This study investigated the relationship between physical performance and PA patterns in prepubertal children by means of EUROFIT tests and high frequency accelerometry measurement. The main finding is that the physical performance was not related to physical activity level in 6- to 12-yr-old children. Only a trivial negative relationship between PA level and body fatness was observed in boys with short bouts of VPA (from 5 to 15s) that made a significant contribution in percentage body fat variation as demonstrated by the multivariate analysis.
From childhood to adolescence, the literature has generally shown a weak to moderate relationship between PA and physical performance . We hypothesized that high frequency accelerometry monitoring (5s epoch) should be able to provide a more accurate measurement of children PA behaviors, notably to capture short bouts of VPA and VHPA that are generally diluted in PA measurement when accelerometer epoch is set at 1 min. In the current study, no relationship was found between PA and physical performance. As observed in adults , these results suggests that habitual activity did not show adequate intensity, volume, and duration to induce positive changes in motor and functional capacities. Indeed, VPA or VHPA are generally considered as the upper boundaries of physical activity domains in the context of health related studies. The cut off used in the present study for VPA and VHPA were derived from counts values associated with velocities of 6.4 km.h-1 (6 METs) and 9.7 km.h-1(9 METs), respectively. However, these intensities of exercise remain largely lower than those reached in the context of performance. For instance, in the present study, the running velocity associated with 20-MST performance (around 10 km.h-1) was certainly not sustained for sufficient long periods to expect an improvement in aerobic fitness. Similarly, a 9.7 km.h-1 velocity (VHPA) remained two times lower than a sprint velocity in such a population that is around 20 km.h-1. Thus, it could be hypothesized that systematic training and not time spent in VPA or VHPA is needed to increase physical performances such as SBJ, SHR or 20-MST.
Only for boys, time spent in LPA was positively correlated with body fatness, whereas VPA and VHPA were negatively associated. Rowlands et al.  reported a negative relationship between fatness and PA in 8- to 10-yr-old boys and girls, whilst Dencker and Andersen  reported only low to moderate inverse relationships between moderate to vigorous PA and body fatness with a comparable population. Using accelerometry, Abbott and Davies  and Dencker et al.  found significant relationships between VPA and VHPA and body fatness, but no relationship was reported with MPA. These findings agree with the present study, where trivial relationships between VPA and VHPA and body fatness were also observed. Conversely, Ness et al.  reported higher correlations between moderate to vigorous PA and body fatness in boys of comparable ages. These differences could be explained by an acute measure of fat mass (dual x-ray emission absorptiometry) in the latter study, while skinfold measurement, as used in the present study, remains a less robust method. Nevertheless, the weak relationship found between LPA and percentage body fat in boys only, should also be carefully interpreted. The cut-offs used  in the present study do not allow to differentiate between LPA and sedentary activity, the latter being included in LPA. Body fatness in boys may thus be more related to sedentary activity than to LPA. Few recent studies proposed thresholds to discriminate between LPA and sedentary activity [30–32] but none of them used a 5s- epoch. It is possible that the 5-s epoch might not be appropriate to assess "true" sedentary activities. Indeed, this type of activity would better assessed with longer epochs such as 1-min rather than 5-s epoch. The latter could lead to an overestimation of sedentary activity. For example, during spontaneous PA, where children alternate short bouts of PA with short recovery periods, the recovery bout would be classified as sedentary activity, whereas it is not.
Relationships between waist circumference and body composition seem to be associated with cardiovascular disease risk and suggested that PA may have a beneficial effect, notably with respect to overweight [33, 34]. In the present study, no relationship was found between hip and waist circumferences and times spent from LPA to VHPA. Using accelerometry, Hussey et al.  reported a significant negative correlation between waist circumference and time spent in VPA in boys, but not in girls. Similar conclusions were drawn by Ness et al. .
In the literature, positive relationships between PA and aerobic fitness have generally been observed in children [25, 26]. However, correlations reported in the literature are moderate, therefore Dencker et al.  suggested that the aerobically fitter children were not obligatorily the more active. The results found in the current study showed that aerobic fitness was not associated with overall physical activity. Differences between experimental protocols (epoch length and/or field vs laboratory tests) may explain these discrepancies.
Children's PA is characterized by rapid changes from rest to PA including vigorous intensities. Investigating PA patterns related to health gives information about how to promote PA during childhood. To the best of our knowledge, only one study  has investigated the relationship between PA patterns and physical fitness. They reported that children's sporadic activity was negatively related to waist circumference and aerobic fitness. In the present study, short bouts of VPA in boys were positively related to lower body fatness, whereas LPA and sedentary long bouts were negatively associated. Our results show that PA intensity has a positive influence on the maintenance of a healthy body composition and suggest that children should be engaged in longer VPA sessions. As the majority of children's VPA and VHPA is accumulated via short intermittent bouts rather than long continuous bouts, intermittent activity seems to be efficient to promote healthy body composition and health-related fitness. However, there is an urgent need for more longitudinal studies on children in which PA, physical fitness (or performance), and health are measured repeatedly in the same individual over an extended period of time [36, 37].
More than ninety-five percent of the boys and 77.5% of the girls accumulated 60 or more minutes of moderate to very high PA per day. These results are higher than reported on UK children , but lower compared with some American or European studies [21, 38]. However, the proportion of children reaching PA guidelines should be interpreted with caution. The use of higher accelerometer thresholds to classify intensity or the use of a smaller epoch to monitor PA might explain these discrepancies between studies.
Times spent from MPA to VHPA were lower than reported in the literature. Girls spent more time in LPA than boys (2.1%, p < 0.001), whilst the latter spent more time in MPA (1.5%, p < 0.001) and in VPA (0.4%, p < 0.05). A sex-related difference was only found for 5 to 10-s VHPA bouts (p < 0.001). Hussey et al.  assessed PA during 4 days on Irish children aged 7 to 10. They observed that boys spent twice as much VPA and VHPA as girls (64.3 min vs 37 min each day, p < 0.001). Trost et al.  reported that boys spent about 120 min in MPA, whereas girls spent 100 min per day (p < 0.05). These differences may be explained by the use of different accelerometers [35, 40]; different epochs [21, 39]; number of the monitoring days  and different cut-offs . In the present study a 5-s epoch was used to assess more accurately children's PA patterns in free-living conditions. Vigorous PA and VHPA are captured and not diluted in MPA or LPA. Using a 2-s epoch, Rowlands et al.  reported that differences in PA were largely due to the intensity of the most frequent bouts of activity and the frequency of the most intense bouts. McClain et al.  also reported that shorter epoch lengths should be used to minimize error among individual estimates.