The prevalence of overweightness (19.9%) in our study was in agreement with findings in general population. A prevalence of 17.6% and 21.6% has been previously reported in 11–17 yr schoolchildren in France and Greece respectively
and 23% in schoolchildren 16–19 yr in Sweden 
. Therefore, it is indicated that overweightness affects young soccer players in a similar extent as it does in general population. The participation in a sport per se cannot guarantee that an adolescent is not overweight and proper family-, sport club-, and school-based exercise interventions should target weight and body fat control. The increase in body mass results from the misbalance between energy intake (nutrition) and energy expenditure (physical activity), and an optimal intervention should consider both parameters, as well as genetic and environmental factors [20, 23]
. Even if the application of BMI in sport population has been questioned
, it might be useful in youth soccer. Unlike sports where BMI could misclassify elite athletes as overweight or obese (e.g. American football 
, soccer is a sport that is characterized by normal values of BMI (about 23 kg.
in adults), as it has been shown by research on four elite European leagues 
. In this context, the excess of body mass observed in our study was unexpected. Consequently, the current values of BMI found in our study should not be attributed to sport-specific physiological adaptations. It is unlikely that the high BMI in our study is due to a healthy increase in muscle mass alone and it may not be without health consequences. The high prevalence of overweightness in our sample warrants further investigation to determine the consequences of excessive weight in adolescent soccer players and to develop exercise intervention targeting weight management
The results of this study indicate that BMI accounts for a large proportion of between-individual differences in BF; 59.8% of the variance in BF was explained by BMI. An important consideration was whether BF could be predicted from BMI in youth soccer. The strong relationship between BMI and BF, and the acceptable standard error of estimate of the former based on the latter, suggests the further use of BMI in adolescent soccer players. This correlation was comparable with findings in 5–19 yr boys (r
. It was also similar with what was found in Japanese general population aged 9–10 yr (r
<0.001) and 12–13 yr (r
; and higher than in the case of 4–11 yr boys (r
. Thus, the use of BMI in youth soccer as a surrogate measure of BF is further recommended
Based on previous studies on general population, it was hypothesized that there was an inverse relationship between BMI, BF and physical fitness in soccer players. The negative values of the correlation coefficient between these parameters confirmed our hypothesis. The most interesting finding derived from the comparison between BMI and BF quartiles (), which revealed that the highest BF quartile scored lower in all the tests. This suggests a threshold exists in BF, over which physical fitness is affected in great extent. These results emphasized the role of adiposity, but supported the role of BMI in youth soccer. We found associations between physical fitness and overweightness, whereby boys in higher BMI quartiles demonstrated reduced fitness compared with those in the lower quartiles. These relationships were even stronger between physical fitness and adiposity. These associations were not similar for all the parameters of physical fitness that were examined. Flexibility was a parameter that found to be in low or non-significant correlation with overweightness and adiposity, which came to terms with previous observations 
Fig. 2 Physical working capacity in heart rate 170 beats*min−1 (P170; upper row), mean power during the Wingate anaerobic test (Pmean; middle row) and maximal power estimated by a Force-velocity test (Pmax; lower row) for quartiles (Q1, Q2, Q3 and Q4) (more ...)
Our findings are encouraging and should be validated in countries with different prevalence of overweigh/obesity than Greece. Future work should focus on the identification of those sports in which children and adolescents are in risk to present such disease. An important question for future studies is to determine whether an optimal BMI for sport performance exists. Despite its inherent limitations (i.e. inability to discriminate between fat and fat-free mass), BMI association with physical fitness could be examined without difficulty by coaches and trainers in a variety of sport settings.