World-wide the prevalence of overweight and obesity has increased dramatically during the last three decades not only in adults but also in children. Intense research efforts have been undertaken to clarify mechanisms involved in the development of overweight and obesity in humans; however, despite these efforts, mechanisms involved are still poorly understood and universally accepted therapies are still lacking. Results of several epidemiologic studies indicated that dietary pattern of adults not only differ markedly between sexes [10, 16], but also between overweight and normal weight individuals . However, similar studies have rarely been performed in children under the age of 10 years and results are contradictory [4, 5]. In the present study, we tested the hypothesis that dietary intake differs between normal and overweight children and between boys and girls. It was shown before by others that a long-term overnutrition of ~100 kcal/d may cause overweight . In the present study, total energy and macronutrient intake did not differ between normal weight and overweight children and all children being overweight or even obese, reached the recommendation to spend one hour per day with sportive activities regardless of weight status . However, leisure time spent with sedentary activity like watching TV was markedly longer in overweight children regardless of sex whereas time spent physical active (e.g. playing soccer) did not differ between groups regardless of sex, suggesting that nutritional intake (e.g. total energy intake) of overweight children may have been too high in comparison to energy expenditure. Indeed in the present study, sedentary activity was identified as an independent risk factor for the development of overweight in children. These findings are in line with those of other groups, who found a positive association between screen time or total media consumption and the degree of overweight . In addition, results of a meta-analysis suggest that media consumption often replaces physical activity (e.g. sport) during leisure time . Contrary to the findings in adults , in the present study, intake of sweets was markedly higher in normal weight children than in overweight or obese children. Still, intake of different food groups e.g. vegetables and fruits as well as meat intake of children in general did not meet the recommendations of the German research institute for nutrition of children (FKE). These findings are in line with those of others. Indeed, results of a meta-analysis of studies performed in 34 different countries revealed that normal weight children quite frequently eat more sweets than overweight children . Furthermore, Ortega et al. also reported that total energy intake between overweight and normal weight adolescents did not differ , whereas in the studies of Hassapidou et al.  and Rocandio et al. , total energy intake of normal weight adolescents was even higher than that of overweight study participants. In line with our findings, in the study of Kersting et al. it was also found that vegetable and meat intake of German children do not meet the recommendations of the German research institute for nutrition of children (FKE) .
Interestingly, whereas BMI of fathers did not differ between normal weight and overweight children, mean BMI of mothers of overweight children was markedly higher than that of the normal weight children. This difference was even more pronounced in boys than in girls, which might be explained with the markedly smaller difference in BMI-SDS between overweight and normal weight girls (Δ boys: 1.62 vs. Δ girls: 1.31). Furthermore, multiple regression analysis revealed BMI of mothers as independent risk factor for elevated BMI-SDS of children. Danielzik et al. also found a closer association between children’s BMI and maternal BMI than paternal BMI . The association with maternal overweight may have different reasons and may even be multifactorial. Indeed, results of studies performed in twins suggest that genetic factors may have a strong effect on the variation of BMI . In addition, parents and herein particularly mothers’ behaviour may also play an important role, because children’s food environment is usually influenced more by mothers than by fathers . Furthermore, the results of Fogelholm et al. indicated that physical activity pattern of children are also associated with that of parents . In the present study ~75% of study participants claimed their mothers to be the primary attachment figure. In addition, the results of the report of the German association of nutrition released in 2008 indicates that only 17% of students, aged 14 years eat lunch at school, indicating that lunch at home is still important in Germany . However, as neither physical activity nor dietary pattern and nutritional intake or genetic polymorphisms of parents were assessed in the present study, possible causes of the positive association of maternal BMI with that of children remains to be determined. Taken together, results of the present study suggest that neither a marked overnutrition nor an “abnormal” dietary pattern are responsible for the elevated body weight in overweight children, but rather leisure time activity as well as maternal BMI may be risk factors for the development of obesity. However, reasons for this altered physical activity pattern (e.g. negative parental model, genetic backgrounds) but also the maternal influence on children’ s body weight remain to be determined.
Sex-specific differences in dietary pattern
Sex-specific differences in nutritional intake and dietary pattern have frequently been reported for adults but also for adolescents (for overview see [6, 7, 28]). In the present study, we found that boys, regardless of BMI, consumed markedly more cheese while girls, in general, ate more vegetables. In adults, sex-specific differences in dietary pattern and particularly the higher intake of fruit and vegetable were claimed to be a result of the higher interest of women in nutrition  and their better knowledge of the relation of eating pattern and disease prevention . However, during childhood parents and herein particularly mothers have a prominent responsibility for children’s diet and it seems to be rather unlikely that parents alter dietary pattern in consideration of sex . Causes of these sex-specific differences in dietary pattern therefore remain to be determined. As already described in other studies particularly for adolescents and adults [16, 31], we found that total energy and macronutrient intake in boys was markedly higher than in girls, despite only slight differences in weight status. One explanation for this could be the differences in leisure time activities, as girls spent significantly more time with sedentary activities than boys. This is in line with the findings of Riddoch et al. who reported sex-specific differences in sportive activities . Furthermore, several studies demonstrated a lower fat mass and higher bone-free lean tissue mass in boys already before the onset of puberty , which may result in an increased energy expenditure. However, exact mechanisms underlying these sex-specific differences have to be determined. Taken together, the results of the present study suggest that dietary pattern differs between boys and girls, aged 5–8 years, regardless of body weight status. Furthermore, the results also indicate that sex-specific differences in leisure time activities are responsible for the fact that no differences in body weight status were found between boys and girls, despite the higher total energy and macronutrient intake in boys.
Our study has some limitations that need to be considered when interpreting the data. First, nutritional intake was self-reported and might be influenced by recording errors or under-reporting. This is a problem often described in the context of nutritional interviews, especially when performed in overweight adults and children [20, 34]. Additionally, it cannot be ruled out that overweight and obese parents tended to underestimate dietary intake of their children. However, as mentioned before, our findings are in line with those of other larger studies performed in Europe. Second, leisure time activities were also self-reported and not assessed by objective measurements such as accelerometers. Indeed, over-reporting of physical activity is also a problem often found in overweight people . Third, the results are not representative for the whole population, as the study sample consists mainly of Caucasian children and parents with an educational level higher than 10 years and it was a rather small sample. Forth, data of this cross-sectional study only represent a short window in time, whereas the development of obesity is clearly a long-term issue. Furthermore, the possibility of statistical type II error cannot be ruled out.