Thinness as Major Underlying Problem Among Adolescents of Northeast India

Background: Undernutrition is a major public health concern in many of the developing countries of Asia. Due to immense population size, socioeconomic disparities, illiteracy and inadequate access to health facilities prevalence is very high in India. The objective of this study was to determine the prevalence of age-sex specific thinness (low BMI-for-age) among adolescents residing in rural regions in India. Materials and Methods: This cross-sectional study has conducted among 1165 adolescent (602 boys; 563 girls) aged 10-18 years of Darjeeling district, West Bengal, India. Anthropometric measurements were height and weight obtained and BMI (weight/height2, kg/ m2) was calculated. The prevalence of thinness was assessed using newly proposed age-sex specific cut-offs of Cole et al. The data were analyzed using chi-square, ANOVA and Least Mean and Square (L,M and S) model approach. Results: Prevalence of overall thinness is 49.10% (51.16% boys, 46.89% girls) among rural adolescents. The boys are found to be more sufferer than girls in the different thinness grades include mild (grade I; 27.41% vs. 27.11%), moderate (grade II; 14.62% vs. 12.08%) and severe (grade III; 9.14% vs. 8.70%) (p>0.05). The age and sex specific smooth percentile curves of BMI were derived using L,M and S model approach for further evaluation of nutritional status. Conclusion: The existence of high prevalence of thinness among adolescents indicates nutritional deprivation among rural Indian adolescents. There is an urgent need of appropriate nutritional intervention program to address the public health problem related to undernutrition especially among nutritionally vulnerable segments of adolescents residing in rural regions to ameliorate the nutritional status


Introduction
T he prevalence of undernutri on is a major public health concern in many of the developing countries.The prevalence is very high due to immense popula on size, socioeconomic dispari es, illiteracy and inadequate access to healthcare facili es in India 1,2,3,4,5 .Undernutri on con nues to be the principal cause of ill-health condi on and premature mortality and morbidity in developing countries 3 .In India, one of the important aims of nutri onal research is to improve the undernutri on status and its related consequences among most nutri onally vulnerable segments especially children, adolescents and women.Furthermore, the foremost objec ve of nutri onal assessment study is to improvement of human health condi on 6 .Therefore, the assessments became a priority and bears a poten al role in formula ng developmental strategies and interven on programme to the nutri onally vulnerable popula ons in India.The adolescent period is a very important phase in the lifespan of an individual.The WHO has been defi ned by the transi on period between childhood to adulthood is referred to as 'adolescence' of 10-19 years aged and characterized by an excep onally rapid physical growth 6 .Nutri onal status of adolescents needs to be monitored closely as they cons tutes the next genera on of the parenthoods.The nutri onal status of the adolescent girls, the future mothers, bear special signifi cance as they contribute signifi cantly to nutri onal and health status of the popula on 2,7,8,9 .Currently, it is es mated that adolescents (aged 10-19 years) contributes to 1.20 billion of the total world popula on and belonging to the largest popula on of adolescent cohort of 243 million (20.00%) in India 2,10 .Most importantly, a very high propor on of adolescent girls get married at an early age and even before comple on of their physical growth 7,11 .Recent na onal survey report of Na onal Family Health Survey (NFHS-3) has shown that 50.4% adolescents (aged 15-24 years) are married and among them 39.0% female either have a child or are currently pregnant 2 .
Anthropometry has become a widely used non-invasive and inexpensive prac cal technique of assessing nutri onal status of individual and/or popula ons especially in clinical and epidemiological studies 3,4,6,12,13,14,15 .The prevalence of undernutri on is generally assessed by using conven onal anthropometric measures of stun ng (height-forage), underweight (weight-for-age), was ng (weightfor-height), thinness {Body Mass Index (BMI)-forage and Triceps skinfold-for-age comparing with interna onal and interna onal standards 3,6,16,17 .Moreover, these conven onal anthropometric indices (e.g., stun ng, underweight and was ng) are generally overlapping in nature and markedly underes mate the actual magnitudes of undernutri on prevalence in popula on 3,5,15 .Furthermore, there are diffi cul es in the use of diff erent anthropometric cut-off s points to determine the prevalence of underweight/thinness due to between-country comparisons 18,19 .The BMI as measured by weight in kilogram (kg) divided by height in meter square (m 2 ) is an anthropometric measure that has been extensively u lized to assess nutri onal status {e.g., chronic energy defi ciency (CED) and/or overweight-obesity} 6,16,17 .The use of BMI in rela on to age (e.g., thinness) was recommended as the best surrogate anthropometric measure of thinness and overweight during adolescents 6,16 .The WHO has already recommended cut-off s are the <5 th and the >85 th percen le of the Nutri onal Health and Nutri on Examina on Survey (NHANES) references to assess the thinness and overweight respec vely 6 .Several studies have already reported the undernutri on among Indian adolescents residing in both rural and urban regions using the WHO recommended references 7,8,9,12,14,20,21 .Very recently, interna onal age-sex specifi c cutoff s have been proposed to assess thinness 16 and overweight and obesity 17 among children and adolescents (aged 2-18 years).The informa on on the thinness prevalence among adolescents of developing countries like India is needed to be generated due to its immense popula on size and socio-economic and cultural dispari es for the na onal and interna onal comparisons.The present study has been conducted to assess the prevalence of thinness among adolescents aged 10 to 18 years residing an economically and socially backward rural areas of Darjeeling district, West Bengal, India.Further, to compare the results with available popula on specifi c thinness prevalence records to provide more insight into the nutri onally vulnerable popula ons residing in rural and sub-urban regions of India.However, the recent na onal study has already advocated that the prevalence of thinness is serious concern among Indian adolescents 2 .Therefore, such assessment of the prevalence will enable the government and non-governmental agencies to formulate policies and ini ate developmental strategies for well-being of the nutri onally vulnerable segments of adolescent popula ons in India.

Materials and Methods
The present community based cross-sec onal study was conducted among 1165 adolescents (602 boy and 563 girls) of aged 10-18 years residing in rural areas of Phansidewa Block (La tude 26° 34' 59'' N, Longitude 88° 22' 0'' E) under Siliguri subdivision of Darjeeling district, West Bengal, India.This community block covers an area of 308.65 km 2 and a total popula on of 1,71,508 individuals (males: 87,945; females: 83,563) with a literacy rate of 41.59% (males: 51.85%; females: 30.80%).The region is situated near the Indo-Bangladesh border region, situated approximately 35-40 km distant from the sub-divisional town of Siliguri.The community block consists of predominant ethnic communi es viz., Rajbanshi, Bengalee Muslim popula on (BMP) and Proto-australoid (e.g., Santal, Munda and Oraon) ethnic popula ons 4 .There is some government healthcare and supplementa on programs in place in the area under study includes Integrated Child Development Service and midday meal facility with primary and junior high schools by Government of India 14 .Ini ally subjects are iden fi ed in the present study are belonging to the major dominant ethnic groups of this region includes Rajbanshi and BMP.The subjects belonging to the adolescent (aged 10-18 years) were iden fi ed by u lizing stra fi ed random sampling.Age of the subjects was collected using their birth cer fi cates and relevant offi cial records issued by the local government offi cials.The data were collected from February 2009 to March 2011.The minimum number of subjects required to reliably es mate the prevalence of thinness was calculated u lizing the standard sample size es ma on procedures 22 .In this method, the an cipated popula on propor on of 50%, absolute precision of 5% and confi dence interval of 99% were taken into considera on.Generally, the prevalence of thinness has been observed >50% among children and adolescents in India 8,9,14,23,24 .Hence, the minimum sample size in the present study was es mated to be 661 subjects.A structured schedule was completed by house visits and interviewing the subjects and their parents.The villages of three Gram Panchayets (GP) (the local level government authority) viz., Chathat GP, Phansidewa GP and Jalashnijamtara GP were covered.The socioeconomic status (SES) was evaluated using a modifi ed version of the scale of Kuppuswamy's was proposed by Mishra and Singh 25 .The scale determines the SES based on a score calculated from educa on, occupa on and monthly income.It was subsequently observed that all selected children were belonged to a lower SES group.All the subjects were free from any physical deformi es, systema c diseases, nutri onal defi ciencies and previous histories related to medical and surgical episodes.An informed consent was obtained local village level authori es, subjects and their parents prior to the data collec on.The par cipa on of the subject in this study is completely voluntary in nature, and the study was conducted in accordance with the ethical guidelines for human experiments as laid down in the Helsinki Declara on 26 .

Anthropometric measurements obtained
The anthropometric measurement of height and weight of the subjects was recorded using the standard procedure 27 .Height of the subjects was recorded to the nearest 0.10 cm with the help of an anthropometer (GPM type, Galaxy Informa cs, New Delhi) with the head held in the Frankfort horizontal plane.The weight of the subjects, wearing minimum clothing and with bare feet was taken using a portable weighing scale (Libra®, Edryl-India, Tiswadi, Goa) to the nearest 100 gm.The BMI was calculated by using the standard equa on: BMI= weight (kg) /height (m 2 ) 6 .The diff erences in anthropometric measurements were calculated for tes ng the co-effi cient of reliability [R= {1− (TEM) 2 /SD 2 }, the SD= standard devia on of all measurements] using the technical error measurement {TEM=√ (ΣD 2 /2N), D=diff erence between the measurements, N= number of individuals measured} following the method of Ulijaszek and Kerr 28 .Very high values of R (>0.98) were obtained for height and weight using TEM analysis and values were found within the cut-off values 28 .

Assessment of nutritional status (Thinness)
The prevalence thinness was determined following the newly proposed interna onal age-sex specifi c BMI reference and cut-off points of Cole et al 16 .The BMI values were used to determine the defi nite grades of thinness includes Grade-III (severe), Grade-II (moderate) and Grade-I (mild) 16 .The above classifi ca ons are similar to the adult CED classifi ca on as proposed by WHO 6 .This reference popula on and cut-off points have already been u lized to assess such prevalence among Indian children and adolescents 13,29,30,31 .

Statistical analysis
The data were sta s cally analyzed using the Sta s cal Package for Social Sciences (SPSS, Inc., Chicago, IL; version 15.0).The descrip ve sta s cal analysis of the data was depicted in mean and standard devia on.One way analysis of variance (ANOVA) using the Scheff e procedure was done to assess age and sex diff erences.Chi-square analysis was u lized to assess the diff erences in prevalence of thinness.A p-value of less than 0.05 was considered to be sta s cally signifi cant.The Least Mean Square (L, M and S) model approach was u lized to convert the measurements for a child of known age-and sex to evaluate the cen le and standard devia on score or z-scores 32,33 .The L, M and S model approach take into the account the degree of skewness (L), central tendency (M; Median) and dispersion or the generalized coeffi cient of varia on (S) for the conversion.This method was used to derive the age-sex specifi c percen le reference curves of BMI.The cen le curves of 3 rd , 10 th , 25 th , 50 th , 75 th , 90 th and 97 th were derived as reference data.The method summarizes percen les at each age based on the power of age-specifi c Box-Cox power transforma ons used to normalize data.The LMS Chart Maker so ware program (The Ins tute of Child Health, London) was used to obtain the smooth cen le curves that fi ed smooth cen le curves to the reference data.

Results
The age-sex specifi c subject distribu on, descrip ve sta s cs of anthropometric variables and prevalence of diff erent grades of thinness among adolescents are depicted in Table 1.The age specifi c mean values of weight, height and BMI values were observed to be progressively increased with age among both sexes but an excep on was observed among girls in 18 years (in height) and 17 years (in BMI).The overall mean BMI was found to be signifi cantly greater among girls (17.08±2.57kg/m 2 ) than boys (16.p<0.05).The age-specifi c diff erences were also found sta s cally signifi cant (p<0.01) with height (F-value =117.20,d.f. 8, 601), weight (F-value =144.83,d.f. 8, 601) and BMI (F-value =60.95, d.f. 8, 601) among boys, and age and height (F-value =75.46, d.f. 8, 562), weight (F-value =99.75 d.f. 8, 562) and BMI (F-value =45.26, d.f. 8, 562) among girls using ANOVA.The age-and sexspecifi c smooth percen le curves of 3 rd , 10 th , 25 th , 50 th , 75 th , 90 th and 97 th of BMI were derived for the further evalua on using L, M and S parameter in the model approach sta s cal procedures are depicted in Figure 1.

Prevalence of Thinness (BMI-for-age)
The overall prevalence of thinness was found to be 49.10% (51.16% boys, 46.89% girls) among rural adolescents (p>0.05).The prevalence of thinness showed that boys were found to be more suff erers in the diff erent thinness grades include mild (grade I; 27.41% vs. 27.11%),moderate (grade II; 14.62% vs. 12.08%) and severe (grade III; 9.14% vs. 8.70%) grades were found greater among boys than girls (p>0.05)(Table 1).Age specifi c trends in thinness prevalence was absent among both sexes but the magnitude is greater among early age groups (10-14 years).Age-specifi c overall thinness was found greater among 10 years and 14 years old boys (63.64%) and 11 year old girls (61.54%), while lower prevalence was observed among 18 years old boys (27.78%) and 16 years old girls (36.84%).The prevalence of thinness in diff erent grades was ranged 3.64% to 36.36% (in boys) and 4.23% to 32.69% (girls).Age-specifi c prevalence in diff erent grades of thinness (p>0.05) was found sta s cally insignifi cant in both sexes using chi-square analysis (p>0.05).Ageand sex-specifi c prevalence of overall thinness among adolescents is also depicted in Figure 2.

Discussion
The prevalence of undernutri on among children and adolescent are considered as a serious public health problem in India 1,3,4,5,7,8,9 .It is well known fact that contemporary India consists of a sizable number of ethnic and indigenous elements with enormous amounts of ethnic and gene c diversity 34 .Therefore, the nutri onal assessment is necessary for the overall improvement of health situa ons, nutri onal status and thereby development of the popula on concern and provides an indicator of the well-being of the individuals and/or popula ons.Therefore, the assessment of nutri onal status among children and adolescents belonging to vulnerable segments of the developing countries should be emphasized, not only for the iden fi ca on the nutri onal risks and also to improve health situa on.The present study determined the prevalence of thinness using newly proposed reference 16 .This reference was suggested to encourage direct comparison of global trends in thinness prevalence and to provide a classifi ca on for public health purposes.Prior to this report, there were no such suitable references for this age group (e.g., 2-18 years) for interna onal comparison except the Na onal Centre for Health Sta s cs and NHANES 6 .Further, several researchers have already raised their specula on about the suitability of western popula on based references for the nutri onal status assessment among Indian children and adolescents due to large ethnic and/or popula on physical growth varia on 8,9 .Moreover, the newly proposed references 16 was derived using the mul centre data of the United States, Great Britain, Hong Kong and the Netherlands including data from Brazil (developing country).Several studies have already assessed the prevalence of thinness u lizing this newly proposed reference 16 among Indian children and adolescents 13,29,30,31,35,36,37,38,39 .
The comparison with the NHANES 6 and WHO 40 showed that the majority of the adolescents refl ec ng inappropriate nutri onal status and poor physical growth a ainments than their normal counterparts (Figure 3).Majority of the boys and girls were found <15 th percen le, while boys among 10 years and 13 years (<5 th percen le) and 18 years and girls among 15-18 years were found >15 th percen le of the NHANES 7 .Whereas, comparison with WHO 40 reveals that the majority of the subjects was founded <15 th percen le, while excep on (<25 th percen le) was observed in 14-18 years (in girls) (Figure 3).Several studies have already reported that children and adolescents residing in rural areas were found more nutri onally vulnerable than their urban counterparts.The results reveal that there is a very high prevalence of overall thinness (49.10%) observed among boys  1).The adolescents belonging to the early ages (e.g., 10-14 years) found signifi cantly greater thinness prevalence than older ages (e.g., 17-18 years) (Figure 2).4] In the present study, the boys were found to be a more suff erer than the girl counterparts (p>0.05).It has also been observed that boys were more suff erer than girls in thinness among tribal and non-tribal children and adolescents residing in rural and sub-urban regions in India 7,13,14,31,35,37,38,39 .The comparison with the present study showed that the prevalence of overall thinness was found lower than the children and adolescents belonging to Bengalee 35 (62.26%), rural adolescents of Darjeeling 14 (63.40%),Kora-Modi 38 (67.23%),Santal 37 (56.40%) and greater than Nepali 36 (45.15%), rural adolescents of Midnapore 30 (20.80%) and Sonowal Kachari 31 (25.99%).Recent study u lizing data of NFHS-3 showed that a total of 40.20% (in boys) and 46.80% (in girls) of the adolescents aged 15-19 years were found suff erer from thinness in India 2 .Such varia on in the prevalence of thinness can be a ributed due to the large ethnic varia on, socioeconomic disparity and diverse socio-cultural and healthcare prac ces across the popula ons.However, it is now generally accepted facts that there is a high prevalence of thinness among rural Indian popula ons with more than 50.00% of Indian adolescents 8,9,14 .
Several studies have already advocated that the prevalence of undernutri on is be er assessed as thinness rather than was ng 6,16 .Available data suggested that the problem of thinness is persistent transversely among diff erent Indian popula ons with consistent propor ons especially among Indian children and adolescents 8,9,31,35,36,37,38,39 .The undernutri on during childhood and adolescence have not only delayed growth a ainment but aff ected overall the linear mechanism of growth processes 2,20 .The adolescents were suff ering from thinness is more likely to develop into thin adults with a lower BMI (e.g., CED) that would have an impact on their reduce physical work produc vity as well as lead to greater morbidity and mortality in popula on 6,41 .These delicate consequences of the adolescence undernutri on were subsequently confi rmed by the researchers among diff erent Indian communi es 1,7,42 .Moreover, introduc on of physical training could substan ally improve the work capacity, but inac vity leads to rapid and substan al reduc ons in the ability to sustain heavy work.Prevalence of thinness is frequently associated with nutri onal defi ciencies, menstrual irregularity and ea ng disorders, 43,44 and adverse reproduc ve outcomes 1,6 among adolescent girls and young women.Studies have also confi rmed that undernutri on (e.g., CED or thinness) among young women of childbearing age is a risk factor for adverse pregnancy outcomes, such as intrauterine growth-restricted or low-birth weight infants in Indian popula ons 1,45,46 .Moreover, marriage and early confi nement among adolescents are a long established custom in Indian society, with poverty and ignorance magnifying such problems 2,7,11,47 .The high prevalence undernutri on related to the thinness among adolescents could be the major a ribute for such manifesta ons that indicate to be the thinness is actually being more frequent than overweight-obesity among Indian adolescents 2,8,9,14 .

Conclusion
The adolescent popula ons of India have a great nutri onal risks in term of thinness which is more pronounced among early adolescents.The results of the present study will be useful for policy makers in their endeavour to formulate various developmental, interven on and healthcare programs.Further studies should be conducted using of newly proposed reference popula on especially to iden fy the actual propor on of nutri onally vulnerable popula ons for Na onal and interna onal comparison and to develop of new reference popula on for future use.Such studies would also help to reveal the enhanced usefulness and eff ec veness of the launched and/or ongoing interven on or supplementary programs at targeted popula ons.There is an urgent need of appropriate nutri onal interven on program to address the public health problem related to undernutri on especially in adolescents residing in rural areas to necessary ameliorate the nutri onal status.

Fig 1 :Fig 2 :
Fig 1: Age and sex specifi c smooth percen le curves among adolescent boys (A) and girls (B) using L, M and S model approach parameters

Fig 3 :
Fig 3: Comparison of age and sex specifi c mean BMI values of NHANES (WHO, 1995) and WHO (2007) with boys (A) and girls (B) author gratefully acknowledges the help and co-opera on of the Block, Gram Panchayet, local village level authori es and par cipants during the study.The extended help of Department of Anthropology, North Bengal University is also acknowledged.The fi nancial assistance in the form of UGC-NET-Junior Research Fellowship by the University Grants Commission, Government of India [UGC-Ref.No. 129/(NET-June 2007)] is also acknowledged.Funding: The fi nancial assistance in the form of UGC-NET-Junior Research Fellowship by the University Grants Commission, Government of India [UGC-Ref.No. 129/(NET-June 2007)] Confl ict of Interest: None Permission from IRB: Yes

Table 1 :
Age-sex specifi c subject distribu on, descrip ve sta s cs (mean ±SD) and prevalence of thinness among adolescents