Despite our increasing knowledge of the positive impact of physical activity during childhood for long term health, children are becoming less active (Boreham and Riddoch, 2001; Tremblay et al., 2011a). Childhood sedentary behaviour is strongly associated with obesity (Must and Tybor, 2005; Pearson and Biddle, 2011) and is also known to predict overweight in adolescence and adulthood (Guo et al., 2002; Magarey et al., 2003). In addition, sedentary behaviour is related to other chronic metabolic diseases, such as osteoporosis, type 2 diabetes and ischaemic heart disease (Dunstan et al., 2010; Katzmarzyk et al., 2009; Tremblay et al., 2011a). Increasing physical activity participation in school is notionally a very practical method to improve the health of children at the population level (Boreham and Riddoch, 2001). An exercise regime that would effectively improve the health of multiple body systems, however, is yet to be described. For instance, it is well known that increasing the duration of physical activity is an effective strategy to enhance cardiovascular function (Haskell et al., 2007), but not the skeletal system. To stimulate the latter system, short duration, high intensity loading is required (Lanyon and Rubin, 1984).

Exercise prescriptions can be manipulated in terms of frequency (exercise bouts per week), intensity (metabolic and musculoskeletal load), duration (length of exercise bout), and exercise type (Must and Tybor, 2005). The current physical activity recommendation for children includes the accumulation of 60 minutes per day of moderate to vigorous exercise in addition to activities that strengthen bone and muscle on at least three days per week (Janssen, 2007; Strong et al., 2005; Tremblay et al., 2011b; Twisk, 2001). Measures of intensity traditionally focus on cardiovascular or metabolic load and are classified in terms of heart rate and estimates of energy expenditure (EE) (Haskell et al., 2007). Such measures, however, fail to capture characteristics of mechanical intensity that are vital to the musculoskeletal response. In order to identify childhood activities that are broadly beneficial to multiple systems, it is important to know the intensity of both cardiovascular/metabolic load and musculoskeletal load. Furthermore, from a psychosocial standpoint, variety and enjoyment are critically important elements of an exercise program that contribute to uptake and ongoing engagement (Richard et al., 1997).

Accelerometry is a recognised technique to track physical activity in bone and metabolic research (Janz et al., 2003). Accelerometry-derived weight-bearing movements can predict bone mass and density in young children (Janz et al., 2001; Specker et al., 2001). Biomechanical characteristics of weight-bearing exercises are typically estimated by measuring ground reaction forces (GRF) (Prapavessis and McNair, 1999).

The CAPO Kids program was a recent brief and enjoyable in-school exercise intervention designed to improve the musculoskeletal and metabolic health of pre and peripubertal children (Nogueira et al., 2014; 2015). The goal of the program was to simultaneously apply a moderate to vigorous aerobic load and high intensity impact loading. Exercises were based on capoeira, a Brazilian sport that combines martial art, dance and music, presenting a new and interesting activity for the participants. Capoeira was supplemented with high impact activities including a variety of jumps and upper limb loading activities. The program improved bone health and metabolic factors such as waist circumference, resting heart rate and maximal oxygen uptake over a nine-month period (Nogueira et al., 2014; 2015).

The aim of the present study was to characterise the biomechanical and metabolic loads of the CAPO Kids exercise program, in boys and girls. Those data will allow others to make an informed judgement with regards to the potential of the program to produce metabolic and musculoskeletal benefit or other outcomes.

We conducted a cross-sectional descriptive study of metabolic and mechanical loads experienced by participants in the CAPO Kids trial. Ethical approval was obtained from the Griffith University Human Research Ethics Committee (PES/35/12/HREC), and all participants provided informed consent. Parents had the option to override and decline to consent.

A sample of the Year 5 and 6 primary school children (9.7-11.4 years of age) who were participating in the 9-month CAPO Kids exercise intervention were recruited for the current study. Participants were included if they were healthy, ambulant, and enrolled in the exercise arm of the trial. Participants were excluded if they had a metabolic, endocrine or renal condition; were taking medications known to affect bone, muscle or fat metabolism; were recovering from a serious lower limb fracture or other immobilising injury in the past six months; or were affected by any condition not compatible with short bouts of physical activity. Specific details of school and participant recruitment are available in previous publications (Nogueira et al., 2014; 2015).

Participants were invited to wear a SenseWear Armband (SWA, BodyMedia, Pittsburgh, PA, USA) for the entirety of one CAPO Kids exercise bout in order to have parameters of metabolic intensity measured. The same participants were also invited to attend a single session of testing at the Biomechanics Laboratory at Griffith University in order to have the ground reaction forces (GRF) associated with intervention activities measured on a force platform. Participation in the GRF session was optional. An a priori sample size estimate was not conducted.

Statistical analyses were performed with SPSS version 22.0 for Windows (IBM, Chicago, IL). Descriptive statistics were calculated, including means and standard deviations for participant characteristics and means, standard deviations and 95% confidence intervals for all dependent variables derived from SWA and GRF collection. Independent t tests were used to examine sex differences for metabolic data. Statistical significance was determined to be ≤ 0.05.

A total of 48 children (28 boys; 20 girls) participated in the study, representing approximately 15% of the CAPO Kids group cohort. All participants completed SWA data collection, while a subsample of 11 children (6 boys; 5 girls) completed GRF data collection. Descriptive characteristics are provided in Table 2. Age, weight, standing height, BMI and estimated VO2 max were similar between boys and girls included in the study (Table 2).

Results for metabolic measures including EE (kcal and METs), number of steps and intensity of PA for combined and sex-specific samples are presented in Table 3. The average duration of the CAPO Kids activity was 10.2 ± 0.6 minutes. Both boys and girls had similar EE (38.7 ± 7.3 vs. 41.2 ± 11.8 kcal, p = 0.371), METs (5.5 ± 0.9 vs. 5.6 ± 0.9 METs, p = 0.641); total steps (690 ± 92 vs. 636 ± 97, p = 0.056) and duration in each PA intensity zone. The average EE was 4.0 kcal·min-1 for both groups, which is classified as moderate intensity. Moderate PA intensity lasted an average of 6.9 ± 2.3 min for boys vs. 6.5 ± 2.9 min for girls (p = 0.602); vigorous intensity lasted 3.0 ± 0.8 min for boys vs. 2.6 ± 1.7 min for girls (p = 0.431); and very vigorous intensity lasted 0.5 ± 0.9 min vs. 0.7 ± 0.8 min (p = 0.510) for boys and girls respectively.

Table 4 presents GRF measures including the average of vGRF magnitude and rates of application for each activity performed during the CAPO Kids intervention. The 360° jumps and tuck jumps produced the highest vGRFs (5.4 ± 2.3 and 5.2 ± 2.0 BW, respectively), followed by jump plus martelo (4.8 ± 1.2 BW), and then hop and hop plus martelo (4.0 ± 1.0 and 4.0 ± 1.2 BW, respectively). The smallest vGRFs were recorded during the upper body exercises, cartwheel and handstand (1.3 ± 0.2 and 1.4 ± 0.2 BW, respectively), followed by the ginga and jump-lunge (1.4 ± 0.4 and 1.5 ± 0.6 BW, respectively). The fastest rate of force application was recorded for the jump-lunge (143.6 ± 144.9 BW·s-1) followed by the 360° jump (132.4 ± 78.6 BW·s-1) and the jump plus martelo (121.8 ± 93.5 BW·s-1).

Our aim was to determine the metabolic and musculoskeletal load intensity of the CAPO Kids program; an in-school exercise intervention designed to improve the health of pre and early pubertal children. Specifically, our goal was to quantify energy expenditure and vertical ground reaction forces and rates of force application associated with a typical 10-minute intervention session of capoeira plus jumping.

We used SenseWear armbands to estimate energy expenditure of children participating in the school-based CAPO Kids exercise intervention program. We found that a single 10-minute bout of the intervention induced an absolute EE of 39.7 ± 9.3 kcal or 5.5 ± 0.9 METs, which represents an average EE of 4.0 kcal·min-1 (moderate intensity). Previous pediatric studies using the validated SWA algorithms to assess physical activity in children have reported EE for a variety of activities (Arvidsson et al., 2007; Calabró et al., 2009). For instance, the average EE for walking at 4 km·h-1 was 2.5 kcal·min-1 (low intensity), while the average for cycling was 3.0 kcal·min-1 (moderate intensity); both activities being less vigorous than the CAPO Kids intervention (Calabró et al., 2009). Activities such as basketball, running at 10 km·h-1, and trampolining have been observed to consume more than 7 kcal/min, representing activities performed at a vigorous level (Arvidsson et al., 2007). While the average EE of a single bout of CAPO Kids was only 4.0 kcal·min-1 (moderate intensity), including short periods of rest, 34% of each 10-min session was spent performing activities at a vigorous or very vigorous level, suggesting the CAPO Kids program has the potential to provide a beneficial metabolic stimulus.

The maximal vGRFs produced by participants ranged from 1.3 BW for cartwheels to 5.4 BW for jumps and capoeira manoeuvres. The rates of force application varied from 15 BW·s-1 for cartwheels and ginga to around 140 BW·s-1 for jump-lunges and 360° jumps. Although the vGRF data presented represent a range of magnitudes and rates, five of the 11 measured activities produced vGRFs greater than 4.0 BW and six applied forces at more than 90 BW·s-1. Ground reaction forces have been reported for other common activities such as walking, running, and drop jumps from different heights (Weeks and Beck, 2008, McKay et al., 2005), as well as specific high impact sports such as gymnastics and volleyball (Daly et al., 1999, Salci et al., 2004). Low impact activities generally produce lower GRFs, such as walking (i.e. 1.1 to 1.5 BW) and running (i.e. 2.0 to 3.0 BW), whereas high impact activities such as maximal jumping produce higher peak forces (i.e. 3.0 to 5.0 BW) (Weeks and Beck, 2008). Landings from heights above ground level generate even higher peaks that may reach more than 8.0 BW (Fuchs et al., 2002) with peak forces of up to 10.0 BW recorded for elite athletes from high impact sports such as volleyball and gymnastics (Daly et al., 1999, Salci et al., 2004).

Our GRF results were similar to those reported from other in-school pediatric jumping interventions, with beneficial musculoskeletal outcomes (McKay et al., 2005; Weeks et al., 2008). Those studies reported vGRFs between 3.0 and 5.0 BW from jumping activities such as hops, lunges and jump-squats in their intervention programs for children. Thus, the peak forces and rates of force application produced by the CAPO Kids exercises appear to be sufficiently high to be osteogenic.

The upper extremities are relatively unaccustomed to habitual impact loads and, accordingly, there are very few reports of the GRFs associated with upper limb weight-bearing activities. Forearm fractures, however, are highly prevalent in children (Khosla et al., 2003). As the bone response to mechanical loading is site-specific (Bass et al., 2002; Johannsen et al., 2003), it is important to specifically load the upper extremity to achieve adaptive benefits. We therefore also quantified the loads to which the upper limbs were exposed while weight-bearing during the CAPO Kids program. The upper extremity weight-bearing manoeuvres of handstands and cartwheels in the current study produced vGRFs of 1.3 and 1.4 BW, respectively. Although upper extremity GRFs of gymnasts performing elite level acrobatic movements have been reported up to 3.6 BW (Daly et al., 1999), our values nevertheless represent a large increase in loading to a region of the skeleton that is not typically exposed to weight bearing. Such loading was very simple and feasible for all of the children and may translate to considerable benefits in fracture prevention.

For practical reasons, we used a light-weight, non-invasive, portable device to estimate EE, (i.e. armband sensor). We recognise that direct and indirect calorimetry remains the gold standard to estimate energy expenditure. Similarly, highly invasive techniques of direct measurement of bone strain were not feasible in the current cohort. GRFs were the best practical surrogate of musculoskeletal loading, but may not provide an entirely accurate estimate of the precise loads experiences by the skeleton.

We determined the metabolic and musculoskeletal load intensity of the CAPO Kids exercise intervention. We found that the CAPO Kids program has the ability to apply significant load intensity to both the metabolic and musculoskeletal systems of pre and early pubertal children. We conclude that the CAPO Kids program is likely to be an appealing, feasible and effective strategy to load the musculoskeletal and metabolic systems of children via in-school programming.