Gender moderated the association between weight status and mental HRQoL score. Among girls, mental HRQoL was higher for thin girls and decreased with increasing BMI. In contrast, mental HRQoL was lower for thin, overweight and obese rather than normal-weight boys.
Only one study investigated the effect of gender on the association between weight status and HRQoL in a community sample including older adolescents (14–18 years old)  who are affected by physical and psychological changes due to puberty. Keating et al. showed that girls who were overweight and obese had significantly lower physical functioning scores than normal-weight girls, but the same was not true for boys. Similarly, girls who were obese had significantly lower school and emotional functioning scores than did normal-weight girls, but the same was not true in boys. Keating et al. excluded thin adolescents, included younger and older adolescents (11–18 years), and used the PedsQL to estimate HRQoL scores. In our study, thin adolescents were not excluded, only older adolescents were included (14–18 years), and the adolescent Duke Health Profile was used. Like Keating et al., the present investigation showed that overweight and obese girls had poorer mental HRQoL score than did normal-weight girls. However, weight status (including thin) and mental HRQoL score were linearly and negatively related in girls but in not boys, thin girls having a better mental HRQoL score than normal-weight girls. Thus, it appears important to include thin adolescents when investigating weight status.
Our results suggest that the perception of ideal body size and shape differs between adolescent girls and boys; thinness represented the ideal body image for girls, which could explain the better mental HRQoL score in thin girls . The representation of the ideal body shape may come from a collective vision influenced by television, magazines, advertisements and the social stigma attached to obesity and the “fat phobia” that pervades our daily life. Girls seek to be thin and fit, not only to be healthy but also to be perceived by themselves and others as having desirable personal qualities. Girls may be more attuned to or aware of their bodies and their health than boys. Girls consistently report greater body dissatisfaction than do boys [36, 37]. Some studies showed that girls were more likely to perceive themselves as being overweight than boys [36, 38–40]. In a study by Lawler et al. , all overweight girls revealed a desire to weigh less, as compared with only 78.6% of overweight boys. Average-weight girls wanted to be lighter, whereas average-weight boys were satisfied with their bodies or wanted to be bigger . Boys find a greater variety of body shapes socially acceptable compared to girls, and girls have a narrower range of what is considered the ideal body image . Even if some men strive to lose weight to conform to today’s ideal body shape, the ideal image for most boys is muscular and strong, the main characteristics of virility, so boys with an athletic build may be more popular among their peers and more difficult to victimize . This observation may explain the non-linear association between BMI and HRQoL observed in boys.
Cross-sectional community surveys reflect the subjects’ HRQoL at one time, which means that temporality cannot be established  and it remains unclear whether BMI determines HRQoL or vice versa. Some studies merged data for adolescents and younger children , but findings in children cannot be extrapolated to adolescents. Adolescence is the period of transition to adulthood and sexual differentiation. This stage presents several characteristics that justify a particular interest in terms of obesity, as well as at the level of prevention, screening and care. Indeed, obese adolescents are more likely than their normal-weight counterparts to remain obese into adulthood (78% for men vs. 63% for women) . Moreover, the large sample size gave the study considerable statistical power. In addition, the quality of anthropometric measurements reported by qualified and trained nurses in our study attested to the validity of the data and minimized biased measurements. As linearity was not verified, BMI was considered as o nominal variable with four classes including thinness. Therefore, HRQoL score was estimated for thinness, and strongly differed between girls and boys.
Implications for research and public health
The association between weight status and HRQoL in adolescents must be studied by gender and the results taken into account when developing health programs, in order to identify the best strategy with which to address these issues. The present investigation found that girls had significantly lower mental HRQoL than did boys and their decrease in HRQoL was higher when their body size increased. This information could help educators implementing health programs that focus on the educational system and are tailor-made to meet the specific needs of target groups. Finally, thinness in boys affected HRQoL in the same way as overweight and obesity in boys. Thinness may be an issue to consider in public health programs for adolescents.