O

Md

ACa

b

c

d

e

f

a

AAA

KBISA

1

tdbb

gg

Pf

1d

Joint Bone Spine 79 (2012) 403–408

Available online at

www.sciencedirect.com

riginal article

uscle strength and soccer practice as major determinants of bone mineralensity in adolescents

ndré Seabraa,∗, Elisa Marquesb, João Britoa, Peter Krustrupc,d, Sandra Abreub, José Oliveirab,arla Rêgoe,f, Jorge Motab, António Rebeloa

Faculty of Sport, University of Porto, Centre of Research, Education, Innovation and Intervention in Sport, Porto, PortugalFaculty of Sport, University of Porto, Research Centre in Physical Activity, Health and Leisure, Porto, PortugalUniversity of Copenhagen, Section of Human Physiology, Department of Exercise and Sport Sciences, Copenhagen, DenmarkUniversity of Exeter, St. Luke’s Campus, College of Life and Environmental Sciences, Sport and Health Sciences, Exeter, United KingdomFaculty of Medicine, University of Porto, Center for Research in Health Technologies and Information Systems, Porto, PortugalChildren and Adolescent Center, CUF - Hospital, Porto, Portugal

r t i c l e i n f o

rticle history:ccepted 21 September 2011vailable online 8 November 2011

eywords:one mineral density

sokinetic strengthoccer practicedolescence

a b s t r a c t

Objectives: To analyse the relationship between isokinetic strength of the lower limb muscles and bonemineral density and content (BMD, BMC) of adolescent male soccer players and age-matched controlsnot involved in sport (12–15 years).Methods: A random sample of 151 young males was divided into soccer players (SG; n = 117) and controlsubjects (CG; n = 34). Peak torque of knee extensors (PTE) and flexors (PTF) was measured during isokineticknee joint movement (90◦/s) of the dominant and non-dominant lower limbs. BMD and BMC of thewhole-body, lumbar spine, dominant/non-dominant lower limb were determined by dual-energy X-rayabsorptiometry. Physical activity was monitored with accelerometers during 5 days. Estimated maturityoffset was used as an indicator of biological maturity status.Results: Whole-body BMD (1.03 ± 0.01 vs. 0.98 ± 0.01 g/cm2, P = 0.003) and dominant (1.09 ± 0.01 vs.1.02 ± 0.01 g/cm2, P < 0.001) and non-dominant (1.09 ± 0.01 vs. 1.01 ± 0.01 g/cm2, P < 0.001) lower limbBMD was greater in SG compared to CG. No significant differences were found for BMC. Compared toCG, SG performed better in the YY-IE2 test (780 ± 40 vs. 625 ± 31 m), exhibited higher PTE (dominantlimb: 155.2 ± 30.3 vs. 123.4 ± 37.0 N m; non-dominant limb: 156.2 ± 36.1 vs. 120.4 ± 41.1 N m) and PTF

muscles (dominant limb: 79.0 ± 25.3 vs. 57.1 ± 25.3 Nm; non-dominant limb: 73.3 ± 20.7 vs. 57.0 ± 24.2 Nm). Moreover, the PTE, soccer participation and maturity status were positively associated with the BMDat all body sites (r2 = 0.57–0.73, P < 0.05).Conclusions: Muscle strength of knee extensors is associated with BMD and BMC at all body sites. Muscle-skeletal structures respond positively to the weight-bearing and impact-loading imposed by soccer

o be ancais

practice. Soccer seemed t© 2011 Société fra

. Introduction

Data from prospective and retrospective cohort studies sus-ained that physically active children have higher bone mineralensity (BMD) than sedentary controls [1] and sports practice haseen reported to have a great developing impact during growth,oth in BMD [2] and strength [3].

It was hypothesized that higher BMD may be a function ofreater muscle strength [4]. The muscle–bone relationship duringrowth could be explained by the mechanostat theory [5], as bigger

∗ Corresponding author. Faculdade de Desporto, Universidade do Porto, R. Dr.lácido Costa, 91, 4200 450 Porto, Portugal. Tel.: +35 122 507 4771;ax: +35 122 550 0689.

E-mail address: aseabra@fade.up.pt (A. Seabra).

297-319X/$ – see front matter © 2011 Société francaise de rhumatologie. Published by Eoi:10.1016/j.jbspin.2011.09.003

multilateral balanced sport activity.e de rhumatologie. Published by Elsevier Masson SAS. All rights reserved.

muscles exert higher tensile forces on the bones they are attached.The exercise takes two separate ways of developing bone resis-tance: one that occurs from high impact loads from sports; and anindirect osteogenic way, by the development of muscles in a spe-cific area, making a higher tension on the bone attached to them[6,7]. Since the skeletal muscle is the primary component of leanmass, participation in sport would not have a single direct, but alsoan indirect osteogenic effect, by increasing muscle mass and hencethe tensions generated on bones [7].

Beneficial effects for bone mass may be achieved by promotingparticipation in high impact sports, which submit the growing

skeleton to frequent strains in different directions [7–9]. Soccer is aweight-bearing and intense intermittent sport with billions of par-ticipants around the world, both at competitive and recreationallevel [10]. The forces generated while rapidly changing direction,

lsevier Masson SAS. All rights reserved.

4 ne Sp

scbhdtsia

pdlaeapipl2m

ictns

ioisB

2

2

cnolwsss(ej3wvp4oimw

owwp

04 A. Seabra et al. / Joint Bo

topping and landing, as well as during jumping and kicking mayonfer excellent osteogenic properties to soccer, at least in weight-earing bone sites [11]. In fact, male soccer players showed 10–20%igher BMD than controls in the lower limb [12,13]. In the lastecade, few interventions studies have been conducted to analysehe effect of regular participation in soccer training on healthtatus [10]. In these studies, bone mineral content (BMC) and BMDncreased after 3 and 15 months of soccer training in men [14] andfter 14 weeks and 16 months of soccer participation in women [8].

There is strong empirical evidence showing that sports partici-ation could affect the normal development of muscle and boneuring growth. Prepubertal male soccer players showed higher

umbar spine and femoral BMD than their non-physically activege-, weight- and height-matched control counterparts [11]. How-ver, to our knowledge, no studies so far were performed within thedolescent population. Adolescence, defined as the period betweenuberty and maturity, provides a “window of opportunity” for pos-

tive skeletal adaptations to mechanical loading unlike any othereriod in life [15]. In fact, the peak velocity of bone mass accumu-

ation occurs around the age of 12–14 years in boys [16] and at least5% of the adult BMC is attained in just a 2-year period of fast boneineral accrual during growth [17].Data from cross-sectional studies also suggest that unilateral

ndividual sports practice may lead to bilateral differences in mus-le strength and bone accrual [18,19]. Differences in BMD seemo be site-specific and may be associated with the type and mag-itude of the weight-bearing and impacts acting on each boneite.

The aim of this study was to analyse the relationship betweensokinetic strength of the lower limb muscles, and BMD and BMCf adolescent male soccer players and age-matched controls notnvolved in sport. It was also investigated in what extent theoccer practice might induce bilateral differences in BMD andMC.

. Methods

.1. Participants

The adolescent Caucasian males (n = 151) analysed in this studyame from two different sources. The soccer players group (SG;= 117) were recruited in different five sports clubs from the cityf Porto on a voluntary basis. All the players had completed ateast 11 months of training and playing an official match every

eek for the last 9 months. In addition, they were involved inoccer training for at least 3 years, performing 3–5 training ses-ions per week (≈300 min per week). In general, soccer trainingessions lasted 60–90 min and included a 15–20 min warm-up10 min of low intensity running, 10 min of stretching and strengthxercises), 25–30 min of soccer skill exercises (kicking, dribbling,umping and running with fast accelerations and decelerations) and0–40 min of formal game practice. The control group (CG; n = 34)ere recruited from a secondary school of the city of Porto on a

oluntary basis. All the adolescents only participated in the com-ulsory physical education classes at school (two sessions per week,5–90 min each) and did not participate in any kind of regular orrganized sport activity for at least 3 years. Neither of the partic-pants had recent fractures, chronic diseases, or had been taking

edication known to affect bone metabolism (e.g., corticosteroids)ithin 6 months of the testing date.

The experimental protocol followed the Declaration of Helsinki

f the World Medical Association for research with humans andas approved by the Institutional Review Board. All participantsere fully informed about the aims, experimental protocol androcedures and their parents provided written informed consent.

ine 79 (2012) 403–408

2.2. Anthropometry

Height and sitting height was measured with a fixed stadiome-ter (Holtain Ltd.) (± 0.1 cm) and body mass was estimated with abody fat monitor (Tanita®, BC-418MA) (± 0.1 kg). The body massindex (BMI) was calculated using the standard formula: body mass(kg)/height2 (m).

2.3. Body composition

Whole BMC (g) and BMD (g/cm2), as well as body fat percent-age and lean body mass were determined by dual-energy X-rayabsorptiometry (DXA; Hologic QDR 4500A, Hologic Inc., Waltham,MA, USA). The equipment was calibrated according to the manu-factures instruction and well-trained technician made the exams.Subjects were scanned in supine position and the scans were per-formed in high resolution. BMC and BMD were measured for thewhole body and the lumbar spine (L1–L4) using standard protocolsand the dominant and non-dominant lower limb, using a regionof interest program. The same investigator analysed all total bodyscans. The principles behind body composition analyses with DXAare explained elsewhere [20].

2.4. Biological maturity status

Maturity offset, that is, time before or after PHV, was pre-dicted with the equation of Mirwald et al. [21]: maturityoffset = –9.376 + 0.0001882 × leg length and sitting eight interac-tion + 0.0022 × age and leg length interaction + 0.005841 × age andsitting height interaction – 0.002658 × age and weight interaction+ 0.07693 × weight by height ratio, where r = 0.94, r2 = 0.89 andSEE = 0.57. Length measurements are in centimetres and weightmeasurements are in kilograms; the weight by height ratio ismultiplied by 100.

2.4.1. Socioeconomic status (SES)SES was assessed by the occupation of the head of household.

Head of household was defined in terms of who had the dominantoccupational position (mother or father) as suggested by Kunst et al.[22]. This indicator of SES was coded from questions asking adoles-cents about their parents jobs. The occupation was categorised intothree levels: high (professionals), medium (intermediate skilledworkers) and low (semi-skilled and unskilled manual workers).

2.5. Isokinetic dynamometry

In order to evaluate players’ lower limb muscle strength, maxi-mal gravity corrected concentric peak torque of the knee extensorsand flexors were measured during isokinetic knee joint movement(Biodex System 2, NY, USA) of the dominant and non-dominantlower limbs, at the angular velocity of 90◦/s. After individualself-report, the dominant leg was determined by a routine visualinspection in a simple target-kicking test requiring accuracy.Prior to muscle function measurements, subjects performed astandardized warm-up consisting of 5-min period on a cycleergometer. Players were then seated on the dynamometer chairat 85◦ of inclination (external angle from the horizontal) withstabilization straps at the trunk, abdomen and thigh to preventinaccurate joint movements. The knee to be tested was positionedat 90◦ of flexion (0◦ = full extension) and the axis of the dynamome-ter lever arm was aligned with the distal point of the lateral femoral

condyle. Subjects were instructed to hold their arms comfortablyacross their chest to further isolate knee joint flexion and exten-sion movements. All subjects performed a specific sub-maximalwarm-up protocol on the Biodex device in order to familiarize

ne Spine 79 (2012) 403–408 405

wr

2

bcafra

2

(o5aow

2

tecciw(ibiiag

2

staiatwbidp

3

TbdI2dico

Table 1Anthropometric, body composition, biological maturation, calcium intake and phys-ical fitness of soccer group (SG) and control group (CG).

Characteristics SG (n = 117) CG (n = 34) P

Age (years) 13.8 (1.5) 13.3 (1.3) 0.082

SES 0.233Low n (%) 29 (24.8) 13 (38.2)Medium n (%) 22 (18.8) 7 (20.6)High n (%) 66 (56.4) 14 (41.2)

Height (m) 161.2 (10.1) 160.6 (10.1) 0.775

Weight (kg) 55.1 (12.0) 53.7 (12.7) 0.573

BMI (kg/m2) 20.4 (2.8) 21.3 (4.0) 0.141

% fat 19.7 (6.6) 20.5 (10.0) 0.645

Lean body mass (kg)Total 45.9 (10.8) 40.2 (8.9) 0.348Trunk 19.0 (5.3) 18.4 (5.3) 0.708Right arm 2.4 (0.7) 2.6 (0.7) 0.129Left arm 2.2 (0.7) 2.5 (0.6) 0.128Lower limb dominant 7.5 (2.1) 7.2 (1.6) 0.304Lower limb non-dominant 7.5 (2.1) 7.3 (1.6) 0.460

Maturity offset (years) 1.1 (1.8) 0.5 (1.4) 0.049

Peak torque (N m)Qdom 155.2 (30.3) 123.4 (37.0) < 0.001Qndom 156.2 (36.1) 120.4 (41.1) < 0.001Hdom 79.0 (25.3) 57.1 (25.3) < 0.001Hndom 73.3 (20.7) 57.0 (24.2) 0.011

YY-IE2 (m) 780 (40) 625 (31) 0.023

Calcium intake (mg) 970 (80) 1024 (69) 0.802

MVPA (minutes) 146.1 (22.8) 157.3 (77.6) 0.139

Stepwise multiple linear regression analysis demonstrateda significant positive association between bone variables andbiological maturity status and peak torque of the extensors mus-cles of the dominant lower limb (Table 3). BMD variables and

Table 2Adjusted regional bone mineral density (BMD) and bone mineral content (BMC) insoccer group (SG) and control group (CG).

Characteristics SG (n = 117) CG (n = 34) P

BMD (g/cm2)Whole-body 1.03 (0.01) 0.98 (0.01) 0.006Lumbar spine 0.89 (0.01) 0.85 (0.01) 0.079Dominant lower limb 1.14 (0.01) 1.06 (0.02) < 0.001Non-dominant lower limb 1.15 (0.01) 1.05 (0.02) < 0.001

BMC (g)Whole-body 1844.51 (21.94) 1815.82 (40.41) 0.538

A. Seabra et al. / Joint Bo

ith the isokinetic device and test procedure. Three maximalepetitions at an angular velocity 90◦/s were therefore carried out.

.6. Yo-Yo intermittent endurance test – level 2 (YY-IE2)

The YY-IE2 consists of repeated 2 × 20-m runs back and forthetween the start and finish line at a progressively increased speedontrolled by audio bleeps from a tape-recorder, with a 5-s periodctive recovery between runs [23]. The aim of the test was to per-orm as many runs as possible. When the players failed twice toeach the finish line in time, the distance covered was recordednd used as the test result. Only one trial was given for this test.

.7. Daily physical activity

Daily physical activity was assessed using GT1 M accelerometersActigraph, 72 Pensacola, Florida). All participants provided greaterr equal to 5 days of accelerometer data with greater or equal to00 min of data per day. To provide an indication of the volumend intensity of the participants’ physical activity, mean minutesf moderate to vigorous intensity physical activity per day (MVPA)ere obtained.

.7.1. Calcium intakeEach subject completed questionnaires regarding medical and

raining history. Daily intake of dairy products was obtained fromvery subject by a 24-h nutritional recall interview, in order to cal-ulate the calcium intake. During the 24-h nutritional recall, eachhild was asked to recall all food and beverages consumed dur-ng the past 24-h. Portion sizes of foods and beverages consumed

ere estimated, using food models and photos and other propscups, glasses, food wrappers or containers) as an aid in determin-ng serving sizes. The conversion of food into nutrients was doney the Food Processor SQL® (ESHA Research, USA). The 24-h recall

s the most commonly used dietary assessment method because its easy to administer and can be used to assess adequacy of energynd nutrient intakes [24]. All questionnaires were completed underuidance of the same investigator.

.8. Statistical analysis

Results were expressed either as means (or proportions) andtandard deviations for the anthropometric, SES, body composi-ion, biological maturity status, calcium intake, physical activitynd physical fitness characteristics. Sports participation differencesn all the characteristics were tested using an independent t testnd Pearson’s Chi2 test. Analysis of covariance (ANCOVA) was usedo compare BMD and BMC adjusted for maturity offset. A for-ard stepwise multiple linear regression analyses were conducted

etween maturity offset, sport participation and strength tests asndependent variables and each of the bone mass measures asependent variable. Significance level was set at 0.05. Statisticalrocedures were done using SPSS 18.0.

. Results

Descriptive statistics by sport participation are summarized inable 1. Soccer players were significantly more matured than CG,ut there were no sports participation differences in the otheremographic and biological variables. SG presented 25% higher YY-

E2 test performance than CG. Knee extensors peak torque was6 and 30% higher in SG than CG for the dominant and non-

ominant lower limb, respectively, with corresponding differences

n hamstring peak torque of 38 and 29%, respectively. No signifi-ant bilateral differences were found for SG or CG for peak torquef knee extensors and flexors.

Data are presented as means (standard deviation). LBM: lean body mass; Q: kneeextensors; H: knee flexors; dom: dominant lower limb; ndom: non-dominant lowerlimb. MVPA: moderate to vigorous intensity physical activity.

Given that there were significant differences in biologicalmaturation between the two groups, total and regional BMD andBMC comparisons were calculated with adjustments for maturityoffset (Table 2). After adjustment, the SG had significantly higherwhole-body BMD and higher dominant and non-dominant lowerlimb BMD than CG (Table 2, P < 0.05). No significant differenceswere found for lumbar spine BMD and BMC variables between thetwo groups (Table 2).

A significant positive relationship was observed between thepeak torque of knee extensors and flexors (P < 0.05) and BMD of thedominant lower limb in both groups, with r values ranging from0.41–0.64 (Figs. 1 and 2).

Lumbar spine 45.21 (0.71) 46.87 (1.32) 0.274Dominant lower limb 373.00 (5.32) 352.74 (9.79) 0.074Non-dominant lower limb 368.88 (5.31) 356.29 (9.79) 0.067

Data are presented as means (standard error). *Adjusted to maturity offset.

406 A. Seabra et al. / Joint Bone Spine 79 (2012) 403–408

Table 3Multiple linear regression and adjusted coefficient estimates and 95% confidence intervals (CI) with bone mineral density (BMD) and bone mineral content (BMC) as dependentvariables.

Dependent variables Independent variables � (95% CI) P r2 adjusted

BMD (g/cm2)Whole-body Maturity offset 0.03 (0.01–0.05) 0.001 0.57

Sports participation 0.08 (0.04–0.13) 0.001 0.61Peak torque Qdom 0.001 (0.000–0.002) 0.001 0.66

Lumbar spine Maturity offset 0.05 (0.03–0.07) < 0.001 0.58Sports participation 0.07 (0.02–0.13) 0.013 0.60Peak torque Qdom 0.001 (0.000–0.002) 0.024 0.62

Dominant lower limb Maturity offset 0.04 (0.02–0.07) 0.001 0.60Sports participation 0.15 (0.09–0.21) < 0.001 0.68Peak torque Qdom 0.001 (0.000–0.002) 0.004 0.71

Non-dominant lower limb Maturity offset 0.05 (0.02–0.07) < 0.001 0.62Sports participation 0.16 (0.10–0.22) < 0.001 0.71Peak torque Qdom 0.001 (0.000–0.002) 0.007 0.73

BMC (g)Whole-body Maturity offset 0.001 (0.000–0.002) < 0.001 0.70

Peak torque Qdom 0.11 (0.06–0.15) < 0.001 0.76Lumbar spine Maturity offset 5.86 (4.28–7.44) < 0.001 0.66

Peak torque Qdom 0.10 (0.04–0.17) 0.002 0.70Dominant lower limb Maturity offset 35.98 (22.63–49.33) < 0.001 0.69

Peak torque Qdom 0.88 (0.42–1.34) < 0.001 0.73Sports participation 52.94 (21.50–84.38) 0.001 0.76

Non-dominant lower limb Maturity offsetPeak torque Qdom

Sports participation

Fig. 1. Correlations between peak torque of the dominant lower limb of soccer group(SG) (n = 117) knee extensors and bone mineral density (BMD) of the control group(

drwa

Fs(

CG) (n = 34).

ominant and non-dominant lower limb BMC was positivelyelated to sport participation. No association was found betweenhole-body and lumbar spine BMC and sport participation among

dolescents.

ig. 2. Correlations between peak torque of the knee flexors and bone mineral den-ity (BMD) of dominant lower limb of soccer group (SG) (n = 117) and control groupCG) (n = 34).

37.33 (24.08–50.57) < 0.001 0.700.81 (0.36–1.27) 0.001 0.7354.05 (22.86–85.23) 0.001 0.76

4. Discussion

This study aimed to analyse the relationship between bonemass and knee muscle strength in adolescent male soccer play-ers and age-matched controls not involved in sport. Compared tonon-exercising group, soccer players performed better in the YY-IE2 test, were early matured, exhibited higher peak torque of theknee extensor and flexor muscles, and had higher whole-body anddominant and non-dominant lower limb BMD. The peak torque ofthe knee extensors of the global sample was positively associatedwith BMD and BMC at all analysed body sites. Moreover, soccerpractice was associated with BMD in all sites and dominant andnon-dominant lower limb BMC.

4.1. Strength, YY-IE test performance and bone characteristics

The results of the present study showed that SG had higher peaktorque of the knee extensor muscles than their matched-controls.In general, soccer players have more muscle strength than inactivepeople, mainly in the lower limb, independently of age and gender.Previous studies in female and male soccer players showed similarresults. Indeed, in a previous study it was found that youngfemale soccer players had significantly higher concentric andeccentric peak torque of the thigh muscles than controls [25]. Inother study, while young male soccer players conventionally- orresistance-trained showed higher values of isokinetic concentricand eccentric strength of the lower limb extensor and flexormuscles of the knee joint of the dominant and non-dominant limbthan non-soccer players [26]. Moreover, our results showed thatSG covered a higher distance in the YY-IE2 test than the controls,which was in accordance the statement that soccer practice highlytaxes the aerobic and anaerobic metabolism and the ability toperform high-intensity intermittent exercise [23].

The SG presented a significantly higher whole-body andlower limbs BMD than the CG, which corroborates previouscross-sectional studies that compared youth soccer players with

non-active controls of both sexes. Soderman et al. [25] showed thatadolescent female soccer players had significantly higher whole-body BMD (2.7%), lumbar spine BMD (6.1%) and dominant andnon-dominant hip BMD than non-active girls. Similar results were

ne Sp

anaawc

fasst[billtlbmt

4d

bsnsewTlldrsibfi

4d

btpoisip3gipi1oga2a

A. Seabra et al. / Joint Bo

lso observed in pubertal male soccer players that showed sig-ificantly higher whole-body, lumbar spine, femoral neck, pelvisnd lower limb BMD than their counterpart controls [27]. Also, instudy with 9-year old children, femoral and lumbar spine BMDas found to be greater in soccer players than in their age-matched

ontrols [11].It is well known that differences in bone mass between groups

rom different sports are site-specific and associated with the typend magnitude of the loading during training [28]. Thus, previoustudies analyzing muscle and bone properties in single-handedports (e.g. tennis and golf) found higher strength and BMD inhe arm used to throw the ball than in the contralateral arm18,19]. Because no differences in muscle strength and bone massetween dominant and non-dominant lower limbs were observed

n SG, our results support the concept that soccer imposes highevels of muscle contractions and weight-bearing forces, on bothower limbs. In fact, soccer stimulates the dominant lower limb inechnical movements such as the kick on the ball and the oppositeower limb is fundamental for body support in several soccerody actions. These results suggest that soccer is a well-balancedultilateral activity that strengthens both, muscles and bones of

he both sides of the body.

.2. Relationship between muscle strength and bone mineralensity (BMD)

Our results showed a significant positive association betweenone variables, biological maturation and peak torque of the exten-or muscles of the dominant lower limb, which is in line with aumber of previous studies examining the muscle–bone relation-hip in children and adolescents [29]. In fact, Vicente-Rodriguezt al. [11] found that maximal isometric force had a high correlationith total and regional BMC and BMD in male prepubertal players.

hey also demonstrated that femoral and lumbar BMC, and, to aower extent, BMD were also correlated with the muscle mass of theower limb. Thus, the results of the present study pinpoint that boneevelopment is influenced by muscle development, which corrobo-ates the well-accepted mechanostat theory. In brief, the mechano-tat theory postulates the key point that developmental changesn bone strength are secondary to the increasing loads imposedy larger muscle forces [5] which is strictly related with ourndings.

.3. Relationship between soccer practice and bone mineralensity (BMD)

Physical activity can contribute to the close associationetween muscle and bone, since both factors respond favourablyo increased loading [29]. In the present study, soccer practice wasositively associated with lower limb BMD. These results corrob-rate cross-sectional studies that described that soccer practicenduces the strengthening of bone mass [30–32]. Interventiontudies also have evidenced a positive influence of soccer practicen bone mass. In a study by Vicente-Rodriguez et al. [33], a group ofrepubertal soccer players (Tanner < 2) was followed throughout a-year period. At the end of the intervention, the players exhibitedreater BMC in the legs and BMD in all bone-loaded. Also, a signif-cant bone gain was observed in weight-bearing bones of soccerlayers in comparison with controls. Similar results were observed

n the adult population. Adult women (9–49 years) increased by.3 and 2.3% whole-body BMD after 4 and 16 months, respectively,f soccer training, while no changes were observed in a running

roup and in controls [14]. Bone mass of the lower extremitieslso increased after 12 weeks of training, in a group of men aged0–43 years that played soccer, but not in those of the runningnd control groups [34]. Additionally, adult women averaging 36

ine 79 (2012) 403–408 407

years old playing recreational soccer during 14 weeks improvedvolumetric BMD in the distal tibia, suggesting a decreased riskof fracture, due to stronger bones, and a reduced risk of falling[8]. Similarly, leg bone mass (3.5%) and density (2.0%) of malerecreational soccer players (20–43 years) were increased after 64weeks of training [35].

Contrasting with a study that found that soccer was associatedwith higher bone mass only in the lower extremities [32], thepresent study showed that soccer practice was positively asso-ciated with lumbar spine and whole-body BMD. However, itmust be underlined that the sample analysed in the study byZouch et al. [32] included prepubescent boys, with less time ofaccumulated soccer practice. Our results indicate that soccer hasthe potential to strengthen the bone not only in the body segments(lower limbs) more stimulated during the practices, but also inthe other body compartments. This might be linked to the factthat soccer is an exercise activity that activates a wide range ofbody muscles in order to fulfil the physical demands of the game,such as sudden changes of speed running, kicking the ball andcorporal contacts against the opponent players. Furthermore, assoccer involves a high number of jumps and other unorthodoxmovements requiring abrupt changes of the position of the bodycentre of mass, it should be considered a typical weight-bearingsport. This might be revealed since high impact forces may be ofgreater importance in regulating bone mass. However, it mightno ne neglected that, during puberty, bone modification resultingfrom high-impact soccer activity may be mediated or promotedby high levels of somatotropic hormones, since a relationshipbetween the somatotropic hormonal axis (GH, IGF-1, and IGFBP-3)and whole-body BMD was observed in soccer players aged 13years, but not in their non-physically active counterparts [27].

The present study provides several important limitations thatshould be addressed. The cross-sectional design of the currentstudy is a first limitation. This may limit conclusions regardingthe direction of relations between soccer participation, isokineticstrength and each of bone variables, because it is difficult todetermine if bone variables are influenced by a correlate or viceversa. These results should therefore be interpreted with sufficientcaution and require replication in others cross-sectional and lon-gitudinal studies before any definite conclusions can be drawn.Given the lack of such studies at this moment, however, we fellthat these results provide important, although limited evidenceabout the relative contribution of soccer participation and isoki-netic strength in adolescent’s bone variables. The sample size of thisstudy is a second limitation. The sample size was small, particularlythe non-exercising group, which might have reduced the statisti-cal power on the group comparisons and hence the generalizabilityof the results may be limited. However, our post hoc statisticalpower tests for detecting BMD and BMC differences among thetwo groups ranged from 61% to 99%. Other limitations include theethnic and socioeconomic status homogeneity of the participants,which reduces generalizability and eliminates the ability to assessthese relationships in different ethnic and socioeconomic popula-tions. Despite the limitations, this study has several strengths. First,the study used rigorous instrument for measuring BMD and BMC.Second, the study includes a large number of adolescent soccerplayers compared to other studies [9,11,25–27,31–33]. Third, thestudy establish important and unique information regarding therelationship between, soccer participation and isokinetic strengthof the lower limb muscles and BMD and BMC of adolescent males.

The present study showed that soccer participation was pos-itively associated with improved physical fitness and increased

BMD in the lower limb, lumbar spine and whole-body, as wellas with higher leg muscle strength, indicating that the muscle-skeletal structures respond positively to the weight-bearing andimpact-loading imposed by soccer practice. Additionally, since no

4 ne Sp

dasbtam

D

c

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[[

[

[

[

[

[

[

[

[

[

08 A. Seabra et al. / Joint Bo

ifferences in muscle strength were found between the dominantnd non-dominant lower limbs of the soccer players and as theyhowed higher bone mass values than controls, soccer seemed toe a multilateral balanced sport activity. These results reinforcehe existing guidelines for sports participation and bone health,nd highlight the positive contribution of soccer to augment boneineral accrual during growth.

isclosure of interest

The authors declare that they have no conflicts of interest con-erning this article.

eferences

[1] Pitukcheewanont P, Punyasavatsut N, Feuille M. Physical activity and bonehealth in children and adolescents. Ped Endocrin Rev 2010;7:275–82.

[2] Daly RM. The effect of exercise on bone mass and structural geometry duringgrowth. Med Sport Sci 2007;51:33–49.

[3] Kellis S, Kellis E, Manou V, et al. Prediction of knee extensor and flexorisokinetic strength in young male soccer players. J Orthop Sports Phys Ther2000;30:693–701.

[4] Fricke O, Schoenau E. The “Functional Muscle-Bone Unit”: probing the rele-vance of mechanical signals for bone development in children and adolescents.Growth Horm IGF Res 2007;17:1–9.

[5] Rauch F, Bailey DA, Baxter-Jones A, et al. The “muscle-bone unit” during thepubertal growth spurt. Bone 2004;34:771–5.

[6] Rahnama N, Lees A, Bambaecichi E. Comparison of muscle strength and flexi-bility between the preferred and non-preferred leg in English soccer players.Ergonomics 2005;48:1568–75.

[7] Vicente-Rodriguez G. How does exercise affect bone development duringgrowth? Sports Med 2006;36:561–9.

[8] Helge EW, Aagaard P, Jakobsen MD, et al. Recreational football trainingdecreases risk factors for bone fractures in untrained premenopausal women.Scand J Med Sci Sports 2010;20:31–9.

[9] Vicente-Rodriguez G, Ara I, Perez-Gomez J, et al. Muscular development andphysical activity as major determinants of femoral bone mass acquisition dur-ing growth. Br J Sports Med 2005;39:611–6.

10] Krustrup P, Aagaard P, Nybo L, et al. Recreational football as a health-promotingactivity: a topical review. Scand J Med Sci Sports 2010;20:1–13.

11] Vicente-Rodriguez G, Jimenez-Ramirez J, Ara I, et al. Enhanced bone mass andphysical fitness in prepubescent footballers. Bone 2003;33:853–9.

12] Wittich A, Mautalen CA, Oliveri MB, et al. Professional football (soccer) playershave a markedly greater skeletal mineral content, density and size than age-

and BMI-matched controls. Calcif Tissue Int 1998;63:112–7.

13] Karlsson MK, Linden C, Karlsson C, et al. Exercise during growth and bonemineral density and fractures in old age. Lancet 2000;355:469–70.

14] Krustrup P, Nielsen JJ, Krustrup BR, et al. Recreational soccer is an effectivehealth-promoting activity for untrained men. Br J Sports Med 2009;43:825–31.

[

ine 79 (2012) 403–408

15] Greene DA, Naughton GA. Adaptive skeletal responses to mechanical loadingduring adolescence. Sports Med 2006;36:723–32.

16] Bailey DA, McKay HA, Mirwald RL, et al. A six-year longitudinal study of therelationship of physical activity to bone mineral accrual in growing children:the university of Saskatchewan bone mineral accrual study. J Bone Miner Res1999;1410:1672–9.

17] Bailey DA, Martin AD, McKay HA, et al. Calcium accretion in girls and boysduring puberty: a longitudinal analysis. J Bone Miner Res 2000;15:2245–50.

18] Calbet JA, Moysi JS, Dorado C, et al. Bone mineral content and density in pro-fessional tennis players. Calcif Tissue Int 1998;62:491–6.

19] Dorado C, Sanchis Moysi J, Vicente G, et al. Bone mass, bone mineral densityand muscle mass in professional golfers. J Sports Sci 2002;20:591–7.

20] Kelly TL, Berger N, Richardson TL. DXA body composition: theory and practice.Appl Radiat Isot 1998;49:511–3.

21] Mirwald RL, Baxter-Jones AD, Bailey DA, et al. An assessment of maturity fromanthropometric measurements. Med Sci Sports Exerc 2002;34:689–94.

22] Kunst AE, Bos V, Mackenbach JP. Monitoring socioeconomic inequalities inhealth in the European Union: guidelines and illustrations. A report for theHealth Monitoring Program of the European Comission. Rotterdam: ErasmusUniversity; 2001.

23] Bangsbo J. Fitness training in football: a scientific approach. Bagsvaerd:HO + Storm; 1994.

24] Willet W. Nutritional epidemiology. 2nd ed. Oxford University press; 1998.25] Soderman K, Bergstrom E, Lorentzon R, et al. Bone mass and muscle strength

in young female soccer players. Calcif Tissue Int 2000;67:297–303.26] Iga J, George K, Lees A, et al. Cross-sectional investigation of indices of isokinetic

leg strength in youth soccer players and untrained individuals. Scand J Med SciSports 2009;19:714–9.

27] Nebigh A, Rebai H, Elloumi M, et al. Bone mineral density of young boy soccerplayers at different pubertal stages: relationships with hormonal concentra-tion. Joint Bone Spine 2009;76:63–9.

28] Pettersson U, Nordstrom P, Lorentzon R. A comparison of bone mineral densityand muscle strength in young male adults with different exercise level. CalcifTissue Int 1999;64:490–8.

29] Daly RM, Stenevi-Lundgren S, Linden C, et al. Muscle determinants of bonemass, geometry and strength in prepubertal girls. Med Sci Sports Exerc2008;40:1135–41.

30] Falk B, Galili Y, Zigel L, et al. A cumulative effect of physical training on bonestrength in males. Int J Sports Med 2007;28:449–55.

31] Falk B, Braid S, Moore M, et al. Bone properties in child and adolescent malehockey and soccer players. J Sci Med Sport 2010;13:387–91.

32] Zouch M, Jaffre C, Thomas T, et al. Long-term soccer practice increases bonemineral content gain in prepubescent boys. Joint Bone Spine 2008;75:41–9.

33] Vicente-Rodriguez G, Ara I, Perez-Gomez J, et al. High femoral bone min-eral density accretion in prepubertal soccer players. Med Sci Sports Exerc2004;36:1789–95.

34] Krustrup P, Hansen PR, Andersen LJ, et al. Long-term musculoskeletal and car-diac health effects of recreational football and running for premenopausal

women. Scand J Med Sci Sports 2010;20:58–71.

35] Randers MB, Nielsen J, Krustrup BR, et al. Positive performance and healtheffects of a football training program over 12 weeks can be maintained overa 1-year period with reduced training frequency. Scand J Med Sci Sports2010;20:80–9.