Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
March 2006

Association of Depression and Anxiety Disorders With Weight Change in a Prospective Community-Based Study of Children Followed Up Into Adulthood

Author Affiliations

Author Affiliations: Gerald J. and Dorothy R. Friedman School of Nutrition Science and Policy, Tufts University (Ms Anderson and Dr Must), and Department of Public Health and Family Medicine, Tufts University School of Medicine (Drs Naumova and Must), Boston, Mass; and Department of Psychiatry, College of Physicians and Surgeons, Columbia University, and Department of Epidemiology, New York State Psychiatric Institute, New York, NY (Dr Cohen).


Copyright 2006 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2006

Arch Pediatr Adolesc Med. 2006;160(3):285-291. doi:10.1001/archpedi.160.3.285

Objective  To investigate childhood to adulthood weight change associated with anxiety and depression.

Design  The Children in the Community Study. A prospective longitudinal investigation.

Setting  Albany and Saratoga Counties, New York.

Participants  Eight hundred twenty individuals (403 females and 417 males) assessed at 4 time points: in 1983 when they were 9 to 18 years old (n = 776), in 1985 to 1986 when they were 11 to 22 years old (n = 775), in 1991 to 1994 when they were 17 to 28 years old (n = 776), and in 2001 to 2003 when they were 28 to 40 years old (n = 661).

Main Exposures  Anxiety disorders and depression assessed by structured diagnostic interview.

Main Outcome Measures  Centers for Disease Control and Prevention body mass index z score (BMIz), a measure of weight status; and association of anxiety and depression with BMIz level and annual change.

Results  In females, anxiety disorders were associated with higher weight status, a BMIz of 0.13 (95% confidence interval, 0.01-0.25) units higher compared with females without anxiety disorders. Female depression was associated with a gain in BMIz of 0.09 units/y (95% confidence interval, 0.03-0.15 units/y), modified by the age when depression was first observed, such that early depression onset was associated with a higher subsequent BMIz than depression onset at older ages. In males, childhood depression was associated with a lower BMIz (−0.46; 95% confidence interval, −0.93 to 0.02 units lower at the age of 9 years), but BMIz trajectories for males with or without depression converged in adulthood; male anxiety disorders were not substantively associated with weight status.

Conclusions  Anxiety disorders and depression were associated with a higher BMIz in females, whereas these disorders in males were not associated with a higher BMIz. These results, if causal and confirmed in other prospective studies, support treating female anxiety and depression as part of comprehensive obesity prevention efforts.

The prevalence of obesity among US children, adolescents, and adults represents a public health crisis.1 Obesity often has roots early in life, and if present in adolescence is likely to persist into adulthood.2 Understanding the predictive factors associated with weight gain is important for obesity prevention; obesity treatment will be most effective when targeted, and informed by understanding of etiologic pathways and correlates.

Evidence suggests that a child's social and psychological environment contributes to obesity risk. Psychosocial difficulty during childhood is associated with increased risk for adult obesity,3 and with rapid weight gain in children.4 Physical or verbal abuse,5 and parental neglect,6 are associated with a higher obesity risk, as is poor quality of a child's home environment.7 Thus, childhood adversity can increase risk for later obesity; these associations may be mediated by negative psychological symptoms.

That psychopathological dysfunction and obesity are associated has long been hypothesized, but studies have yielded conflicting results. Some cross-sectional population-based studies817 show associations between poorer psychological functioning and obesity, while others1824 do not. Similar variability exists among studies19,21,2426 of pediatric populations.

Evidence from the few prospective studies2734 conducted suggests that psychological disorders are associated with weight gain. Limitations of these studies include having observations at only 2 time points,2732 relatively short periods of follow-up,27,30,33,34 and clinical rather than population-based samples.28

Our objectives were to evaluate the association of anxiety disorders and major depressive disorder (depression), defined based on Diagnostic and Statistical Manual of Mental Disorders (DSM) diagnostic criteria, with weight trajectory in a community-based cohort of children followed up into adulthood.


The Children in the Community Study is a prospective cohort study of determinants and correlates of psychological health. In these analyses, we have examined the influence of depression and anxiety disorders on childhood to adulthood weight trajectory. Children in the Community Study design and operations have been described previously.35,36 Briefly, 976 families with children born between 1965 and 1974, residing in Albany and Saratoga Counties in Upstate New York, were sampled in 1975. The sample was demographically representative of the area, and was primarily of white race/ethnicity (91.5%). Participants were assessed in 1983 (wave 1, n = 776) when they were 9 to 18 years old, in 1985 to 1986 (wave 2, n = 775) when they were 11 to 22 years old, in 1991 to 1994 (wave 3, n = 776) when they were 17 to 28 years old, and in 2001 to 2003 (wave 4, n = 661) when they were 28 to 40 years old. Trained interviewers conducted in-home interviews with participants at waves 1 to 4, and a parent (usually the mother) at waves 1 to 3. Informed consent (or assent for children) was obtained at each wave. The institutional review boards of the New York State Psychiatric Institute and Tufts–New England Medical Center, Boston, approved this study. Eight hundred twenty individuals (403 females and 417 males) contributed data to these analyses: 593 (72.3%) with data for 4 waves, 176 (21.5%) with data for 3 waves, 37 (4.5%) with data for 2 waves, and 14 (1.7%) with data for 1 wave.


Depression and anxiety disorders were assessed using structured diagnostic interviews for DSM disorders. The Diagnostic Interview Schedule for Children37 was administered, separately, to a parent and the participant at waves 1 and 2. Parents and children provide unique nonoverlapping information,38 and in accordance with current practice, disorders were identified if either reported a minimum number of items meeting DSM-III criteria and the combined parent-child symptom score was at least 1 SD above the mean. At wave 3, the Diagnostic Interview Schedule for Children, with minor adjustments made for age appropriateness and comparability across DSM revisions,35 was administered to participants. At wave 4, disorders were assessed using the Structured Clinical Interview for DSM-IV. The Diagnostic Interview Schedule for Children and the Structured Clinical Interview for DSM-IV assessed whether an individual had symptoms consistent with a DSM disorder at any point during the year leading up to the interview. In addition, during the wave 2 assessment, disorders were identified if they occurred at any point between waves 1 and 2.

For each participant, we determined if the participant was observed to have met the diagnostic criteria for an anxiety disorder or depression at any of the 4 interviews. If the diagnostic criteria were met, the age of the participant at the interview during which the disorder was first identified was designated as the age of observed onset. For example, if a participant did not meet the diagnostic criteria for depression at wave 1 or 2, and did meet the diagnostic criteria for depression at wave 3, we defined the age we first observed depression in this individual as his or her age at wave 3.

The specific disorders that could be considered “anxiety” were not identical across waves. We defined anxiety as follows: at waves 1 and 2, separation anxiety, social phobia, or overanxious disorder; and at waves 3 and 4 (when participants were adults), social phobia, generalized anxiety disorder, obsessive-compulsive disorder, panic disorder, or agoraphobia.

Depression is often accompanied by symptoms of anxiety, some of which may be severe enough to meet the diagnostic criteria for an anxiety disorder. Thus, we do not attempt to separate depression from comorbid anxiety.


Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters, from height and weight reported at each wave. At wave 1, a parent reported the participant's height and weight. At wave 2, a parent and the participant reported the participant's height and weight. At waves 3 and 4, participants self-reported their height and weight. Agreement between participant- and parent-reported height and weight at wave 2 was high (intraclass correlation coefficient for height, 0.92; and for weight, 0.95). At wave 2, we used the participant's report of height and weight if the participant was older than 13 years; otherwise, we used the parent report. Reported height for individuals across waves was examined for consistency and implausible values, such as decreases in height over time. Approximately 30% of individuals reported a decrease in height; however, in all except 64 individuals, the difference was 1 inch or less (≤2.54 cm). Height decreases were adjusted by substituting a “stable adult height” if the participant reported 2 or more identical heights; otherwise, the mean of reported adult heights was used. Adult was defined herein as follows: for males, those older than 17 years; and for females, those older than 15 years. Similar methods were used to stabilize height for 18 individuals reporting growth of greater than 1 inch (>2.54 cm) between 2 waves after the age of 20 years. Observations (<1%) were not included in analyses if implausible and not accompanied by stable estimates.

Participants were observed during childhood and adulthood, necessitating a measure of weight status comparable throughout this period. Body mass index is commonly used to assess weight status in adults, but BMI is not a good measure for comparison of relative weight status among children of different ages because it increases with age, independent of adiposity, during childhood. Instead, in children, it is recommended that weight status be determined relative to age- and sex-specific reference data.39,40 Therefore, we used the Centers for Disease Control and Prevention BMI-for-age reference data,41 which are available for children and adolescents aged 2 to 20 years, to calculate a BMI z score (BMIz) for each participant at each wave. The BMIz scores correspond with growth chart percentiles and allow for tracking a child's relative weight through childhood and adolescence. A BMIz of 0 equals the 50th percentile; 1.04, the 85th percentile; and 1.65, the 95th percentile. To provide continuity from childhood to adulthood, we used the age 20 reference to calculate BMIz when individuals were 20 years or older, as have others.42 Because BMIz is a continuous measure of relative weight adjusted for age and sex, we were able to assess the association of depression and anxiety disorders on weight trajectory over an age range spanning childhood, adolescence, and adulthood.


Socioeconomic status was defined by an index (mean, 10; SD, 1) of family income, parental education, work status, occupation, and receipt of public assistance.36 Potential confounding variables included whether medications were taken for emotional or behavioral problems before the age of 21 years, assessed by self-report at wave 4, and whether the participant reported being a current smoker, assessed by self-report at each wave. Analyses were conducted separately for males and females because we hypothesized that the association between anxiety or depression and weight status would depend on sex.


We used linear mixed models43 to estimate differences in BMIz level and annual change in BMIz associated with observed onset of anxiety disorder or depression, compared with participants who were never observed as having an anxiety disorder or depression. We used age (in years) as the measure of time in these models.

Our model-building strategy was established a priori and used to assess whether observed onset of female anxiety, male anxiety, female depression, or male depression was associated with BMIz level or annual change in BMIz; we used the term trajectory when referring simultaneously to BMIz level and annual change. We controlled for socioeconomic status in all models. The BMIz trajectory was defined as a cubic function of age. We determined the effect of anxiety disorders and depression on BMIz level and annual change in BMIz, and assessed whether the participant's age at the wave in which anxiety or depression was first recognized modified these associations. We assessed potential confounding variables (current smoking and early medication use) by determining whether their inclusion changed model estimates. We identified those individuals with an anxiety disorder who did not meet criteria (at any point in the study) for depression, and evaluated the result of using this more restricted definition of anxiety.


The mean BMIz for males and females increased by greater than 0.5 units between waves 3 and 4 (Table 1). During the study, 310 individuals (119 males and 191 females) met the diagnostic criteria for an anxiety disorder; of these individuals, 86 males and 116 females did not also meet the criteria for depression. The mean age at the wave in which an anxiety disorder was first recognized was 17.5 years in females (range, 9.4-38.0 years) and 16.8 years in males (range, 9.3-39.4 years). One hundred forty-eight individuals (50 males and 98 females) met the diagnostic criteria for depression during the study. The mean age at the wave in which depression was first identified was 23.8 years in females (range, 9.7-38.4 years) and 23.5 years in males (range, 10.1-37.5 years).

Table 1. 
Image not available
Age, Weight Status, Anxiety, and Depression of Participants Studied in 1983, 1985 to 1986, 1991 to 1994, and 2001 to 2003*

In females, anxiety disorders were associated with higher weight status; our model estimated that the mean BMIz for females observed as having had an anxiety disorder was 0.13 units higher compared with females of the same age and socioeconomic status who were not observed as having had an anxiety disorder (Table 2). This difference of 0.13 units, comparing females with an anxiety disorder with females without an anxiety disorder, was unrelated to the number of years elapsed since observed anxiety disorder onset (ie, annual change in BMIz was estimated to be essentially the same, a difference in annual BMIz change of −0.0002 units/y, regardless of whether the participant had been recognized as having an anxiety disorder). In males, anxiety disorders were not associated with a substantive or statistically significant difference in BMIz level or annual change in BMIz compared with males without an anxiety disorder (Table 2). We found no evidence that the association of BMIz with anxiety disorders in either females or males was related to age at first recognition of anxiety. Controlling for cigarette smoking or early medication use did not substantively change results.

Table 2. 
Image not available
Estimated Difference in BMIz Trajectory (Level and Annual Change) Associated With Anxiety Disorder and Depression*

Using a definition of anxiety disorder restricted to females who had never met the diagnostic criteria for depression increased the estimated mean difference in BMIz associated with anxiety to 0.18 units higher compared with other females (Table 2). Our results for males with anxiety disorders that were not comorbid with depression were virtually the same as those for male anxiety disorders overall (Table 2).


For females, a history of depression was associated with greater yearly gains in BMIz compared with females without a history of depression; the annual change in BMIz for females observed as having had depression was greater by 0.09 units/y than the annual change in BMIz for females who had never been identified as having had depression. The magnitude of this annual BMIz change was modified by age at first recognition of depression and by the number of years elapsed since depression was recognized; the estimated yearly gain in BMIz was reduced by 0.003 units for each additional year of age a female was when depression was first recognized and each year elapsed since then (Table 2 and Figure 1). In males, the first recognition of depression was of borderline statistical significance (P=.06) and predicted a lower BMIz compared with males without depression. Childhood depression was associated with the greatest estimated difference in BMIz (a 9-year-old boy with depression would be estimated to have a BMIz of 0.46 units lower than a boy of the same age without depression). The BMIz trajectories converged with increasing age for males with and without a history of depression; each additional year of age reduced the difference between estimated BMIz for males with depression compared with males without depression by 0.02 units (Table 2 and Figure 2). Controlling for cigarette smoking or early medication use did not substantively change results.

Figure 1.
Image not available

Mean difference calculated from model estimates of body mass index z score (BMIz) with age for females with depression first recognized at the ages of 9, 14, or 18 years compared with females without depression (reference line at 0). The linear mixed-effects model was as follows: BMIz = 1.77 − [0.14(Age − 9)] + {0.01[(Age − 9)2]} − {0.0002[(Age − 9)3]} − [0.12(Socioeconomic Status)] − [0.07(Depression)] + [0.09(Years of Depression)] − {0.003[(Age − 9) (Years of Depression)]}.

Figure 2.
Image not available

Mean difference calculated from model estimates of body mass index z score (BMIz) with age for males with depression first recognized at the ages of 9, 14, or 18 years compared with males without depression (reference line at 0). The linear mixed-effects model was as follows: BMIz = 1.78 − [0.13(Age − 9)] + {0.01[(Age − 9)2]} − {0.0002[(Age − 9)3]} − [0.11(Socioeconomic Status)] − [0.46(Depression)] − [0.005(Years of Depression)] + {0.02[(Age − 9) (Depression)]}.


We found that DSM disorders of anxiety or depression were associated in females with a higher BMIz. The BMIz was predicted to be 0.13 or 0.18 units higher, depending on depression comorbidity, for females recognized as having had an anxiety disorder compared with the BMIz level of similar females without an anxiety disorder. The estimated annual change in BMIz was essentially the same for females irrespective of whether they had been observed as having had an anxiety disorder and, thus, the estimated mean difference of 0.13 or 0.18 units was maintained over time. A 0.18-unit difference in BMIz would translate, depending on initial BMI, to a difference in adult BMI of, for example, approximately 25 to 26, 28 to 30, or 30 to 32. For a woman with a height of 64 inches (163 cm) (the average height of US women aged 20-40 years),45 these BMI changes correspond to 2.7-, 4.2-, and 5.3-kg weight differences, respectively. Although these average weight differences are not large, obesity results from incremental increases in weight, and successful prevention is likely to require interventions targeted toward many factors, no one of which, alone, is sufficient to prevent obesity.

The association of female depression with BMIz depended on age at first recognition of depression and the number of years since elapsed. Our model predicted that a 30-year-old woman first recognized with depression at the age of 14 years would have a BMIz that was 0.34 units higher than a similar woman without depression; if depression were first recognized at the age of 18 years, the estimated difference in BMIz at the age of 30 years would be reduced to 0.23 units. A 0.34-unit difference in BMIz translates to a difference in BMI of approximately 24 to 26, 25 to 27, or 27 to 30; for a woman of average height, these values represent weight differences of 4.8, 5.3, and 7.4 kg, respectively. Our model estimated that differences in BMIz were largest for adolescents and young adults when depression was present at an early age (Figure 1); irrespective of the age of depression recognition, differences between BMIz trajectories for women with and without a history of depression lessened as women approached their 30s. However, statistical power at later ages was not as great as at younger ages.

In males, the association of anxiety disorders with BMIz trajectory was small and not statistically significant. Thus, our results suggest that anxiety disorders do not greatly influence weight status in males. Childhood depression for males was associated with lower weight; the magnitude of the difference was inversely related to age. Our results for male depression were of borderline statistical significance. Our model predicted that a 14-year-old male with depression would have a BMIz 0.36 units lower than a similar male without depression; however, by the age of 30 years, the BMIz difference between these men would be reduced to 0.12 units, and by the age of 35 years, to less than 0.05 units. In adult males, a 0.12-unit difference in BMIz translates to a difference in BMI of approximately 25 to 24.5 or 29 to 28; for a man with a height of 69 inches (176 cm) (the average height of US men aged 20-40 years),45 these BMI differences represent weight differences of 1.5 and 2.2 kg, respectively.

The literature describing associations between obesity and psychological disorders is replete with cross-sectional and clinical studies,46,47 but contains fewer prospective studies.2734 Several of these studies27,30,31 use symptom scales that are not directly comparable to DSM criteria, and others2732 are limited to observations at 2 time points. Our analyses of the association of anxiety and depression with weight are unique in using structured diagnostic interviews to assess DSM disorders in a community-based cohort studied at 4 occasions spanning childhood and early adulthood.

Our results are broadly consistent with other prospective studies28,30 in finding that psychological distress, especially when present in childhood, predicts higher weight. The study33 most methodologically comparable to ours found that adolescent depression increased risk for later obesity in girls, but not boys. However, anxiety disorders were not studied, and the study outcome was obesity at the age of 26 years. In an earlier analysis29 that used data from the 1983 and 1991 to 1994 waves of the Children in the Community Study, young adulthood BMI was associated with symptoms of conduct disorder and depression.

The relationship of depression to BMIz we observed in males was fundamentally different from that observed in females. Our results are consistent with those of other investigators9,10,14,16,17,33 in reporting null or inverse associations between depressive symptoms and weight in males. Differences in results in these mainly cross-sectional studies could be due to influences of age on the association of depression with weight.

Longitudinal studies of anxiety and weight are rare. Mustillo et al34 observed no association between children's weight trajectory and DSM anxiety disorders; however, comparability to our results is limited by differences in objective and approach. Cross-sectional studies of anxiety and weight provide inconsistent results; in 2 studies,9,18 anxiety symptoms were inversely associated with weight. Like ours, more recent studies13,16 observed positive associations between obesity and anxiety, particularly in women.

Depression and anxiety are often comorbid disorders.48,49 An estimated 60% of adults with depression have had an anxiety disorder,49 and up to 40% of adults with anxiety disorders have had depression.50 In our study, 66.0% of males and 76.5% of females with depression also met the diagnostic criteria for an anxiety disorder. Attempting to separate the influence of depression on weight from the influence of anxiety on weight is counterproductive because “pure” depression is not common. In contrast, only 27.7% of males and 39.3% of females with an anxiety disorder met the diagnostic criteria for depression; that results of our female anxiety model were strengthened in this subset counters the potential argument that our observed association between anxiety and BMIz was due primarily to depression.

By early adulthood, females with depression outnumber males by approximately 2:1.51 Atypical depression, characterized by increased appetite, hypersomnia, and decreased activity level, is more common in women than men,52 suggesting a potential mechanism for the association we observed.

Our analyses provide evidence that anxiety disorders and depression are associated with higher weight in females. A strength of our study is assessment of depression and anxiety disorders consistent with DSM criteria. In contrast to symptom scales, use of DSM diagnostic criteria facilitates interpretability and generalizability of results.

There were several limitations of our analyses. First, we modeled the association of first recognition of anxiety or depression with BMIz, but it is unlikely that we captured true disorder onset, and some individuals with depression or anxiety were likely not identified (eg, if a disorder began after wave 3 and remitted well before wave 4). Also, because we did not continuously assess weight and do not know the true age of disorder onset, we cannot decisively establish whether weight change preceded or followed anxiety or depression onset. Second, height and weight were self-reported; self-reported height and weight have been shown to be accurate in adults and older teenagers.53,54 In younger adolescents (aged 12-16 years), accuracy is related to age, with values for younger youth more likely to be inaccurate.55 Third, it is possible that missing data biased our results. However, we assessed 94.6% of participants at waves 1, 2, and 3 and 80.6% of participants at wave 4; participants missing data at wave 4 did not differ substantively in age or BMIz at waves 1, 2, or 3. Finally, although we controlled for socioeconomic status, smoking, and medication use, it is possible that our results are biased by residual confounding due to imperfect measurement of these variables or the lack of measurement of other unknown confounders.

In conclusion, our analysis of a community-based cohort studied from childhood until adulthood provides evidence that, in females, anxiety disorders and depression are associated with higher weight. The potential for sex and age to influence whether an association between weight and anxiety disorders or depression is seen cross-sectionally or in short-term studies underscores the necessity of applying a life-course approach. Our results suggest that efforts to improve mental health in populations may also help prevent female obesity; consideration of the potential for psychological antecedents and correlates of obesity could improve prevention and treatment.

Back to top
Article Information

Correspondence: Sarah E. Anderson, MS, Tufts University, 136 Harrison Ave, Boston, MA 02111 (

Accepted for Publication: October 20, 2005.

Author Contributions: Ms Anderson and Dr Must had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Funding/Support: This study was supported by grants T32 DK62032-11 and R21 DK64254 from the National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, Md; grant HD-40685 from the National Institute of Child Health and Human Development, Bethesda; and grants MH-36971, MH-38916, and MH-49191 from the National Institute of Mental Health, Rockville, Md.

Role of the Sponsor: The funding bodies had no role in data extraction and analyses, in the writing of the manuscript, or in the decision to submit the manuscript for publication.

Acknowledgment: We thank Paul Jacques, PhD, for his helpful comments on an earlier draft of the manuscript.

Hedley  AAOgden  CLJohnson  CLCarroll  MDCurtin  LRFlegal  KM Prevalence of overweight and obesity among US children, adolescents, and adults, 1999-2002 JAMA 2004;2912847- 2850
Guo  SWu  WChumlea  WCRoche  AF Predicting overweight and obesity in adulthood from body mass index values in childhood and adolescence Am J Clin Nutr 2002;76653- 658
Johnson  JCohen  PKasen  SBrook  J Childhood adversities associated with risk for eating disorders or weight problems during adolescence or early adulthood Am J Psychiatry 2002;159394- 400
Mellbin  TVuille  JC Rapidly developing overweight in school children as an indicator of psychosocial stress Acta Paediatr Scand 1989;78568- 575
Williamson  DFThompson  TJAnda  RFDietz  WHFelitti  VJ Body weight and obesity in adults and self-reported abuse in childhood Int J Obes Relat Metab Disord 2002;261075- 1082
Lissau  ISorensen  TI Parental neglect during childhood and increased risk of obesity in young adulthood Lancet 1994;343324- 327
Strauss  RSKnight  J Influence of the home environment on the development of obesity in children Pediatrics 1999;103e85Available at August 12, 2004
Moore  MEStunkard  ASrole  L Obesity, social class, and mental illness JAMA 1962;181962- 966Article
Crisp  AHMcGuiness  B Jolly fat: relation between obesity and psychoneurosis in general population BMJ 1976;17- 9
Istvan  JZavela  KWeidner  G Body weight and psychological distress in NHANES I Int J Obes Relat Metab Disord 1992;16999- 1003
Carpenter  KMHasin  DSAllison  DBFaith  MS Relationships between obesity and DSM-IV major depressive disorder, suicide ideation, and suicide attempts: results from a general population study Am J Public Health 2000;90251- 257
Onyike  CUCrum  RMLee  HBLyketsos  CGEaton  WW Is obesity associated with major depression? results from the Third National Health and Nutrition Examination Survey Am J Epidemiol 2003;1581139- 1147
Becker  ESMargraf  JTurke  VSoeder  UNeumer  S Obesity and mental illness in a representative sample of young women Int J Obes Relat Metab Disord 2001;25(suppl 1)S5- S9
Kawada  TSuzuki  S Obesity has an inverse relationship with a depressive state J Occup Health 2001;43371- 374Article
Dong  CSanchez  LEPrice  RA Relationship of obesity to depression: a family-based study Int J Obes Relat Metab Disord 2004;28790- 795
Jorm  AFKorten  AEChristensen  HJacomb  PARodgers  BParslow  RA Association of obesity with anxiety, depression and emotional well-being: a community survey Aust N Z J Public Health 2003;27434- 440
Palinkas  LAWingard  DLBarrett-Connor  E Depressive symptoms in overweight and obese older adults: a test of the “jolly fat” hypothesis J Psychosom Res 1996;4059- 66
Stewart  ABrook  R Effects of being overweight Am J Public Health 1983;73171- 178
Sallade  J A comparison of the psychological adjustment of obese vs non-obese children J Psychosom Res 1973;1789- 96
Hallstrom  TNoppa  H Obesity in women in relation to mental illness, social factors and personality traits J Psychosom Res 1981;2575- 82
Renman  CEngstrom  ISilfverdal  SAAman  J Mental health and psychosocial characteristics in adolescent obesity: a population-based case-control study Acta Paediatr 1999;88998- 1003
Hayes  DRoss  CE Body and mind: the effect of exercise, overweight, and physical health on psychological well-being J Health Soc Behav 1986;27387- 400
Han  TSTijhuis  MARLean  MEJSeidell  JC Quality of life in relation to overweight and body fat distribution Am J Public Health 1998;881814- 1820
Lamertz  CMJacobi  CYassouridis  AArnold  KHenkel  AW Are obese adolescents and young adults at higher risk for mental disorders? a community survey Obes Res 2002;101152- 1160
Erickson  SJRobinson  TNHaydel  KFKillen  JD Are overweight children unhappy? body mass index, depressive symptoms, and overweight concerns in elementary school children Arch Pediatr Adolesc Med 2000;154931- 935
Braet  CMervielde  IVandereycken  W Psychological aspects of childhood obesity: a controlled study in a clinical and nonclinical sample J Pediatr Psychol 1997;2259- 71
Lumeng  JCGannon  KCabral  HJFrank  DAZucherman  B Association between clinically meaningful behavior problems and overweight in children Pediatrics 2003;1121138- 1145
Pine  DSGoldstein  RBWolk  SWeissman  MM The association between childhood depression and adulthood body mass index Pediatrics 2001;1071049- 1056
Pine  DSCohen  PBrook  JCoplan  JD Psychiatric symptoms in adolescence as predictors of obesity in early adulthood: a longitudinal study Am J Public Health 1997;871303- 1310
Goodman  EWhitaker  RC A prospective study of the role of depression in the development and persistence of adolescent obesity Pediatrics 2002;110497- 504
DiPietro  LAnda  RFWilliamson  DFStunkard  AJ Depressive symptoms and weight change in a national cohort of adults Int J Obes Relat Metab Disord 1992;16745- 753
Barefoot  JCHeitmann  BLHelms  MJWilliams  RBSurwit  RSSiegler  IC Symptoms of depression and changes in body weight from adolescence to mid-life Int J Obes Relat Metab Disord 1998;22688- 694
Richardson  LPDavis  RPoulton  R  et al.  A longitudinal evaluation of adolescent depression and adult obesity Arch Pediatr Adolesc Med 2003;157739- 745
Mustillo  SWorthman  CErkanli  AKeeler  GAngold  ACostello  EJ Obesity and psychiatric disorder: developmental trajectories Pediatrics 2003;111851- 859
Cohen  PCohen  JKasen  S  et al.  An epidemiological study of disorders in late childhood and adolescence–I: age- and gender-specific prevalence J Child Psychol Psychiatry 1993;34851- 867
Cohen  PCohen  J Life Values and Adolescent Mental Health  Mahwah, NJ Lawrence A Erlbaum Associates1996;
Costello  EJEdelbrock  CSDuncan  MKKalas  R Testing of the NIMH Diagnostic Interview Schedule for Children (DISC) in a Clinical Population: Final Report to the Center for Epidemiological Studies, NIHM.  Pittsburgh, Pa University of Pittsburgh1984;
Costello  EJEdelbrock  CSCostello  AJ Validity of the NIMH Diagnostic Interview Schedule for Children: a comparison between psychiatric and pediatric referrals J Abnorm Child Psychol 1985;13579- 595
American Academy of Pediatrics' Committee on Nutrition, Prevention of pediatric overweight and obesity Pediatrics 2003;112424- 430
Barlow  SEDietz  WH Obesity evaluation and treatment: Expert Committee recommendations: the Maternal and Child Health Bureau, Health Resources and Services Administration and the Department of Health and Human Services Pediatrics 1998;102e29.Available at July 22, 2004
Ogden  CLKuczmarski  RLFlegal  KM  et al.  Centers for Disease Control and Prevention 2000 growth charts for the United States: improvements to the 1977 National Center for Health Statistics version Pediatrics 2002;10945- 60
Wright  CMParker  LLamont  DCraft  AW Implications of childhood obesity for adult health: findings from thousand families cohort study BMJ 2001;3231280- 1284
Singer  JDWillett  JBSinger  JDedWillett  JBed Modeling discontinuous and nonlinear change Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence New York, NY Oxford University Press Inc2003;189- 242
Singer  JDWillett  JBSinger  JDedWillett  JBed Doing data analysis with the multilevel model for change Applied Longitudinal Data Analysis: Modeling Change and Event Occurrence New York, NY Oxford University Press Inc2003;113- 116
McDowell  MAFryar  CDHirsch  ROgden  CL Anthropometric Reference Data for Children and Adults: US Population, 1999-2002  Atlanta, Ga Centers for Disease Control and Prevention, US Dept of Health and Human Services2005;Advance Data From Vital and Health Statistics, No. 361
Friedman  MABrownell  KD Psychological correlates of obesity: moving to the next research generation Psychol Bull 1995;1173- 20
Faith  MSMatz  PEJorge  MA Obesity-depression associations in the population J Psychosom Res 2002;53935- 942
Kessler  RCChiu  WTDemler  OWalters  EE Prevalence, severity, and comorbidity of 12-month DSM-IV disorders in the National Comorbidity Survey Replication Arch Gen Psychiatry 2005;62617- 627
Kessler  RCNelson  CBMcGonagle  KALiu  JSwartz  MBlazer  DG Comorbidity of DSM-III-R major depressive disorder in the general population: results from the US National Comorbidity Survey Br J Psychiatry Suppl 1996; ((30)) 17- 30
Overbeek  TGriez  EJLGriez  EJLedFaravelli  CedNutt  DJedZohar  Jed At the interface of depression and anxiety: comorbidity and concepts Mood Disorders: Clinical Management and Research Issues Hoboken, NJ John Wiley & Sons2005;503- 526
Saluja  GIachan  RScheidt  POverpeck  MDSun  WGiedd  J Prevalence of and risk factors for depressive symptoms among young adolescents Arch Pediatr Adolesc Med 2004;158760- 765
American Psychiatric Association, Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition  Washington, DC American Psychiatric Association1994;
Stewart  AL The reliability and validity of self-reported weight and height J Chronic Dis 1982;35295- 309
Brener  NDMcManus  TGaluska  DALowry  RWechsler  H Reliability and validity of self-reported height and weight among high school students J Adolesc Health 2003;32281- 287
Strauss  RS Comparison of measured and self-reported weight and height in a cross-sectional sample of young adolescents Int J Obes Relat Metab Disord 1999;23904- 908