As expected from the previous research mentioned in the introduction we found significant associations between ADHD symptoms, low dietary quality, and high total energy intake. We could demonstrate that these associations were not entirely explained by the investigated parental variables or other psychopathology of the children. The association with the SDQ-HI scores was more pronounced for food quality than for food volume. In detail, the significant positive association between SDQ-HI scores and energy intake from beverages was due to both increased volume and increased energy density of the beverages. The association between SDQ-HI scores and energy intake from food was explained by the energy density of food but not by food volume. This might indicate that ADHD symptoms may be associated with poor food selection rather than overeating in terms of volume. This is in line with the results of a recent cross-sectional study showing that Australian children with ADHD symptoms were more likely to consume a “Western” diet high in saturated and total fat and refined sugar while low in fiber, than to consume a Mediterranean diet rich in fish, vegetables, fruit, legumes, and whole-grain foods 
We found that girls showed a significantly stronger association between ADHD symptoms and dietary quality, food volume, food energy intake, and (marginally) total energy intake than boys. This corresponds with our previous results based on the same population 
, where we found a significant independent association between ADHD symptoms and overweight only in adolescent girls. The success of treatment for obesity is reduced by ADHD symptoms or impulsivity in most studies 
. Girls are under stronger societal pressure to reduce weight and food intake than boys. In those motivated to do so, the success in reducing intake of high energy food may depend on attention and impulse control and thus be inversely related to ADHD symptoms.
We could confirm the independent association between ADHD symptoms and television and video exposure 
. This is in line with the recently published results of a cross-sectional study of 68,634 children from the National Survey of Children’s Health in the United States 
. The authors reported that an average TV usage during weekdays of 1 hour or more was significantly associated with a diagnosis of ADHD even after adjusting for very similar parental variables as used in our study. The design of our study does not allow causative interpretations, but there is growing evidence in the literature that exposure to television and video games in childhood may actually be associated with increased subsequent attention problems 
The weak but significant positive association between ADHD symptoms and medium to high intensity physical exercise in the 11–17 year old participants is in line with studies using actometry 
but contradicts the results of large US population based studies which found that children with ADHD symptoms were less likely to participate in vigorous physical activity and organized sports 
. However, we did not assess participation in organized sports and clubs. The negative association between participation in organized sports and ADHD symptoms found in the US studies could be due to the fact that children and adolescents with ADHD symptoms might have difficulties in sports teams because of their behavioral problems. In addition, different age ranges and predictor variables (ADHD symptoms vs. ADHD diagnosis) as well as possible cultural differences might also explain this discrepancy.
Parental variables such as socioeconomic status, BMI, and smoking have been shown to be strongly associated with ADHD symptoms, television exposure, and dietary quality. However, these confounding variables did not entirely explain the positive association between ADHD symptoms and television exposure, dietary quality, and energy intake in our sample. In addition, the associations between the SDQ-HI subscale scores and the health behaviors remained significant for all but food energy intake when adjusting for the other SDQ-subscales and are, thus, largely independent of peer relations, emotional problems and other behavioral problems assessed with the SDQ.
Strengths and Limitations
The main strengths of our report lie in the large representative study on which it is based, and the multivariate analyses that allowed looking for independent effects of ADHD symptoms. ADHD symptoms were used as a dimensional and not a categorical variable. In our effort to control for parental confounding variables, we included a sophisticated socioeconomic status score, migrant status, and parental smoking in addition to parental BMI, age, and sex.
However, given the cross-sectional nature of the data, we cannot establish the direction of the detected relationships. In addition, the health behavior variables used in our study were based on self-reported data and are thus liable to social desirability and recall bias. Also, we have combined parent-rated and self-rated television exposure time because only the combined variable spans the whole age range of 6–17 years used in our study. Fortunately, both ratings were significantly associated with SDQ-HI scores, also after adjusting for parental confounders as well as sex and age.
Finally, while the associations were statistically significant, they were generally weak, explaining less than 1% (derived from the partial eta2 values given in multiplied by 100%) of the variance of each of the variables adjusted for age, sex, and parental confounders. The imprecision of measuring behavior by self-report and parental report may have mitigated the observed associations.
There is evidence that overweight/obese children have a significantly higher risk for ADHD symptoms than normal weight children. Our study adds to this finding in that poor nutrition and high television exposure time also seem to directly be associated with ADHD symptoms even after adjusting for potential confounding variables. Clinicians should be aware that children and adolescents with ADHD symptoms should be monitored with regard to food intake and television/video exposure. Environmental control measures and parental monitoring may be required to improve the dietary patterns and to reduce TV usage time of children and adolescents with ADHD symptoms, as previous studies have shown that these individuals profit less from behavioral weight loss interventions 
. Also preventive efforts to promote healthy behaviors will have to address the network of associations between ADHD symptoms, health behaviors, overweight/obesity, and parental variables.
Finally, further research based on large longitudinal cohorts is needed to address the direction of causality.