SEARCH

SEARCH BY CITATION

Keywords:

  • socioeconomics;
  • overweight;
  • television;
  • physical activity;
  • adolescents

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Objective: The problem of overweight and obesity is not confined only to developed countries but is also widely prevalent in developing countries. The objective of this study was to assess the prevalence of overweight and obesity as defined by the International Obesity Task Force (IOTF) among school-age children in Hyderabad, India, and identify its associated factors.

Research Methods and Procedures: A cross-sectional and institutional study, adopting a multistage stratified cluster sampling procedure, was carried out during 2003 on adolescents 12 to 17 years of age of both sexes from Hyderabad, India.

Results: The overall prevalence of overweight was 6.1% [95% confidence interval (CI): 4.2, 8.0] among boys and 8.2% among girls (CI: 6.0, 10.4); 1.6% and 1.0% were obese, respectively. The prevalence was significantly higher (p < 0.05) among adolescents who watched television ≥3 h/d (10.4%) or belonged to a high socioeconomic background (14.9%, p < 0.001), whereas it was significantly lower among those participating regularly in outdoor games ≥6 h/wk (3.1%, p < 0.004) and household activities ≥3 h/d (4.7%, p < 0.001). The logistic regression analysis revealed that the prevalence of overweight was 4 times higher among the adolescents of high socioeconomic status [odds ratio (OR): 4.1; CI: 2.25, 7.52], 3 times higher in those not participating in outdoor games (OR: 2.75; CI: 1.56, 4.72), and 1.92 times higher in those watching television ≥3 h/d (OR: 1.92; CI: 1.16, 3.18).

Discussion: This study confirmed the findings of earlier studies carried out in Western countries and emphasizes that regular physical exercise, doing household activities, regulated television viewing, and healthy eating behaviors could contribute to controlling overweight and obesity.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Overweight and obesity and their health consequences have been recognized as major public health problems worldwide. A significant increasing trend in the prevalence of overweight and obesity among children and adolescents has been documented over the last few decades in developed and in developing countries (1)(2). The most significant long-term consequences of childhood and adolescent overweight and obesity are their persistence into adulthood with all of the attendant health risks (3)(4), such as dyslipidemia, hyperinsulinemia, type 2 diabetes, hypertension, cardiovascular diseases (5), arthritis, and behavioral problems. Obesity in children and adolescents is gradually becoming a major public health problem in many developing countries, including India (6). The prevalence is higher in urban than in rural areas (7). The results of studies among adolescents from parts of Punjab, Maharashtra, Delhi, and South India revealed that the prevalence of overweight and obesity was high (11% to 29%) (7). In Ludhiana, Punjab, urban children in the age group of 11 to 17 years of age were more overweight (11.6%) than their rural counterparts (4.7%) (7). In Pune, Maharashtra, studies among 1228 boys in the age group of 10 to 15 years indicated that ∼20% were overweight, whereas 5.7% were obese (7). A study carried out in Ludhiana, Punjab, on school children in the age group of 9 to 15 years revealed that the overall prevalences of overweight and obesity were 11% and 14%, respectively (7). Another study carried out in Delhi, India, among 5000 private school children in the age group of 4 to 18 years in 2002 by the Nutrition Foundation of India revealed that the prevalence of overweight was 29% (7). A similar study conducted in Chennai, in South India, showed that the prevalence of overweight was ∼17% and of obesity was 3% (7). In the studies in Punjab, Delhi, and Maharashtra, BMI cut-off points suggested by James et al. (8) were used for the definitions of overweight and obesity, as well as age- and sex-specific percentiles of BMI (National Health and Nutrition Examination Study), and the subjects involved belonged only to highly affluent families. In the study in South India, the adolescents belonged to high, middle, and low socioeconomic status (SES).1 The age groups included in these studies varied from one another. Several cross-sectional studies in Western countries have shown that overweight and obese adolescents are less physically active than non-obese subjects, and physical inactivity, high socioeconomic background, and dietary transition were found to be major factors (9). However, in this study, the role of factors such as participation in sports and games, household chores, physical inactivities such as television viewing and playing computer/video games, and consumption of junk foods were also studied.

Therefore, this study was undertaken to estimate the overall prevalence of overweight (≥85th percentile) among adolescents of Hyderabad in South India and to examine associated factors such as SES, occupation, literacy status of parents, physical activities such as participation in sports and games, sedentary activities such as watching television and playing computer/video games, and dietary behaviors.

Research Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

Study Design and Sample Size

This was a cross-sectional and institutional study and adopted a multistage stratified cluster sampling procedure. The sample size was calculated assuming the prevalence of overweight and obesity of 9% reported in an earlier study (6)(7) and 95% confidence interval (CI), with 20% relative precision and 1.2 design effect. The sample size arrived at was 1145 to 1150. For the selection of schools, the list of all schools (1147) belonging to different categories (government, semi-private, and private) was obtained from the school authorities of the local government. Generally, government schools cater to poorer sections of the population, whereas semi-private schools cater to lower middle and middle income groups. Children from upper middle and high income groups attend private schools. Twenty-three institutions were selected based on proportion to population size from each of the above categories. It was assumed that from each institution, at least 50 subjects would be recruited from one or two classes. The classes were selected randomly. A total of 1208 subjects were used proportionately from three categories, i.e., from government (36.2%), semi-private (19.3%), and private (44.5%) schools. This distribution was about the same as the distribution of the population (10).

Subjects

The subjects were adolescents 12 to 17 years of age in the city of Hyderabad, Andhra Pradesh, Southern State of India. Hyderabad is the capital city of Andhra Pradesh, established 400 years ago (1591 to 1592) on Deccan plateau, 541 m above sea level and sprawled over an area of 260 km2. The population of ∼3.69 million consists of different ethnic, language, and socioeconomic groups. The majority of the population are Hindus and Muslims. Other communities such as Christians, Jains, and Sikhs are represented in smaller numbers. The State of Andhra Pradesh is one of the largest provinces of the Indian union, with a total population of ∼76.2 million. Approximately 73% of the population live in rural areas, subsisting mainly on agriculture. A large number of heavy industries are located in the state, and this area is recognized for its contribution in the area of information and technology. Educational institutions catering to disabled children were excluded from the study in view of the difficulty in taking accurate anthropometric measurements and because of their heterogeneity. From the selected institutions, adolescents who were in 7th to 10th grade were recruited to get the required age groups (i.e., 12 to 17 years). All of the selected children participated in the study, except those from one private school consisting of ∼40 children, because the head of the institution did not agree to the study.

Data Collection

The Institutional Ethical Review Board approved the study protocol, and written consent was obtained from the heads of the educational institutions. Oral assent was obtained from all of the adolescents. The household socioeconomic and demographic data such as community, literacy status, and occupation of father and mother were collected from the adolescents, and the same was confirmed with school records. The date of birth of each pupil was taken from the school records or from the adolescent.

Information on the following aspects was collected from the adolescents using a pre-tested and validated questionnaire.

Household possession of articles, ownership of parental house, and residential status of adolescents were collected as proxy variables for calculation of socioeconomic index. Information was also collected on physical activity, which included distance of school from the residence and the mode of transport used to go to school and physical activities such as participation in sports and games, aerobic physical exercises, frequency and duration of participation in household activities, time spent watching television and playing computer and video games, perception of body image, diet preferences, and consumption pattern.

Measurements

Trained investigators weighed all of the adolescents without shoes and heavy clothing, using a SECA electronic weighing scale (Seca, Hanover, MD), with an error of ±100 grams. The weighing scale was regularly checked with known standard weights. A portable anthropometric rod was used for measuring height, with an error to the nearest of 0.1 cm, using standard procedures (11). The International Obesity Task Force references were used to define overweight and obesity in this study (12). The survey was carried out during 2003.

Analysis of Data

Data were analyzed using SPSS for Windows version 14.0 (SPSS, Inc., Chicago, IL). Adolescents were categorized into two groups: overweight (≥85th percentile) and non-overweight (<85th percentile) using age- and sex-specific percentiles of BMI (12). Composite socioeconomic index was computed using the proxy income variables and categorized based on the quartile scores (13) into low (<6.0), middle (6.0 to 10.9), upper middle (11.0 to 17.9), and high (≥18.0) socioeconomic groups. Watching television was categorized based on hours of watching per day into 0, <3, and ≥3 h/d. Similarly, other variables such as physical activities and aerobic exercises were also categorized (14). The prevalences of overweight and obesity and 95% CI were calculated, according to age, sex, SES, ethnic group, type of school, and physical activity level. Associations were assessed using χ2 test. Multiple logistic regression analysis was also carried out to examine associations between independent variables and overweight and obesity. For all statistical tests, p < 0.05 was taken as the significant level.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

A total of 1208 adolescents (48.5% boys) in the age group of 12 to 17 years of age, with a mean age of 14.5 years, were studied. Approximately 60% of the sample belonged to communities that are considered socially backward communities, a proportion that is similar to that reported in Census 2001. The major occupation of fathers was either service (48.3%) or business (27.8%). Approximately 45% of adolescents did not participate in outdoor games, whereas only one third of adolescents were participating in any outdoor games for ≥6 h/wk. About two thirds of adolescents were participating in household activities, whereas 91% reported watching television on school days. Approximately 22% of adolescents preferred to consume junk foods because these were their favorite dishes. Approximately 7.9% of adolescents felt that they were either overweight or obese, whereas 70% perceived themselves as normal-weight and 20% as lean. The mean BMI of the sample was 18.0 kg/m2. There were significant differences (p < 0.05) in mean BMI between sexes, communities, SES, and likes and dislikes for junk foods (Table 1). In general, the prevalence of overweight (≥85th percentile) among adolescents was 7.2% (95% CI: 5.7, 8.7). Although the difference was not statistically significant (p > 0.05) between sexes, the proportion of overweight was higher among girls (8.2%; 95% CI: 6.0, 10.4) than among boys (6.1%; 95% CI: 4.2, 8.0). The prevalence of overweight (≥85 percentile) among girls tended to increase from 6.2% at 12 years to 10.8% at the age of 15 years and gradually decreased at 17 years (9.2%), whereas in boys, it was the highest at the age of 14 years (9.2%) and decreased to 5.3% at the age of 17 years.

Table 1.  Basic characteristics of the sample and prevalences of overweight and obesity among adolescents in Hyderabad, India
VariableSub-variablePercentMean BMIOverweight (≥85th to <95th percentile)Obese (≥95th percentile)Overweight and above (≥85th percentile) and 95% CI
  • CI, confidence interval; SES, socioeconomic status. Variation in superscripts indicates significance of difference (p < 0.05) between the groups for overweight and obesity.

  • *

    Age in years.

Mean age (yrs)Pooled14.5*18.0 (17.8, 18.2)5.91.37.2 (5.7, 8.7)
AdolescentsBoys48.517.4a (17.2, 17.7)5.11.06.1a (4.2, 8.0)
 Girls51.518.6b (18.3, 18.8)6.61.68.2a (6.0, 10.4)
Subjects from typeGovernment36.217.2 (17.0, 17.5)3.20.03.2a(1.5, 4.9)
 of schoolSemi-private19.318.1 (17.7, 18.6)6.92.19.0b (5.3, 12.7
 Private44.518.7 (18.3, 19.0)7.62.09.6b (7.1, 12.1)
CommunityScheduled caste/tribes19.317.8 (17.4, 18.1)4.70.04.7a (2.0, 7.4)
 Backward community40.917.6 (17.3, 17.9)4.51.45.9a (3.8, 8.0)
 Others39.818.6 (18.3, 18.9)7.91.99.8b (7.1, 12.5)
Composite SESLow SES23.017.2 (16.9, 17.5)2.90.43.3a (1.2, 5.4)
 Low middle SES24.617.3 (17.3, 17.7)1.70.32.0a (0.4, 3.6)
 Upper middle SES26.617.9 (17.6, 18.3)7.10.67.7b (4.7, 10.7)
 High SES23.919.5 (19.1, 19.9)11.23.714.9a, b (11.0, 18.8)
OccupationService23.918.4 (18.1, 18.7)7.41.79.1a (6.8, 11.4)
 Business48.318.1 (17.7, 18.4)6.01.57.5a (4.7, 10.3)
 Others27.817.2 (16.9, 17.6)2.80.33.1b (1.1, 5.1)
Consumption ofLike21.718.9 (18.5, 19.4)8.83.812.6a (8.6, 16.6)
 junk foodDislike78.317.8 (17.6, 18.0)5.10.65.7b (4.2, 7.2)
Mode of conveyanceCar/scooter/public transportation59.418.3 (17.9, 18.6)8.11.89.9a (6.4, 13.4)
 to schoolWalking/cycling40.618.0 (17.7, 18.2)5.21.26.4b (4.9, 8.1)

Association with Socioeconomic Factors

The prevalence of overweight among the adolescents studying in private schools (9.6%; 95% CI: 7.1, 12.1) was significantly higher (p < 0.05) than among those studying in government schools (3.2%). Similarly, it was significantly higher (p < 0.05) among the adolescents of high SES (15%; 95% CI: 11.0, 18.8) compared with those of low SES (3.3%). The prevalence was also higher among the adolescents whose parents’ occupations were either service (9.1%) or business (7.4%) than other occupations (3.1%) and among those who were fond of junk foods (12.6%; 95% CI: 8.6, 16.6). The prevalence was significantly lower (p < 0.05) among adolescents who either walked to school or came on bicycle (6.4%; 95% CI: 4.9, 8.1) than among the adolescents who used vehicular transport such as motorcycles or cars (9.9%; 95% CI: 6.4, 13.4) (Table 1).

Role of Physical Activity

The prevalence of overweight and obesity (3.1%) was significantly lower among the adolescents who participated in outdoor games (p < 0.004) than among the non-participants (9.7%). It was also significantly higher (p < 0.001) among the adolescents who did not perform any household activities (18.6%) compared with those participating in various household chores (4.7%). Similarly, overweight and obesity were marginally higher among adolescents who were not involved in physical activities such as walking, cycling, and jogging (Table 2). The prevalence of overweight among adolescents who were sedentary, watching television ≥3 h/d, was significantly higher (10.4%) compared with those who watched 0 or <3 h/d (5.9% to 6.3%).

Table 2.  Association between prevalence (%) of overweight and physical activities
VariableDurationNNon-overweightOverweight (≥85th percentile)P
  1. Variation in superscripts indicates significance of difference between the groups.

Participation in household activities (h/d)None22181.418.6a<0.0001
 <323396.13.9b 
 ≥371695.34.7b 
Participation in outdoor games (h/wk)None52690.39.7a<0.004
 <622891.28.8a 
 ≥641696.93.1b 
Participation in indoor games (h/wk)None56991.68.4a>0.05
 <634893.46.6a 
 ≥625394.95.1a 
Walking (h/d)None90593.16.9>0.05
 <325491.88.2 
 ≥311100.00.0 
Jogging (h/d)None102692.67.4>0.05
 <313094.45.6 
 ≥314100.00.0 

To adjust for potentially confounding variables and to study possible mediating factors, a multivariate logistic regression analysis was carried out. In the model, although overweight and obesity were dependent variables, age, sex, ethnicity, occupation, socioeconomic and educational status of parents, physical activity such as participation in sports and games, household activities, aerobic exercises, likes and dislikes of junk foods, and physical inactivity such as watching television and playing computers/video games formed independent variables. This analysis revealed that the risk of overweight was 4 times higher among the adolescents of high SES, 3 times higher among those participating <3 h/wk in outdoor games/sports, 2.7 times higher among those who were not participating in household activities (≥3 h/d), and 2 times higher among those who reported watching television for ≥3 h/d.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

To our knowledge, this is the first comprehensive study in the Indian subcontinent attempting to document the prevalence of overweight and obesity and their associated factors that covered an adequate sample of urban adolescents (12 to 17 years old). The overall prevalence of overweight (≥85th percentile of BMI) among the urban adolescents studied (7.2%) was 10 times higher then that of their rural counterparts (0.6%) reported by the National Nutrition Monitoring Bureau surveys in 2002 (15). However, the prevalence was lower in this study compared with studies carried out in cities such as Ludhiana, Punjab, Pune, Maharashtra, Dehli, Chennai, and Tamil Nadu (7). The reason for the higher prevalence of overweight (26%) and obesity (7.4%) among the adolescent population studied in Delhi and Ludhiana might be that the subjects selected for these studies were affluent. In the Delhi study, the sample was selected from one school only. Umesh Kapil (7) and Nutrition Foundation of India (7) used the cut-off for BMI for overweight and obesity on the basis of the classification of James et al. (8), which may not be suitable for growing children. In other studies, even though the definition of the National Health and Nutrition Examination Study and BMI age- and sex-specific percentiles were used, the sample was drawn from a high socioeconomic stratum. The prevalence was, however, comparable with figures reported in other developing countries (16). The prevalence was marginally higher among girls compared with boys (p > 0.05), as observed in many international studies.

A clear socioeconomic gradient in the prevalence of overweight and obesity was observed in this study, which is consistent with other studies (17)(18). This could be for several reasons that are related to obesity, encountered to a greater extent in higher-income groups. Studies have reported that the rise in sedentary behaviors such as increased use of vehicular transport and decreased physical activity has led to increased prevalence of overweight and obesity.

Overweight and obesity were marginally higher in the pubertal age group, i.e., 13 to 15 years of age, as was observed in other studies in Delhi (7) and Chennai (7), perhaps because of increased adipose tissue and overall body weight in children during puberty. The prevalences of overweight and obesity were marginally less in the postpubertal period (16 to 17 years of age). It has been reported earlier that the number of fat cells increases during periods of rapid growth up to 16 years of age, after which increased fat ordinarily accumulates by increasing size of the fat cells already present (19).

The results clearly revealed that regular physical activity was an important factor in reducing prevalence of overweight and obesity. The prevalence was significantly lower in the children who participated regularly in household chores (p < 0.001), played outdoor games, and performed physical exercise. The diets of the children in the higher socioeconomic group are known for their higher fat content, and the subjects are involved in more sedentary activities. These observations are consistent with results of previous studies (20). In addition, the prevalences of overweight and obesity were higher among children who were involved in sedentary activities such as spending ≥3 h/d on television viewing (13). Klesges et al. (21) also reported the effect of watching television on metabolic rate and overweight and obesity in children. In urban areas, considering the safety of keeping children away from heavy traffic, parents feel more comfortable if their children play indoor games or watch television and, therefore, do not encourage them to participate in outdoor sports and games.

Freedman et al. (5) showed the adverse effects of overweight in their 17-year follow-up study and reported that an early average increase of 0.5 kg/m2 of BMI in children increases the risk for hypertension, dyslipidemia, and type 2 diabetes a decade later. It is interesting to note that ∼8% of adolescents perceived that they were overweight, which indicates that the self-reporting of obesity could also be a good indicator of the problem.

The major conclusion drawn from this study is that low levels of physical activity, watching television, and consuming junk foods are associated with a higher prevalence of overweight. Thus, participation in household activities and regular physical exercise could help in lowering the prevalence of overweight. Therefore, the role of physical activity, games, and sports should be emphasized, and facilities should be provided for outdoor games in schools, with compulsory hours of sports and games. There is an urgent need to educate the urban community on the aspects of healthy food habits and desired lifestyles to prevent overweight/obesity and its associated ill effects.

Acknowledgments

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References

The authors thank the technical staff of the Division of Community Studies, National Institute of Nutrition, Hyderabad, India, who helped carry out the study; our colleagues at the Division of Community Studies, State Education officials, heads of schools, government of Andhra Pradesh, and the parents and children for cooperation and encouragement in carrying out the study; and Director, National Institute of Nutrition, (ICMR), Hyderabad, and Director, Sree Chitra Tirunal Institute for Medical Sciences and Technology, Thiruvananthapuram, Kerala, for support and cooperation in carrying out this study. There was no funding/outside support for this study.

Footnotes
  • 1

    Nonstandard abbreviations: SES, socioeconomic status; CI, confidence interval.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Research Methods and Procedures
  5. Results
  6. Discussion
  7. Acknowledgments
  8. References
  • 1
    Chinn, S., Rona, RJ (2001) Prevalence and trends in overweight and obesity in three cross sectional studies of British children, 1974–94. BMJ 322: 2426.
  • 2
    Martorell, R., Khan, Kettel L., Hughes, M. L., Grummer-Strawn, LM (2000) Overweight and obesity in preschool children from developing countries. Int J Obes Relat Metab Disord. 24: 959967.
  • 3
    Must, A., Strauss, RS (1999) Risks and consequences of childhood and adolescent obesity. Int J Obes Relat Metab Disord. 23 (Suppl 2), S2S11.
  • 4
    Power, C., Lake, J. K., Cole, TJ (1997) Measurement and long-term health risks of child and adolescent fatness. Int J Obes Relat Metab Disord. 21: 507526.
  • 5
    Freedman, D. S., Dietz, W. H., Srinivasan, S. R., Berenson, G. (1999) The relation of overweight to cardiovascular risk factors among children and adolescents. The Bogalusa Heart Study. Pediatrics 103: 11751182.
  • 6
    Popkin, B. M., Doak, CM (1998) The obesity epidemic is a worldwide phenomenon. Nutr Rev. 56: 106114.
  • 7
    Kaur, S., Kapil, U., Singh, P. (2005) Pattern of chronic diseases amongst adolescent obese children in developing countries. Curr Sci. 88: 10521056.
  • 8
    James, W. P., Ferro-Luzzi, A., Waterlow, JC (1988) Definition of chronic energy deficiency in adults. Report of a working party of the International Dietary Energy Consultative Group. Eur J Clin Nutr. 42: 969981.
  • 9
    Eisenmann, J. C., Bartee, R. T., Wang, MQ (2002) Physical activity, TV viewing, and weight in U.S youth: 1999 Youth Risk Behaviour Survey. Obes Res. 10: 379385.
  • 10
    Tooley, J., Dixon, P.. The private sector serving the educational needs of the poor: a case study from India. http:www.ncl.ac.ukegwestpdfs (Accessed October 5, 2006).
  • 11
    Jelliffee, DB (1988) Assessment of Nutritional Status of the Community World Health Organization Geneva, Switzerland.
  • 12
    Cole, T., Bellizzi, M. C., Flegal, K. M., Dietz, WH (2000) Establishing a standard definition for child overweight and obesity: international survey. BMJ 320: 16.
  • 13
    National Council for Applied and Economic Research (2001) National Human Development Report 2001 Oxford University Press Oxford, UK.
  • 14
    Trest, S. G., Kerr, L. M., Ward, D. S., Pate, RR (2001) Physical activity and determinants of physical activity in obese and non-obese children. Int J Obes Relat Metab Disord. 25: 822829.
  • 15
    National Nutrition Monitoring Bureau (2002) Diet and Nutritional Status of Rural Population National Institute of Nutrition, Indian Council of Medical Research Hyderabad, India.
  • 16
    Sakmoto, N., Wansorns, S., Tontisirink, Marni, E. (2001) A social epidemiological study of obesity among preschool children in Thailand. Int J Obes Relat Metab Disord. 25: 389394.
  • 17
    Kasmini, K. (1997) Prevalence of overweight and obesity among school children aged between 7–16 years amongst the 3 major ethnic groups in Kuala Lampur, Malaysia. Asia Pacif J Clin Nutr. 6: 172174.
  • 18
    Chu, NF (2001) Prevalence and trends of obesity among school children in Taiwan–the Taipei Children Heart Study. Int J Obes Relat Metab Disord. 25: 170176.
  • 19
    Hirsch, J. (1995) Fat cell number and size as a determinant of subsequent obesity Winick, M. eds.. Childhood Obesity John Wiley and Sons New York.
  • 20
    Patrick, K., Calfas, G. J., Zabinski, M. F., Cella, J. (2004) Diet, physical activity, and sedentary behaviours as risk factors for overweight in adolescence. Arch Pediatr Adolesce. 158: 385390.
  • 21
    Klesges, R. C., Shelton, M. L., Klesges, LM (1993) Effects of television on metabolic rate: potential implications for childhood obesity. Pediatrics 91: 281286.