The aims of the current study were 1) to investigate the prevalence of physical, sexual and mental abuse among adolescents, and 2) to examine the associations of physical, sexual and mental abuse with BMI status of adolescents.
Results showed that one out of five adolescents reported mental abuse, while one out of ten reported physical or sexual abuse. These prevalences are in line with expectations based on previous conducted research
[22, 23]. All subtypes of abuse were significantly more common among underweight, overweight and obese adolescents compared to their normal weight peers. The adjusted GEE analyses showed significant associations of mental abuse with underweight, overweight and obesity. Physical and sexual abuse were significantly associated with obesity in boys and vocational school students, and with underweight in girls.
In sum, our findings confirm that abuse is associated with underweight, overweight and obesity among adolescents. Suggesting that the hypothesized mechanisms (i.e., distorted eating behavior as an ineffective coping response, developing overweight as a defense mechanism, and elevated HPA-axis activity) through which abuse might cause an aberrant BMI status in adults apply to adolescents as well. However, due to the cross-sectional design of this study, we were not able to test these mechanisms. Hence, intermediate-analyses within longitudinal research are needed to explore the accuracy of these mechanisms. In addition to overeating, future research should also consider physical inactivity and sedentary behavior as potential responses to abuse. Abused individuals generally show signs of isolation from the community
[17, 32, 33] as a way to protect themselves from potential abusers. However, such behavior may cause a decrease in physical activity and an increase in sedentary behavior
 which in turn may contribute to the development of overweight.
Especially in case of physical and mental abuse, the relation with overweight may even be reciprocal. Because of their weight, overweight or obese adolescents are more often victimized, either overt or relational
[17, 35]. Longitudinal data are needed to examine the direction of the relationship between abuse and overweight.
Current adjusted regression results also showed that physical and sexual abuse were significantly associated with obesity in boys, but not in girls. These findings do not agree with earlier finding showing significant associations of physical and sexual abuse with overweight in women
[1, 2, 8, 10, 11]. This discrepancy with our findings may be due to the age difference between the study populations. Fuemmeler and colleagues (2009) hypothesized that, unlike adult women, adolescent girls perceive a strong social pressure to be thin which might overrule the need to start overeating as a coping response towards the perceived abuse
. Since these are the first studies investigating the effects of different forms of abuse on overweight among adolescents, further research is needed to confirm these findings.
Contrary to findings among higher secondary school students, physical abuse was significantly associated with obesity in vocational school students. Although scientific proof is still lacking, Finkelstein et al. (2007) suggested that adolescents with a lower socioeconomic status have fewer psychological resources (e.g., optimism) to properly cope with adverse events
. This might trigger these individuals more easily to adopt ineffective coping style such as overeating, compared to their higher educated peers. More extensive research on the psychological resources and (lifestyle based) coping methods in response to abuse among this subpopulation is desirable.
Finally, mental abuse showed to be associated with underweight in boys and vocational school students (in the adjusted models), while physical and sexual abuse was associated with underweight in girls. This is in line with previous studies that showed distorted eating behaviour in abuse victims
[6, 7, 37, 38], which can lead to overweight or underweight and even anorexia nervosa, especially in girls
Strengths and limitations
This study is distinctive in the age-group of concern as well as the fact that we were able to include different types of abuse. Another strength of this study is the size of the study population. However, some limitations should be noted. First, all data were self-reported by adolescents, therefore it is likely that the prevalence of the different subtypes of abuse are underestimated. However, since adolescents have been told that all answers were strictly confidential, it is hypothesized that the current study provides a good estimation of the extent to which adolescents perceive themselves to be abused. Besides, previous research indicates that when using self-reported height and weight data, the number of adolescents classified as overweight and obese are most likely to be underestimated
[41, 42]. This was confirmed by yearly data, including weight and height of children aged 0 till 19, collected by the youth health care institute in the Netherlands. This institute reported that 18% of all 14-year olds was overweight (including obese) and 4% was obese
, compared to 8% and 1% in the current study. Hence, the association between BMI status and abuse might be even larger than estimated in the current study. Second, the current study population might not be entirely representative for the total Dutch population of adolescents since there were slight differences in the questionnaires that were used by the different Community Health Services and non-response was not registered consequently by all Community Health Services. Besides, as in all voluntary based research, non-response might not be random. It could very well be that adolescents with certain specific characteristics did not (entirely) complete the questionnaire, leading to either an under- or overestimation of the current findings. Since it was not possible to conduct a non-response analysis, neither the demographics of the non-responders nor the possible differences between responders and non-responders concerning these demographics could be tested. Finally, one of the answering options was ‘I do not want to answer this question’ to the questions on the different subtypes of abuse. Since it can only be speculated whether these adolescents were abused or not, these adolescents were excluded from the current analyses. Students who did not want to answer the abuse questions had a significant higher weight status compared to students who answered these questions by ‘yes’ or ‘no’. However, the association of abuse with overweight and obesity did not differ between these sub-groups (results not shown).