Relationship Between Body Mass Index and Body Image Disturbances Among South African Mothers and Their Daughters Living in Soweto, Johannesburg Emmanuel Cohen, PhD; Philippe Jean-Luc Gradidge, PhD; Lisa K. Micklesfield, PhD; Shane A. Norris, PhD In South African families, a phenomenon of mothers’ acceptance of stoutness coexists with their daughters’ appreciation for thinness. A sample of N = 615 mother-and-daughter pairs was recruited to conjointly identify the relationships toward body image and body mass index between both groups by assessing body weight satisfaction, body esteem, and eating disorders risk. We observed higher prevalence of obesity in mothers and higher eating disorders risk in daughters, while mother-daughter relationships were identified for body mass index and psychometric dimensions. The high prevalence of obesity in mothers and their tolerance for stoutness could expose their daughters to eating disorders and obesity. Key words: eating and feeding disorders, mothers/daughters communication, obesity, weight management S TUDIES in high-income countries (HICs) have shown that overweight and obesity are pos- itively associated with psychological disturbances due to body weight dissatisfaction linked with stigmatization of obesity and fear of being “fat.”1-3 As a result, people living in HICs are faced with a complex, discordant situation, where the prevalence of overweight and obesity is increasing even though thinness is the predominant desired body size.4 This paradox has also been observed among black pop- ulations living in HICs for which the prevalence of obesity is high,5 while the so-called traditional Author Affiliations: MRC/Wits Developmental Pathways for Health Research Unit, Department of Paediatrics, Faculty of Health Sciences, University of the Witwatersrand, Parktown, Johannes- burg, South Africa (Drs Cohen, Micklesfield, and Norris); and Centre for Exercise Science and Sports Medicine, Wits Education Cam- pus, Faculty of Health Sciences, University of the Witwatersrand, Parktown, Johannesburg, South Africa (Dr Gradidge). Drs Cohen and Norris are supported by the South African DST/NRF Centre of Excellence in Human development at the University of the Witwatersrand, Johannesburg, South Africa. All authors have read and approved the final manuscript. E.C. analyzed and interpreted the data. E.C. and P.J.G. drafted the manuscript. E.C., P.J.G., L.K.M., and S.A.N. provided critical revi- sions significantly improving the intellectual content. The study was supervised by L.K.M. and S.A.N. The authors declare no conflict of interest. Correspondence: Emmanuel Cohen, PhD, MRC/Wits Develop- mental Pathways for Health Research Unit, Department of Pae- diatrics, Faculty of Health Sciences, University of the Witwater- srand, 7 York Rd, Parktown, Johannesburg 2193, South Africa (emmcohen@outlook.fr). Copyright © 2019 Wolters Kluwer Health, Inc. All rights reserved. DOI: 10.1097/FCH.0000000000000220 valorization of stoutness is questioned by Western norms promoting pronounced thinness in young adults.6-8 Given the increasing urbanization, ex- posure to Western norms, and obesity prevalence in sub-Saharan Africa (SSA),9,10 researchers have started examining this ambivalence among African populations. Studies have identified the sociocultural val- orization of adiposity as a risk factor for obe- sity in African populations,9-13 and a relationship between eating disorders and body weight per- ception among urban-dwelling African adolescents and young adults has been demonstrated, particu- larly recently in Africa.14,15 Furthermore, the media has been acknowledged to influence contemporary thinking toward favoring thinner body sizes among younger adults.16,17 Thus, African cultural values of body size, perceived as a protective factor toward common modern body image disturbances and eat- ing disorders, seems to be diluted in young people through acculturation of Western influence.14,18,19 Hence, in the context of an increasing obesity epi- demic in SSA, it is therefore important to under- stand the acculturation toward thinness occurring in African populations, as well as the persistence of tradition valorization of stoutness.20-22 South Africa is an SSA country experiencing a particularly strong obesity epidemic,23 with ev- idence of coexistence between valorization and devaluation of stoutness differing by age. Older black populations consider stoutness as a symbol of health (countering the thinness stigma associated with human immunodeficiency virus), wealth, and prosperity while younger people consider body fat as an expression of morbidity and sexual Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 140 Family and Community Health April–June 2019 ■ Volume 42 ■ Number 2 mailto:emmcohen@outlook.fr E. Cohen et al. Relationship Between BMI and Body Image Disturbances 141 undesirability.24,25 Despite the acculturation of younger age groups through the urbanization and modernization of South Africa expressed by a devaluation of overweight/obesity,26,27 parents may still represent the traditional African values and could influence the way their children perceive their bodies by encouraging behaviors that favor larger body size.25,28 Studies from various countries such as Thailand, Israel, and the United States, including African ethnic minorities, have shown that parents, and particularly those who are overweight/obese, tend to underestimate their children’s body weight, especially when these ones are overweight.29-32 In addition, other studies have shown that parents, through their dietary intake habits and global parenting style, can mediate and influence the rela- tionship of their offspring with their body weight management.33 For instance, Cutting et al34 have shown that US mothers’ dietary disinhibition me- diates familial similarities in degree of overweight for mothers and daughters. In addition, Mcdonald et al35 demonstrated that overweight in children is associated with maternal obesity in Colombia. Beyond the nutritional status, other studies attested that parents can influence the relationship of their offspring with their body weight perception,36 with a potential intergenerational transmission of body image disturbances and eating disorders.37-41 In South Africa, although recent studies showed that adolescents are exposed to obesity, body image disturbances, and eating disorders in urban areas at an early age,26,42 the influence of mothers as a deter- minant of these disorders in their daughters remains relatively unknown. The relationship between obe- sity and body image disturbances/eating disorders already demonstrated has not been considered con- jointly in South Africa, particularly in their poten- tial for intergenerational transmission from mothers to daughters. Therefore, this study aims to exam- ine the relationship between mothers and daughters on (1) body image disturbances and eating disor- ders and (2) obesity by testing the association be- tween mothers’ and daughters’ body weight per- ceptions, body esteem, eating attitudes, and body mass index (BMI) within families living in a poor- urban South African township, which is exposed to obesity.43 The 3 first dimensions could assess in- tergenerational body weight satisfaction, poor body esteem and eating disorders risk, and the last dimen- sion, overweight and obesity. METHODS Sample For this study, we assembled a sample of 615 mother-daughter pairs, who are part of the Birth to Twenty Plus cohort study (Bt20), a cohort study in- vestigating the determinants of health and develop- ment of South African families predominantly living in Soweto, Johannesburg.44 Data were collected on maternal attitudes and practices when their daugh- ters were 12 years of age (early adolescence), and data on the daughters’ attitudes and practices were collected at 18 years of age (late adolescence). Measures All instruments were measured in both mothers and daughters. Eating attitudes The participants’ risk of eating disorders was mea- sured using the Eating Attitudes Test-26 (EAT-26).45 This questionnaire includes 26 items in the form of a Likert scale and was administered by trained field workers. The individual items were recoded accord- ing to the coding developed by Garner et al.45 Total EAT-26 scores of 20 or greater indicate an increased risk of eating disorders, requiring clinical referral. The validity of this instrument has been reported in other South African study (Cronbach α ranged from 0.75 to 0.79).19 Body esteem scale An adapted questionnaire was used to determine the estimated attitude toward own body image.46 The instrument consisted of 21 questions, divided by components or subscales of body esteem, and in the form of a 5-point Likert scale ranging from options of “never” to “always.” These components included eliciting questions on the participants’ overall feelings toward their own overall body appearance (the appearance subscale: BES-AP), contentment with personal body weight (the weight subscale: BES-WE), and positive perception of personal body image (the attribution subscale: BES-AT). The internal consistency of the tool is high (Cronbach α ranged from 0.86 to 0.86). Body weight perceptions Furthermore, a body figures scale based on the Figure Rating Scale developed by Stunkard et al47 and adapted/validated in South African adolescent and adult females48 was used to determine body weight perceptions. The figures were coded 1 to 8 (underweight: 1-2, normal weight: 3-4, over- weight: 5-6, and obesity: 7-8) and shuffled be- fore being presented individually to participants for each interview.49 Participants were required to se- lect a body figure which they identified as repre- senting their current body size, as well as their ideal body size (IBS). A body weight self-satisfaction index was calculated by the discrepancy between current body size and IBS, which is termed the Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 142 Family and Community Health April–June 2019 ■ Volume 42 ■ Number 2 “Feel minus Ideal Discrepancy” (FID). Negative, zero, and positive scores indicated desire to be fat- ter (FID-gain weight), contentment with body size (FID-satisfied), and a desire to be thinner (FID- lose weight), respectively.11 We also created a body weight self-assessment index by the discrepancy be- tween BMI categories of participants and their cur- rent body size assessed by the Figure Rating Scale adapted to South African populations. Negative, zero, and positive scores indicated an overestima- tion, a correct estimation, and an underestimation of body size, respectively. Then, we created a val- orization of fatness index by the coding of all re- sponses on IBS variable of 5 and greater as an ap- preciation for overweight and obesity. In addition, we completed this assessment pro- tocol on body weight perceptions using a question also assessing weight change attempts (WCAs). Pre- cisely, we asked whether participants have already tried to lose weight by asking this question: “Have you already tried to lose weight during the past?” A positive response coded as “lose weight desire” means a WCA and a negative response coded as “no lose-weight desire” means an absence of WCA. Anthropometry Height was measured using a portable stadiometer (Holtain, Crymych, United Kingdom) and recorded to the nearest millimeter, and weight was measured to the nearest 100 g, using a digital weighing scale. Participants were measured in light clothing without shoes. Using height and weight mea- sures of mothers and daughters (18 years of age), BMI was calculated as weight (kg)/height (m)2. Body mass index categories included underweight (BMI <18.5 kg/m2), normal (BMI ≥18.5-24.9 kg/m2), overweight (BMI ≥25-29.9 kg/m2), obese (BMI ≥30-39.9 kg/m2), and morbidly obese (BMI ≥40 kg/m2). Statistical analysis Statistical analysis was performed using Stata ver- sion 12 (StataCorp, Texas). Stratifying by weight status of participants, we used chi-square and Fisher exact tests to evaluate the prevalence of under- weight, overweight, obesity, and eating disorders risk (EAT-26 scores ≥20), the frequency of WCA and FID scores (FID-gain weight, FID-satisfied, FID- lose weight), and analysis of variance/analysis of co- variance or paired t test to assess means of BMI, EAT-26 scores, the 3 body esteem dimensions (ap- pearance: BES-AP, attribution: BES-AT, and weight: BES-WE), and IBS. We also compared these vari- ables each other between mothers and daughters. We used the Pearson correlation and the Cohen κ: simple kappa (κ) and weighted kappa (κw) to corre- late and assess interrater agreement of BMI and all psychometric variables each other between mothers and daughters. Finally, we used binary logistic re- gression and multivariable linear regression models to identify the predictors (including mothers’ char- acteristics) of WCA (lose weight desire) and BMI in daughters, respectively. Ethics The study was approved by the Human Research Ethics Committee (Medical) of the University of the Witwatersrand. All participants gave written in- formed consent and minor assent was obtained if younger than 18 years before being involved in the study. RESULTS Most of the participants were of black South African ethnicity (85.9%), with whites, Indians, and participants of mixed ethnicity making up the re- mainder of the sample. Mothers were 41.4 ± 8.0 and daughters were 17.9 ± 0.39 years of age. BMI and prevalence The prevalence of underweight, normal weight, overweight, and obesity was 2.2%, 21.5%, 26.9%, and 49.4%, respectively, in mothers and 10.8%, 63.5%, 16.4%, and 9.3% in their daughters (P < .001). There was a significant BMI means difference between both groups: 30.4 ± 7.2 kg/m2 in mothers versus 23.0 ± 4.8 kg/m2 in daughters (P < .001). Mother’s and daughter’s body weight perceptions according to BMI status Ideal body size Mothers and their daughters had significantly dif- ferent IBS and body esteem dimensions according to their BMI status (Table 1). Although IBS in- creased along increasing BMI categories of mothers and daughters, it was situated in the normal BMI range on the Figure Rating Scale: figural stimuli 3 and 4. Indeed, the BMI/IBS correlation in mothers was r = 0.23 (P < .001), and in daughters, r = 0.25 (P < .001). Body esteem Concerning body esteem dimensions, normal- weight and overweight mothers had higher body es- teem than underweight and obese mothers, while the global body esteem was observed to decrease with increasing BMI in the daughters. Body weight self-assessment In addition, among all daughters, 23.8% overes- timated, 61.6% correctly estimated, and 14.6% underestimated their weight versus 4.7%, 35.7%, and 59.7% in all mothers (P < .001). Among underweight/normal-weight subjects, 28.5% of Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. E. Cohen et al. Relationship Between BMI and Body Image Disturbances 143 TABLE 1. Ideal Body Size and Body Esteem Scale Means According to Body Mass Index Underweight Normal Weight Overweight Obesity Pa Mothers Ideal body size 3.7 ± 1.6 3.9 ± 1.0 4.0 ± 0.9 4.4 ± 1.0 b BES appearance 6.6 ± 9.6 10.2 ± 7.1 9.8 ± 7.2 6.2 ± 9.6 b BES attribution 8.3 ± 2.1 9.8 ± 2.2 9.9 ± 2.1 8.9 ± 2.6 b BES weight 11.7 ± 7.3 14.7 ± 6.3 12.5 ± 7.9 9.3 ± 9.0 b Daughters Ideal body size 3.5 ± 1.1 3.7 ± 1.2 4.2 ± 1.2 4.4 ± 1.4 b BES appearance 9.5 ± 6.6 8.3 ± 8.2 5.4 ± 8.6 1.9 ± 10.0 b BES attribution 10.0 ± 1.9 9.7 ± 2.4 9.0 ± 2.6 8.4 ± 2.6 b BES weight 15.5 ± 5.6 14.3 ± 6.8 10.0 ± 9.0 6.2 ± 8.3 b Abbreviation: BES, body esteem scale. aAnalysis of variance between body mass index categories. bP < .001. daughters overestimated, 65.3% correctly esti- mated, and 6.2% underestimated their weight ver- sus 16.8%, 73.3%, and 9.9% in mothers (P < .05). Among overweight/obese subjects, 10.3% of daughters overestimated, 50.7% correctly esti- mated, and 39.0% underestimated their weight ver- sus 0.7%, 23.6%, and 75.7% in mothers (P < .001). Body weight self-satisfaction Table 2 shows a relative concordance between WCA (desire to lose weight) and FID-lose weight in both mothers and daughters. The desire to lose weight increased significantly with increasing BMI, while FID-gain weight decreased significantly with BMI. In addition, among underweight/normal- weight subjects, 22.6% of daughters had a desire to lose weight versus 9.2% in mothers (P < .001), while among overweight/obese subjects, 52.7% of daughters had a desire to lose weight versus 32.2% in mothers (P < .001). Among those subject who were underweight/normal weight, 24.1%, 51.4%, and 24.5% of daughters wanted to gain, maintain, TABLE 2. Weight Change Attempt, Eating Disorder Risk, and Feel-Ideal Discrepancy Frequencies According to Body Mass Index Underweight Normal Weight Overweight Obesity Pa Mothers Weight change attempt 20.0% (2) 8.3% (10) 27.9% (39) 34.5% (92) b Eating disorders risk 0% (0) 10.7% (12) 9.5% (12) 12.3% (31) NS FID-gain weight 55.6% (5) 33.9% (41) 7.6% (31) b FID-satisfied 22.2% (2) 48.8% (59) 28.0% (114) FID-lose weight 22.2% (2) 17.4% (21) 64.4% (262) Daughters Weight change attempt 9.7% (6) 24.8% (90) 41.9% (39) 71.7% (38) b Eating disorders risk 3.2% (2) 9.9% (36) 19.2% (18) 15.1% (8) c FID-gain weight 47.5% (29) 20.0% (71) 3.5% (5) b FID-satisfied 45.9% (28) 52.4% (186) 23.1% (33) FID-lose weight 6.6% (4) 27.6% (98) 73.4% (105) Abbreviations: FID, Feel minus Ideal Discrepancy; NS, not significant. aFisher exact test between body mass index categories. bP < .001. cP < .01. Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 144 Family and Community Health April–June 2019 ■ Volume 42 ■ Number 2 and lose weight, respectively, versus 35.4%, 46.9%, and 17.7% in mothers (P < .05). Conversely, among overweight/obese subjects, 26.6% and 73.4% of daughters wanted to gain/maintain (only when cate- gories aggregated) and lose weight, respectively, ver- sus 35.6% and 64.4% in mothers (P < .05). Finally, we noticed that 32.7% of mothers and 27.0% of daughters who desired to lose weight still valued overweight and obesity on the valorization of fat- ness index. Eating disorders risk The prevalence of eating disorders risk was 11.0% in mothers and 10.8% in daughters, respectively. It was not significantly associated with higher BMI categories in mothers, while in their daughters, eating disorders risk was significantly associated with increasing BMI (Table 2). Finally, among over- weight/obese subjects, 17.7% of daughters had an eating disorders risk versus 11.4% in mothers (P = .053; overweight daughters/mothers: 19.2% vs 9.5%, P < .05; obese daughters/mothers: 15.1% vs 12.3%, not significant). Relationship between mother’s and daughter’s BMI and body weight perceptions Differences between mothers and daughters In Table 3, after controlling for the confounding ef- fect of maternal BMI, it was observed that moth- ers had significantly lower EAT-26 score and higher BES-AP, BES-AT, and BES-WE scores than their daughters, while IBS scores were similar between mothers and their daughters. However, without controlling for maternal BMI, we observed a signif- icant difference between mothers and their daugh- ters for IBS: 4.1 ± 1.0 (in mothers) versus 3.8 ± 1.3 (in daughters) (P < .001). Similarities between mothers and daughters In addition, BMI (r = 0.29, P < .001), EAT-26 (r = 0.10, P < .05), BES-AP (r = 0.19, P < .001), BES-WE (r = 0.15, P < .001), and FID (r = 0.15, P < .01) were significantly correlated between mothers and daughters. Using Cohen κ to ac- curately match the response modalities from all psychometric variables between mothers and their daughters, we observed significant agreement for BMI categories (κ = 0.06, P < .001; κw = 0.14, P < .001), WCA (κ = 0.13, P < .01), BES-AP (κ = 0.03, not significant; κw = 0.12, P < .01), BES-WE (κ = 0.06, P < .05; κw = 0.13, P < .01), and FID (κ = 0.08, P < .01; κw = 0.15, P < .001). Then, we found a significant positive association between mothers and daughters on the valorization of fatness index (P < .01). This index was also asso- TABLE 3. Eating Attitude Test, Body Esteem Scale, and Ideal Body Size Mean Comparisons Between Mothers and Daughters Meana N Pa Eating attitude test Mothers 9.3 ± 8.4 468 Daughters 10.8 ± 8.4 468 b BES appearance Mothers 9.3 ± 9.2 503 Daughters 6.0 ± 9.2 503 c BES attribution Mothers 9.6 ± 2.7 501 Daughters 9.2 ± 2.7 501 d BES weight Mothers 13.0 ± 8.3 503 Daughters 11.3 ± 8.3 503 d Ideal body size Mothers 4.0 ± 1.1 496 Daughters 4.0 ±1.1 496 NS Abbreviations: BES, body esteem scale; NS, not significant. aAnalysis of covariance, adjusted by body mass index of mothers. bP < .05. cP < .001. dP < .01. ciated with overweight and obesity in both groups (mothers: P < .05; daughters: P < .001). Prediction of daughters’ weight change attempt and BMI The results of the binary logistic regression model (Table 4) showed an independent association be- tween mothers’ and their daughters’ desire to lose weight. Furthermore, an independent association between overweight daughters and their desire to lose weight was demonstrated. Univariate analyses showed an association between overweight moth- ers’ and their daughters’ desire to lose weight. The multivariable linear regression model (Table 5) showed an independent positive asso- ciation between mothers’ and daughters’ BMI. Univariate analyses showed a positive association between mothers’ desire to lose weight and daugh- ters’ BMI and a positive association between black ethnicity and daughters’ BMI. DISCUSSION This study analyzed the relationship between African mothers’ and their daughters’ body weight perceptions and management in a poor-urban obe- sogenic environment. As demonstrated in studies from various SSA countries, body image perceptions Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. E. Cohen et al. Relationship Between BMI and Body Image Disturbances 145 TABLE 4. Odds Ratio and 95% Confidence Intervals for Daughters Who Want to Lose Weight, Adjusted by Binomial Logistic Regression Analysis, and Compared With Subjects Who Do Not Wanta Daughters Lose Weight Desire Predictors Categories Odds Ratio 95% Confidence Intervals Ethnicity Nonblackb Black 1.6 0.8-3.3 Mothers age 1.0 1.0-1.0 Mothers’ BMIc No overweightb Overweight/obesity 1.2 0.7-2.0 Mothers lose weight desired Do not wantb Want 1.7 1.0-2.6f Daughters’ BMIe No overweightb Overweight/obesity 3.1 2.0-4.9g N = 488 Abbreviation: BMI, body mass index. aValues in boldface indicate binomial logit analysis significant effects (superscripts f and g): P < .05 and P < .001, respectively. bCategory taken as reference. c,d,eCrude analysis significant effect (P < .05; P < .01; and P < .001, respectively). differ across the age spectrum.9,25,50 In the present study, we observed that mothers and daughters had contrasting and contradictory body weight perceptions. The majority of mothers were over- weight or obese while most of the daughters were normal weight; however, the mothers valued fat- ness more than their daughters. This paradoxical trend between mothers and their daughters has also been observed in other African countries such as Nigeria.15,51 In the current study, mothers had better body esteem and eating attitudes than their daughters, despite most having a larger body weight. Over- weight and obese mothers had higher BES scores than their daughters, and most of them under- estimated their actual body weight.52,53 Mothers seemed to value larger body size because of the overweight/obese category having a significantly higher BES than the normal weight category. In addition, the BES, WCA, and FID scores for the mothers were positively associated with BMI. Mothers who were overweight/obese had a greater desire to lose weight and tended to develop poorer body esteem and higher risk of eating disorders than those who were lean.27,54 However, this trend was observed only with a higher BMI level (around obesity), which could be explained by their better body weight status awareness when diagnosed with TABLE 5. Multivariable Linear Regression to Assess the Factors Associated With Daughters’ BMIa Daughters’ BMI Predictors Coefficients 95% Confidence Intervals Black ethnicityb (0: others, 1: black) 1.21 − 0.04 2.46 Mothers’ age 0.01 − 0.05 0.06 Mothers’ BMIc 0.16 0.10 0.22d Mothers lose weight desireb (0: no desire, 1: desire) 0.73 − 0.20 1.66 Daughters lose weight desirec (0: no desire, 1: desire) 3.51 2.63 4.38d N = 488 Abbreviation: BMI, body mass index. aValues in boldface indicate multivariable linear regression model significant effects (superscript d): P < .001. b,cCrude analysis significant effect (P < .05 and P < .001, respectively). Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 146 Family and Community Health April–June 2019 ■ Volume 42 ■ Number 2 cardiometabolic diseases, as observed in urban Ghana.21 In comparison, daughters showed poorer body esteem than their mothers and were more exposed to eating disorders risk despite most were in nor- mal or underweight BMI categories, a trend in South African young women commented recently by Mchiza.55 Then, the second highest BES mean in daughters was in the underweight category af- ter normal weight. In addition, underweight and normal-weight daughters tended to overestimate their actual weight more than their mothers, and overweight daughters were more exposed to eat- ing disorders risk, as observed in another South African study.56 In addition, WCA and FID scores show that daughters desire to be underweight or normal weight while their mothers expected to be normal weight or overweight, an intergeneration contrast already observed in South Africa,57,58 as al- ready identified in other African countries.20,59 Un- like their mothers, daughters seemed to reject fat- ness for aesthetic reasons since most of them did not present with excess body weight, as also identified in other South African and Nigerian studies.15,19 Indeed, studies of African populations suggest that younger people are more influenced and ac- culturated by Western norms through media, which promotes thinness and could explain this young women depreciation toward fatness.60 Conversely, older people are more influenced by their traditional African culture valuing stoutness, which is per- ceived as a symbol of health, fertility, peacefulness, and prosperity, especially in married women.22,28 In this perspective, the prevalence of overweight and obesity in mothers was similar to those found in other South African studies and therefore strongly higher compared with the prevalence found in many other African and Western countries.61-63 For daughters, we observed that the prevalence of over- weight was lower than the national prevalence in adolescent females but the prevalence of obesity was slightly higher.42 Concerning eating disorders risk, we observed that its prevalence in daughters was between those showed by Le Grange et al64 and Gitau et al.24 The prevalence of eating disorders risk in both mothers and daughters was slightly higher than that found in the rest of the continent and some Western countries65,66 and slightly lower than that found in Nigeria.15 A relationship between mothers and their daugh- ters regarding body weight perceptions and BMI was also observed. First, after controlling for BMI, IBS of both groups was similar, situated in the nor- mal weight category, even though without BMI ad- justment, mothers had an IBS significantly higher than their daughters. We observed a positive cor- relation between BMI and IBS in both groups. This could be explained by a tendency in mothers and daughters to unconsciously “adjust” their IBS to their BMI.59 Since mothers were more overweight and obese than their daughters, this propensity had a strong effect on IBS mean difference. We also found significant correlations and/or agreements be- tween mothers and daughters for EAT-26 and body image variables (WCA, FID, BES-AP, and BES-WE), which could be an accidental convergence caused by 2 different behavioral patterns. Indeed, since the BMI of mothers and daughters was significantly different, most normal-weight daughters are living with mothers who are particu- larly overweight or obese, and a large proportion of them would like to lose weight. Therefore, daugh- ters faced with the reality of their mothers’ morbid BMI status67 could reject being overweight as a self- protective strategy to prevent their potential future weight gain caused by the influence of their mothers’ lifestyle choices, as highlighted in a South African study.40 This preventive reaction at home could also be exacerbated by norms conveyed by media and peers promoting thinness.68 In this specific context, an agreement between mothers and daughters on psychometric variables is understandable, as identi- fied in the binary logistic regression model. It seems that most of mothers do not want to be fat anymore while most of daughters do not want to become like them. This may be illustrated by the BES-AT disagreement between mothers and daughters demonstrating contrasted social attributions of the body, while we found agreements for BES-AP and BES-WE. The correlations and interrater agreements ex- pressed by κ values on BMI were also significant between mothers and daughters. A plausible expla- nation is the partial influence of mothers’ eating attitudes, body weight perceptions, and possibly also genetic inheritance on BMI of some daughters, already identified in previous studies,31,34,69 and probably expressed in the multivariable linear regression model. Since mothers’ BMI is a predictor of daughter’s BMI, the univariate association be- tween mothers’ desire to lose weight and daughters’ BMI, already identified in a previous study,70 could be explained by the association between mothers’ desire to lose weight and their own BMI. Mothers would transmit their own body weight appreciation to their daughters with consequences on their BMI. Beyond these BMI associations, we observed that interrater agreements were weak for BMI as well as psychometric variables. The correlation analyses cannot avoid a potential shift between variables, which does not affect the coefficient values, while κ statistics, more accurate than correlations, Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. E. Cohen et al. Relationship Between BMI and Body Image Disturbances 147 allow matching variable modalities with each other. The relative concordance between mothers and daughters for BMI and psychometric variables, besides the strong differences between both groups on these dimensions, might not be paradoxical but ambivalent, as explained by the possible dou- ble reaction of daughters toward their maternal environment. Some of the daughters (scenario 1) could be influenced by obesogenic mothers’ lifestyle as their relative tolerance for overweight and obesity,71 whereas other daughters (scenario 2), not particularly overweight, try to prevent this influence, exacerbated by an external societal pres- sure, and develop a poorer body esteem associated with the phobia of fatness, as already observed in HICs.33,72,73 At this stage, our study suggests that many daughters experience the second scenario because they are more exposed to eating disorders risk than overweight and obesity. However, the exposure to overweight/obesity in daughters is confirmed, supported by a BMI re- lationship between mothers and daughters, which could be exacerbated with older age of young women, especially by the decrease of peer and me- dia pressures and the beginning of conjugal life. Therefore, we recommend that public health poli- cies focus on the mother-daughter relationship to- ward body image and obesity to prevent eating dis- orders risk and obesity in adolescents and young adults. These policies should also consider the de- layed influence of maternal environment on the BMI of daughters as they age and transition into mar- riage. Finally, future South African studies should integrate simultaneously parental, media, and peer effects on adolescents and young adults to assess their respective influence on body image and eating attitudes.74 CONCLUSIONS The findings of this study demonstrate that despite the variation in body weight perception models be- tween mothers and their daughters, the daughters show lower body fatness appreciation than their mothers; overweight/obese mothers seem to exacer- bate their daughters’ desire to become thinner. The particularly high prevalence of cardiometabolic risk factors in the mothers and their relative tolerance for overweight and obesity could, on the one hand, overexpose daughters to obesity and, on the other hand, motivate many of them to excessive lower weight by exposing them to distorted body images of themselves and increased risk of eating disorders. Since body image disturbances/eating disorders and obesity conjointly have an intergenerational trans- mission from obese mothers to their daughters, South African public health policies should take into account this generational interaction to pre- vent eating disorders and obesity in adolescent and young adult women. REFERENCES 1. Gariepy G, Nitka D, Schmitz N. The association between obesity and anxiety disorders in the population: a system- atic review and meta-analysis. Int J Obes. 2010;34(3):407- 419. 2. Luppino FS, de Wit LM, Bouvy PF, et al. Overweight, obesity, and depression: a systematic review and meta- analysis of longitudinal studies. Arch Gen Psychiatry. 2010;67(3):220-229. 3. Puhl RM, Heuer CA. Obesity stigma: important consider- ations for public health. Am J Public Health. 2010;100(6): 1019-1028. 4. Dave D, Rashad I. Overweight status, self-perception, and suicidal behaviors among adolescents. Soc Sci Med. 2009;68(9):1685-1691. 5. Moore SE, Harris CL, Watson P, Wimberly Y. Do African American mothers accurately estimate their daughters’ weight category? Ethn Dis. 2008;18(2 suppl 2):211- 214. 6. Marques L, Alegria M, Becker AE, et al. Comparative prevalence, correlates of impairment, and service utiliza- tion for eating disorders across US ethnic groups: Impli- cations for reducing ethnic disparities in health care ac- cess for eating disorders. Int J Eat Disord. 2011;44(5):412- 420. 7. Rogers Wood NA, Petrie TA. Body dissatisfaction, ethnic identity, and disordered eating among African American women. J Couns Psychol. 2010;57(2):141. 8. Solmi F, Hatch SL, Hotopf M, Treasure J, Micali N. Preva- lence and correlates of disordered eating in a general population sample: the South East London Community Health (SELCoH) study. Soc Psychiatry Psychiatr Epi- demiol. 2014;49(8):1335-1346. 9. Cohen E, Amougou N, Ponty A, et al. Nutrition transition and biocultural determinants of obesity among Cameroo- nian migrants in urban Cameroon and France. Int J Environ Res Public Health. 2017;14:pii: E696. 10. Flynn K, Fitzgibbon BM. Body images and obesity risk among black females: a review of literature. Ann Behav Med. 1998;20:13-24. 11. Gradidge PJL, Norris SA, Micklesfield LK, Crowther NJ. The role of lifestyle and psycho-social factors in predict- ing changes in body composition in black South African women. PLoS One. 2015;10(7):e0132914. 12. Puoane T, Fourie JM, Shapiro M, Rosling L, Tshaka NC, Oelefse A. “Big is beautiful”—an exploration with urban black community health workers in a South African town- ship. South Afr J Clin Nutr. 2005;18(1):6-15. 13. Thomas S, Ness RB, Thurston RC, Matthews K, Chang CC, Hess R. Racial differences in perception of healthy body weight in mid-life women: results from Do Stage Transitions Result in Detectable Effects (STRIDE) Study. Menopause. 2013;20(3):269-273. 14. Cachelin FM, Rebeck RM, Chung GH, Pelayo E. Does ethnicity influence body-size preference? A comparison of body image and body size. Obes Res. 2002;10(3):158- 166. 15. Fadipe B, Oyelohunnu MA, Olagunju AT, Aina OF, Akinbode AA, Suleiman TF. Disordered eating atti- tudes: demographic and clinico-anthropometric correlates among a sample of Nigerian students. Afr Health Sci. 2017;17(2):513-523. Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. 148 Family and Community Health April–June 2019 ■ Volume 42 ■ Number 2 16. Jones LA, Cook-Cottone C. Media and cultural influences in African-American girls’ eating disorder risk. ISRN Prev Med. 2013;2013:319701. 17. Toselli S, Rinaldo N, Gualdi-Russo E. Body image per- ception of African immigrants in Europe. Glob Health. 2016;12(1):48. 18. Miller MN, Pumariega AJ. Culture and eating disorders: a historical and cross-cultural review. Psychiatry Interpers Biol Process. 2001;64(2):93-110. 19. Szabo CP, Allwood CW. A cross-cultural study of eating attitudes in adolescent South African females. World Psy- chiatry. 2004;3(1):41-44. 20. Appiah CA, Otoo GE, Steiner-Asiedu M. Preferred body size in urban Ghanaian women: implication on the over- weight/obesity problem. Pan Afr Med J. 2016;23:239. 21. Benkeser RM, Biritwum R, Hill AG. Prevalence of over- weight and obesity and perception of healthy and desir- able body size in urban, Ghanaian women. Ghana Med J. 2012;46(2):66-75. 22. Cohen E, Boëtsch G, Palstra FP, Pasquet P. Social val- orisation of stoutness as a determinant of obesity in the context of nutritional transition in Cameroon: the Bamileke case. Soc Sci Med. 2013;96:24-32. 23. Malhotra R, Hoyo C, Ostbye T, et al. Determinants of obe- sity in an urban township of South Africa. South Afr J Clin Nutr. 2008;21(4):315-320. 24. Gitau TM, Micklesfield LK, Pettifor JM, Norris SA. Changes in eating attitudes, body esteem and weight control be- haviours during adolescence in a South African cohort. PLoS One. 2014;9(10):109709. 25. Okop KJ, Mukumbang FC, Mathole T, Levitt N, Puoane T. Perceptions of body size, obesity threat and the willing- ness to lose weight among black South African adults: a qualitative study. BMC Public Health. 2016;16(1):365. 26. Gitau TM, Micklesfield LK, Pettifor JM, Norris SA. Ethnic differences in eating attitudes, body image and self-esteem among adolescent females living in urban. South Afr J Psy- chiatry. 2014;17:468-474. 27. Mchiza ZJ, Parker WA, Makoae M, Sewpaul R, Ku- pamupindi T, Labadarios D. Body image and weight con- trol in South Africans 15 years or older: SANHANES-1. BMC Public Health. 2015;15(1):992. 28. Mvo Z, Dick J, Steyn K. Perceptions of overweight African women about acceptable body size of women and chil- dren. Curationis. 1999;22(2):27-31. 29. Dinkel D, Snyder K, Kyvelidou A, Molfese V. He’s just con- tent to sit: a qualitative study of mothers’ perceptions of infant obesity and physical activity. BMC Public Health. 2017;17(1):585. 30. Hong SA, Peltzer K, Jalayondeja C. Parental misperception of child’s weight and related factors within family norms [published online ahead of print May 22, 2017]. Eat Weight Disord. doi:10.1007/s40519-017-0399-4. 31. Kaufman-Shriqui V, Fraser D, Novack Y, et al. Maternal weight misperceptions and smoking are associated with overweight and obesity in low SES preschoolers. Eur J Clin Nutr. 2012;66(2):216-223. 32. Tschamler JM, Conn KM, Cook SR, Halterman JS. Under- estimation of children’s weight status: views of parents in an urban community. Clin Pediatr (Phila). 2010;49(5):470- 476. 33. Berge JM, Wall M, Loth K, Neumark-Sztainer D. Parent- ing style as a predictor of adolescent weight and weight- related behaviors. J Adolesc Health. 2010;46(4):331-338. 34. Cutting TM, Fisher JO, Grimm-Thomas K, Birch LL. Like mother, like daughter: familial patterns of overweight are mediated by mothers’ dietary disinhibition. Am J Clin Nutr. 1999;69(4):608-613. 35. McDonald CM, Baylin A, Arsenault JE, Mora-Plazas M, Vil- lamor E. Overweight is more prevalent than stunting and is associated with socioeconomic status, maternal obesity, and a snacking dietary pattern in school children from. J Nutr. 2009;139(2):370-376. 36. Martinson LE, Esposito-Smythers C, Blalock DV. The ef- fects of parental mental health and social-emotional cop- ing on adolescent eating disorder attitudes and behaviors. J Adolesc. 2016;52:154-161. 37. Berge JM, MacLehose R, Loth KA, Eisenberg M, Buc- chianeri MM, Neumark-Sztainer D. Parent conversations about healthful eating and weight: associations with adolescent disordered eating behaviors. JAMA Pediatr. 2013;167(8):746-753. 38. Elfhag K, Linné Y. Gender differences in associations of eating pathology between mothers and their adolescent offspring. Obesity. 2005;13(6):1070-1076. 39. Flynn K, Fitzgibbon M. Body image ideals of low-income African American mothers and their preadolescent daugh- ters. J Youth Adolesc. 1996;25(5):615-630. 40. Mchiza ZJ, Goedecke JH, Lambert EV. Intra-familial and ethnic effects on attitudinal and perceptual body image: a cohort of South African mother-daughter dyads. BMC Public Health. 2011;11(1):433. 41. Arroyo A, Segrin C, Andersen KK. Intergenerational trans- mission of disordered eating: direct and indirect mater- nal communication among grandmothers, mothers, and daughters. Body Image. 2017;20:107-115. 42. Reddy SP, Resnicow K, James S, et al. Rapid increases in overweight and obesity among South African adoles- cents: comparison of data from the South African National Youth Risk Behaviour Survey in 2002 and 2008. Am J Pub- lic Health. 2012;102(2):262-268. 43. Alaba O, Chola L. Socioeconomic inequalities in adult obe- sity prevalence in South Africa: a decomposition analysis. Int J Environ Res Public Health. 2014;11(3):3387-3406. 44. Richter LM, Victora CG, Hallal PC, et al. Cohort profile: the consortium of health-orientated research in transition- ing societies. Int J Epidemiol. 2011;41(3):621-626. 45. Garner DM, Olmsted MP, Bohr Y, Garfinkel PE. The eat- ing attitudes test: psychometric features and clinical cor- relates. Psychol Med. 1982;12(4):871-878. 46. Mendelson BK, Mendelson MJ, White DR. Body-esteem scale for adolescents and adults. J Pers Assess. 2001; 76(1):90-106. 47. Stunkard AJ, Sorenson T, Schlusinger T. Use of the Dan- ish adoption registers for the study of obesity. In: Kelly SS, Rowland LP, Sidman RL, Matthys SW, eds. Genet- ics of Neurological and Psychiatric Disorders. New York, NY: Raven Press; 1983:115-129. 48. Mchiza ZJ, Goedecke JH, Steyn NP, et al. Development and validation of instruments measuring body image and body weight dissatisfaction in South African mothers and their daughters. Public Health Nutr. 2005;8(5):509-519. 49. Matoti-Mvalo T, Puoane TB. Perceptions of body size and its association with HIV/AIDS. South Afr J Clin Nutr. 2011;24(1):40-45. 50. Gitau TM, Micklesfield LK, Pettifor JM, Norris SA. Eating at- titudes, body image satisfaction and self-esteem of South African Black and White male adolescents and their per- ception of female body silhouettes. J Child Adolesc Ment Health. 2014;26(3):193-205. 51. Okoro EO, Oyejola BA, Etebu EN, et al. Body size prefer- ence among Yoruba in three Nigerian communities. Eat- ing and Weight Disorders-Studies on Anorexia. Eat Weight Disord. 2014;19(1):77-88. 52. Ettarh R, Van de Vijver S, Oti S, Kyobutungi C. Overweight, obesity, and perception of body image among slum Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited. E. Cohen et al. Relationship Between BMI and Body Image Disturbances 149 residents in Nairobi, Kenya, 2008-2009. Prev Chronic Dis. 2013;10:130198. 53. Phetla MC, Skaal L. Perceptions of healthcare profession- als regarding their own body weight in selected public hos- pitals in Mpumalanga Province, South Africa. S Afr Med J. 2017;107(4):338-341. 54. Faber M, Kruger HS. Dietary intake, perceptions regarding body weight, and attitudes toward weight control of nor- mal weight, overweight, and obese black females in a rural village in South Africa. Ethn Dis. 2005;15(2):238-245. 55. Mchiza ZJ. Eating disorders in South African schools: a public health crisis that needs immediate intervention. South Afr J Clin Nutr. 2014;27(4):185-186. 56. Senekal M, Lasker GL, van Velden L, Laubscher R, Tem- ple NJ. Weight-loss strategies of South African female uni- versity students and comparison of weight management- related characteristics between dieters and non-dieters. BMC Public Health. 2016;16(1):918. 57. Draper CE, Davidowitz KJ, Goedecke JH. Perceptions re- lating to body size, weight loss and weight-loss interven- tions in black South African women: a qualitative study. Public Health Nutr. 2016;19(3):548-556. 58. Pedro TM, Micklesfield LK, Kahn K, Tollman SM, Pettifor JM, Norris SA. Body image satisfaction, eating attitudes and perceptions of female body silhouettes in rural South African adolescents. PLoS One. 2016;11(5):e0154784. 59. Maruf FA, Akinpelu AO, Nwankwo MJ. Perceived body image and weight: discrepancies and gender differ- ences among University undergraduates. Afr Health Sci. 2012;12(4):464-472. 60. Sow AM. Médecine interne et image de soi: canons de beauté ou indicateurs de santé? Médecine Trop. 2002;62(6):577-581. 61. Abrahams Z, Mchiza Z, Steyn NP. Diet and mortality rates in Sub-Saharan Africa: stages in the nutrition transition. BMC Public Health. 2011;11(1):801. 62. Macia E, Cohen E, Gueye L, Boetsch G, Duboz P. Preva- lence of obesity and body size perceptions in urban and rural Senegal: new insight on the epidemiological transi- tion in West Africa. Cardiovasc J Afr. 2017;28(5):324-330. 63. Ng M, Fleming T, Robinson M, et al. Global, regional, and national prevalence of overweight and obesity in chil- dren and adults during 1980-2013: a systematic analy- sis for the Global Burden of Disease Study. Lancet. 2014; 384(9945):766-781. 64. Le Grange D, Louw J, Russell B, Nel T, Silkstone C. Eating attitudes and behaviours in South African adolescents and young adults. Transcult Psychiatry. 2006;43(3):401-417. 65. Smink FR, Van Hoeken D, Hoek HW. Epidemiology of eat- ing disorders: incidence, prevalence and mortality rates. Curr Psychiatry Rep. 2012;14(4):406-414. 66. van Hoeken D, Burns JK, Hoek HW. Epidemiology of eat- ing disorders in Africa. Curr Opin Psychiatry. 2016;29:372- 377. 67. Case A, Menendez A. Sex differences in obesity rates in poor countries: evidence from South Africa. Econ Hum Biol. 2009;7(3):271-282. 68. Carney T, Louw J. Eating disordered behaviors and me- dia exposure. Soc Psychiatry Psychiatr Epidemiol. 2006; 41(12):957-966. 69. Phillips EA, Comeau DL, Pisa PT, Stein AD, Norris SA. Perceptions of diet, physical activity, and obesity-related health among black daughter-mother pairs in Soweto, South Africa: a qualitative study. BMC Public Health. 2016; 16(1):750. 70. Duchin O, Marin C, Mora-Plazas M, Villamor E. Maternal body image dissatisfaction and BMI change in school-age children. Public Health Nutr. 2016;19(2):287-292. 71. Styles JL, Meier A, Sutherland LA, Campbell MK. Par- ents’ and caregivers’ concerns about obesity in young chil- dren: a qualitative study. Fam Community Health. 2007; 30(4):279-295. 72. Diedrichs PC, Atkinson MJ, Garbett KM, et al. Evaluating a Website designed to improve body image and psychoso- cial well-being among adolescent girls and their mothers: a cluster randomised controlled trial with mother-daughter dyads. J Adolesc Health. 2017;60(2):S6. 73. van den Berg PA, Keery H, Eisenberg M, Neumark- Sztainer D. Maternal and adolescent report of mothers’ weight-related concerns and behaviors: longitudinal asso- ciations with adolescent body dissatisfaction and weight control practices. J Pediatr Psychol. 2010;35(10):1093- 1102. 74. Aliyev B, Türkmen A. Parent, peer and media effect on the perception of body image in preadolescent girls and boys. Univers J Psychol. 2014;2(7):224-230. Copyright © 2019 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.