave

tors (age, gender). This was statistically significant

a higher prevalence of thinness followed by the IOTF

reference. Lopes and de Moraes also found thinness

(<0.001). The result was similar for all the BMI refer-

ence standards but the odds of upper social class being

was higher using the IOTF criteria in Portugal and Bra-

overweight or obese was highest with the IOTF refer-

zil. One reason that has been adduced to the higher yield

ence standard. The result is shown in Table 5.

of thinness by the IOTF reference is that the countries

from which the reference population used was drawn,

Table 5: Multivariate logistic Regression Analysis

have a Gross Domestic Product (GDP) which is higher

than the world average. Therefore, when used to assess

Regression

p-value

Odds

95% confi-

coefficient

ratio

dence interval

subjects from less endowed countries, it may overesti-

( β)

for odds ratio

mate the prevalence of underweight in these popula-

WHO

tions. The CDC reference gave the lowest prevalence of

Age(14-16yr)

0.31

0.43

1.364

0.628 to 2.963

overweight and obesity in this study. This may not be

Gender (male)

-0.13

0.63

0.877

0.518 to 1.484

surprising considering that the data used to derive the

Social class I

1.648

<0.001

5.197

2.704 to 9.991

Social class II

0.479

0.165

1.615

0.821 to 3.175

reference was drawn from only the United States as

IOTF

compared to the more inclusive nature of the other refer-

Age (14-16 yr)

0.04

0.92

1.042

0.468 to 2.319

ences.

Gender (male)

-0.43

0.16

0.653

0.362 to 1.177

Social class I

2.003

<0.001

7.409

3.476 to 15.79

Prevalence of overweight and obesity was higher in fe-

Social class II

0.757

0.06

2.13

0.97 to 54.682

males than males using all the criteria. In the studies by

CDC

Gonzalez et al and Banjade et al females had higher

Age (14-16 yr)

0.19

0.67

0.666

0.255 to 1.743

prevalence of overweight, however the males were more

Gender (male)

-0.6

0.07

1.210

0.503 to 2.913

obese only with the use of CDC criteria. It has been ob-

Social class I

1.745

<0.001

5.724

2.643 to 12.396

Social class II

0.601

0.15

1.823

0.812 to 4.093

served that in developing countries, female adolescents

are usually more overweight or obese than their male

counterparts while in developed countries, males are

equally or may be more affected. The risk factors for

obesity including ingestion of high caloric diets, seden-

Discussion

tary lifestyle in addition to increase screen time either

with television and other electronic devices are all

Thinness was the most prevalent abnormality probably

higher in developed countries and may account for

reflective of the fact that a higher proportion of the

this. Also, female adolescents participate less in out-

population were of the lower social class and subject to

door activities and exercise than their male counterparts

chronic undernourishment. Thinness was also the most

in developing countries and this may also be contribu-

prevalent in the study by Ejike et al in Umuahia south-

tory. Additionally, female adolescents also tend to be

east Nigeria but overweight was the prominent abnor-

more conscious of their weight with higher prevalence

mality in the study by Jaja and Alex-Hart from Port

of eating disorders aimed at losing weight amongst them

Harcourt also in southern Nigeria. This may be ex-

in developed countries.

23,24

For the age distribution, it

plained by the fact that the Port Harcourt study was only

was seen that overweight was more prevalent among the

conducted amongst females who were from a cosmo-

mid-adolescents, while obesity was more in the older

politan area and were more likely to be overweight.

adolescents. This finding may be related to pubertal

Other studies from Canada and Brazil reported over-

growth and development.

weight as the more prevalent abnormality in adoles-

cents.

16,17

WHO estimates that three-quarters of all deaths in the

developing world by the year 2020 will be due to non-

The WHO reference standard yielded a higher preva-

communicable diseases (NCDs).

Underweight and

lence of overweight and obesity in this study compared

overweight are NCDs in childhood that are associated

to the IOTF and CDC references. The WHO reference

with inherent complications and may adversely influ-

also yielded a higher prevalence of overweight and obe-

ence the manifestation of other diseases. The magnitude

sity compared to the other two references in studies from

of these problems can be determined by the BMI status

Southern Nigeria by Ejike et al

and Jaja and Alex-

of the population to ascertain the prevalence, trends and

Hart. Similar results were also reported by Gonzalez

determinants of any abnormality detected. This is neces-

and Banjade

from Colombo and India respectively.

sary so as to effectively design public health interven-

The reason for this could be that the WHO reference

tions.

includes more data from developing countries. Reh-

man reported that the WHO reference affords earlier

When assessing BMI-for-age, it is important that the

diagnosis of obesity because the data was derived from a

available reference systems are compared to ascertain

non-obese population before the onset of the obesity

which may be more suitable for a particular population.

epidemic.

The IOTF cut-offs are recommended for researchers and

The IOTF criteria yielded a higher prevalence of thin-

policy makers in different countries for descriptive and

ness in this study which was also similar to results from

comparative purposes while the CDC Growth Charts

Port Harcourt and Umuahia.

9,15

Among Brazilian adoles-

and WHO charts are intended for clinical use in moni-

toring children’s growth.

5,6,8

However, all three systems

However, in some populations, especially in developed

are being used in practice for research purposes and

countries with large slums, those of lower social class

there is no perfect reference system which may be suit-

have been found to be more overweight, due to poor

feeding practices and lack of areas for exercise.

28,29

able globally.

Gen-

erally, indulgence in junk food which have high caloric

In this study, there was more agreement between the

content and insufficient physical activity, behaviours

WHO and IOTF, while there was more agreement be-

that are learned in childhood, are major contributors to

obesity and its resulting health problems.

tween the WHO and CDC in the study by Jaja and Alex-

Hart and in the study from India by Banjade et al.

Other studies from Canada and Portugal showed there

A limitation of the study was that there was no compari-

was more agreement between the CDC and IOTF refer-

son to any standard measurement of body fat. Also, we

ences compared to the WHO in studies.

16,17

There was

did not explore the relationship of BMI distribution with

more agreement between the WHO and IOTF references

ethnicity because a significant proportion of the study

in our study probably because these two references cut

population were of the Hausa-Fulani extraction. With

across different populations and would likely give a

respect to BMI distribution with ethnicity, further stud-

truer prevalence of BMI status than the CDC reference

ies are needed in other parts of Nigeria for comparison.

which was only drawn from American children. Despite

the variations in prevalence of thinness, overweight and

obesity, there was a good agreement among the three

references in this study.

Conclusion

Results from our study also indicated that those of

It is concluded that thinness was more prevalent in the

higher social economic class were more likely to have

study population, and the WHO and IOTF reference

combined overweight and obesity compared to the lower

charts may be more suitably applied to adolescents from

social class, similar to results from the studies by Gon-

Nigeria since they may permit earlier diagnosis of under

zalez and van Vliet. . Though nutrient intake and level

and over-nutrition.

of physical activity were not assessed, those of higher

social class are more likely to indulge in fast foods and

Conflict of interest : None

less likely to participate in daily exercises like trekking

Funding: None

to school, therefore may be more prone to overweight.

References

Caballero B. A nutrition para-

Kuczmarski RJ, Ogden CL,

Jaja T, Alex Hart B. Preva-

dox--underweight and obesity

Guo SS, Grummer-Strawn LM,

lence of overall and central

in developing countries. N Engl

Flegal KM, Mei Z. 2000 CDC

obesity among adolescent

J Med 2005;352:1514-1516.

Growth Charts for the United

girls in Port Harcourt: a com-

Ringback Weitoft G, Eliasson

States: methods and develop-

parison of different methods.

M, Rosen M. Underweight,

ment. Vital and health statistics

Nig J Paed 2016;43:209.

overweight and obesity as risk

Series 11, Data from the na-

10. http://www.latlong.net/place/

factors for mortality and hospi-

tional health survey 2002:1-

sokoto-nigeria-21027.html

talization. Scand J Public

190.

(Accessed 17/05/2017).

Health 2008;36:169-176.

Cole TB, Bellizzi M, Flegal

11. National Population Commis-

Roh L, Braun J, Chiolero A,

MK, Dietz WH. Establishing a

sion. 2006 National Census.

Bopp M, Rohrmann S, Faeh D.

standard definition for child

Federal Republic of Nigeria

Mortality risk associated with

overweight and obesity world-

Official Gazette 2007;94:196.

underweight: a census-linked

wide: International survey.

12. Araoye MO. Sample size cal-

cohort of 31,578 individuals

BMJ 2000;320:1240-1243.

culation. In: Araoye M (Ed):

with up to 32 years of follow-

Cole TJ, Lobstein T. Extended

Research Methodology with

up. BMC Public Health

international (IOTF) body mass

Statistics for Health and So-

2014;14:371-371.

index cut-offs for thinness,

cial Sciences. Nathdex (Publ)

Sweeting HN. Measurement

overweight and obesity. Pedi-

Ilorin 2004; 115-121.

and Definitions of Obesity In

atr Obes 2012;7:284-294.

13. Oyedeji GA. Socioeconomic

Childhood and Adolescence: A

de Onis M, Onyango AW. De-

and Cultural Background of

field guide for the uninitiated.

velopment of a WHO growth

Hospitalized Children in

Nutr J 2007;6:32. Accessed

reference for school-aged chil-

Ilesha. Nig J Paed

12/11/13. DOI: 10.1186/1475-

dren and adolescents: World

1985;12:111-117.

2891-1186-1132.

Health Organization; 2007

14. Cohen JA. A coefficient of

[Accessed 13/11/2013]. Avail-

agreement for nominal scales.

able from: http://

Educ Psychol Meas

www.ncbi.nlm.nih.gov/pmc/

1960;20:37 - 46.

articles/PMC2636412.

15. Ejike CCC, Onukogu IA, On-

20. Rehman A. Growth standards

26. van Vliet JS, Gustafsson PA,

yemairo JN. Co-existence of

and charts; World Health Or-

Duchen K, Nelson N. Social

child and adolescent obesity

ganization. Professional Med J

inequality and age-specific

and thinness in a city in Nige-

2014;21:232-239.

gender differences in over-

ria: Comparison of results de-

21. Sousa Lopes H. Diagnostic

weight and perception of over-

rived from different reference

accuracy of CDC, IOTF and

weight among Swedish chil-

standards. Int J Nutr Pharma-

WHO criteria for obesity clas-

dren and adolescents: a cross-

col Neurol Dis 2013;3:276.

sification, in a Portuguese

sectional study. BMC Public

16. Medehouenou TC, Ayotte P, St

school-aged children popula-

Health 2015;15:628.

-Jean A, Meziou S, Roy C,

tion. Porto: Institute of Public

27. Dinsa GD, Goryakin Y, Fu-

Muckle G, et al. Overweight

Health, University of Porto;

magalli E, Suhrcke M. Obe-

and Obesity Prevalence Among

2012. https://repositorio-

sity and socioeconomic status

School-Aged Nunavik

aberto.up.pt/. (Accessed

in developing countries: a

Inuit Children According to

9/12/2013).

systematic review. Obesity

Three Body Mass Index Classi-

22. de Moraes LI, Nicola TC, de

Reviews 2012;13:1067-1079.

fication Systems. J Adolesc

Jesus JSA, Alves ERB, Gio-

28. Knai C, Lobstein T, Darmon

Health 2015;57:31-36.

vaninni NPB, Marcato DG, et

N, Rutter H, McKee M. So-

17. Minghelli B, Nunes C, Oliveira

al. High blood Pressure in chil-

cioeconomic Patterning of

R. Body mass index and waist

dren and its correlation with

Childhood Overweight Status

circumference to define thin-

three definitions of obesity in

in Europe. Int J Environ Res

ness, overweight and obesity in

childhood. Arquivos Bra-

Public Health 2012;9:1472-

Portuguese adolescents: com-

sileiros de Cardiologia

1489.

parison between CDC, IOTF,

2014;102:175-180.

29. Bhurosy T, Jeewon R. Over-

WHO references. Pediatr En-

23. Pelegrini A, Silva DA, Gaya

weight and Obesity Epidemic

docrinol Rev 2014;12:35-41.

AC, Petroski EL. Comparison

in Developing Countries: A

18. Gonzalez-Casanova I, Sar-

of three criteria for overweight

Problem with Diet, Physical

miento OL, Gazmararian JA,

and obesity classification in

Activity, or Socioeconomic

Cunningham SA, Martorell R,

Brazilian adolescents. Nutr J

Status? Sci World J

Pratt M, et al. Comparing three

2013;12:5.

2014;2014:964236.

body mass index classification

24. Striegel-Moore RH, Rosselli F,

30. Must A, Strauss RS. Risks and

systems to assess overweight

Perrin N, DeBar L, Wilson GT,

consequences of childhood

and obesity in children and

May A, et al. Gender Differ-

and adolescent obesity. Int J

adolescents. Rev Panam Salud

ence in the Prevalence of Eat-

Obesity 1999;23:S2-S11.

Publica 2013;33:349-355.

ing Disorder Symptoms. Int J

19. Banjade B, Naik V, Narasan-

Eat Disord 2009;42:471-474.

navar A. Comparison of CDC,

25. WHO. Non- communicable

WHO and IOTF growth refer-

diseases. Accessed 22/8/2017.

ences in relation to overweight

http://www.who.int/

and obesity in college adoles-

mediacentre/factsheets/fs355/

cents of North Karnataka, In-

en/.

dia. Al Ameen Journal of Medi-

cal Sciences 2015;08:72-76.