**3. Results**

Of the 392 eligible participants 303 children between 6 and 8 years of age at baseline (55.1% male; 9.3% overweight/obese) provided valid measurements for at least five time points, including baseline and the last follow-up measurement. There were no differences in sex distribution and anthropometric characteristics at baseline between children included in the analyses and those excluded due to missing follow-up data. Children with sufficient follow-up measurements, however, displayed better overall physical fitness at baseline compared to those excluded (104.2 ± 5.9 vs. 100.9 ± 6.5, *p* < 0.01).

Descriptive characteristics of participants included in the analyses are shown in Table 1. Boys were taller and heavier compared to girls but there was no difference in BMIPCT. Accordingly, there was also no difference in weight status and sex distribution did not differ across quartiles of BMIPCT. There was also no sex difference in overall physical fitness and sex distribution across quartiles of physical fitness. Absolute performance for the 6 min run, sit ups, standing long jump and 20 m sprint, however, was better in boys, whereas balance and flexibility was better in girls (*p* < 0.05).

**Table 1.** Descriptive characteristics at baseline for the total sample and separately for boys and girls. Values are Mean ± SD.


<sup>1</sup> positive values indicate reaching beyond the toes, while negative values indicate not reaching toes. \* sig. sex difference (*p* < 0.05); \*\* sig. sex difference (*p* < 0.01).

Pearson correlation analyses did not show significant cross-sectional correlations between BMIPCT and components of physical fitness at baseline, except for a low inverse association between BMIPCT and 6 min run performance (Table 2). Nevertheless, the prevalence of overweight/obesity differed significantly across fitness quartiles (Q1: 4.0%, Q2: 3.0%, Q3: 1.7%, Q4: 0.7%; *p* = 0.02). However, during the last follow-up assessment, when children were 10.4 ± 0.5 years of age, significant cross-sectional associations were observed between BMIPCT and all components of physical fitness, except for flexibility. Accordingly, overall physical fitness was negatively associated with BMIPCT (r= −0.43, *p* < 0.01) when participants were in fourth grade, whereas there was no significant correlation between overall physical fitness and BMIPCT in children when they were in first grade. In addition, low significant negative correlations between weight change and change in various components of physical fitness were observed, which also resulted in a negative association of weight change with change in overall physical fitness during the elementary school years (r = −0.25, *p* < 0.01).


*Int. J. Environ. Res. Public Health* **2022**, *19*, x FOR PEER REVIEW 5 of 12

elementary school years (r= −0.25, *p* < 0.01).

coefficients.

**Table 2.** Association between body weight and physical fitness at baseline and last followup, and between change in body weight and change in physical fitness. Values are Pearson correlation coefficients.  **<sup>6</sup> Min Run (m) Sit Ups (Reps) Push Ups (Reps) Long‐ jump (cm) 20 m Sprint Side Jumps Balance (Steps) Stand and Reach (cm)**

**Table 2.** Association between body weight and physical fitness at baseline and last follow‐up, and between change in body weight and change in physical fitness. Values are Pearson correlation

< 0.01) when participants were in fourth grade, whereas there was no significant correlation between overall physical fitness and BMIPCT in children when they were in first grade. In addition, low significant negative correlations between weight change and change in various components of physical fitness were observed, which also resulted in a negative association of weight change with change in overall physical fitness during the

\* *p* < 0.05; \*\* *p* < 0.01; BMIPCT—BMI (body mass index) percentile; reps—repetitions in 40 s for sit ups and push ups, and repetitions in 15 s for sideways jumping. ups and push ups, and repetitions in 15 s for sideways jumping. Longitudinal analyses showed a significant increase in the prevalence of

Longitudinal analyses showed a significant increase in the prevalence of overweight/ obesity from 9.3% in first grade to 17.5% in the fourth grade. Across the entire sample, participants with a higher baseline BMIPCT displayed a lower improvement in all components of physical fitness (*p* for trend < 0.05), except for balance and flexibility (Figure 2). Change in BMIPCT, however, did not differ across quartiles of baseline BMIPCT and baseline fitness. Accordingly, the increase in the prevalence of overweight/obesity did not differ across quartiles of physical fitness (Figure 3). There were also no differences in changes in various components of physical fitness across quartiles of baseline fitness, except for the 20 m sprint (*p* for trend < 0.01), where a more pronounced improvement was detected in participants with lower baseline performance. overweight/obesity from 9.3% in first grade to 17.5% in the fourth grade. Across the entire sample, participants with a higher baseline BMIPCT displayed a lower improvement in all components of physical fitness (*p* for trend < 0.05), except for balance and flexibility (Figure 2). Change in BMIPCT, however, did not differ across quartiles of baseline BMIPCT and baseline fitness. Accordingly, the increase in the prevalence of overweight/obesity did not differ across quartiles of physical fitness (Figure 3). There were also no differences in changes in various components of physical fitness across quartiles of baseline fitness, except for the 20 m sprint (*p* fortrend < 0.01), where a more pronounced improvement was detected in participants with lower baseline performance.

**Figure 2.** Change in physical fitness across quartiles of baseline BMI percentile (Quartile 1 indicates lowest BMI percentiles). Values are Mean with 95% CI.

lowest BMI percentiles). Values are Mean with 95% CI.

**Figure 2.** Change in physical fitness across quartiles of baseline BMI percentile (Quartile 1 indicates

**Figure 3.** Cumulative prevalence of overweight/obesity throughout the observation period across quartiles of baseline physical fitness (Q1 indicates low physical fitness). **Figure 3.** Cumulative prevalence of overweight/obesity throughout the observation period across quartiles of baseline physical fitness (Q1 indicates low physical fitness).

Linear regression analyses further showed a significant negative association between change in BMIPCT and subsequent fitness (Table 3). The detrimental effect of higher weight gain on fitness development was particularly pronounced in participants with higher BMIPCT at baseline. No clear pattern of this association was observed across baseline fitness quartiles. The association between baseline fitness and fitness at the last follow‐up, however, was strongerin participants with lower baseline fitness. These results remained essentially unchanged after including sex as covariate. Linear regression analyses further showed a significant negative association between change in BMIPCT and subsequent fitness (Table 3). The detrimental effect of higher weight gain on fitness development was particularly pronounced in participants with higher BMIPCT at baseline. No clear pattern of this association was observed across baseline fitness quartiles. The association between baseline fitness and fitness at the last follow-up, however, was stronger in participants with lower baseline fitness. These results remained essentially unchanged after including sex as covariate.

**Table 3.** Regression coefficients based on linear regression analysis for overall physical fitness at last measurement. **BL Fitness (β) BL BMIPCT (β) Δ BMIPCT (β) R2 Table 3.** Regression coefficients based on linear regression analysis for overall physical fitness at last measurement.


\*\* *p* < 0.01. Low BMIPCT—BMIPCT < 29; below average BMIPCT—29 ≤ BMIPCT < 50.2; above average BMIPCT—50.2 ≤ BMIPCT < 76; high BMIPCT—BMIPCT ≥ 76. Low fitness—overall fitness < 100; average fitness—100 ≤ overall fitness < 105; above average Fitness—105 ≤ overall fitness < 108; high fitness—overall fitness ≥ 108. BL—baseline; BMIPCT—BMI percentile; ∆—annual change based on linear mixed model. \* *p* < 0.05, \*\* *p* < 0.01. Low BMIPCT—BMIPCT < 29; below average BMIPCT—29 ≤ BMIPCT < 50.2; above average BMIPCT—50.2 ≤ BMIPCT < 76; high BMIPCT—BMIPCT ≥ 76. Low fitness—overall fitness < 100; average fitness—100 ≤ overall fitness < 105; above average Fitness—105 ≤ overall fitness < 108; high fitness overall fitness ≥ 108.

Change in physical fitness was negatively associated with subsequent BMIPCT as well (Table 4). This association was also more pronounced in participants with higher baseline BMIPCT, whereas no clear pattern was observed across baseline fitness quartiles. Further, the association between baseline BMIPCT and BMIPCT at follow‐up was less Change in physical fitness was negatively associated with subsequent BMIPCT as well (Table 4). This association was also more pronounced in participants with higher baseline BMIPCT, whereas no clear pattern was observed across baseline fitness quartiles. Further, the association between baseline BMIPCT and BMIPCT at follow-up was less pronounced in participants with below-average baseline BMIPCT compared to those with higher baseline BMIPCT. The inclusion of sex as additional independent variable in the regression models did not have a significant impact on the previously reported results.


**Table 4.** Regression coefficients based on linear regression analysis for overall BMI percentile at last measurement.

BL—baseline; BMIPCT—BMI percentile; ∆—annual change based on linear mixed model. \* *p* < 0.05, \*\* *p* < 0.01. Low BMIPCT—BMIPCT < 29; below average BMIPCT—29 ≤ BMIPCT < 50.2; above average BMIPCT —50.2 ≤ BMIPCT < 76; high BMIPCT—BMIPCT ≥ 76. Low fitness—overall fitness < 100; average fitness—100 ≤ overall fitness < 105; above average Fitness—105 ≤ overall fitness < 108; high fitness —overall fitness ≥ 108.
