Disordered Eating Behaviors: What About Boys?

Journal of Adolescent Health 44 (2009) 111–117

Original article

Disordered Eating Behaviors: What About Boys? Franc¸oise Domine´, M.D., Andre´ Berchtold, Ph.D., Christina Akre´, M.A., Pierre-Andre´ Michaud, M.D., and Joan-Carles Suris, M.D., M.P.H., Ph.D.* Research Group on Adolescent Health, Institute of Social and Preventive Medicine, University of Lausanne, Lausanne, Switzerland Manuscript received April 22, 2008; manuscript accepted July 31, 2008

See Editorial p. 99 Abstract

Purpose: To determine the characteristics specific to boys with disordered eating behaviors (DEB) and the general context in which these DEB occur. Method: Data were drawn from the SMASH02 database, a survey carried out among post–mandatory school students in Switzerland aged 16–20 years in 2002. Only males (N ¼ 3890) were included, and were classified into into one of four groups based on their level of concern about weight/food and on their eating behaviors, as follows: group 1: one concern without behavior (N ¼ 862); group 2: more than one concern without behavior (N ¼ 361); group 3: at least one behavior (N ¼ 798); and a control group (N ¼ 1869), according to previously validated items. Groups were compared for personal, family, school, experience of violence, and health-compromising behaviors variables on the bivariate level. All significant variables were included in a multinomial logistic regression using Stata 9 software. Results: About one-half of the boys reported either a concern or unhealthy eating behavior. Compared with the control group, boys from the three groups were more likely to be students and to report a history of sexual abuse, delinquency, depression, and feeling fat. In addition, boys from group 3 were more likely to report a history of dieting, early puberty, peer teasing, having experienced violence, frequent inebriation, and being overweight. Conclusion: DEB concern adolescent males more frequently than thought and seem to be integrated in a general dysfunctional context, in which violence is predominant. Adolescent males also need to be screened for DEB. Moreover, prevention programs should target the increasing social and media pressure regarding boys ideal body shape and raise public consciousness about this phenomenon. Ó 2009 Society for Adolescent Medicine. All rights reserved.

Keywords:

Eating behavior; Adolescent behavior; Male; Violence; Co-morbidity; Switzerland

Although anorexia nervosa is common mostly in girls, occurring in 10 or more girls in proportion to one boy, bulimia nervosa and atypical eating disorders (ED) could be underestimated among boys. Some authors estimate that 5 to 15% of cases of anorexia nervosa and bulimia nervosa, and 40% of binge eating disorder cases affect boys [1]. Data for the United States published in 2000 indicate that 20% of boys reported dieting at the moment of the survey and 7% reported purging or binge eating episodes (4% once a day) [2]. *Address correspondence to: Joan-Carles Suris, M.D., M.P.H., Ph.D., Research Group on Adolescent Health, Institute of Social and Preventive Medicine, Bugnon 17, 1005-Lausanne, Switzerland. Fax: þ41 21 314 7244. E-mail address: [email protected]

However, most research about ED in adolescents has studied girls, and very few studies have focused exclusively on boys [3–8]. Adolescence is a critical period for these pathologic conditions, because of pubertal changes and psychological turmoil [9]. Some adolescents adapt to their body changes with difficulty, sometimes perceiving a distortion of their body and experiencing low self-esteem [4,10]. Body image is known to be a crucial predictive factor for dieting [11], which increases the risk of ED [11,12]. Behavior changes are also characteristic of that period: for many teenagers, engaging in some new exploratory behaviors is a required transition stage toward adult status [13]. Jessor’s theory of problem behavior postulates that the clustering of health-compromising

1054-139X/09/$ – see front matter Ó 2009 Society for Adolescent Medicine. All rights reserved. doi:10.1016/j.jadohealth.2008.07.019

112

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117

behaviors in adolescents could be the result of interactions between their personal characteristics and their environment [14]. Some studies have shown the association between ED and other health-compromising behaviors [15], but ED are also considered as risk behaviors or coping strategies in a context of more general lack of well-being [13,16]. Engaging in some health-compromising behaviors, like ED, is mainly a consequence of the interaction of several biographical and sociocultural factors [4,10,11,17], as well as some psychological traits such as depression [4,10,17]. Variations of pubertal timing could increase the risk of engaging precociously in such behaviors, although the link between puberty and ED has been widely studied and contested. Pubertal timing has been studied mainly in the case of girls because of the ease of menarche age evaluation [18]. Regarding substance use, the association between ED and smoking as a weight-control method is unclear [19], whereas alcohol misuse [20] or dependency [21], together with cannabis use [21], seem to be associated with ED [22]. Again, most of these observations were made among girls, and, as Ackard et al [3] highlighted in a recent paper, little attention is paid to the characteristics of ED proper among adolescent boys. The aim of this exploratory research is to determine factors specifically associated to disordered eating behaviors (DEB) in adolescent boys. We hypothesize that the characteristics of boys and girls affected by DEB present dissimilarities and could more often be associated with other healthcompromising behaviors, such as substance use, unsafe sex, or suicidal behaviors, among boys. Methods Data were drawn from the 2002 Swiss Multicenter Adolescent Survey on Health (SMASH02) data base, a survey including 7548 (3890 males, 51.5%) adolescents in post– mandatory school aged 16–20 years. In Switzerland school is mandatory up to the age 16 years. Afterward, about 30% of adolescents follow high school (students), 60% vocational school (apprentices, who have 1 or 2 days of class per week and spend the rest of the time working in a company related to their field of study), and 10% either do not continue or delay their education. The survey was carried out through an anonymous self-administered questionnaire that was filled out in the classroom and was approved by the ethical committee of the Medicine Faculty in Lausanne. A description of the questionnaire and sampling method has been published elsewhere [23]. Dependent variable We separated surveyed boys (n ¼ 3890) into four groups, based on a set of six items related to eating concerns and behaviors. Although a single question does not allow one to diagnose DEB, reported thoughts or conducts linked to food can enable to identify at-risk subjects [24]. We used four items from the Weight and Eating Concerns Inventory (WECI) [25], a validated evaluation scale, worded as

follows: ‘‘For the last few months, has it happened to you. 1) to be afraid of putting on weight?, 2) to feel unattractive when you have eaten too much?, 3) to often think about anything related to food?, 4) to love the feeling of having an empty stomach?’’ This set of questions have proved its efficacy in previous studies as a good indicator of DEB (25). Two items were added, enabling the screening of some health-compromising weight-control methods, often reflecting pathologic conditions such as bulimia nervosa or binge eating: ‘‘For the last few months, has it happened to you . 5) to eat a lot and to have difficulty to stop?, 6) to make yourself vomit?’’ The possible answers to the six items were ‘‘never,’’ ‘‘once a week,’’ ‘‘several times a week,’’ and ‘‘every day.’’ Every item’s answer was dichotomized into at least once a week (including any answer but never) and never. Subjects who answered at least once a week to one of the first four items but not to the last two items (group 1) were called ‘‘1 concern without behavior’’ (n ¼ 862, 22.16%); subjects who answered at least once a week to at least two of the first four items but not to the last two items (group 2) were included in the group with ‘‘>1 concern without behavior’’ (n ¼ 361, 9.28%); subjects who answered at least once a week to any of the last two items, with or without positive responses to the first four items (group 3), were included in the group with ‘‘at least 1 behavior’’ (n ¼ 798, 20.51%); and subjects who answered ‘‘never’’ to all items were included in the control group (n ¼ 1869, 48.05%). Independent variables Personal variables included age, body dissatisfaction (yes/ no), having ever been on a diet (yes/no), feeling fat (yes/no), body mass index (BMI), pubertal timing, depression, and history of suicide attempt (yes/no). BMI was calculated from self-reported height and weight. The German percentile curves for BMI used in Switzerland [26] defined the cutoffs for leanness (<10th percentile), average weight (10th and  90th percentile) and overweight (>90th percentile). Self-assessment of pubertal timing was determined by the question ‘‘If you think about the age you began puberty, compared with your same age peers, would you say that you were.’’ with five possible answers: ‘‘very advanced / quite advanced / like them / quite late / very late.’’ Subjects who answered ‘‘very or quite advanced’’ were defined as ‘‘early pubertal timing’’; those who answered ‘‘like them’’ were defined as ‘‘average pubertal timing’’; and those who answered ‘‘quite or very late’’ were defined as ‘‘late pubertal timing.’’ This approach has showed a 94% correlation rate with physician evaluations in previous studies [27]. To measure depression we used the Depressive Tendencies Scale, an instrument proved to be valid and reliable [28]. The scale includes eight items and covers depressed symptomatology, feelings of sadness, hopelessness, and unhappiness. In the present study Cronbach’s alpha was r ¼ .89. Family variables included family structure (parents living together or not), parents’ socioeconomic status (SES) and

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117

parent–adolescent relationship. SES was assessed using father’s and mother’s education level (mandatory schooling or lower versus higher education) as a proxy. We considered a low SES when both parents had a low education level (mandatory school or less). To assess the parents-adolescent relationship, a six-item inventory was used. Five items were taken from the Inventory of Parent and Peer Attachment [29]. These items tapped adolescents’ perceptions of their parents’ acceptance, understanding, trustfulness, and sensitivity to their emotional states, as well as their own use of their parents as confidants. In addition, an item was created tapping adolescents’ perception of how much their parents trusted them. Overall, Cronbach’s alpha in the present study was r ¼ .85. School variables included academic track (apprentice/student), school context, and relationship with peers. School context was measured with five items used in previous studies [30]. These items tapped adolescents’ perception of their general school life and school connectedness. Cronbach’s alpha for the present study was r ¼ .61. To measure the quality of the relationship with peers, we used a four-item inventory. We took all four items from the Inventory of Parent and Peer Attachment [29]. These items tapped adolescents’ perceptions of peer acceptance, trustfulness, and sensitivity to their emotional states, as well as their use of peers as confidants. Overall, Cronbach’s alpha in the present study was r ¼ .77. Previous studies have shown that peer teasing is often associated with low self-esteem and self-devaluation, which are closely related to ED [4]. We assessed peer teasing with the question ‘‘During the last 12 months, did your peers laugh at you or insult you?’’ and we considered as positive the subjects who answered ‘‘About once a week or more.’’ Violence included experienced violence (having been a victim of physical violence, robbery, and/or racket at least once in the past 12 months), delinquency (having committed at least one of the following in the previous 12 months: attacking an adult, stealing with violence, carrying a weapon, using a weapon in a fight, vandalism, burglary, setting fire to something) and a history of sexual abuse (yes/no). Health-compromising behaviors included smoking (no/ occasionally/regularly), inebriation in the past 30 days (no / 1 or 2 times / 3 times or more), cannabis use in the past 30 days (no / 1 or 2 times / 3 times or more), use of other illegal drugs in the previous 30 days (no / at least once), and unsafe sexual practices. Unsafe sexual practices were defined by two or more positive answers to the following: having had sex before 15 years of age, having had more than three partners in their lifetime, not having used a condom at last intercourse, and having made a partner pregnant.

113

Chi-square tests and analysis of variance to compare each individual variable between the four groups. Results are given as point prevalence and means. In a second step, all statistically significant variables (p < .05) at the individual level were included in a multinomial regression analysis. Results are given as adjusted relative risk ratios (RRR) and 95% confidence interval (CI). Results Table 1 shows the associations between assessed factors and boys’ eating concerns/disturbed behaviors. Factors for which associations strengthen the level of concern were poor relationship with parents, experience of violence, and a history of sexual abuse, as well as academic track (student), early pubertal timing, depression, suicide attempt, peer teasing, frequent inebriation, body dissatisfaction, feeling fat, and a history of dieting. The three groups also report significantly more delinquency than the control group. Regarding the BMI distribution, the overweight/obese group significantly increased with the DEB level. The results of the multinomial regression are reported in Table 2. The differences between boys in group 1 and those in the control group are that the former are more likely to be students (RRR ¼ 1.29 [1.01–1.64]) and to report depression (RRR ¼ 1.30 [1.03–1.64]), a history of sexual abuse (RRR ¼ 3.47 [1.18–10.17]), a delinquent behavior (RRR ¼ 1.78 [1.29–2.46]) and feeling fat (RRR ¼ 1.73 [1.23–2.43]. All factors associated to group 1 are also associated to group 2. However, boys in group 2 also report more often a poor relationship with parents (RRR ¼ 1.46 [1.12–1.89]), an early puberty (RRR ¼ 1.50 [1.04–2.15]) and are less likely to report frequent cannabis use in the past 30 days (RRR ¼ 0.67 [0.45–0.99]). A stronger association with a history of sexual abuse than in group 1 is also noted (RRR ¼ 5.02 [1.64– 15.37]). By contrast, the number of factors associated to disturbed eating behaviors (group 3) increases dramatically. Boys from group 3 report the same factors than group 2, except that there is no difference for cannabis use. However, in addition, they are more likely to report peer teasing (RRR ¼ 1.56 [1.06–2.27]), a history of dieting (RRR ¼ 2.77 [1.72– 4.47]) and three or more episodes of inebriation in the past 30 days (RRR ¼ 1.60 [1.09–2.36]). The violent context is even more emphasized in this Group, with experience of violence (RRR ¼ 1.33 [1.01–1.74]), which adds to delinquent behaviors (RRR ¼ 1.43 [1.10–1.87]) and a history of sexual abuse (RRR ¼ 3.67 [1.34–10.05]). These boys are also more likely to be overweight or obese (RRR ¼ 1.76 [1.16–2.66]). Discussion

Statistical analysis Statistical analysis were performed with Stata 9.2 software, which allows computing coefficient estimates and variances taking into account the sampling weights, clustering, and stratification procedure. In a first step, we used

According to our classification, more than half of the surveyed adolescent boys reported at least one concern, and about 20% presented with at least one unhealthy eating behavior. Moreover, it is important to note that one of every 10 boys in the control group was dissatisfied with his body.

114

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117

Table 1 Main characteristics of the four subgroups reported as percentages (95% confidence intervals [CIs]) Characteristic

Distribution (%) Mean ages (CI) Personal variables BMI <10 percentile 10–90 percentile (reference category) >90 percentile Pubertal timing Early Average (reference category) Late Depression [scale 0 / 3]* Suicide attempt Body dissatisfaction Feeling fat History of dieting Familial variables Living with both parents Low educational level of both parents Poor relationship with parents [ scale 1 / 4]* School variables Academic track: student School Context [scale 1 / 4]* Poor relationship with peers Peer teasing at least once a week Violence Experienced violence in past 12 months Delinquency History of sexual abuse Health-compromising behaviors Unsafe sex Substances use Smoking No (reference category) Occasionally Regularly Had a drunkenness in the past 30 days No (reference category) 1–2 times 3 times Cannabis use in past 30 days None (reference category) 1–2 times 3 times Illicit drug use in the past 30 days

1 Concern w/out behavior

>1 Concern w/out behavior

1 Behavior

(n ¼ 1869)

(n ¼ 862)

(n ¼ 361)

(n ¼ 798)

Control group

Group 1

Group 2

Group 3

48.05% 17.94 [17.82–18.05]

22.16% 17.96 [17.80–18.11]

9.28% 18.04 [17.86–18.21]

20.51% 17.99 [17.88–18.11]

5.4% [4.2–7.0] 89.7% [87.7–91.3] 4.9% [3.9–6.2]

3.8% [2.6–5.7] 90.5% [87.9–92.7] 5.6% [4.0–7.8]

2.0% [1.0–4.0] 81.0% [73.6–86.7] 17.0% [11.5–24.5]

3.0% [1.9–4.8] 78.7% [74.1–82.7] 18.3% [14.4–22.8]

20.9% [18.4–23.5] 64.0% [60.3–67.6] 15.1% [12.4–18.4] .48 [.44–.53] 2.1% [1.4–3.2] 10.6% [7.4–15.0] 11.4% [9.7–13.4] 2.9% [2.1–4.0]

24.2% [20.8–28.1] 59.4% [54.8–63.9] 16.3% [13.6–19.5] .59 [.53–.65] 3.4% [2.2–5.3] 12.3% [9.5–15.9] 17.7% [14.7–21.2] 3.2% [2.2–4.7]

31.5% [26.1–37.5] 53.1% [46.6–59.4] 15.4% [18.6–22.5] .72 [.63–.80] 3.6% [1.9–6.8] 20.5% [16.0–25.9] 44.0% [37.5–50.7] 10.2% [7.0–14.2]

31.3% [26.9–36.2] 50.1% [45.4–54.9] 18.6% [15.2–22.5] .78 [.71–.84] 5.7% [3.8–8.4] 21.7% [18.0–25.8] 36.4% [32.0–41.1] 15.1% [11.8–19.0]

.006 .004 .000 .000 .000

80.0% [77.3–82.4] 17.2% [14.9–19.9] 1.60 [1.56–1.64]

75.3% [71.2–79.0] 15.8% [12.5–19.8] 1.70 [1.64–1.76]

78.0% [72.3–82.9] 20.8% [15.1–28.0] 1.81 [1.73–1.89]

73.8% [69.2–77.9] 16.2% [13.0–19.1] 1.77 [1.68–1.86]

.039 .494 .013

19.1% [17.0–21.4] 2.10 [2.06–2.15] 1.9% [1.1–3.4] 10.5% [7.9–13.7]

21.9% [18.9–25.2] 2.08 [2.02–2.13] 1.4% [.8–2.6] 14.6% [11.7–18.1]

22.6% [18.1–27.8] 2.12 [1.95–2.29] 3.3% [1.6–6.7] 18.0% [12.7–24.9]

25.4% [22.1–28.9] 2.19 [2.13–2.25] 2.8% [1.7–4.3] 21.5% [17.7–25.9]

.015 .462 .274 .000

20.1% [17.6–22.8] 42.6% [38.8–46.7] .47% [.17–1.27]

24.6% [21.0–28.6] 59.1% [54.5–63.7] 2.18% [1.24–3.81]

26.6% [20.6–33.7] 57.1% [50.2–63.7] 3.45% [1.69–6.93]

32.6% [28.0–37.5] 58.8% [54.2–63.3] 3.52% [2.18–5.66]

.000 .000 .000

12.8% [10.9–15.0]

14.5% [9.8–20.9]

16.5% [11.2–23.7]

15.8% [12.6–19.8]

.532

52.7% [48.6–56.7] 15.1% [11.7–19.2] 32.3% [29.0–35.7]

48.5% [43.6–53.5] 12.1% [9.7–15.1] 39.3% [34.2–44.8]

53.7% [46.9–60.4] 18.7% [14.0–24.4] 27.6% [21.9–34.2]

49.4% [44.6–54.1] 16.3% [12.8–20.4] 34.4% [30.1–39.0]

.087

64.5% [61.0–67.9] 25.5% [22.7–28.5] 10.0% [8.3–12.0]

57.0% [52.2–61.8] 27.2% [23.4–31.4] 15.8% [12.6–19.5]

56.8% [50.1–63.2] 29.0% [23.6–35.0] 14.2% [10.3–19.4]

56.3% [51.6–61.0] 25.0% [21.5–28.9] 18.7% [14.8–23.3]

62.8% [58.8–66.6] 13.4% [10.2–17.4] 23.8% [21.1–26.8] 9.0% [7.2–11.1]

55.4% [50.6–60.2] 11.9% [9.6–14.8] 32.6% [28.5–37.1] 13.0% [9.7–17.2]

63.0% [56.7–69.0] 13.7% [10.2–18.2] 23.3% [18.5–28.8] 7.6% [4.7–12.3]

56.5% [51.7–61.2] 10.0% [8.0–12.3] 33.5% [28.9–38.5] 15.5% [11.6–20.4]

Control group

p

.000

.000

.000

.001

.005

BMI ¼ body mass index. Last column provides p values for the bivariate associations. *Mean.

Even if our figures are more striking than those in the literature [3,31], the questions that remain are why the reported prevalence is generally low and why it might still be underestimated. Body and weight concerns are socially considered as ‘‘typical women worries,’’ and boys could be reluctant or ashamed to admit them [1,4]. Moreover, adolescent boys and girls have quite different self-perceptions [1,4] and it is possible that boys do not perceive their eating behavior as

problematic. In addition, these aspects might still not be screened for and diagnosed enough among boys by health care providers. The most prominent observation of our results has to do with the violence context. Boys with DEB are more likely to report a history of sexual abuse or violence and to present with delinquent behaviors. Miotto et al [32] have already discussed the association between DEB and aggressiveness,

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117

115

Table 2 Multinomial regression: Relative risk ratios (95% confidence intervals for subgroups 1–3 compared with the control group Characteristic

Personal variables Body mass index >90 percentile Early puberty Depression (scale) Concerns Feeling fat History of dieting Familial variables Poor relation with parents (scale) Schooling Academic track: student Peer teasing > 1 epidode/week Violence Experienced violence in past 12 months Delinquency History of sexual abuse Health-compromising behaviors 3 Inebriation episodes in past 30 days 3 Cannabis use episodes in past 30 days

1 Concern without behavior

>1 Concern without behavior

1 Behavior

Group 1

Group 2

Group 3

n ¼ 862

n ¼ 361

n ¼ 798

– – 1.30 [1.03–1.64]

– 1.50 [1.04–2.15] 1.43 [1.13–1.81]

1.76 [1.16–2.66] 1.34 [1.01–1.76] 1.54 [1.26–1.88]

1.73 [1.23–2.43] –

5.21 [3.43–7.91] –

2.95 [2.13–4.09] 2.77 [1.72–4.47]

–

1.46 [1.12–1.89]

1.29 [1.05–1.58]

1.29 [1.01–1.64] –

1.54 [1.12–2.13] –

1.85 [1.44–2.37] 1.56 [1.06–2.27]

– 1.78 [1.29–2.46] 3.47 [1.18–10.17]

– 1.88 [1.32–2.66] 5.02 [1.64–15.37]

1.33 [1.01–1.74] 1.43 [1.10–1.87] 3.67 [1.34–10.05]

– –

– .67 [.45–.99]

1.60 [1.09–2.36] –

pointing out impulsivity as the common factor. Some behaviors such as binge eating, binge drinking, or suicide attempts could result from impulsivity but also from aggressiveness and acts of violence. Impulsivity in some teenagers could also lead to conflicts with parents and could damage their relationship with them, as our study indicates. According to our results, all boys with DEB are more likely to report sexual abuse, whereas other types of violence are limited to those with DEB (group 3). The link between DEB and sexual abuse has been discussed in the literature but remains unclear [16,33]. In fact, even though DEB (mostly bulimia nervosa) tend to be more often reported by subjects with a history of sexual abuse, it seems to be integrated in a complex posttraumatic disorder after the abuse, as a kind of coping strategy [16,34]. One study has shown that the association between DEB and physical or sexual abuse strongly increases when the two types of abuse are combined [35]. We find a similar observation for group 3. The literature reports a co-morbidity between ED and depression [4,31] and in our study, depression is indeed reported by all groups. The link between DEB, depression, and experienced violence in our study can be explained by the fact that mistreatments generate low self-esteem (34), which has been shown to be associated with ED and depression [33]. The number of factors associated to boys’ DEB increases progressively throughout the groups, but augments dramatically in group 3. Moreover, there are more subjects in this group than in group 2, suggesting boys’ tendency to act more than to intellectualize to cope with difficulties when they interfere too much in their lives. Boys with DEB (group 3) who report more violence, are also more likely to report

inebriation episodes than the control group. Two hypotheses could explain this association. First, similarly to alcohol use, binge eating could be considered as an addictive/dependency disorder, based on similarities in symptoms such as lack of control or futile efforts to reduce ingested amounts [36]. Second, the link between alcohol and other health-compromising behaviors such as delinquency has been well demonstrated previously [37], and some DEB may be considered as health-compromising conducts (binge eating, self-induced vomiting). These observations corroborate our hypothesis referring to Jessor’s theory [14,38]: DEB may be anchored in a global ‘‘life dysfunction,’’ bringing together general lack of well-being, eating problems, experience with violence, and other health-compromising behaviors. However, this hypothesis is only confirmed for alcohol and delinquency, given that in our sample there is no association between DEB and smoking, other illicit drug use, and unsafe sex, and given that boys with eating concerns (group 2) are less likely to be regular cannabis users. Contrary to literature reports [2,39], in our study SES does not seem to play a role. Nevertheless, boys in the three groups are more likely to be students than those in the control group (in Switzerland, those following a ‘‘student’’ track are the best pupils) and to report more peer teasing. These characteristics probably refer to those usually assigned to girls with ED: perfectionism, studious temperament, good school results, and sometimes social isolation. If these traits currently tend to fade out for girls, they seem to be present for boys. Boys with unhealthy eating behaviors (group 3) were more likely to be overweight or obese and to report a history of dieting, as reported in the literature [31]. Even though they

116

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117

were more likely to feel fat, boys in the three groups did not report body dissatisfaction more often than those in the control group, according to the multivariate analysis, although this variable is known as a predictor of DEB [11,39]. Boys from groups 2 and 3 were more likely to report early puberty. A possible explanation is that they had more difficulties accepting their early physical changes compared with peers (e.g., they were taller or larger than peers) and thus tried to control these changes. Another possible explanation is that their more mature looks led them to relate to older boys, acting like them and sometimes getting prematurely involved in potentially risky exploratory behaviors.

Acknowledgments Franc¸oise Domine´ was supported by a scholarship of the Fond National de la Recherche Scientifique in Belgium and given a grant from the Belgium Study Group for Pediatric Endocrinology. This survey was carried out with the financial support of the Swiss Federal Office of Public Health (contract 00.001721 / 2.24.02.-81) and the participating cantons. The survey was run within a multicenter multidisciplinary group from the Institute of social & preventive medicine in Lausanne, Institute for Psychology, Psychology of Development and developmental Disorders, University of Berne, Switzerland and the Sezione Sanitaria, Dipartimento della sanita` e della socialita`, Canton Ticino.

Study strengths and limitations The main strengths of our study are that it is based on a large nationally representative sample, focuses exclusively on males and examines the global context in which DEB tend to develop. However, some limitations need to be mentioned. First, the cross-sectional nature of our survey prevents any conclusions regarding causality. Second, our results are based on self-reports; however, the fact that the SMASH02 questionnaire was anonymous should limit potential bias. Third, the items included in the questionnaire do not allow specific ED diagnosis. Fourth, our database does not include adolescents who have dropped out of school (w10% in Switzerland) and who are more likely to engage in multiple risk behaviors and to report depression [40].

Conclusion In conclusion, based on our study findings it appears that about half of adolescent boys may be affected by some kind of eating concern or involved in unhealthy eating behaviors. It is known that media and social pressure to be thin have a great influence on girls’ weight concerns and eating behaviors [11]. In recent years, both the scientific literature and the media have given a lot of attention to ED among girls, sensitizing both health professionals and parents. However, we can observe the current spreading of media pressure about male ideal body shape, concurrent to the considerable increase in ED among boys. It is important to keep in mind that boys are also vulnerable to DEB and need to be attentively screened. In particular, those with depression traits, overweight or obesity, an early puberty onset or a history of dieting, and those engaged in other health-compromising behaviors require more attention. From this point of view, prevention programs should also target boys and warn them against the increasing media and social pressure. Such programs should also and foremost sensitize adults in contact with adolescents—particularly parents, educators and health providers—as has been the case for girls. More research is needed to confirm our results and to specify the characteristics of ED in boys, particularly for anorexia nervosa and bulimia nervosa.

References [1] Muise AM, Stein DG, Arbess G. Eating disorders in adolescent boys: A review of the adolescent and young adult literature. J Adolesc Health 2003;33:427–35. [2] Neumark-Sztainer D, Hannan P. Weight-related behaviors among adolescent girls and boys: Results from a national survey. Arch Pediatr Adolesc Med 2000;154:569–77. [3] Ackard DMP, Fedio GM, Neumark-Sztainer DPM, Britt HRP. Factors associated with disordered eating among sexually active adolescent males: Gender and number of sexual partners. Psychosom Med 2008; 70:238. [4] Ricciardelli LA, McCabe MP. A biopsychosocial model of disordered eating and the pursuit of muscularity in adolescent boys. Psychol Bull 2004;130:179–205. [5] Crisp A. Anorexia nervosa in males: Similarities and differences to anorexia nervosa in females. Eur Eat Disord Rev 2006;14:163–7. [6] Gila A, Castro J, Cesena J, Toro J. anorexia nervosa in male adolescents: Body image, eating attitudes and psychological traits. J Adolesc Health 2005;36:221–6. [7] Peixoto Labre M. Adolescent boys and the muscular male body ideal. J Adolesc Health 2002;30:233–42. [8] Carlat DJM, Camargo CAJ, Herzog DBM. Eating disorders in males: A report on 135 patients. Am J Psychiatry 1997;154:1127–32. [9] Lewinsohn PM, Hops H, Roberts RE, Seeley JR, Andrews JA. Adolescent psychopathology: I. Prevalence and incidence of depression and other DSM-III-R disorders in high school students. J Abnorm Psychol 1993;102:133–44. [10] Leon GR, Fulkerson JA, Perry CL, et al. Three to four year prospective evaluation of personality and behavioral risk factors for later disordered eating in adolescent girls and boys. J Youth Adolesc 1999;28: 181–96. [11] Stice E, Mazotti L, Krebs M, Martin S. Predictors of adolescent dieting behaviors: A longitudinal study. Psychol Addict Behav 1998;12: 195–205. [12] Patton GC, Selzer R, Coffey C, Carlin JB, Wolfe R. Onset of adolescent eating disorders: Population based cohort study over 3 years. Br Med J 1999;318:765–8. [13] Brener ND, Collins JL. Co-occurrence of health-risk behaviors among adolescents in the United States. J Adolesc Health 1998;22:209–13. [14] Jessor R. Risk behavior in adolescence: A psychosocial framework for understanding and action. J Adolesc Health 1991;12:597–605. [15] Neumark-Sztainer D, Story M, Dixon LB, Murray D. Adolescents engaging in unhealthy weight control behaviors: Are they at risk for other health-compromising behaviors? Am J Public Health 1998;88:952–5. [16] Wonderlich SA, Brewerton TD, Jocic ZM, et al. Relationship of childhood sexual abuse and eating disorders. J Am Acad Child Adolesc Psychiatry 1997;36:1107–15.

F. Domine´ et al. / Journal of Adolescent Health 44 (2009) 111–117 [17] Neumark-Sztainer D, Wall MM, Story M, Perry CL. Correlates of unhealthy weight-control behaviors among adolescents: Implications for prevention programs. Health Psychol 2003;22:88–98. [18] Stice E. Risk and maintenance factors for eating pathology: A metaanalytic review. Psychol Bull 2002;128:825–48. [19] Maldonado-Molina MM, Komro KA, Prado G. Prospective association between dieting and smoking initiation among adolescents. Am J Health Promot 2007;22:25–32. [20] Beary MD, Lacey JH, Merry J. Alcoholism and eating disorders in women of fertile age. Br J Addict 1986;81:685–9. [21] Stock SL, Goldberg E, Corbett S, Katzman DK. Substance use in female adolescents with eating disorders. J Adolesc Health 2002;31:176–82. [22] Conason AH, Brunstein KA, Sher L. Recognizing alcohol and drug abuse in patients with eating disorders. Q J Med 2006;99:335–9. [23] Jeannin A, Narring F, Tschumper A, et al. Self-reported health needs and use of primary health care services by adolescents enrolled in post-mandatory schools or vocational training programmes in Switzerland. Swiss Med Wkly 2005 Jan 8;135:11–8. [24] Garner DM, Bemis KM. A cognitive-behavioral approach to anorexia nervosa. Cognit Ther Res 1982;6:123–50. [25] Lau B, Alsaker FD. Assessment of weight and eating concerns in Norwegian adolescents. Scand J Psychol 2000;41:189–96. [26] Kromeyer-Hauschild K, Wabitsch M, Kunze D, et al. Perzentile fu¨r den Body-mass-Index fu¨r dans Kindes- und Jugendalter unter Heranziehung verschiedener deutscher Stichproben. Monatsschrift Kinderheilkunde 2001;149:807–18. [27] Berg-Kelly K, Erdes L. Self-assessment of sexual maturity by midadolescents based on a global question. Acta Paediatr 1997;86:10–7. [28] Holsen I, Kraft P, Vitterso J. Stability in depressed mood in adolescence: Results from a 6-year longitudinal panel study. J Youth Adolesc 2000;29:61–78.

117

[29] Armsden GC, Greenberg MT. The Inventory of Parent and Peer Attachment: Individual differences and their relationship to psychological well-being in adolescence. J Youth Adolesc 1987;16:427–54. [30] Flammer A, Alsaker FD, Grob A. Fragebogen Alltag und Schule von Kindern und Jugendlichen in der Schweiz und in Norwegen. Schweiz: Institut fu¨r Psychologie, Universita¨t Bern, 1994. [31] Chamay-Weber C, Narring F, Michaud PA. Partial eating disorders among adolescents: A review. J Adolesc Health 2005;37:417–27. [32] Miotto P, De Coppi M, Frezza M, Petretto DR, Masala C, Preti A. Eating disorders and aggressiveness among adolescents. Acta Psychiatr Scand 2003;108:183–9. [33] Striegel-Moore RH, Dohm FA, Pike KMP, Wilfley DEPh, Fairburn CGM. Abuse, bullying, and discrimination as risk factors for binge eating disorder. Am J Psychiatry 2002;159:1902–7. [34] Wonderlich SA, Wilsnack RWP, Wilsnack SCP, Harris TRP. Childhood Sexual Abuse and Bulimic Behavior in a Nationally Representative Sample. Am J Public Health 1996 Aug;86:1082–6. [35] Rayworth BB, Wise LA, Harlow BL. Childhood Abuse and Risk of Eating Disorders in Women. Epidemiology 2004 May;15:271–8. [36] Cassin SE, von Ranson KM. Is binge eating experienced as an addiction? Appetite 2007;49:687–90. [37] Swahn MH, Donovan JE. Correlates and predictors of violent behavior among adolescent drinkers. J Adolesc Health 2004;34:480–92. [38] Donovan JE, Jessor R, Costa FM. Syndrome of problem behavior in adolescence: A replication. J Consult Clin Psychol 1988;56:762–5. [39] Morande G, Celada J, Casas JJ. Prevalence of eating disorders in a Spanish school-age population. J Adolesc Health 1999;24:212–9. [40] Piot Delbos, Narring F, Michaud P-A. La sante´ des jeunes hors du syste`me de formation: Comparaison entre jeunes hors formation et en formation dans le cadre de l’enqueˆte sur la sante´ et les styles de vie des 15–20 ans en Suisse Romande. Sante´ Publique 1995;1:59–72.