Q1

Q3

12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455

http://dx.doi.org1550-7289/r 20

*CorrespondPublic HealthTomtebodavägeE-mail: Finn.Ra

Surgery for Obesity and Related Diseases ] (2014) 00–00

Original article

Children’s weight status, body esteem, and self-concept after maternalgastric bypass surgery

Mikaela Willmer, R.D.a, Daniel Berglind, M.Sc.a, Per Tynelius, M.Sc.a, Ata Ghaderi, PSYCH,Ph.D.b, Erik Näslund, M.D., Ph.D.c, Finn Rasmussen, M.D., PhD.a,*

aDepartment of Public Health Sciences, Karolinska Institutet, Stockholm, SwedenbDepartment of Clinical Neuroscience, Karolinska Institutet, Stockholm, Sweden

cDepartment of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden

Received August 27, 2014; accepted December 3, 2014

Abstract Background: There is little research on the effects of Roux-en-Y gastric bypass (RYGB) surgery

/10.1014 A

ence:Epiden 18smuss

on the patients’ offspring’s weight and psychosocial well-being. The present study investigates howchildren are affected in terms of weight status, body esteem, and self-concept after maternal RYGB.Methods: Sixty-nine women and their families were recruited from RYGB waiting lists at 5Swedish hospitals. Data was collected during home visits 3 months before and 12 months afterRYGB. Anthropometrical measures were taken, and the children completed the Body Esteem Scale(BES) and the Beck Self-Concept Inventory (BYI-S).Results: Prevalence or relative risk in weight status between the time points were explored usingPoisson regression models (General Estimating Equations). We found stronger statistical evidencefor a decreased risk of overweight (RR ¼ .85, 57.0% versus 48.6%, P ¼ .048), than for obesity (RR¼ .87, 18.2% versus 15.9%, P ¼ .447). The boys’ body esteem increased slightly, whilst the girls’did not. There were no significant differences in mean BYI-S scores.Conclusions: The reasons behind the statistically significant decrease in overweight but not obesityrisk in children after maternal RYGB may be caused by insufficient statistical power to detectchanges in obesity risk alone. Body esteem seems to be slightly positively affected in boys 9 monthsafter maternal RYGB. Longer follow-up times and larger samples would be useful in futureresearch. (Surg Obes Relat Dis 2014;]:00–00.) r 2014 American Society for Metabolic andBariatric Surgery. All rights reserved.

Keywords: ■■■

5657585960616263

Laparoscopic Roux-en-Y gastric bypass (RYGB)accounts for 90% of all bariatric procedures performed inSweden [1], and induces major and permanent changes inthe individual’s eating behavior and food choices.Bariatric surgery has been shown to positively affect

psychosocial aspects such as body image and health-related

64656667

16/j.soard.2014.12.003merican Society for Metabolic and Bariatric Surgery. All r

Finn Rasmussen, Professor, Child and Adolescentmiology, Department of Public Health Sciences,A, Widerströmska huset, 171 77 Stockholm.en@ki.se

quality of life [2,3]. Since some (but not all) studies havefound a correlation between maternal and child body esteemand eating habits [4–6], changes in maternal body imageand eating habits after RYGB may also affect the patients’children. However, changes in maternal eating behaviorafter RYGB may also have a negative impact on children’sbody image, as some previous research has shown thatchildren who perceive their mothers to strive for thinnerfigures may have higher levels of body dissatisfactionthemselves [7]. Furthermore, if the mother changes her

686970ights reserved.

7172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125

126127128129130131132133134135136137138139140141142143144145146147148149

M. Willmer et al. / Surgery for Obesity and Related Diseases ] (2014) 00–002

eating habits after surgery, and no longer buys energy-dense, nutrient-poor foods for the household, this mayaffect her child’s weight status.We have found 2 other published studies concerning the

effects of parental bariatric surgery on the weight status ofthe patients’ children. Woodard et al. could not find anysignificant Body Mass Index (BMI) or waist circumferencechanges in 15 children of RYGB patients, but did see anincrease in the number of children who stated that they werecurrently on a diet (25% before versus 50% after) [8].Hirsch et al. conducted a register-based case-control

study, comparing changes in BMI in 128 children livingwith an adult undergoing bariatric surgery with those of 384children matched for age, gender, and BMI and living withan adult with obesity who did not have bariatric surgery.They found no differences between cases and controls 24months after bariatric surgery, except for boys with over-weight, where those living in bariatric surgery householdshad a lower than expected BMI, and those in controlhousehold had a higher than expected BMI [9].The aim of the present study was to explore how children

are affected in terms of weight development, body esteem,and self-concept when their mothers undergo gastric bypasssurgery for obesity.

150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180

Methods

Recruitment

Bariatric surgery, like all healthcare in Sweden, is largelytax-funded and available at no (or very small) costs to theindividual patient [10]. Women eligible for participationwere identified from the waiting list for RYGB at 5 surgicalclinics in Sweden. The inclusion criteria were defined asbeing cleared for RYGB, having at least one child betweenthe ages of 7 and 14, and being able to speak andunderstand Swedish. All patients gave oral and writtenconsent and the study was approved by the StockholmRegional Ethical Review Board (no. 2009/1472-31/3).Between April 2011 and October 2012, 69 families were

recruited to the study. The data collection took place during2 home visits to the families, usually in the evenings, withthe first one carried out approximately 3 months beforesurgery and the second one approximately 9 months aftersurgery. All data was collected by 2 of the authors (MWand DB).At the time for the follow-up home visit a year later, one

family declined to continue participation in the study, 2women had decided not to undergo the RYGB surgery, and3 families had moved and proved impossible to locate.Additionally, a few children did not want to be weighed atone or both data collection time points. Thus, the data setconsists of 61 families with a total of 81 children whoparticipated in both data collections (both home visits andthe collection of anthropometrical data).

Data collection

All family members taking part were weighed usingcalibrated scales (VB2-200-EC, Vetek AB, Väddö, Swe-den), their height was measured using a portable stadiom-eter (Seca 213, Seca, Chino, CA, USA) and their waistcircumference measured. They were also given question-naires to complete after the visits and return by post. Thechildren’s questionnaire included the Body Esteem Scale(BES), which measures 3 dimensions of weight and bodyconcerns; weight-related (BES-W), appearance-related(BES-APP) and attribution-related (BES-ATT). It contains23 statement items, to which the subject responds by usinga 5-point Likert-type scale ranging from “Never” to“Always.” The BES has been validated and shown to bea reliable measure of body esteem, including in Swedishchildren [11].The children also completed the Beck Self-Concept

Inventory (BYI-S), which is designed for use with childrenbetween the ages of 7 and 14 [12]. Like the BES, it hasgood reliability and validity [12]. It consists of 20 statementitems, and the subject responds to these using a 4-pointLikert-type scale ranging from “Never” to “Always.” Theresponses are scored, and the scores are then transformedinto gender-specific percentile ranks based on a representa-tive sample of 2400 Swedish school children. A scorebetween the 26th and the 89th percentiles is consideredaverage [13].Despite receiving at least 3 reminders, 7 mothers and 14

children failed to return their questionnaires for the first datacollection. For the second data collection, the numbers were7 mothers and 13 children.

Statistical analysis

The children were categorized as normal weight, over-weight, and obese, using age- and gender-specific BMI cut-off points as specified by Cole et al. [14]. We then exploredchanges in prevalence (proportion) or relative risk in weightstatus (binary outcomes) between the 2 time points usingage- and sex-adjusted modified Poisson regression modelswith robust variance [15] estimated with generalizedestimating equations (GEE) to account for within-individual correlation [16]. We also conducted the analysesstratified on sex with an added interaction term for sex*timepoint to test whether any changes found were differentbetween girls and boys in the sample.Differences in mean BES and BYI-S scores (continuous

outcomes) were analyzed using linear regression modelswith robust variance (GEE). We also used the outcomestransformed to standard deviation scores (SD scores) toexpress effect sizes. Interaction terms were added for sexand weight status at baseline to test whether any changesdiffered between groups. All models were estimated usingSTATA 12.1 (Stata Corp, College Station, Texas, USA).

T1

T2

T3

181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235

236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290

Table 1Descriptive characteristics of the participating mothers and children

Mothers n 3 mo before RYGB(SD)

9 mo after RYGB(SD)

Age 61 39.2 (5.4) 40.3 (5.5)BMI 61 39.1 (3.3) 27.5 (3.1)Weight 61 107.3 (12.5) 75.1 (10.1)Waist circumference 60 117.6 (9.7) 88.9 (8.4)ChildrenAge 81 10.0 (2.1) 11.1 (2.1)Females 81 40 (49.4%) 40 (49.4%)BMI 81 20.5 (4.04) 20.9 (4.0)Waist circumference 80 71.6 (12.6) 73.6 (11.3)Overweight 81 45 (55.6%) 41 (50.6%)Obese 81 15 (18.5%) 15 (18.5%)BES-W 60 17.8 (6.5) 17.0 (6.3)BES-APP 55 36.7 (11.5) 34.5 (11.0)BES-ATT 53 8.3 (2.4) 8.2 (2.6)BYI-S percentile 56 75.0 (21.1) 73.0 (25.3)

RYGB ¼ Roux-en-Y Gastric Bypass; SD ¼ Standard Deviation; BMI ¼Body Mass Index; BES-W ¼ Body Esteem Scale - Weight; BES-APP ¼Body Esteem Scale - Appearance; BES-ATT ¼ Body Esteem Scale -Attribution; BYI-S: ¼ Beck Youth Inventories - Self-Concept.

Table 2Adjusted prevalences and relative risks of overweight and obesity in theparticipating children 3 mo pre- and 9 mo postmaternal RYGB estimatedusing Poisson regression models with robust variance (General EstimatingEquations)

Variable n Prevalence % (before);prevalence % (after)

Relative risk (95%confidence intervals)

P

Overweight* 81 57.0; 48.6 .85 (.73; 1.00) .048Obesity* 81 18.2; 15.9 .87 (.62; 1.24) .447Overweight –girls†

40 56.6; 45.7 .81 (.66; .98) .031

Overweight –boys†

41 57.4; 51.4 .90 (.72; 1.11) .311

Differencebetweengroups

�.8; �5.7 1.11 (.86; 1.43) .429

Obesity –

girls†40 13.1; 11.4 .87 (.48; 1.60) .660

Obesity –

boys†41 25.1; 21.9 .87 (.61; 1.25) .458

Differencebetweengroups

�12.0; -10.5 1.00 (.54; 1.86) .996

RYGB ¼ Roux-en-Y gastric bypass.*Adjusted for age and sex.†Adjusted for age.

Familial Effects of Gastric Bypass Surgery / Surgery for Obesity and Related Diseases ] (2014) 00–00 3

Results

Characteristics of the participating women and childrenare shown in Table 1. The women lost a mean 11.6 BMIunits, or 84.3% of their excess BMI, after RYGB. Fifty-eight of the 61 participating women reported their level ofeducation, and 9 (15.5%) of these reported having apostsecondary education, compared with 45% of the generalSwedish female population [17].

Overweight and obesity

The prevalence of overweight and obesity in our sampleof children was very high compared with the generalSwedish population of children – almost 56% in oursample, compared with approximately 18% in the generalpopulation [18,19]. Adjusted for sex and age, we foundstronger statistical evidence for a decreased risk of over-weight and obesity together (RR = .85, 57.0% versus48.6%, P = .048), than for obesity alone (RR = .87,18.2% versus 15.9%, P = .447) after maternal RYGB(Table 2). When stratifying the analysis by sex, the decreasewas only significant for the girls in the sample.

Body esteem and self-concept

When stratifying the adjusted BES scores by weightstatus and by sex, the nonsignificant and small decrease onthe BES-W subscale for girls (difference ¼ -.84, P ¼ .307)compared with a trend toward a significant increase amongboys (difference ¼ 1.80, P ¼ .058), indicating morepositive weight-related body esteem, resulted in a signifi-cant group difference (difference ¼ 2.64, P ¼ .016) with an

almost moderate effect size (SD score ¼ .41) (Table 3). Thesame pattern was seen for BES-APP (the girls’ scoresdecreased whilst the boys’ increased), although the groupdifference was not significant, and the magnitude of effectswas small (SD score ¼ .11). When it came to the BYI-S, theage- and sex-adjusted difference in the children’s percentilerank was less than one percentage point, with both estimatesin the “average” category for self-concept.To check for any bias introduced by between-sibling

correlation in the case of the 18 families with 2 childrentaking part in the study, we also performed multilevelanalyses that took this into account. This did not change theresults, however.

Discussion

Changes in relative risk of overweight and obesity

The present study found a decreased risk of overweightafter maternal RYGB. This is partly in accordance with theresults reported by Woodard et al, who found a tendencytoward a lower-than-expected BMI in obese children ofRYGB patients [8].As for the mechanisms behind this decrease in over-

weight risk, it might be that the mothers, once they can nolonger tolerate certain calorie-dense foods, also stop buyingthem for the rest of the family. However, a previous studyby Walters-Bugbee et al. compared the food items availablein the homes of bariatric surgery patients with those inthe homes of women who were waiting to undergo the

291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345

346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381

Table 3Adjusted differences in the children’s Body Esteem Scale and Beck Self-Concept Inventory scores between 3 mo pre- and 9 mo postmaternalRYGB, estimated with linear regression models (General EstimatingEquations)

Variable n Difference in score(95% ConfidenceIntervals)

P Difference in standarddeviation score (95%Confidence Intervals)

BES-W –

all*60 .53 (�.85; 1.90) .454 .08 (�.13; .30)

BES-W –

NW atbaseline*

27 �.46 (�2.19; 1.27) .591 �.07 (�.34; .20)

BES-W –

OW atbaseline*

33 .99 (�.74; 2.72) .268 .16 (�.12; .43)

Differencebetweengroups

1.45 (�.78; 3.68) .203 .23 (�.12; .58)

BES-W –

girls†29 �.84 (�2.46; .77) .307 �.13 (�.39; .12)

BES-W –

boys†31 1.80 (�.06; 3.66) .058 .28 (.01; .57)

Differencebetweengroups

2.64 (.50; 4.79) .016 .41 (.08; .75)

BES-APP– all*

55 .30 (�2.33; 2.93) .822 .03 (�.21; .26)

BES-APP– NW atbaseline†

25 �.64 (�3.73; 2.45) .685 �.06 (�.33; .22)

BES-APP– OW atbaseline†

30 .61 (�2.80; 4.02) .726 .05 (�.25; .36)

Differencebetweengroups

1.25 (�2.95; 5.45) .560 .11 (�.26; .49)

BES-APP– girls†

28 �.48 (�3.62; 2.66) .766 �.04 (�.32; .24)

BES-APP– boys†

27 1.10 (�2.54; 4.74) .554 .10 (�.23; .42)

Differencebetweengroups

1.58 (�2.65; 5.81) .465 .14 (�.24; .52)

BES-ATT– all*

53 �.09 (�.74; .57) .796 .04 (�.30; .23)

BES- ATT– NW atbaseline†

21 .01 (�.96; .98) .986 .00 (�.39; .40)

BES- ATT- OW atbaseline†

32 �.21 (�1.02; .60) .613 �.09 (�.42; .25)

Differencebetweengroups

�.22 (�1.42; .99) .722 �.09 (�.58; .40)

BES- ATT- girls†

30 .05 (�.79; .88) .912 .02 (�.32; .36)

BES- ATT- boys†

23 �.26 (�1.18; .66) .580 �.11 (�.48; .27)

Differencebetweengroups

�.31 (�1.48; .87) .609 �.12 (�.60; .35)

BYI-S –

all*57 �.84 (�7.04; 5.37) .792 �.04 (�.30; .23)

Table 3Continued.

Variable n Difference in score(95% ConfidenceIntervals)

P Difference in standarddeviation score (95%Confidence Intervals)

BYI-S –

NW atbaseline†

27 �1.68 (�8.97;5.61)

.651 �.07 (�.39; .24)

BYI-S –

OW atbaseline†

30 �.93 (�9.97; 8.11) .841 �.04 (�.43; .35)

Differencebetweengroups

.76 (�1.11; 11.62) .892 .03 (�.44; .50)

BYI-S –

girls†28 �1.74 (�1.57;

7.09).699 �.08 (�.46; .31)

BYI-S –

boys†29 .04 (�7.67; 7.75) .992 .00 (�.33; .34)

Differencebetweengroups

1.78 (�9.18; 12.73) .675 .08 (�.40; .55)

RYGB ¼ Roux-en-Y Gastric Bypass; BES-W ¼ Body Esteem Scale –

Weight; BES-APP ¼ Body Esteem Scale – Appearance; BES-ATT ¼Body Esteem Scale – Attribution; BYI-S ¼ Beck Youth Inventories – Self-Concept; NW ¼ Normal weight; OW ¼ Overweight/obese.

*Adjusted for age and sex.†Adjusted for age.

M. Willmer et al. / Surgery for Obesity and Related Diseases ] (2014) 00–004

procedure, and found no significant differences between thegroups [20].It is not clear why we only found a statistically significant

decrease in overweight but not obesity risk. The relativelysmall sample size, especially for children with obesity,means that the resulting insufficient statistical power may bethe cause behind the nonsignificant decrease in obesity.Alternatively, the present study reflects relatively short-termeffects of maternal RYGB, and it may take longer for theseeffects to change the weight status of children with obesity.

382383384385386387388389390391392393394395396397398399400

Changes in body esteem and self-concept

The concept of body esteem has been defined as the“self-evaluation of one’s body or appearance” [21]. It hasalso been described as the physical counterpart of self-esteem [22]. It is often used synonymously with bodyimage, although the latter is in fact a somewhat broaderconstruct [23].Although the inherent power problems of the present

study make it somewhat risky to draw any firm conclusionsfrom the results, it is possible that girls are influenceddifferently from boys in terms of weight concerns. Beyondthe issue of significance, the magnitude of effect wassomewhat larger for boys than girls. It is possible that theweight concerns of the girls in the present study wereaffected negatively by what they perceived as increaseddieting behavior in their mothers, as this association hasbeen found in previous studies [7,24].

401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455

456457458459460461462463464465466467468469470471472473474475476477478479480481482

Familial Effects of Gastric Bypass Surgery / Surgery for Obesity and Related Diseases ] (2014) 00–00 5

We compared the results of the present study with thosereported by Ivarsson et al, who administered the BES to 405Swedish adolescents with a mean age of 14.7 years, andfound that the children in our study scored higher on theBES-APP (34.0 compared with 27.1) but lower in the BES-W (16.1 compared with 20.9) and BES-ATT (8.0 comparedwith 11.3) dimensions [25]. It is somewhat worrying thatthe children in our study report lower scores on the BES-Wand BES-ATT dimensions than children who were a mean 3years older, as scores usually decrease with age [26,27].This may be due to the large proportion of overweight andobesity in our sample, as this has been shown to beassociated with lower body esteem both in the study byIvarsson et al. and in others [25,28].Self-concept is often used interchangeably with self-

esteem, and has been described as “the totality of anindividual’s thoughts and feelings having reference tohimself as an object” [29]. Self-esteem forms the evaluativepart of the self-concept, and is thus a slightly differentconcept, although often used interchangeably [30]. Therewere no meaningful changes in the children’s BYI-S scores,indicating unchanged self-concept. Their mean percentilebefore surgery was approximately 75, which is classified asaverage. As the children’s self-concept was already withinthe normal range for children of their age, it is perhaps notsurprising that it did not improve further during the timeperiod.

483484485486487488489490491492493494495496497498499500501502503504505506507508509510

Strengths and limitations

This study has several strengths and limitations. The datacollection was conducted in the participants’ home environ-ment, and all home visits were conducted by the sameresearchers (MW and DB). All anthropometrical data wascollected by the researchers, rather than self-reported. Thisis important as it has been shown that individuals withobesity are more likely to underreport their weight thannormal-weight individuals. There was also little drop-outbetween data collection time points, and those families whodid drop out did so predominantly because they movedaway and could not be relocated, which should notcontribute to any systematic bias.When it comes to limitations, the relatively small sample

size and short follow-up time should be mentioned,although the sample is in fact considerably larger than theonly other comparable study, that by Woodard et al. [8].However, as already mentioned, it does create powerproblems, especially in the subgroup analyses.Another potential limitation is the fact that RYGB

patients who did not understand spoken and writtenSwedish were excluded from the study, limiting the generalapplications of the results. However, it would have beenethically unacceptable to include participants who wereunable to read and understand the questionnaires and thewritten consent forms.

Conclusions

The present study found that children of female RYGBpatients may have a reduced risk of overweight 9 monthsafter maternal surgery. Self-concept remained unchanged,whilst weight-related body esteem seems to be positivelyaffected in boys, but not girls. Future research should focuson whether these results can be replicated in larger samples,and on exploring the specific mechanisms behind them. Thefinding that overweight and obese children seem to benefitfrom maternal RYGB may, if replicated in further stu-dies, provide an additional rationale for clinical practitionersregarding which patients to recommend for bariatricsurgery.

Disclosures

The authors have no commercial associations that mightbe a conflict of interest in relation to this article.

Acknowledgments

The present study was funded by a grant to FinnRasmussen from the Swedish Council for Working Lifeand Social Research (grant number: 2008-0654), StockholmCounty Council (ALF), NovoNordisk fund, KarolinskaInstitutet Diabetes Theme center, the National School inCaring Sciences at Karolinska Institutet (NFV), and throughthe Erling-Persson Family Foundation.

References

[1] Stenberg E, Szabo E, Agren G, et al. Early Complications AfterLaparoscopic Gastric Bypass Surgery: Results From the ScandinavianObesity Surgery Registry. Ann Surg 2013;260:1040–7.

[2] Sarwer DB, Wadden TA, Moore RH, Eisenberg MH, Raper SE,Williams NN. Changes in quality of life and body image after gastricbypass surgery. Surg Obes Relat Dis 2010;6:608–14.

[3] Karlsson J, Sjostrom L, Sullivan M. Swedish obese subjects (SOS)—an intervention study of obesity. Two-year follow-up of health-relatedquality of life (HRQL) and eating behavior after gastric surgery forsevere obesity. Int J Obes Relat Metab Disord 1998;22:113–26.

[4] Edlund BH, K, Sjödén PO. Eating behaviours, and attitudes to eating,dieting and body image in 7-year-old Swedish girls. Eur Eating DisRev 1994;4:40–53.

[5] Savage JS, Fisher JO, Birch LL. Parental influence on eatingbehavior: conception to adolescence. J Law Med Ethics 2007;35:22–34.

[6] Elfhag K, Tynelius P, Rasmussen F. Family links of eating behaviourin normal weight and overweight children. Int J Pediatr Obes 2010;5:491–500.

[7] Lowes J, Tiggemann M. Body dissatisfaction, dieting awareness andthe impact of parental influence in young children. Br J HealthPsychol 2003;8:135–47.

[8] Woodard GA, Encarnacion B, Peraza J, Hernandez-Boussard T,Morton J. Halo effect for bariatric surgery: collateral weight loss inpatients' family members. Arch Surg 2011;146:1185–90.

[9] Hirsch AG, Wood GC, Bailey-Davis L, Lent MR, Gerhard GS, StillCD. Collateral weight loss in children living with adult bariatricsurgery patients: a case control study. Obesity (Silver Spring)2014;22:2224–9.

511512513514515516517518519520521522523524525526527528529530531532533534

535536537538539540541542543544545546547548549550551552553554555556557558

M. Willmer et al. / Surgery for Obesity and Related Diseases ] (2014) 00–006

[10] Anell A, Glenngard AH, Merkur S. Sweden health system review.Health Syst Transit 2012;14:1–159.

[11] Erling A, Hwang CP. Body-esteem in Swedish 10-year-old children.Percept Mot Skills 2004;99:437–44.

[12] Bose-Deakins J, Floyd R. A review of the Beck Youth Inventories ofEmotional and Social Impairment. J School Psychol 2004;42:333–40.

[13] Tideman E. Manual of the Swedish Version of the Beck YouthInventories. Stockholm, Sweden: The Psychological Corporation:2004.

[14] Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standarddefinition for child overweight and obesity worldwide: internationalsurvey. BMJ 2000;320:1240–3.

[15] Zou G. A modified poisson regression approach to prospective studieswith binary data. Am J Epidemiol 2004;159:702–6.

[16] Diggle PJ HP, Liang KY, Zeger SL. Analysis of Longitudinal Data.Oxford, UK: Oxford University Press; 2002.

[17] SCB. Yearbook of Education Statistics 2014/UtbildningsstatistiskÅrsbok 2014: Statistics Sweden; 2013.

[18] Olds T, Maher C, Zumin S, et al. Evidence that the prevalence ofchildhood overweight is plateauing: data from nine countries. Int JPediatr Obes 2011;6:342–60.

[19] Sundblom E, Petzold M, Rasmussen F, Callmer E, Lissner L.Childhood overweight and obesity prevalences levelling off inStockholm but socioeconomic differences persist. Int J Obes (Lond)2008;32:1525–30.

[20] Walters-Bugbee SE, McClure KS, Kral TV, Sarwer DB. Maternalchild feeding practices and eating behaviors of women with extremeobesity and those who have undergone bariatric surgery. Surg ObesRelat Dis 2012;8:784–91.

[21] Mendelson BK, Mendelson MJ, White DR. Body-esteem scale foradolescents and adults. J Pers Assess 2001;76:90–106.

[22] Mendelson BK, White DR. Development of self-body-esteem inoverweight youngsters. Dev Psychol 1985;21:90–6.

[23] Chen W, Swalm RL. Chinese and American college students' body-image: perceived body shape and body affect. Percept Mot Skills1998;87:395–403.

[24] Usmiani S, Daniluk J. Mothers and their adolescent daughters:Relationship between self-esteem, gender role identity, and bodyimage. J Youth Adolescence 1997;26:45–62.

[25] Ivarsson T, Svalander P, Litlere O, Nevonen L. Weight concerns,body image, depression and anxiety in Swedish adolescents. EatBehav 2006;7:161–75.

[26] Byely L, Archibald AB, Graber J, Brooks-Gunn J. A prospectivestudy of familial and social influences on girls’ body image anddieting. Int J Eat Disord 2000;28:155–64.

[27] Packard P, Krogstrand KS. Half of rural girls aged 8 to 17 yearsreport weight concerns and dietary changes, with both more prevalentwith increased age. J Am Diet Assoc 2002;102:672–7.

[28] Shriver LH, Harrist AW, Page M, Hubbs-Tait L, Moulton M, TophamG. Differences in body esteem by weight status, gender, and physicalactivity among young elementary school-aged children. Body Image2013;10:78–84.

[29] Gecas V. The self-concept. Annu Rev Sociol 1982;8:1–33.[30] Wadden TA, Foster GD, Brownell KD, Finley E. Self-concept in

obese and normal-weight children. J Consult Clin Psych 1984;52:1104–5.