The contribution of growth and maturation in the functional capacity and skill performance of male adolescent handball players

543Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

accepted after revision November 18 , 2011

BibliographyDOI http://dx.doi.org/10.1055/s-0031-1298000Published online: May 4, 2012Int J Sports Med 2012; 33:543–549 © Georg ThiemeVerlag KG Stuttgart · New YorkISSN 0172-4622

Correspondence Stijn PJ Matthys, MSc Ghent University Movement and Sports Sciences Watersportlaan 2 9000 Ghent Belgium Tel.: + 32/9/264 86 36 Fax: + 32/9/264 64 84 [email protected]

Key words ! " talent selection ! " talent identi cation ! " sport-speci c skills ! " maturity-o# set ! " training

The Contribution of Growth and Maturation in the Functional Capacity and Skill Performance of Male Adolescent Handball Players

players is a challenging task and it is well known that within the same age group, individual chronological age can di! er by as much as 4 or 5 years from biological age [ 21 , 25 ] . In addition, performance advantages associated with early maturation in boys are especially the case for size, strength, power and speed [ 21 ] . In contrast, several studies in team ball sports showed that players of varying maturity status do not di! er consistently in sport-speci c skills [ 6 , 10 , 22 ] . Identifying talent for a team sport at an early age is therefore complex and from the available lit-erature, little work has been conducted on talent identi cation in youth team handball. As team handball is a collision sport with many tough contacts, tall and strong players are at an advantage. Therefore coaches and talent scouts possibly tend to favour early maturing players for selection purposes. Research in soccer revealed that early maturing soccer players were overrep-resented in elite youth teams and coaches tend to focus on biological and physical maturation [ 11 , 15 , 19 , 32 ] . The physical and functional char-acteristics associated with variance in growth

Introduction ! Handball is a team sport, characterised by run-ning, jumping, sprinting, throwing, hitting, blocking and pushing [ 12 ] . Besides technical and tactical skills, anthropometric characteristics, speed, strength and power are important factors when participating in elite levels of handball [ 13 , 30 ] . Quantifying these factors in young team handball players is of great interest for talent identi cation and development purposes [ 36 ] . However, talent in adolescent players is recog-nised within an interaction of innate abilities, demonstration of mature play patterns at an early age and demonstration of highly sport-spe-ci c skills [ 28 ] . Due to di! erences in growth (an increase in the size of the body as a whole or the size attained by speci c parts of the body in the rst 2 decades of human life), maturation (the timing and tempo of progress toward the mature biological state) and development (the acquisi-tion and re nement of skilful performance in a variety of motor activities), identifying perform-ance-related characteristics in young handball

Authors S. P. J. Matthys 1 , R. Vaeyens 1 , M. J. Coelho-e-Silva 2 , M. Lenoir 1 , R. Philippaerts 1

A! liations 1 Ghent University , Movement and Sports Sciences , Ghent , Belgium 2 University of Coimbra , Faculdade de Ciências do desporto e Educação Física , Coimbra , Portugal

Abstract ! The present study determined to what extent the variance in performance might be explained by chronological age, biological maturation, training load and anthropometry in 168 Belgian male handball players aged 14 years: anthro-pometric, strength, speed and sport-speci c skills were assessed. MANOVA tested the e! ect of chronological age and biological maturity, whereas MANCOVA was used to compare matu-rity groups controlling for chronological age and training load. In addition, canonical correlation analysis was used between age, maturity-o! set and anthropometry, on one side, and perform-ance and sport-speci c skills, on the other side.

Results revealed signi cant di! erences between early, on-time and late maturity groups for anthropometry ( p < 0.001), strength ( p < 0.001) and sprint 20-m ( p < 0.05) in favour of the early maturing players. The di! erence between the mean values of the extreme groups for height was 24.8 cm, for weight 33.2 kg and, for body fat 6.5 %; for handgrip 20.2 kg, for 5-jump test 1.1 m and for 20-m sprint 0.20 s. Maturity status had no e! ect on sport-speci c skills. Canonical cor-relations indicated that poorer scores in sport-speci c skills were related to fatness and lack of training. In parallel, a substantial relationship was found between early maturity-o! set, body size, strength and 20-m sprint.

544 Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

and maturation are purported to hold social stimulus value for those involved in youth sports (i. e. coaches and scouts), in u-encing the quality and nature of athletes socialization experi-ences [ 33 ] . For example, gymnastics coaches react more favourably (i. e. greater encouragement and instruction, and less punishment) to gymnasts who are lighter, shorter and carry less body mass, irrespective of performance or ability [ 9 ] . Another consequence of this selection strategy in soccer is that gifted late maturing players may drop out early because they cannot com-pete with more mature, stronger boys [ 11 ] . Moreover, it is plau-sible that these late maturing players are also excluded from elite levels, because of deselections, discouragements or injuries [ 19 , 32 ] . Conversely, these inter-individual di! erences in size and physical performance disappear when late maturers catch up with their early maturing counterparts in young adulthood [ 2 , 16 ] . Those few late maturers who do persevere are likely to eventually surpass the performances of the early maturers [ 35 ] . Earlier studies on youth team handball [ 17 , 27 ] have not system-atically investigated the young athlete from a multidisciplinary perspective, i. e. anthropometric dimensions and physical per-formance components together with sport-speci c skills in rela-tion to maturity status. It is well documented, although not in handball, that early maturing players are often overrepresented in youth selection teams [ 11 , 15 , 19 , 32 ] . However, the advantage of early maturing handball players in sport-speci c skills is more questionable. Therefore, the purpose of this study was 2-fold: (1) to test the e! ect of biological maturation on body size, com-position, physical performance measures and sport-speci c skills and (2) to determine to what extent the variance in per-formance might be explained by chronological age, biological maturation, training load and anthropometry.

Methods ! The Ghent Youth Handball Project (GYHP) was a 3-year study of talent identi cation in youth handball. The Ethics Committee of the Ghent University Hospital approved this study. Informed parental consent and player assent were obtained. Parents and players were also informed that participation was voluntary and that they could withdraw at any time [ 14 ] . During 3 competitive handball seasons (2007–2008; 2008–2009; 2009–2010), players from 15 di! erent clubs, from the Flemish Top Sport Academy and from the Belgian national team took part in this study. In this period, 715 players were tested, aged 11–18 years. The present sample included 168 male youth handball players, 14.00–14.99 years of age, because from this age on the process of maturation started, and moreover, from this age players can be selected for an elite team (regional selec-tion or national selection team). The training load of the players was obtained through a questionnaire in which players described the amount of training hours per week they participated in, either in their clubs, in the Academy or in the national team. All the players included in this study were of European ancestry and none of these players had reached the fully mature status (according to the estimation of Sherar et al. [ 31 ] ). The test battery was scheduled instead of a regular training ses-sion. The team of investigators who took these tests was speci -cally trained for this. The anthropometric assessments and questionnaires were conducted prior to a standardised 10-min warm up. Following the warm up, the physical performance

tests (strength, speed and sport-speci c skills) were completed. Between 4–6 min of recovery was provided between each test. The following anthropometric measurements were taken using standardised protocols [ 18 ] : height (0.1 cm, Harpenden Portable Stadiometer, Holtain, UK), sitting height (0.1 cm, Harpenden Sit-ting Height Table, Holtain, UK), body mass (0.1 kg) and body fat percentage (0.1 %) with a digital balance scale, with bioelectrical impedance analysis indirectly measuring body composition (TANITA BC-420, Japan). Sitting height was subtracted from height to estimate leg length (0.1 cm). The same researcher took all anthropometric measures. The intra-class correlation coe" -cient for test-retest reliability and technical error of measure-ment (test-retest period of 1 h) in 40 adolescents was 1.00 ( p < 0.001) and 0.49 cm for height and 0.99 ( p < 0.001) and 0.47 cm for sitting height. In order to estimate maturity status of the handball players, a non-invasive technique based upon chronological age (decimal age) and anthropometric variables, was used [ 26 ] . The biological maturation index predicts years from peak height velocity (PHV) as a measure of maturity-o! set according to the following equa-tion: maturity-o! set = # 9.236 + 0.0002708 (leg length × sitting height) – 0.001663 (age × leg length) + 0.007216 (age × sitting height) + 0.02292 (weight by height ratio), where R = 0.94, R 2 = 0.891, and SEE = 0.592. Length measurements are in centi-meters and weight measurements are in kilograms; the weight by height ratio is multiplied by 100. Consequently, age at peak height velocity (APHV) was calculated as the di! erence between chronological age and the predicted years from PHV. Estimated ages at peak height velocity were classi ed relative to the 3 samples upon which the protocol was developed [ 26 ] . The APHV for the 3 samples adjusted for sample sizes was 13.81 years. The standard deviation for the combined sample was esti-mated as 0.83 years, which was similar to standard deviations for estimated ages at peak height velocity in several longitudinal studies [ 21 ] . Taking into account the cross-sectional data of the present study and previous literature with young athletes it was decided to assume a 1-year band to classify maturity groups. The decision is in accordance with previous literature with young athletes [ 1 ] . On-time was de ned as within a band of 1 year from mean APHV derived from the 3 original samples (13.81 years); late was de ned as an APHV > 14.81 years; early was de ned as an APHV < 12.81 years. The physical performance measures consisted of strength, speed and sport-speci c skills. A 5-jump test for distance (0.01 m) was executed in order to evaluate lower limb explosive power [ 4 ] and static strength of the throwing hand was measured by hand-grip [ 8 ] . 2 maximal sprints of 20-m were performed with split times at 5-m and 20-m, with the fastest sprint used for analysis. The sprint test was recorded with MicroGate Racetime2 chro-nometry and Polifemo Light photocells (0.01 s) (MicroGate, Italy). The sport-speci c skills of the handball players were assessed trough 3 di! erent performance tests: a cross hopping test, a handball speci c shuttle run and a slalom dribbling test. The cross hopping test was conducted as a measure of general coordination. In this test, players jumped clockwise with both feet together out and back in a 30 cm square in 4 directions, with the number of total cycles in 30 s as the nal score. Speci c handball movements were tested in a handball speci c shuttle run (0.01 s) [ 27 ] . The player was instructed to run forward in a straight line and touch a spot marked on the oor 3 m from the starting position with one foot. Then the player had to slide diagonally and backwards to the stand, which was positioned

545Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

2.5 m to the right of the starting position. After touching the stand with one hand, the player had to move back to the starting position and repeat the same cycle on the left side. When return-ing to the starting position, a second cycle was to be performed immediately. The time to nish the 2 cycles was recorded as the player’s performance. The player’s agility with the ball was measured by the slalom-dribbling test (0.01 s) [ 17 ] . The player ran a distance of 15 m, back and forth, dribbling a handball around 5 cones. The distance between the starting line and the rst cone, as well as between the other cones, was 3 m. A sche-matic representation of the handball speci c shuttle run and the slalom-dribbling test can be found in ! " Fig. 1 . The intra-class correlation coe" cients for test-retest reliability was 0.91 ( p < 0.001) for 5-jump test, 0.81 ( p < 0.001) for handgrip, 0.92 ( p < 0.001) for sprint 20-m, 0.74 ( p < 0.001) for cross hopping, 0.81 ( p < 0.001) for handball speci c SHR and 0.76 ( p < 0.001) for slalom dribble test. Descriptive statistics (mean and standard deviation) were deter-mined for 2 age groups (14.00–14.49 years and 14.50–14.99 years) and maturity groups (early, on-time, late). Multivariate analysis of variance (MANOVA) was performed to test the e! ect of chronological age (decimal age) and biological maturation on anthropometry, strength, speed and sport-speci c skills. Post hoc tests, with Bonferroni adjustment for multiple comparisons, were used in case of a signi cant main e! ect. In addition, multi-variate analyses of covariance (MANCOVA) with chronological age and training load as covariates were performed in order to re-examine the potential e! ect of maturation controlling for the mentioned confounding factors. Cohen’s d e! ect sizes ( ! 2 ) were used to indicate the magnitude of the di! erences (small e! ect: 0.2–0.3; medium e! ect: 0.5–0.8; large e! ect: 0.8 and above) [ 7 ] . Since the classi cation of early, on-time and late maturing hand-ball players is based on arbitrary adoption of a 1-year age band and taking into account the magnitude of standard deviations from the literature, it was decided to use canonical correlations to examine the relationship between a set of variables: chrono-logical age, training load, maturity-o! set, height, weight, per-centage of body fat, on one side; handgrip strength, 5-jump test, 5-m sprint, 20-m sprint, handball speci c shuttle run, and the 3 sport-speci c skills, on the other side. Each domain in the analy-

sis was collapsed into a canonical variate (linear combination of variables that was derived to maximize the relationship between domains: X = a 1 X 1 + a 2 X 2 + a 3 X 3 + a 4 X 4 + a 5 X 5 + a 6 X 6 ; Y = a 1 Y 1 + a 2 Y 2 + a 3 Y 3 + a 4 Y 4 + a 5 Y 5 + a 6 Y 6 + a 7 Y 7 ). The canonical coe" cient (r c = r x,y ) measures the magnitude of the association between the 2 variates and its squared value (r c 2 ) corresponds to shared vari-ance between the 2 sets. The analytical protocol calculates bivariate correlations between each original variable and its canonical variate (r X,X1 ; r X,X2 ; …; r X,Xn ; and r X,X1 ; r X,X2 ; …; r X,Xn ). The coe" cients inform the contribution of a single variable to the observed multivariate association. It is possible to extract as many pairs of canonical variates (r Xa,Ya , r Xb,Yb , … r Xi,Yi ) as the number of variables in the smallest set of variables ( i ) in the analysis. In the present study, the rst pair of canonical variates was extracted after performing the Wilk’s Lambda and the test was repeated before the extraction of subsequent canonical cor-relations. The percentage of shared variance between the 2 sets that was extracted by each canonical correlation was calculated. Additionally, for each pair of extracted roots and in particular for each of the 2 variates the following information was determined: percentage of shared variance between the indicators and the respective linear combination (r 2 X,(a1 X 1 + … + anXn) ; r 2 Y,(a1Y1 + … + anYn) ), percentage of shared variance between the indicators and the linear combination of variables generated in the other side of the equation (r 2 (a1 X 1 + … + anXn), Y ; r 2 (a1Y1 + … + anYn), X ). All analyses were performed using SPSS 15.0 with the minimal level of signi -cance set at p < 0.05.

Results ! ! " Table 1 summarises the anthropometric characteristics, phys-ical performance measures and sport-speci c skills per chrono-logical age groups and also by contrasting maturity groups. Training load and estimated APHV did not di! er between oldest and youngest groups. Similarly, training load and chronological age were equivalent between early, on-time and late maturing participants. In contrast to biological maturation, chronological age did not correspond to a signi cant e! ect on body size (height, weight) and estimated percentage of body fat. Maturity

Fig. 1 Schematic representation of a the hand-ball speci c shuttle run (Mohamed et al., 2009) and b the slalom dribbling test (Lidor et al., 2005).

Handball specific shuttle run:

2.5 m 2.5 m 3

2 1, 4 5

6

3.0 m

Distances

15.0 m

3.0 m3.0 m 3.0 m 3.0 m 3.0 m

Slalom dribble test:

Running sequences:1, 2, 3, ...

a

b

546 Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

groups di! ered in height (F = 73.29, p < 0.01, ! 2 = 0.47), weight (F = 72.79, p < 0.01, ! 2 = 0.47) and % body fat (F = 9.74, p < 0.01, ! 2 = 0.11). Early maturing players were 24.8 cm taller, 33.2 kg heavier and presented 6.5 % more body fat, compared to their late maturing peers. Regarding strength characteristics, the 5-jump test was signi -cantly a! ected by chronological age (F = 9.60, p < 0.01, ! 2 = 0.06) and biological maturation (F = 7.33, p < 0.01, ! 2 = 0.08), while for the handgrip static strength test the unique e! ect was produced by the main e! ect of biological maturation (F = 31.77, p < 0.01, ! 2 = 0.28). The oldest players jumped about 0.40 m further, com-pared to their younger peers aged 14.00–14.49. The gradient for maturity groups is early > on time > late with di! erences of 20.2 kg and 1.1 m between the extreme groups, respectively for

the handgrip and the jumps. The 5-m sprint was a! ected neither by age nor by maturation, in contrast to the longer distance. Players aged 14.50–14.99 years performed better (# 0.10 s) on the 20-m speed test (F = 9.15, p < 0.01, ! 2 = 0.05). Maturation was also a signi cant source of inter-individual variability (F = 4.23, p < 0.05, ! 2 = 0.05). Again the gradient was early > on time > late. Note that a better performance corresponds to a lower score given in seconds. Sport-speci c skills appeared to be independ-ent from the e! ect of biological maturation. In parallel, slalom dribble test was the unique skill test signi cantly a! ected by chronological age (F = 5.46, p < 0.01, ! 2 = 0.03). Again, older par-ticipants attained better performances (# 0.36 s). Multivariate analyses of covariance (MANCOVA) were used to re-examine the e! ect of maturity groups controlling for chrono-logical age and training load (see ! " Table 2 ). When comparisons between maturity groups are adjusted for di! erences in chrono-logical age and training load, the results are similar to the previ-ously described e! ect of biological maturation in ! " Table 1 . Between the 2 sets of variables, it was possible to extract 2 signi cant canonical correlations (see ! " Table 3 ). The rst explained 63.8 % (r 1c = 0.75, r 1c 2 = 0.56) of the variance and the second 25.4 % (r 2c = 0.58, r 2c 2 = 0.34). The rst pair of linear func-tions (see ! " Fig. 2 ) corresponds to a direct relationship between estimated age at peak height velocity (# 0.93), height (+ 0.91) and weight (+ 0.89), on one side; and handgrip (+ 0.97), 20-m sprint (+ 0.50) and 5-jump test (+ 0.48), on the other side. This means that there is an association between early maturation (negative value), height (the taller the better; positive value) and weight (the heavier the better; positive value) of the players and better performances (positive values) on the aforementioned tests. The second canonical correlation (see ! " Fig. 3 ) emerged from train-ing load (# 0.66), percent of body fat (+ 0.56), and poor results in several tests: 5-jump test (# 0.81), slalom dribble (# 0.70), 20-m sprint (# 0.51), 5-m sprint (# 0.48), handball speci c shuttle run (# 0.48) and cross hopping (# 0.37). This shows that a low train-ing load (negative value) is related to a higher fat percentage

Table 1 Descriptive statistics and results of MANOVA to test the e# ect of chronological age and maturity groups in inter-variability of U15 youth handball players.

Chronological age groups Maturity groups E" ect of chrono-

logical age

E" ect of biological

maturity

14.00–14.49

(n = 72)

14.50–14.99

(n = 96)

Early

(n = 13)

On-time

(n = 135)

Late

(n = 20)

F p # 2 F p # 2

training load, hrs-week $ 1 5.35 ± 3.73 5.13 ± 3.26 6.35 ± 4.98 5.31 ± 3.50 3.90 ± 0.75 0.18 0.67 0.01 2.22 0.11 0.03 chronological age, years 14.26 ± 0.13 14.74 ± 0.14 14.57 ± 0.32 14.52 ± 0.27 14.56 ± 0.27 534.76 0.00 0.76 0.31 0.74 0.00 maturity o# set 0.46 ± 0.67 0.82 ± 0.74 1.85 ± 0.60 a 0.73 ± 0.51 b $ 0.52 ± 0.33 c 10.40 0.00 0.06 92.26 0.00 0.53 APHV, years 13.80 ± 0.68 13.92 ± 0.72 12.72 ± 0.65 a 13.80 ± 0.47 b 15.09 ± 0.25 c 1.32 0.25 0.01 109.93 0.00 0.57 anthropometry height, cm 170.0 ± 8.2 171.4 ± 8.1 182.6 ± 7.6 a 171.6 ± 5.9 b 157.8 ± 5.9 c 1.18 0.28 0.01 73.29 0.00 0.47 weight, kg 57.4 ± 10.5 58.4 ± 10.8 76.2 ± 8.9 a 58.4 ± 8.0 b 43.0 ± 5.2 c 0.42 0.52 0.00 72.79 0.00 0.47 body fat, % 11.9 ± 4.5 11.6 ± 4.3 15.6 ± 3.4 a 11.7 ± 4.2 b 9.1 ± 4.4 c 0.18 0.67 0.00 9.74 0.00 0.11 strength handgrip, kg 40.0 ± 8.6 42.5 ± 8.2 52.0 ± 9.2 a 41.8 ± 7.1 b 31.8 ± 7.1 c 3.43 0.07 0.02 31.77 0.00 0.28 5 jump test, m 10.5 ± 0.9 10.9 ± 1.0 11.0 ± 1.0 a 10.8 ± 0.9 a 9.9 ± 0.9 b 9.60 0.00 0.06 7.33 0.00 0.08 speed sprint 5 m, s 1.18 ± 0.09 1.16 ± 0.07 1.13 ± 0.07 1.17 ± 0.08 1.19 ± 0.07 2.81 0.10 0.02 2.21 0.11 0.03 sprint 20 m, s 3.49 ± 0.23 3.39 ± 0.20 3.36 ± 0.22 a 3.42 ± 0.22 a 3.56 ± 0.20 b 9.15 0.00 0.05 4.23 0.02 0.05 sport-speci c skills handball speci c SHR, s 13.48 ± 1.37 13.36 ± 1.27 13.71 ± 1.47 13.35 ± 1.34 13.64 ± 0.92 0.36 0.55 0.00 0.79 0.45 0.01 cross hopping, n 10.64 ± 2.15 10.81 ± 2.07 10.23 ± 2.77 10.83 ± 1.99 10.45 ± 2.35 0.28 0.60 0.00 0.69 0.50 0.01 slalom dribble test, s 9.06 ± 1.04 8.70 ± 0.96 9.17 ± 1.27 8.79 ± 0.97 9.04 ± 1.09 5.46 0.02 0.03 1.22 0.30 0.02 Note: means in the same row having the same subscript are not signi cantly di# erent at p < 0.05

Table 2 Results of MANCOVA (left: chronological age as covariate; right: training load as covariate) to test the e# ect of maturity groups in inter-varia-bility of U15 youth handball players.

Controlling for

chronological age

Controlling for

training load

F p ! 2 F p ! 2

anthropometry height, cm 75.15 0.00 0.48 71.65 0.00 0.47 weight, kg 73.25 0.00 0.47 73.10 0.00 0.47 body fat, % 9.66 0.00 0.11 10.42 0.00 0.11 strength handgrip, kg 33.20 0.00 0.29 31.65 0.00 0.28 5 jump test, m 8.47 0.00 0.09 5.64 0.00 0.06 speed sprint 5 m, s 2.26 0.11 0.03 1.65 0.20 0.02 sprint 20 m, s 4.93 0.01 0.06 3.53 0.03 0.04 sport-speci c skills handball speci c SHR, s 0.84 0.43 0.01 0.81 0.45 0.01 cross hopping, n 0.73 0.48 0.01 0.74 0.48 0.01 slalom dribble test, s 1.49 0.23 0.02 1.58 0.21 0.02

547Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

(positive value) and weaker performances (negative values) on the sport-speci c skills, the speed-tests and the 5-jump test.

Discussion ! The rst purpose of the present study was to determine matu-rity-related variation in body size, fat mass and physical per-formance measures. Early maturing players surpassed the late maturing peers in body size, strength and speed (the exception was the shortest distance dash: sprint 5-m). The early maturing players outperformed their peers, even when controlled for chronological age and training load. However, sport-speci c skills revealed to be substantially independent from the e! ect of

biological maturity. The second purpose was to determine to what extent the variance in performance is explained by chron-ological age, maturity-o! set, training load and anthropometry. Being more advanced in biological maturation (thus showing low values in maturity-o! set), and having a larger body size were strongly related to better scores in strength as well as in speed as measured by the 20-m sprint. Additionally, the subse-quent extraction of linear functions between the 2 sets of vari-ables corresponded to lower values in training load, higher percentage of body fat, on one side, and poor scores in sport-speci c skills and speed, on the other side. Maturity-related di! erences in anthropometry among male adolescent athletes are well documented e. g. [ 10 , 20 , 24 , 27 , 29 , 32 ] . Moreover, several authors consider anthropometric characteristics as important determinants in youth and adult team handball [ 17 , 27 , 34 , 36 ] . In that case early and on-time maturing players are at an advantage, due to their more favourable anthropometric dimensions compared to their later maturing peers. Furthermore, early maturation in boys is also associated with performance advantages [ 21 , 29 ] . In this study, performances on strength and speed were signi cantly di! erent between the maturation groups, with the better scores for the early maturing handball players (e! ect sizes showed small e! ects for anthropometry, strength and speed). The results con rm that body size and maturation have an impact on strength and speed performances, even when accounted for training load. This is similar to the results of the rst canonical correlations, indicating a strong association between biological maturation and body size (height and weight), strength (hand-grip and 5-jump test) and sprint 20-m. However, other studies have shown that, after puberty, late maturing boys can catch up with or even surpass their early and on-time maturing counter-parts in anthropometric characteristics and physical perform-ance [ 2 , 16 , 28 , 35 ] . In many sports, sport-speci c skills are a central component in the development of young athletes [ 22 ] . The outcomes of the sport-speci c skill tests in the present study revealed that matu-ration does not a! ect these skills and the rst canonical correla-tions indicated a relatively low association between the studied sport-speci c skills and biological maturation. This is in line with earlier studies in soccer and basketball [ 5 , 10 , 22 ] . A possi-ble explanation is that many factors other than body size and maturity status (e. g. neural control of movement, deliberate

Table 3 Results of canonical correlations (r c ) between chronological age, maturity o# set, body size plus percentage of body fat, and functional vari-ables (strength, speed and skills).

Canonical correlations

1 st 2 nd

r c 0.75 0.58 r c 2 0.56 0.34 eigenvalue 1.27 0.51 Wilks’ Lambda 0.24 0.54 F 6.21 3.46 p < 0.01 < 0.01 extracted variance (X,Z) 63.8 % 25.4 % X 1 : training load (TL) $ 0.02 $ 0.66 X 2 : chronological age (CA) + 0.13 $ 0.44 X 3 : age at PHV (APHV) $ 0.93 + 0.04 X 4 : height (H) + 0.91 $ 0.18 X 5 : weight (W) + 0.89 + 0.24 X 6 : percentage of body fat (%Fat) + 0.32 + 0.56 explained variance in the X-side 43.8 % 17.3 % explained variance in the Y-side 24.5 % 5.8 % Y 1 : handgrip (HGR) + 0.97 $ 0.07 Y 2 : 5-Jump test (5JT) + 0.48 $ 0.81 Y 3 : 5-m sprint (5 m sprint) + 0.37 $ 0.48 Y 4 : 20-m sprint (20 m sprint) + 0.50 $ 0.51 Y 5 : handball speci c shuttle run (Hb-SHR) + 0.11 $ 0.48 Y 6 : cross hopping (C-hp) $ 0.06 $ 0.37 Y 7 : slalom dribble test (Sl-db) $ 0.03 $ 0.70 explained variance in the Y-side 22.4 % 28.7 % explained variance in the X-side 12.5 % 9.6 %

Fig. 2 Correlations of training load, chronologi-cal age, plus morphology (height, weight and % fat) and functional items with their respective rst canonical variates.

1.00TL

0.13– 0.93

0.91 0.89 0.32 0.97

0.480.37

0.50

0.11

– 0.06 – 0.03– 0.02

CA APHV H W

TL = training load; CA = chronological age; APHV = age at peak height velocity; H = height; W = weight; % Fat = percentage of bodyfat; HGR = handgrip; 5JT = 5 - jump test; Hb - SHR = handball specific shuttle run; C- Hp = cross hopping; Sl - db = slalom dribble test

% Fat HGR 5JT5m

sprint20m

sprint Hb - SHR C - Hp SI - db

0.80

0.60

0.40

0.20

0.00

–0.20

–0.40

–0.60

–0.80

–1.00

548 Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

play and practice, and training content) in uence performances on sport-speci c skills [ 22 ] . Coaches and scouts should also use sport-speci c skills for selection purposes, because body size and maturation seem to have no in uence on these tests. Fur-thermore, sport-speci c skill tests are probably most useful for a long-term development of athletes, ensuring to retain those later maturing players within the sport. However, the studied sport-speci c skill tests might be too general, future studies should include more handball-speci c tests and should also investigate the in uence of training history on these skills. Addi-tionally, federations have the di" cult but challenging task to motivate coaches to focus their attention also on late maturing players, in order to prevent them from injuries, discouragements and dropout. It is plausible that late maturing players receive fewer (and less qualitative) opportunities to train and play, resulting in a lower training load compared to their early and on-time maturing peers and causes problems for talent develop-ment purposes. The second canonical correlations suggested the crucial importance of training to avoid low performances on strength (except handgrip which strongly correlates with bio-logical maturation), speed and sport-speci c skills. Furthermore, coaches might choose to place late maturing handball players on eld positions where their maturity-associated limitations are of less importance for short-term team success (i. e. wing-posi-tion). The non-invasive estimate of maturity status by Mirwald et al. [ 26 ] has some limitations. It has not been externally validated on an independent longitudinal sample of athletes for which APHV is known or with other established maturity indicators during adolescence (i. e. skeletal age, secondary sex characteristics). However Mirwald et al. [ 26 ] reported a correlation coe" cient of 0.83 between skeletal age-o! set and PHV maturity-o! set based on a longitudinal sample of Canadian schoolchildren. Therefore, application to individual athletes needs to be made with care. In contrast with the used method, other methods for assessing maturity can be expensive (i. e. wrist radiographs), intrusive (i. e. the assessment of sexual maturity), unethical (i. e. longitudinal assessment of RX-radiographs) or require longitudinal observa-tions. The technique used to categorise players into early, on-time and late maturity status has already been used in earlier studies [ 10 , 23 ] . However, predicted APHV based on maturity-

o! set shows relatively small standard deviations, especially in small children such as female gymnasts [ 23 ] . However, this technique has proven to be valuable, but the usability has to be considered per population. Meantime, in the context of youth sports, a relevant contribution to the discussion of the potential interaction between training and biological maturation assumed a 1-year age band from mean APHV to classify early and late maturing gymnasts [ 1 ] . As maturity status in uences performance, coaches and scouts who are involved in the talent identi cation, selection and development processes should be aware of this phenomenon. Including the studied sport-speci c skill tests can assist coaches in their decisions for selection purposes because these tests are not a! ected by biological maturity and encourage a long-term development of players. Experts should also focus on later maturing players since they can catch up with their early matur-ing peers within a few years [ 2 , 16 , 28 , 35 ] . Therefore, coaches should be encouraged to consider not only the present perform-ance, but try to judge players’ potential to improve this perform-ance. The current study recommends coaches and scouts to be patient towards keeping gifted players who are short and delayed in maturation in the sport. Ideally, due to large di! er-ences between chronological and biological age, federations should group players by biological age rather than chronological age. However, this is di" cult to realise for ethical, practical and economical issues. Additionally, relatively small nations do not have su" ciently large populations to rely on a trial-and-error approach that eventually allows some athletes to reach the elite standard by a process known as natural or self-selection [ 3 ] . The above-mentioned recommendations can help small nations to recruit potentially top handball players from a larger sample. In order to have a full understanding of youth team handball and the concept of maturation, future research might look at track-ing the impact of maturation on selection and retention in the sport, the impact of maturation at varying age levels and stand-ards of play, the extent to which coaches perceive and react dif-ferently to players of varying maturity status, to check for secular trends in elite handball populations over time, also more hand-ball-speci c skill tests (e. g. ball-throwing test, one-leg jump test, …) should be included and the deliberate practice or train-ing history can give useful information on the development of the players and the road to elite performances.

Fig. 3 Correlations of training load, chronological age, plus morphology (height, weight and % fat) and functional items with their respective second canonical variates.

1.00

0.04

– 0.66

– 0.44

– 0.18

– 0.07

– 0.81

TL = training load; CA = chronological age; APHV = age at peak height velocity; H = height; W = weight; % Fat = percentage of bodyfat; HGR = handgrip; 5JT = 5 - jump test;Hb - SHR = handball specific shuttle run; C - Hp = cross hopping; Sl - db = slalom dribble test

– 0.48 – 0.51 – 0.48– 0.37

– 0.70

0.24

0.56

TL CA APHV H W % Fat HGR 5JT5m

sprint20m

sprintHb -SHR C - Hp SI - db

0.80

0.60

0.40

0.20

0.00

– 0.20

– 0.40

– 0.60

– 0.80

– 1.00

549Training & Testing

Matthys SPJ et al. The Contribution of Growth and maturation … Int J Sports Med 2012; 33: 543–549

Acknowledgements ! The Flemish Policy Research Centre for Culture, Youth and Sports nancially supported this study. The authors thank the VHV (Flemish Handball Federation) for their cooperation. The authors would like to thank Job Fransen, Dieter Deprez, Joric Vanden-driessche, Barbara Vandorpe and Johan Pion for their assistance in data collection and appreciated their helpful comments dur-ing the writing of the manuscript.

References1 Baxter-Jones AD G , Ma" ulli N , Mirwald R L . Does elite competition

inhibit growth and delay maturation in some gymnasts? Probably not . Pediatr Exerc Sci 2003 ; 15 : 373 – 382

2 Beunen G P , Malina R M , Lefevre J , Claessens A L , Renson R , Simons J . Prediction of adult stature and noninvasive assessment of biological maturation . Med Sci Sports Exerc 1997 ; 29 : 225 – 230

3 Bloom eld J , Ackland T R , Elliott B C . Applied Anatomy and Biomechan-ics in Sport . Melbourne, VIC : Blackwell Scienti c , 1994

4 Chamari K , Chaouachi A , Hambli M , Kaouech F , Wislo" U , Castagna C . The ve-jump test for distance as a eld test to assess lower limb explo-sive power in soccer players . J Strength Cond Res 2008 ; 22 : 944 – 950

5 Coelho e Silva M J , Figueiredo A J , Carvalho H M , Malina R M . Functional capacities and sport-speci c skills of 14- to 15-year-old male basket-ball players . Eur J Sport Sci 2008 ; 8 : 277 – 285

6 Coelho e Silva M J , Moreiro Carvalho H , Goncalves C E , Figueiredo A J , Elferink-Gemser M T , Philippaerts R M , Malina R M . Growth, matura-tion, functional capacities and sport-speci c skills in 12–13-year-old-basketball players . J Sports Med Phys Fitness 2010 ; 50 : 174 – 181

7 Cohen J . Statistical Power Analysis for the Behavioural Sciences . Hills-dale NJ : Lawrence Erlbaum Associates , 1988

8 Council of Europe . Euro t: Handbook for the EUROFIT Tests of Physical Fitness . Rome : Secretariat of the Committee for the Development of Sport within the Council of Europe , 1988

9 Cumming S P , Eisenmann J C , Smoll F L , Smith R E , Malina R M . Body size and perceptions of coaching behaviors in female adolescent athletes . Psychol Sport Exerc 2006 ; 6 : 693 – 705

10 Figueiredo A J , Goncalves C E , Coelho e Silva M J , Malina R M . Youth soccer players, 11–14 years: maturity, size, function, skill and goal orienta-tion . Ann Hum Biol 2009 ; 36 : 60 – 73

11 Figueiredo A J , Goncalves C E , Coelho e Silva M J , Malina R M . Charac-teristics of youth soccer players who drop out, persist or move up . J Sports Sci 2009 ; 27 : 883 – 891

12 Gorostiaga E M , Granados C , Ibanez J , Gonzalez-Badillo J J , Izquierdo M . E! ects of an entire season on physical tness changes in elite male handball players . Med Sci Sports Exerc 2006 ; 38 : 357 – 366

13 Gorostiaga E M , Granados C , Ibanez J , Izquierdo M . Di! erences in physi-cal tness and throwing velocity among elite and amateur male hand-ball players . Int J Sports Med 2005 ; 26 : 225 – 232

14 Harriss D J , Atkinson G . Update – Ethical standards in sport and exer-cise science research . Int J Sports Med 2011 ; 32 : 819 – 821

15 Hirose N . Relationship among birth-month distribution, skeletal age and anthropometric characteristics in adolescent elite soccer players . J Sports Sci 2009 ; 27 : 1159 – 1166

16 Lefevre J , Beunen G , Steens G , Claessens A , Renson R . Motor perform-ance during adolescence and age thirty as related to age at peak height velocity . Ann Hum Biol 1990 ; 17 : 423 – 435

17 Lidor R , Falk A , Arnon M , Cohen Y , Segal G , Lander Y . Measurement of talent in team handball: The questionable use of motor and physical tests . J Strength Cond Res 2005 ; 19 : 318 – 325

18 Lohman T G , Roche A F , Martorell R . Anthropometric Standardization Reference Manual . Champaign IL : Human Kinetics , 1988

19 Malina R M , Pena Reyes M E , Eisenmann J C , Horta L , Rodrigues J , Miller R . Height, mass and skeletal maturity of elite Portuguese soccer players aged 11–16 years . J Sports Sci 2000 ; 18 : 685 – 693

20 Malina R M , Eisenmann J C , Cumming S P , Ribeiro B , Aroso J . Maturity-associated variation in the growth and functional capacities of youth football (soccer) players 13–15 years . Eur J Appl Physiol 2004 ; 91 : 555 – 562

21 Malina R M , Bouchard C , Bar-Or O . Growth, Maturation and Physical Activity . Champaign, Illinois : Human Kinetics , 2004

22 Malina R M , Cumming S P , Kontos A P , Eisenmann J C , Ribeiro B , Aroso J . Maturity-associated variation in sport-speci c skills of youth soccer players aged 13–15 years . J Sports Sci 2005 ; 23 : 515 – 522

23 Malina R M , Claessens A L , Van Aken K , Thomis M , Lefevre J , Philippaerts R , Beunen G P . Maturity-o! set in gymnasts: Application of a predicted equation . Med Sci Sports Exerc 2006 ; 38 : 1342 – 1347

24 Malina R M , Ribeiro B , Aroso J , Cumming S P . Characteristics of youth soccer players aged 13–15 years classi ed by skill level . Br J Sports Med 2007 ; 41 : 290 – 295

25 Malina R M , Pena Reyes M E , Figueiredo A J , Coelho e Silva M J , Horta L , Miller R , Chamorro M , Serratosa L , Morate F . Skeletal age in youth soccer players: implications for age veri cation . Clin J Sport Med 2010 ; 469 – 474

26 Mirwald R L , Baxter-Jones AD G , Bailey D A , Beunen G P . An assessment of maturity from anthropometric measurements . Med Sci Sports Exerc 2002 ; 34 : 689 – 694

27 Mohamed H , Vaeyens R , Matthys S , Multael M , Lefevre J , Lenoir M , Philippaerts R . Anthropometric and performance measures for the development of a talent detection and identi cation model in youth handball . J Sports Sci 2009 ; 27 : 257 – 266

28 Pearson D T , Naughton G A , Torode M . Predictability of physiological testing and the role of maturation in talent identi cation for adoles-cent team sports . J Sci Med Sport 2006 ; 9 : 277 – 287

29 Philippaerts R M , Vaeyens R , Janssens M , Van Renterghem B , Matthys D , Craen R , Bourgois J , Vrijens J , Beunen G , Malina R M . The relationship between peak height velocity and physical performance in youth soc-cer players . J Sports Sci 2006 ; 24 : 221 – 230

30 Rannou F , Prioux J , Zouhal H , Gratas-Delamarche A , Delamarche P . Physiological pro le of handball players . J Sports Med Phys Fitness 2001 ; 41 : 349 – 353

31 Sherar L B , Mirwald R L , Baxter-Jones AD G , Thomis M . Prediction of adult height using maturity-based cumulative height velocity curves . J Pediatr 2005 ; 147 : 508 – 514

32 Sherar L B , Baxter-Jones AD G , Faulkner R A , Russell K W . Do physical maturity and birth date predict talent in male youth ice hockey play-ers? J Sports Sci 2007 ; 25 : 879 – 886

33 Sherar L B , Cumming S P , Eisenmann J C , Baxter-Jones AD G , Malina R M . Adolescent biological maturity and physical activity: Biology meets behavior . Pediatr Exerc Sci 2010 ; 22 : 332 – 349

34 Skoufas D , Kotzaminidis C , Hatzikotoylas K , Bebetsos G , Patikas D . The relationship between anthropometric variables and the throwing per-formance in handball . J Hum Mov Stud 2003 ; 45 : 469 – 484

35 Wolstencroft E . Talent Identi cation and Development: An Academic Review . Edinburgh : Sportscotland , 2002

36 Ziv G , Lidor R . Physical characteristics, physiological attributes, and on-court performances of handball players: A review . Eur J Sport Sci 2009 ; 9 : 375 – 386