Overweight during childhood and internalizing symptoms in early adolescence: The mediating role of peer victimization and the desire to be thinner

Journal of Affective Disorders 202 (2016) 203–209

Contents lists available at ScienceDirect

Journal of Affective Disorders journal homepage: www.elsevier.com/locate/jad

Research paper

Overweight during childhood and internalizing symptoms in early adolescence: The mediating role of peer victimization and the desire to be thinner Laura Pryor a,b,c,n, Mara Brendgen b,c,d, Michel Boivin b,e,f, Lise Dubois g, Christa Japel b,h, Bruno Falissard i, Richard E. Tremblay j,k, Sylvana M. Côté a,b,c,f,l a

Department of Social and Preventive Medicine, University of Montreal, Canada Research Unit on Children's Psychosocial Maladjustment, University of Montreal and Laval University, Canada Ste Justine Hospital Research Center, Montreal, Canada d Department of Psychology, University of Quebec in Montreal, Montreal, Canada e School of Psychology, Laval University, Quebec City, Canada f Institute of Genetic, Neurobiological, and Social Foundations of Child Development, Tomsk State University, Russian Federation g School of Epidemiology, Public Health and Preventive Medicine, University of Ottawa, Canada h Department of Special Education and Training, University of Quebec in Montreal, Canada i CESP, Paris-Sud University, UVSQ, INSERM, Paris-Saclay University, Villejuif, France j School of Public Health, Physiotherapy and Sports Science, University College Dublin, Ireland k Departments of Pediatrics and Psychology, University of Montreal, Canada l INSERM U1219, Bordeaux University, Bordeaux, France b c

art ic l e i nf o

a b s t r a c t

Article history: Received 9 December 2015 Received in revised form 20 April 2016 Accepted 13 May 2016 Available online 21 May 2016

Background: Overweight is associated with depression and anxiety among adults. It is unclear whether this association begins in childhood. Overweight among children is associated with a higher risk of peer victimization, and may mediate an association between overweight and internalizing symptoms. No study has tested this hypothesis in a longitudinal population-based sample using developmental trajectories of overweight in middle childhood. Methods: Data was drawn from the population-based Quebec Longitudinal Study of Child Development. A three-group trajectory model of overweight development (6–12 years) was previously identified using a semi-parametric group-based approach (n¼ 1678): “early-onset” (11.0%), “late-onset” (16.6%) and “never overweight” (72.5%). Mediation models tested the link between overweight status and child-reported depression and anxiety at 13 years via peer victimization and body dissatisfaction. Results: Children on an early-onset overweight trajectory were at increased risk for both depression (B ¼.318, 95% CI¼ .141;.496) and anxiety (B ¼ .262, 95% CI ¼.09;.44) at 13 years. These direct associations were mediated by peer victimization and subsequent desire to be thinner. Children on a late-onset childhood overweight trajectory were at increased risk for both depression (B¼ .332, 95% CI ¼.187;.477) and anxiety (B ¼.215; 95% CI ¼.072;.358) at 13 years, mediated by the desire to be thinner. Limitations: We were unable to control for previous levels of body dissatisfaction. Our measure of peer victimization was not specific to weight-based teasing. Conclusions: Overweight during middle childhood increases risk of early adolescence internalizing symptoms. Peer victimization and body dissatisfaction are partly responsible for this link. & 2016 Elsevier B.V. All rights reserved.

Keywords: Depression Anxiety Child overweight Peer victimization Body dissatisfaction Developmental trajectories

1. Introduction Childhood overweight and obesity1 has become a national n Correspondence to: Social Epidemiology Research Group (ERES), Pierre Louis Institute for Epidemiology and Public Health (iPLESP), INSERM UMR_S 1136 & Sorbonne Universities, Pierre et Marie Curie, 184 rue du Faubourg Saint-Antoine, Paris, France. E-mail address: [email protected] (L. Pryor). 1 Hereafter, the term "overweight" will designate overweight and obese categorizations.

http://dx.doi.org/10.1016/j.jad.2016.05.022 0165-0327/& 2016 Elsevier B.V. All rights reserved.

public health priority for several countries worldwide, and research and intervention efforts have risen sharply in an attempt to counteract what is now documented as an extremely complex phenomenon. A large proportion of work has been aimed at the physical health problems associated with child overweight and we still lack information on its mental health impact. While several studies have shown that overweight adolescents and adults are at higher risk for depression and anxiety (Atlantis and Baker, 2008; Goldfield et al., 2010; Luppino et al., 2010; Mustillo et al., 2003; ter

204

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

Bogt et al., 2006), studies examining this question in childhood have thus far elicited modest or mixed results, particularly among community samples (Bradley et al., 2008; Datar and Sturm, 2004; Drukker et al., 2009; Harriger and Thompson, 2012; Luppino et al., 2010; Wardle and Cooke, 2005). One potential reason behind the inconsistent findings is that studies have not often been able to evaluate the longitudinal development of weight status across childhood. There exist differential trajectories of weight status development during childhood and it is plausible that only certain overweight experiences lead to internalizing symptoms (e.g: depression and anxiety). For example, a child who becomes overweight early and persistently may be at a greater risk for negative psychosocial outcomes than a child who experiences weight problems more transiently (Mustillo et al., 2003). Alternately, the inconclusive or modest associations identified in the literature thus far could be indicative of the existence of indirect (or mediating) pathways. That is, overweight and internalizing symptoms may be associated via a third factor, such as peer victimization. Indeed, weight-based peer victimization has been identified as one of the most prominent forms of bullying in the school setting (Bradshaw et al., 2011; Puhl et al., 2011). It is documented as a source of major distress in the lives of those who experience it (Cramer and Steinwert, 1998; MacLean et al., 2009; Puhl and Latner, 2007), with negative scars that may last into adulthood (Puhl and Latner, 2007). However, given a lack of longitudinal studies, it is still unclear whether peer victimization mediates the predictive association between childhood overweight development and subsequent internalizing symptoms among youth in community samples. Past research informs us that it is likely not a child's weight status per se, but rather the extent to which they are dissatisfied with their weight status, that is linked to negative psychological outcomes (Mond et al., 2011). Children as young as 9 years of age with a history of being at-risk for overweight have been identified as showing signs of body dissatisfaction (Shunk and Birch, 2004). Overweight youth may exhibit such dissatisfaction because at some level the outside world has informed them their body shape or size does not fit the ideal. They may internalize societal norms and ideals of beauty, attractiveness, and health directly; or indirectly, such as when faced by teasing, bullying or other forms of rejection by peers (Bearman et al., 2006). Our aim for the present study was to test the predictive association between various weight status trajectories during middle childhood (assessed objectively between 6 and 12 years) and internalizing symptoms in early adolescence (13 years), and whether these associations occurred via peer victimization and/or the child's desire to be thinner. We relied on a large population-based sample (n ¼1221) assessed 11 times over the first 13 years of the child's life.

2. Methods 2.1. Participants Participants were from the Québec Longitudinal Study of Child Development (QLSCD; 1998–2011). A random population sample of 2940 families with a 5-month-old singleton infant in 1998 was identified through the provincial master list of birth registries. Participants were selected if they spoke English or French (official languages of Canada). The sample was reduced to 2120 due to non-response, inability to contact, or not meeting study criteria. Trained research assistants conducted home visits with the child and primary caregiver (mostly mothers) every year until age 8 and at 10, 12 and 13 years. The present sample includes 1221 children for whom height and weight data were available and who self-

Table 1. Sociodemographic Characteristics of Sample (n¼ 1221).

Sex of child (female) Mother born outside Canada Mother had a lifetime occurrence of depression Mother did not obtain high school diploma Parents separated between 6 and 12 years Family Income (at age 6) o $30,000 $30,000–60,000 460,000 Family had insufficient income (at least once) from 6 to 12 years Mother's age when child was 5 months (in years)

%

n

54 7.1 21.6 16.2 32.3

661 87 261 198 393

12.6 32.9 54.4 26.5 Mean 29.5

139 364 601 323 SD 5.2

reported their internalizing symptoms at 13 years. Within this dataset, simple imputation based on Estimation Maximization (EM) was used to complete occasional missing data (1.3% of all values). Table 1 presents sample characteristics. We compared the socio-demographic characteristics of the 1221 families included in our sample analyses to that of the 899 families excluded from the present analyses. No significant difference was found for the proportion of non-intact families or maternal depression occurrences. Significant differences were found for maternal education (79.2% of excluded vs. 83.7% of included mothers obtained a high school diploma), family income o30 000$ at age 5 years (17.5% of excluded vs. 12.6% of included families earned o30 000$/year) and maternal immigrant status (15.8% of excluded vs. 7.1% of included were born outside Canada). At each data collection, informed written consent was obtained from the child's primary caregiver. The study was approved by the Health Research Ethics Committees of the Québec Statistics Institute and the University of Montreal. 2.2. Measures 2.2.1. Independent variable The independent variable (IV) was a categorical variable representing the longitudinal developmental of overweight status from 6 to 12 years. This developmental trajectory model was previously identified (Pryor et al., 2015) using a semi-parametric modelling method in the SAS Proc Traj program (Nagin and Tremblay, 1999; Nagin and Odgers, 2010) according to the following method: Trained interviewers obtained measures of height and weight as part of a standardized protocol when children were 6, 7, 8, 10 and 12 years (Lavallée, 2004). Body Mass Index (BMI) was calculated as the child's weight in kilograms divided height in meters squared (BMI¼ kg/m2). Children's weight status was classified as overweight (¼1) vs. not overweight ( ¼0) according to International Obesity Task Force BMI cut-off criteria which account for sex and age in months (Cole et al., 2000). This binary measure was used to model group-based developmental trajectories of weight status. In cases where a binary measure is modeled (in our case, weight status) using the SAS Proc Traj procedure, the technique yields a probability of being overweight at each age, for each group. Children with at least one valid data point from 6 to 12 years were included in the analysis (n¼ 1678). Of these 1678 children, 48.9% had data at all 5 time points, 23.4% had data for four time points, 12% had data for three time points, 7.3% for two time points and 8.2% for one time point. Model selection was guided by a maximized Bayesian Information Criterion (BIC) (Nagin and Tremblay, 1999) and by considering groups that were large enough to maintain power in

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

205

children who responded to at least seven of the ten items (alpha ¼0.8432). Anxiety was measured using a four-item scale from the Child Behavior Questionnaire (Tremblay et al., 1991). Children rated whether they never (0), sometimes (1), or often (2) exhibited the following symptoms over the last 6 months: Were very fearful or anxious; Worried; Cried a lot; Were nervous, high strung or tense (alpha ¼0.7661). These items correspond to items from the Stait Trait Anxiety Inventory (Spielberger, 1973; Spielberger et al., 1970). 2.6. Confounding variables We selected control variables that could theoretically provide alternative explanations of our mediation model. These have been grouped into whether they represent family adversity variables or child mental and physical health characteristics. Fig. 1. Group-based trajectories of the probability of overweight from 6 to 12 yrs (n ¼1678).

subsequent analyses and substantively meaningful in terms of identifiable qualitative group differences. The model identified as having the best fit was composed of three groups: “never overweight” (72.5% of children), “late-onset overweight” (16.6%), and “early-onset overweight” (11.0%) (Fig. 1). The early-onset group included children who had high probabilities of overweight throughout the assessment period, whereas the probability for late-onset overweight began to rise at approximately age 7–8 years. 2.3. Potential mediating variables 2.3.1. Peer victimization At 6, 7, 8, 10, and 12 years, teachers completed questionnaires regarding the target child. The following questions were asked about peer victimization experiences over the past 6 months: How often would you say the target child was (1) made fun of by other children, (2) hit or pushed by other children, (3) called names by other children? (Boivin et al., 2010; Perry et al., 1988). Items were rated on a 3-point scale ranging from 0¼never to 2¼often (mean Cronbach's alpha: 0.66). Within each year, individual item scores were averaged to obtain a global rating for the child's level of peer victimization. Scores were standardized to a 10-point scale and the mean of the 5 time points was calculated. 2.3.2. Desire to be thinner (body dissatisfaction) At age 12 years, children were presented with the children's pictorial body image scale (Collins, 1991; Daraganova, 2014), a series of seven sex-specific, rank-ordered images depicting body types ranging in size/shape (thinnest to largest, from left to right) and were asked to indicate a) the picture that looks most like you (what you look like now)? and b) the picture you want to look like the most? The difference between these two responses was calculated, and the desire to be thinner was coded as an ordinal variable ranging from 0 (¼ does not want to be thinner) to 4 (¼wants to be much thinner). 2.4. Moderating variable Child sex.

2.6.1. Family adversity Family insufficient income was measured using Statistics Canada's low-income cutoffs, which consider family income in the past year, number of individuals in the household, and family location of residence (6, 7, 8, 10 and 12 years) (Giles, 2004; Seguin et al., 2007). The total number of years the family experienced insufficient income from age 6–12 was calculated and included in the model as a continuous variable. Maternal education indicated whether the mother obtained (¼0) or not (¼1) a high school diploma. Maternal lifetime depression categorized mothers as having experienced a major depressive episode (¼ 1) or not ( ¼0) (Cote et al., 2009). Family functioning (6, 7, 8, 10, 12 years) was assessed with the 12-item General Functioning Subscale of the McMaster Family Assessment Device (Byles et al., 1988; Epstein et al., 1983) and standardized to a 10-point scale, with higher values on the scale representing less functional households (mean Chronbach's alpha: 0.82). Within each year, individual item scores were averaged to a combined 10-point scale and the mean score across the five years was calculated. Families where parents separated at least once when children were aged between 6 and 12 years ( ¼1) were differentiated from others (¼ 0). 2.6.2. Child mental and physical health The following mental health symptom scores were drawn from the Child Behavior Questionnaire (Tremblay et al., 1991). Respondents were mothers during early childhood (age 1.5, 2.5, 3.5, 4.5, 5 years) and teachers during middle childhood (age 6, 7, 8, 10, 12 years). Within each year, item scores were summed and standardized to a 10-point scale (Cote et al., 2009). A mean score was calculated for each construct: depression and anxiety symptoms (DAS) in middle childhood (Cote et al., 2009) (6–12 years) (mean Cronbach's alpha: 0.73), Physical aggression in middle childhood (Pingault et al., 2013) (mean Cronbach’s alpha: 0.79), Early peer victimization (Barker et al., 2008) at 3.5, 4.5, 5 years (mean Cronbach's alpha: 0.52). Birthweight in grams was obtained from medical records. Pubertal stage at age 12 was assessed with Tanner Pubertal Scale (Tanner, 1962) images presented to the child. Girls circled the image best corresponding to their stage of breast development whereas boys circled the image best corresponding to their development of external genitalia. Stages 1–5 were recoded as 0 through 4 and treated as an ordinal variable.

2.5. Outcome variables: child internalizing symptoms (depression and anxiety) at 13 years

2.7. Analyses

Depression was measured using the short version of the Kovacs Child Depression Inventory (Kovacs, 1985). For each of 10 items, children selected from 3 statements describing their feelings over the past 2 weeks (coded as 0, 1, 2). CDI scores were calculated for

Analyses were performed using IBM SPSS Statistics 20 and the PROCESS macro (Hayes, 2013). First, bivariate correlations were analyzed between the variables of interest and potential confounders. Potential confounders that were significantly related

206

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

(p ¼0.1 level) to any two of the main variables of interest (overweight, peer victimization, desire to be thinner and depression or anxiety) were retained for multivariate analysis. The following interactions were tested in mediation models: child sex*overweight status, child sex* peer victimization and child sex* desire to be thinner. These were not significant; therefore sex was modeled as a control variable in further analyses. We then tested four serial mediation models. Indirect paths were tested from either the early or late-onset overweight trajectory (IVs) to peer victimization (Mediator 1) to desire to be thinner (Mediator 2) to internalizing symptoms (depression or anxiety: DVs), as well as from each IV to each DV through each mediator individually. All continuous variables were z-standardized to obtain more readily interpretable path coefficients. Mediation analyses involved estimating a series of Ordinary Least Squares (OLS) regression equations for each path and indirect effects were estimated using bootstrapping, with 10,000 bootstrapped samples generated (Hayes, 2013).

Table 2. Direct effects of each of childhood overweight status on depression and anxiety at age 13, and indirect effects via peer victimization and the desire to be thinner. Depression outcome 1. Early-onset OW to DEP Direct effects

B

SE

CI

B

SE

CI

Total effect model Cb (R2 ¼ .08, F¼ 11.26, po .001) Direct effect model C’c (R2 ¼ .11, F¼ 13.11, po .001)

.32

.09

.14;.50

.33

.07

.19;.48

.07

.10

 .12;. 26

.16

.08

.003;.31

.03 .21 .003

.05 .05 .003

.01;.35 .13;.32 .001;. 01

.004 .17 .0004

.01 .04 .0008

 .01;.02 .10;.26  .001;. 003

Indirect effects OW-VIC-DEP OW-DT-DEP OW-VIC-DT-DEP

3. Results

Anxiety Outcome

We report the results in 4 sections: 1) Variables selected as confounders; 2) Direct associations between IVs and DVs, without mediators; 3) Results from full mediation models; 4) Results of a supplementary analysis examining the victimization patterns of the overweight trajectories.

Direct effects models

1. Variables Selected as Confounders The following variables were selected as control variables due to their significant association with at least two main variables: Child Sex, Early Peer Victimization (3-5 yrs), Internalizing Symptoms (6-12 years), Family Functioning, and Pubertal Stage (12 years). 2. Direct Associations between Overweight and Outcomes Table 2 presents the direct associations between each overweight status trajectory and depression and anxiety, respectively, both without (C) and with (C’) consideration for the effect of the putative mediators. Parameter estimates and model summary statistics are listed. For each of the four C paths, confidence intervals (CI) around the regression coefficients (B) do not include zero, indicating significant direct associations between each overweight trajectory and each internalizing outcome when mediators were not taken into account in the models. 3. Mediation Models Table 2 also lists the regression coefficients (B) and confidence intervals (CI) for the C’ paths. As compared to the C paths, the direct associations are reduced and, in all models but one (lateonset overweight to depression), no longer statistically significant when the mediators are included in the model. Figs. 2 and 3 present the regressions coefficients (B) and significance levels for each path in each of the four putative mediation models. Depression and anxiety outcomes were merged into one figure. The lower half of Table 2 presents significant indirect associations (i.e. 95% confidence intervals that do not include zero). Results indicate a child who became overweight early on was more likely to be depressed and anxious at 13 years via the experience of peer victimization during middle childhood, followed by the increased desire to be thinner at 12 years. He or she was also more likely to experience depression and anxiety at 13 years via either the experience of peer victimization or the desire to be thinner individually. A child who became overweight later during middle childhood also had an increased risk of depression and anxiety at 13 years, but these paths were mediated by an increased desire to be

2. Late-onset OW to DEP

a

Total effect model Cb (R2 ¼ .11, F¼ 16.14, po .001) Direct effect model C′c (R2 ¼ .12, F¼ 14.39, po .001) Indirect effects OW-VIC-ANX OW-DT-ANX OW-VIC-DT-ANX

3. Early-onset OW to ANX B SE CI

4. Late-onset OW to ANX B SE CI

.26

.09

.09;.44

.22

.07

.07;.36

.13

.10

 .06;.33

.13

.08

 .03;.28

.02 .11 .002

.012 .042 .001

.001;.05 .03;.20 .0001; 0.001

.003 0.08 .0002

.01 0.03 .0005

 .01;.02 .02;.16  .0003;.002

Abbreviations: OW, overweight status; DEP, depression; ANX, anxiety; VIC, peer victimization; DT, desire to be thinner; SE, standard error; CI, confidence interval. *po .05, **p o .01, ***p o.001 a Regression coefficient. All continuous variables were standardized to facilitate interpretation of effects. b Mediators not included in this model. c Mediators taken into account in this model.

Fig. 2. Mediation model: childhood peer victimization and the desire to be thinner as mediators between early-onset childhood overweight status and depression (or anxiety) at 13 years. *po 0.5,** p o .01,***p o .001,.05 o bpo 0.07.

thinner only. Mediation Model: Childhood Peer Victimization and the Desire to be Thinner as Mediators Between Early-Onset Childhood Overweight Status and Depression (or Anxiety) at 13 years. Mediation Model: Childhood Peer Victimization and Desire to

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

Fig. 3. Mediation model: childhood peer victimization and desire to be thinner as mediators between late-onset childhood overweight status and depression (or anxiety) at 13 years. *p o0.5,** po .01,***po .001,.05 o bp o 0.07.

be Thinner as Mediators between Late-Onset Childhood Overweight Status and Depression (or Anxiety) at 13 years. 4. Supplementary Analysis To examine whether the three overweight trajectory groups differed in terms of the developmental course of peer victimization across middle childhood, a repeated-measures ANOVA was conducted with Overweight Trajectory Group (three levels) as the Between-Subjects factor, Time (five levels) as the Within-Subjects factor and Peer Victimization as the dependent variable. Occasional missing data (29.87% of data points) were included through the use of imputations (EM method). The means of peer victimization across time (ages 6–12) by overweight status are presented in Fig. 4. As can be seen, the pattern of change of victimization closely follows that of weight status. ANOVA results indicated a significant interaction between victimization across time and the overweight trajectories. Those who were never overweight maintained generally low levels of victimization. Those with late-onset overweight showed a significant linear increasing trend in their victimization levels. Lastly, those with early-onset overweight showed a quadratic trend (victimization was low at age 6, increased from ages 7 through 10, then decreased slightly at age 12).

4. Discussion Controlling for several potential confounders, we found children characterized by an early-onset overweight trajectory between 6 and 12 years were at increased risk for both anxiety and depression at 13 years. Increased levels of peer victimization and the desire to be thinner mediated these associations. Children on a late-onset childhood overweight trajectory were also at increased risk for both anxiety and depression at 13 years. These associations were mediated by the desire to be thinner. These results, which vary depending on the child's weight status history, suggest two types of mechanisms linking 2 1.8

Victimization (mean)

1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 6

7

8

10

12

Age (yr) Never overweight

Late-onset

Early-onset

Fig. 4. Mean level of peer victimization in each trajectory at each time.

207

overweight with later internalizing problems. In the case of earlyonset overweight children, chronic ridicule and harassment from peers may eventually trigger the desire to be thinner (Harriger and Thompson, 2012; McCormack et al., 2011) which in turn may lead to internalizing problems (Mond et al., 2011). In the case of lateonset overweight children; in the absence of long-term peer victimization, children may incur negative views about their weight increase over time which in turn increases the likelihood for both depression and anxiety. The literature highlights certain sociocultural factors, such as social comparison against one's peers or media role models (Jeffers et al., 2013; Luppino et al., 2010; Mond et al., 2011), as a potential explanation for the latter pathway. We did not find sex differences in the above-mentioned paths. Past results have been mixed (Erickson et al., 2000; Mond et al., 2011; Mustillo et al., 2003), but it is likely the age group examined here (mostly before puberty onset) is too young to notice any significant differences for internalizing outcomes (Herpertz-Dahlmann et al., 2013). The relatively small effect sizes and the fact that the direct association between late-onset overweight and depression remained significant even when the desire to be thinner was accounted for, suggests the existence of other mediating mechanisms. Previous evidence points to plausible alternate behavioral, biological (Lopresti and Drummond, 2013; Luppino et al., 2010), socio-cultural and psychological mechanisms. For example, being overweight and/or victimized due to one's weight may dissuade children from engaging in physical activities, which in turn may promote depressive symptoms (Reeves et al., 2008). Other potential mediating mechanisms we did not examine include the child's felt pressure to be thin (Chaiton et al., 2009) and eating disorder symptoms including dieting and food restriction (Goldschmidt et al., 2008; Shunk and Birch, 2004). Underlying biological mechanisms could also involve the hypothalamo-pituitary-adrenal (HPA) axis. The experience of being victimized by one's peers is documented as being extremely stressful, and this chronic stress may disrupt HPA axis functioning, which also plays a role in the onset of depression (Hamilton et al., 2008). Further studies examining how these and other putative mechanisms act together given differing developmental histories of overweight and obesity among children would be needed in order to clarify the nature of the associations identified in this paper. 4.1. Strengths and limitations To our knowledge, our study is the first to examine whether differing trajectories of weight status during childhood lead to internalizing symptoms in adolescence via peer victimization and body dissatisfaction (measured as the desire to be thinner). Specific strengths are the use of measured height and weight values, as well as multiple informants, in a 13-year longitudinal study. Importantly, we controlled for depression and anxiety symptoms during the six years prior to the outcome as well as for prior levels of peer victimization. In doing so, we attempted to rule out the possibility that victimization or earlier internalizing symptoms were the cause of weight gain in early childhood. Our findings extend the current literature examining the pathways between overweight, peer victimization, body dissatisfaction and internalizing symptoms to a non-US, longitudinal, community sample of youth spanning from childhood to early adolescence. This study must be interpreted in light of its limitations. Although the initial QLSCD represented families with a newborn child born in Quebec in 1997, due to attrition, the subsample analyzed here is not representative of the population. In addition, our measure of peer victimization was not specific to weightbased teasing, and our measures of overweight and victimization overlapped in time. However, our supplementary analysis showed

208

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

the trends of peer victimization closely followed those of the overweight trajectories across time. Finally, we were not able to control for previous levels of the desire to be thinner. Some studies point to elevated levels of binge or “loss of control” eating (Braet et al., 1997; Puder and Munsch, 2010; Tanofsky-Kraff et al., 2004), which could result from the desire to be thinner and lead to subsequent overweight. However, this process has generally been identified among older adolescents and adults, and our lack of control for prior body dissatisfaction is unlikely to be a confounding factor for the onset of overweight among children as early as 6–8 years. In conclusion, we identified overweight during middle childhood as prospectively associated with depression and anxiety symptoms in early adolescence via the mediating mechanisms of increased levels of peer victimization and the desire to be thinner, a social and a cognitive pathway, respectively. These mechanisms were shown to differ depending on the child's developmental trajectories of weight status. Given the current high prevalence of overweight among children and the apparent poor mental health outcomes, early prevention of overweight is needed, but action on the mediating mechanisms is also crucial. Further research is needed to identify how best to prevent peer victimization of overweight youth and to work towards broader societal weight stigma reduction measures, as well as promote a healthy lifestyle among youth that does not compromise body image.

Acknowledgments

The authors have no potential conflicts of interest to disclose. The first author had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. We thank the children and families whose ongoing participation made this study possible, as well as the Québec Institute of Statistics and the GRIP staff for data collection and management. This research, as part of the Quebec Longitudinal Study of Child Development (QLSCD), was supported by the Québec Government's Ministry of Health and Ministry of Family Affaires, The Lucie and André Chagnon Foundation, the Québec Health Research Fund (FQRS) (grant number 16031), The Québec Research Fund for Society and Culture (FQRSC) (grant number 2009-RG-124779), Canada's Social Science and Humanities Research Council (grant number 412-94-0016 and 839-2000-1008), the Canadian Institutes for Health Research (CIHR), (grant number MOP-246147). In addition, the first author obtained research scholarships from the Canadian Institute of Health Research (CIHR) and the Ste Justine Hospital Research Center specifically for this research project as part of a PhD dissertation.

References Atlantis, E., Baker, M., 2008. Obesity effects on depression: systematic review of epidemiological studies. Int. J. Obes. 32, 881–891. Barker, E.D., Boivin, M., Brendgen, M., Fontaine, N., Arseneault, L., Vitaro, F., Bissonnette, C., Tremblay, R.E., 2008. Predictive validity and early predictors of peer-victimization trajectories in preschool. Arch. Gen. Psychiatry 65, 1185–1192. Bearman, S.K., Martinez, E., Stice, E., Presnell, K., 2006. The skinny on body dissatisfaction: a longitudinal study of adolescent girls and boys. J. Youth Adolesc. 35, 217–229. Boivin, M., Petitclerc, A., Feng, B., Barker, E., 2010. The Developmental Trajectories of Peer Victimization in Middle to Late Childhood and the Changing Nature of their Behavioral Correlates. 56. Merrill Palmer Quarterly (Wayne State University Press), pp. 231–260. Bradley, R.H., Houts, R., Nader, P.R., O’Brien, M., Belsky, J., Crosnoe, R., 2008. The relationship between body mass index and behavior in children. J. Pediatr. 153, 629–634. Bradshaw, C.P., Waasdorp, T.E., O’Brennan, L.M., Gulemetova, M., 2011. Findings from the National Education Association's Nationwide Study of Bullying:

Teachers' and Education Support Professionals' Perspectives. National Education Association Research Department., Atlanta, GA, p. 58. Braet, C., Mervielde, I., Vandereycken, W., 1997. Psychological aspects of childhood obesity: a controlled study in a clinical and nonclinical sample. J. Pediatr. Psychol. 22, 59–71. Byles, J., Byrne, C., Boyle, M., Offor, D., 1988. Ontario child health study: reliability and validity of the general functioning subscale of the McMaster family assessment device. Fam. Process 27, 97–104. Chaiton, M., Sabiston, C., O’Loughlin, J., McGrath, J.J., Maximova, K., Lambert, M., 2009. A structural equation model relating adiposity, psychosocial indicators of body image and depressive symptoms among adolescents. Int. J. Obes. 33, 588–596. Cole, T.J., Bellizzi, M.C., Flegal, K.M., Dietz, W.H., 2000. Establishing a standard definition for child overweight and obesity worldwide: international survey. BMJ 320, 1240–1243. Collins, M.E., 1991. Body figure perceptions and preferences among preadolescent children. Int. J. Eat. Disord. 10, 199–208. Cote, S.M., Boivin, M., Liu, X., Nagin, D.S., Zoccolillo, M., Tremblay, R.E., 2009. Depression and anxiety symptoms: onset, developmental course and risk factors during early childhood. J. Child Psychol. Psychiatry Allied Discip. 50, 1201–1208. Cramer, P., Steinwert, T., 1998. Thin is good, fat is bad: how early does it begin? J. Appl. Dev. Psychol. 19, 429–451. Daraganova, G., 2014. Body image of primary school children, the longitudinal study of Australian children annual statistical report 2013. Australian Institute of Family Studies Melbourne, pp. 111–133 (Chapter 7). Datar, A., Sturm, R., 2004. Childhood overweight and parent- and teacher-reported behavior problems: evidence from a prospective study of kindergartners. Arch. Pediatr. Adolesc. Med. 158, 804–810. Drukker, M., Wojciechowski, F., Feron, F.J., Mengelers, R., Van Os, J., 2009. A community study of psychosocial functioning and weight in young children and adolescents. Int. J. Pediatr. Obes. Off. J. Int. Assoc. Study Obes. 4, 91–97. Epstein, N.B., Baldwin, L.M., Bishop, D.S., 1983. The McMaster family assessment device. J. Marital Fam. Ther. 9, 171–180. Erickson, S.J., Robinson, T.N., Haydel, K.F., Killen, J.D., 2000. Are overweight children unhappy? body mass index, depressive symptoms, and overweight concerns in elementary school children. Arch. Pediatr. Adolesc. Med. 154, 931–935. Giles, P., 2004. Low Income Measurement in Canada. Minister of Industry, Canada, Ottawa, Ontario. Goldfield, G.S., Moore, C., Henderson, K., Buchholz, A., Obeid, N., Flament, M.F., 2010. Body dissatisfaction, dietary restraint, depression, and weight status in adolescents. J. Sch. Health 80, 186–192. Goldschmidt, A.B., Aspen, V.P., Sinton, M.M., Tanofsky-Kraff, M., Wilfley, D.E., 2008. Disordered eating attitudes and behaviors in overweight youth. Obesity 16, 257–264. Hamilton, L.D., Newman, M.L., Delville, C.L., Delville, Y., 2008. Physiological stress response of young adults exposed to bullying during adolescence. Physiol. Behav. 95, 617–624. Harriger, J.A., Thompson, J.K., 2012. Psychological consequences of obesity: weight bias and body image in overweight and obese youth. Int. Rev. Psychiatry 24, 247–253. Hayes, A.F., 2013. Introduction to Mediation, Moderation, and Conditional Process Analysis: A Regression-Based Approach. Guilford Press., New York. Herpertz-Dahlmann, B., Buhren, K., Remschmidt, H., 2013. Growing up is hard: mental disorders in adolescence. Dtsch. Arzteblatt Int. 110, 432–439, quiz 440. Jeffers, A.J., Cotter, E.W., Snipes, D.J., Benotsch, E.G., 2013. BMI and depressive symptoms: the role of media pressures. Eat. Behav. 14, 468–471. Kovacs, M., 1985. The children's depression, inventory (CDI). Psychopharmacol. Bull. 21, 995–998. Lavallée, C., 2004. Enquête Sociale et de Santé Auprès des Cnfants et des Adolescents Québécois: Volet Nutrition. Institution de la Statistique du Québec, Québec, p. 166. Lopresti, A.L., Drummond, P.D., 2013. Obesity and psychiatric disorders: commonalities in dysregulated biological pathways and their implications for treatment. Prog. Neuro-Psychopharmacol. Biol. Psychiatry 45, 92–99. Luppino, F.S., de Wit, L.M., Bouvy, P.F., Stijnen, T., Cuijpers, P., Penninx, B.W., Zitman, F.G., 2010. Overweight, obesity, and depression: a systematic review and metaanalysis of longitudinal studies. Arch. Gen. Psychiatry 67, 220–229. MacLean, L., Edwards, N., Garrard, M., Sims-Jones, N., Clinton, K., Ashley, L., 2009. Obesity, stigma and public health planning. Health Promot. Int. 24, 88–93. McCormack, L.A., Laska, M.N., Gray, C., Veblen-Mortenson, S., Barr-Anderson, D., Story, M., 2011. Weight-related teasing in a racially diverse sample of sixthgrade children. J. Am. Diet. Assoc. 111, 431–436. Mond, J., van den Berg, P., Boutelle, K., Hannan, P., Neumark-Sztainer, D., 2011. Obesity, body dissatisfaction, and emotional well-being in early and late adolescence: findings from the project eat study. J. Adolesc. Health: Off. Publ. Soc. Adolesc. Med. 48, 373–378. Mustillo, S., Worthman, C., Erkanli, A., Keeler, G., Angold, A., Costello, E.J., 2003. Obesity and psychiatric disorder: developmental trajectories. Pediatrics 111, 851–859. Nagin, D., Tremblay, R.E., 1999. Trajectories of boys' physical aggression, opposition, and hyperactivity on the path to physically violent and nonviolent juvenile delinquency. Child Dev. 70, 1181–1196. Nagin, D.S., Odgers, C.L., 2010. Group-based trajectory modeling in clinical research. Annu. Rev. Clin. Psychol. 6, 109–138. Perry, D.G., Kusel, S.J., Perry, L.C., 1988. Victims of peer agression. Dev. Psychol. 24, 807–814. Pingault, J.B., Cote, S.M., Lacourse, E., Galera, C., Vitaro, F., Tremblay, R.E., 2013.

L. Pryor et al. / Journal of Affective Disorders 202 (2016) 203–209

Childhood hyperactivity, physical aggression and criminality: a 19-year prospective population-based study. PloS One 8, e62594. Pryor, L.E., Brendgen, M., Tremblay, R.E., Pingault, J.B., Liu, X., Dubois, L., Touchette, E., Falissard, B., Boivin, M., Cote, S.M., 2015. Early risk factors of overweight developmental trajectories during middle childhood. PloS One 10, e0131231. Puder, J.J., Munsch, S., 2010. Psychological correlates of childhood obesity. Int. J. Obes. 34 Suppl 2, S37–S43. Puhl, R.M., Latner, J.D., 2007. Stigma, obesity, and the health of the nation's children. Psychol. Bull. 133, 557–580. Puhl, R.M., Luedicke, J., Heuer, C., 2011. Weight-based victimization toward overweight adolescents: observations and reactions of peers. J. Sch. Health 81, 696–703. Reeves, G.M., Postolache, T.T., Snitker, S., 2008. Childhood obesity and depression: connection between these growing problems in growing children. Int. J. Child Health Hum. Dev.: IJCHD 1, 103–114. Seguin, L., Nikiema, B., Gauvin, L., Zunzunegui, M.V., Xu, Q., 2007. Duration of poverty and child health in the Quebec longitudinal study of child development: longitudinal analysis of a birth cohort. Pediatrics 119, e1063–e1070. Shunk, J.A., Birch, L.L., 2004. Girls at risk for overweight at age 5 are at risk for dietary restraint, disinhibited overeating, weight concerns, and greater weight

209

gain from 5 to 9 years. J. Am. Diet. Assoc. 104, 1120–1126. Spielberger, C.D., 1973. State-Trait Anxiety Inventory for Children (STAIC): Professional Manual, Redwood City. Spielberger, C.D., Gorsuch, R.L., Lushene, R.E., 1970. Manual for the State-Trait Anxiety Inventory (Self-Evaluation Questionnaire), Palo Alto. Tanner, J.M., 1962. Growth at Adolescence, 2nd (ed.) Blackwell Scientific Publications., Oxford. Tanofsky-Kraff, M., Yanovski, S.Z., Wilfley, D.E., Marmarosh, C., Morgan, C.M., Yanovski, J.A., 2004. Eating-disordered behaviors, body fat, and psychopathology in overweight and normal-weight children. J. Consult. Clin. Psychol. 72, 53–61. ter Bogt, T.F., van Dorsselaer, S.A., Monshouwer, K., Verdurmen, J.E., Engels, R.C., Vollebergh, W.A., 2006. Body mass index and body weight perception as risk factors for internalizing and externalizing problem behavior among adolescents. J. Adolesc. Health: Off. Publ. Soc. Adolesc. Med. 39, 27–34. Tremblay, R.E., Loeber, R., Gagnon, C., Charlebois, S., Larivee, S., LeBlanc, M., 1991. Disruptive boys with stable and unstable fighting behaviour patterns during elementary school. J. Abnorm. Child Psychol., 19. Wardle, J., Cooke, L., 2005. The impact of obesity on psychological well-being. Best Pract. Res. Clin. Endocrinol. Metab. 19, 421–440.