Beyond BMI: waist circumference and social environment is associated with motor performance ability in kindergartners

The present study is based on data of the Children Health Study of Mainz (CHSM) [45]. Parents of children attending one out of 34 of the 35 regional public kindergartens were informed and invited to participate in the CHSM. The purpose of the study was to collect data about health determinants associated with overweight/obesity in pre-school children and their related caregivers [45]. Children were also invited to participate in assessments of anthropometry measurements and MPA tests.

Of 869 parents contacted, 558 replied to our invitation and filled out a questionnaire. Six children were excluded because they lived with their grandparents, foster or adoptive parents. A further 81 were excluded due to incomplete information about body height and weight and their habitual physical activity scores. In addition, thirty-seven children were excluded due to inability to perform the MPA tests (n = 35) and the absence of WC measurements (n = 2). Thus, complete data was available from 434 parent-child pairs. For inclusion in the present study, children had to be free from acute and infectious diseases, chronic diseases and restrictions related to the active and passive musculoskeletal system. A detailed description of the active recruitment strategy is presented elsewhere [45]. Participation in this study was voluntary and all parents of participants provided written informed consent. To increase the level of anonymity, personal data collection was reduced to a minimum.

One of the child’s primary guardians completed a self-constructed standardized paper and pencil questionnaire. During informative meetings about the aims and procedures of the study in each kindergarten, parents were informed about the MPA screening of their child, about the health benefits of regularly conducted physical activity, especially in the family environment, and furthermore about the potential early life determinants of overweight/obesity [3]. Essentially, the questionnaire consisted of two main parts. The first part asked about children’s sociodemographic and anthropometric data like age, gender, height and weight at birth as well as questions about the current health status and language which is primarily spoken by the child and the parents. According to Rapp et al. [46], participants responded when they picked up their child from kindergarten to calculate daily time children attended the kindergarten. Additional questions were asked about physical activity habits and screen time activity (characterized with daily TV time and using the PC/Internet on weekdays/weekends) [47, 48]. Children’s physical activity level was assessed using a modified Baecke Habitual Physical Activity Questionnaire (HPAQ) [49] in which parents reported about their child’s physical activity behavior [50, 51]. Questions about screen time activity have already been used in the German Health Interview and Examination Survey for Children and Adolescents (KiGGS) [48, 52]. Detailed information of the raised questions and all classifications regarding habitual physical activity level and screen time activity were presented in a previous publication [45].

The second part focused on parental anthropometric characteristics (age, gender, weight, height), cultural background (country of birth, nationality), educational and employment level, alcohol intake, and smoking status. Parental BMI was calculated according to the WHO criteria [53] and education level was used to represent the individual socio-economic position (SEP) and was categorized into three levels: low educational level (primary school or none), middle educational level (vocational secondary modern school or equivalent), as well as high educational level (university or vocational postsecondary school) [54].

Children’s height and weight was measured by trained staff with a calibrated portable stadiometer (SECA 217 [SECA, Hamburg, Germany]) and with a calibrated flat scale (SECA 803). Children were weighed without shoes and wore light clothes. WC was measured according to proposed international proceedings [55‐58] using an inelastic tape (SECA 203).

Overweight/obesity were defined according to the IOTF international proposed age- and gender-specific cut-offs. The IOTF cut-offs and international thinness cut-offs are used to classify children aged 2–18 years as thin, normal weight, overweight or obese, based on adult BMI cut-offs at 18 years [59]. BMI and WC SDS were calculated according to the international commonly used UK 1990 reference [60‐62]. WC risk groups were categorized using German WC reference values obtained in 3–11 years old children [57]. Due to the fact that WC is a highly sensitive marker for abdominal obesity in childhood [44, 57, 63] and other anthropometric characteristics (i.e. weight, height, BMI) tend to underestimate obesity in youth [64], two modified risk groups in accordance with Schwandt et al. [57] were used: (1) No risk (<90th percentile) and (2) High risk of having multiple cardio-metabolic risk factors (MCRF; 90th - <97th percentile).

MPA of coordination, speed strength, muscular endurance and speed was assessed with multiple test items (One leg stand, standing long jump, lateral jump) of the modified Karlsruher Motor Ability Screening Test (KMS 3–6) [35], the Motorik Modul of the KiGGS-Study [18] and the KiMo project [16]. Reliability and validity of the test items (r ranging from 0.8 to 0.9) have been shown previously [35].

The following criteria have been applied for selection of test items [15]:

Additionally, the “shuttle run” test, already conducted by Krombholz in 2004 [36], was employed to depict the MPA of speed. Due to the possibility of methodical difficulties which may arise in carrying out the “one leg stand” on a bar with a width of 3.0 cm at pre-school age, one leg stand was performed on a bar with a width of 4.5 cm in accordance with Krombholz [36] and De Toia et al. [16]. All test results were classified according to the summarized age and gender-specific reference values [16, 35, 36]. Due to the fact that there were no significant differences between normal weight and overweight/obese children in the test item “sit and reach” in the proposed age groups [20], this test item was omitted.

MPA tests were conducted in each kindergarten during mid-morning and each child was instructed by one trained member of the staff.

A detailed description of the test items with information about material, test construction and realization was published elsewhere [65]. Data collection generally took 10–12 min per child and each staff member demonstrated the tests before the child performed the test. There was no special order for the test items and up to four children performed the tests simultaneously, depending on the respective spatial conditions in the test room.

Regarding the social environment, we examined if kindergartners who attended kindergartens which are in socially deprived city districts were more likely to be overweight or obese. As part of a social city district analysis conducted by Pfeiffer et al., [66] comprehensive information about general conditions (land use, current living situation, demography, budgetary structure, employment status of the family and income maintenance) and social structures (highly problematic family status and educational level) were collected from each city district of the city of Mainz. This led to the computation of a Life Situation Index (LSI) based on employment status/working life, education, social situation/heterogeneity and home environment, summarized to one scale and which we used in this study. Detailed description and method of calculation for demonstrating socio-demographic and socio-economic differences are presented in our previous studies [45, 47].

Descriptive anthropometric characteristics were presented as means, frequencies and percentages with standard deviations (SD), respectively. The results of the MPA test items of the children were calculated for the total sample for gender, BMI groups, WC risk groups and LSI groups and were presented as medians and interquartile range (IQR). The software package LMSgrowth, Version 2.77, was used to calculate age and gender-specific standard deviation scores of height, weight, BMI and WC. For details of the methods we refer to previous publications [59, 67, 68].

For the MPA test results, we conducted group comparisons between boys and girls and normal weight and at least overweight children, as well as between WC risk groups and the LSI groups of each kindergarten using the Mann-Whitney U-Test.

In addition, descriptive characteristics of the final parental reports were employed and stratified by gender. Product-moment correlation coefficients, adjusted for age, BMI SDS and WC SDS, migration background, LSI and parental BMI were calculated to analyze the relationship between the MPA test items and the indices of children’s physical activity and investigate the associations between the indices of children’s physical activity and the indices of physical activity of their parents.

Table 1 presents the descriptive characteristics of the study sample. Mean age (± SD) was 4.9 ± 1.0 years and 55.9% were boys. Girls were significantly older (p = 0.009) than boys, whereas BMI was higher in boys (p = 0.043). Prevalence of overweight/obesity was significantly higher in boys than in girls (p = 0.027). In summary and according to the IOTF cutoffs, 18.1% were overweight/obese.

Total available parental characteristics concerning age, body height, body weight, calculated BMI with respective BMI groups, smoking status, country of birth, nationality and SEP from the questionnaire by males and females are presented elsewhere [45]. Results of the final parental sample descriptions of the items of habitual physical activity, screen time activity as well as the overweight/obesity status and the perception of weight status and associated health risks stratified by gender are also presented elsewhere [45]. The majority of parents analyzed in the final sample (n = 434) were female (85.7%). Indices of habitual physical activity did not differ between genders. Prevalence of overweight was higher in responding fathers compared to mothers (53.3 to 35.4%; p = 0.008). Moreover, responding fathers had higher PC/Internet use on weekends than mothers (p = 0.007).

Below-average MPA was found in 46% of the total sample, while this was present in 59.2% of the children for the one leg stand, in 27.4% of the children for the standing long jump, in 31.6% of the children for lateral jumping, and finally in 46.0% of the children for the shuttle run test (see Fig. 1).

As shown in Table 2, girls had significantly less floor contacts than boys in the test item “one leg stand” (P < 0.001). No differences were found in the test items “standing long jump”, “lateral jumping” and “shuttle run”. Regarding the relationship of MPA and BMI status, Table 3 shows that significant BMI group differences were only observed within the test item “lateral jumping” between children classified into normal weight and overweight/obesity. Overweight/obese children performed significantly less jumps than children in the normal weight group (21.6 ± 12.1 vs. 29.4 ± 12.2; p = 0.039).

In contrast to this, significant differences between the no risk WC group and the group with high risk of having multiple cardio-metabolic risk factors in all test items except in the “shuttle run” were observed (see Table 4).

Table 5 shows the relationship of motor ability performance and the social environment of the district of the kindergarten expressed as two LSI groups. Significant group differences between LSI groups were observed in the test items “one leg stand” (p = 0.044) and “standing long jump” (p = 0.043).

Determinants of below-average motor ability from multiple regression analysis are shown in Table 6. The final regression model indicated that BMI was not integrated in the final model and it seems that in kindergartners WC is a better predictor of motor ability than BMI. Analyses furthermore indicated that older children performed better in motor ability tests than their younger counterparts; hence, younger children were at higher risk for below-average motor ability than older children. Additionally, the social environment of the district of the kindergarten was significantly positive related to below-average motor ability. Finally, there were additional positive associations with parental sport index as well as with children’s sport and leisure-time index (data not shown).