Development of muscle strength in relation to training

level and testosterone in young male soccer players
L. HANSEN,1 J. BANGSBO,1 J. TWISK,2 AND K. KLAUSEN1
of Human Physiology, Institute of Exercise and Sport Sciences, University
of Copenhagen, DK-2100 Copenhagen, Denmark; and 2Institute for Research in Extramural
Medicine, Vrije Universiteit, 1081 Amsterdam, The Netherlands

1Department

Hansen, L., J. Bangsbo, J. Twisk, and K. Klausen.
Development of muscle strength in relation to training level
and testosterone in young male soccer players. J. Appl.
Physiol. 87(3): 1141–1147, 1999.—Isometric and functional
strength of ninety-eight 11-yr-old male soccer players at an
elite (E) and nonelite (NE) level were determined (3–4 times)
through a 2-yr period, and the changes were related to growth
and maturation. The initial isometric strength for extension
with dominant leg [1,502 ⫾ 35 (E) vs. 1,309 ⫾ 39 (NE) N],
extension with nondominant leg (1,438 ⫾ 37 vs. 1,267 ⫾ 45
N), extension with both legs (2,113 ⫾ 76 vs. 1,915 ⫾ 72 N),
back muscles (487 ⫾ 11 vs. 414 ⫾ 10 N), abdominal muscles
(320 ⫾ 9 vs. 294 ⫾ 8 N), and handgrip (304 ⫾ 10 vs. 259 ⫾ 8 N)
increased by 15–40% during the period. Broad jump increased (P ⬍ 0.05) by 15 (E) and 10% (NE). The E players had
higher (P ⬍ 0.05) initial isometric strength and broad jump
performance compared with NE players, and differences were
maintained throughout the period (multiple ANOVA for repeated measures) also when adjustment was made for age,
dimensions, testosterone, and insulin-like growth factor I
(generalized estimating equations analyses). The development of strength for both E and NE players together was
significantly (P ⬍ 0.001) related to changes in serum testosterone concentrations. The present data indicate that testosterone is important for development of strength in young boys
and that, independent of serum testosterone concentration, E
players have developed greater muscle strength compared
with NE players.
elite and nonelite players; age; dimensions

IN ADULTS, substantial knowledge is present about
factors that determine muscle strength and its change
with training (e.g. Refs. 13, 15), but less information is
available about development of muscle strength in
children. Rochcongar et al. (22) found that young
French elite soccer players had greater isokinetic leg
strength compared with high school students, indicating that soccer training has an effect on the development of muscle strength. In contrast, Maffulli et al. (18)
found that athletic boys (including soccer players) until
the age of 15 yr had similar isometric quadriceps
strength as did nonathletic boys, and after this age the
strength of the athletic group was significantly higher
compared with nonathletic boys. The latter finding may
indicate that the training responses are affected by
maturation.

The costs of publication of this article were defrayed in part by the
payment of page charges. The article must therefore be hereby
marked ‘‘advertisement’’ in accordance with 18 U.S.C. Section 1734
solely to indicate this fact.
http://www.jap.org

Muscular strength increases more or less linearly
with age from early childhood in boys. Strength is
known to be related to the physiological cross-sectional
area of the muscle and hence, according to a dimensional analysis, related to the second power of body
height. During growth the cross-sectional area would
then be expected to increase, with the square of the
increase in the linear dimension. Some studies have
shown that strength development in boys improves
more than can be explained by increase in height
squared (1, 8), indicating that factors other than quantitative changes play a role in the development of
strength. Thus the marked acceleration of muscle
strength during puberty observed in boys is possibly
related to the elevated levels of circulating androgen
hormones in adolescents. In a cross-sectional study of
11- to 13-yr-old athletic boys, Mero et al. (20) found that
muscle fiber area correlated well with serum testosterone. Maturation of the metabolic response to exercise
might be related to the hormonal changes [increases in
testosterone, estradiol, growth hormone, and insulinlike growth factor I (IGF-I)] occurring during puberty
(9, 19). On the basis of a difference in increase in
strength between boys and girls, Parker et al. (21)
suggested that testosterone may stimulate muscle
growth. The growth-promoting effect of growth hormone is mediated by somatomedins, particularly IGF-I
(16). However, no longitudinal study has measured
changes in blood hormone concentrations and related
them to changes in muscle size and strength of children
and adolescents (23).
As indicated by Rochcongar et al. (22), soccer training
at an elite level might increase leg strength, but it is
also possible that the boys selected for the elite level are
stronger because of higher levels of circulating hormones.
The aim of this investigation was to study the
development in strength of boys playing soccer at an
elite and a nonelite level and to examine the association
between the development in strength and testosterone
concentration. Both questions were addressed with
correction for age, body size dimensions, and IGF-I.
METHODS

Subjects. One hundred and ten young male soccer players
from seven successful clubs in the area of Copenhagen, at the
highest level in their age category, were recruited as subjects.
All participants and their parents gave their informed consent, and the study was approved by the Ethics Committee of
Copenhagen, Denmark (KF 01–132/95). Clubs that had at
least four teams in the same age categories were selected (to
ensure differences between elite and nonelite players). The

8750-7587/99 $5.00 Copyright r 1999 the American Physiological Society

1141

1 kg by using a spring balance.23 nmol/l.4 (9) 10.2* 5. Thus 87 boys were tested three times. The subjects were all familarized with the testing procedures as well as with the investigators before the test.3 (elite) and 4.0 (20) 6.8 11. The leisure time activity registered consisted mostly of soccer with friends but also roller-skating and high-activity play as well as participation in other organized sports. 1) In the first analysis. After that. Statistics. Values less than assay sensitivity were assigned the value of assay sensitivity.4 (26) 6. Eight boys were excluded from the study because they were not members of the same team during the whole study. Table 1. including participation in other sports Test Round 1 Elite Nonelite 2 3 4 Organized Leisure Organized Leisure Organized Leisure Organized Leisure 6. For the elite and nonelite players. Strength measurements. Thus 98 subjects were included with 48 boys in the elite group and 50 boys in the nonelite group. and body weight was measured to the nearest 0. Standing and sitting height were measured by a stadiometer to the nearest 0. Blood samples were drawn from an antecubital vein between 1600 and 1730 and were centrifuged. Bicipital. As shown in Table 1. two because they did not want to participate after the first test and two because they moved to another part of the country. P ⬍ 0. and the best of three recorded trials was used as the performance score (cm). the differences in longitudinal development of strength measurements between elite and nonelite soccer players were analyzed with multiple ANOVA (MANOVA) for repeated measures (SPSS. The players included in the study had participated in organized soccer for 6. and the best of three attempts was accepted as maximal. All subjects started with the broad jump followed by measurements of isometric strength. One-half of the boys were recruited from the best team to which they were selected by the coach (elite players). and the sum of these four skinfolds was calculated. The competition intensity was evaluated from heart rate measurements by a heart rate monitor (Polar) both during competition and while the subjects ran on a treadmill with simultaneous measurements of oxygen uptake. Amount of time with organized soccer training (including competitions) and with physical activity in leisure time. GEE analysis was also carried out to analyze the longitudinal relationship between strength parameters and testosterone concentration. a univariate analysis was carried out first. The distance from the takeoff to the point where the nearest heel touched the mat was measured. 39% was technical training.3 vs. The maximal voluntary isometric strength [maximal voluntary contraction (MVC)] of the leg extensors was measured by using a strain-gauge dynamometer in a standardized seated position with support of the back (4).3 ⫾ 10%. subscapular. and the landing was on a 2-cm-thick mat on which the subjects were instructed to land on their feet.05) between the groups. Broad jump was performed as a two-foot takeoff and landing. The elite players had higher relative oxygen uptake in competition compared with the nonelite group (79.4 ⫾ 5. the longitudinal relationship between strength and soccer ability (elite or nonelite) was analyzed with generalized estimating equations (GEEs) (30).1* 5.05. separate analyses were carried out correcting for age.7* (19) 6. before the strength measurements.5-yr intervals for all the boys. To measure MVC of the trunk muscles in a standing position.6 (9) 7. Measurements were taken three times at 0. it allows a correction for both time-dependent and time-independent covariates and the method takes into account that the repeated observations on each individual are not independent. * Significant different from nonelite players. Serum was stored at ⫺20°C and later analyzed for levels of testosterone and IGF-I. For all GEE analysis. The training regimen and competition intensity were evaluated for a subgroup of the original subjects (n ⫽ 30.041 µg/l (12).2 10. the elite players were playing soccer for more hours per week and were in general more physically active compared with the nonelite players. and 12 vs. 10 teams).1 11.1142 STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS boys were included in the study at the age of 10–12 yr according to the selection age at the competition levels. and the other one-half were recruited from the lowest ranked team (nonelite players) from the same club. a strain-gauge dynamometer was connected to a frame placed around the trunk 20 cm below the shoulders by the use of two straps (4). 7% of the training consisted of fitness training (sprint run. in addition. The sensitivity of the assay for IGF-I was 0. All subjects had a standardized warm-up period. 12 nonelite) were also studied a fourth time. 24). The pubertal developmental stages were recorded by one experienced pediatric endocrinologist on the basis of assessment of secondary sex characteristics by using the criteria of Tanner (25) and from testicular volume estimated from measurements of the size of the testes by using a Prader orchidometer (29). tricipital. and suprailiacal skinfolds were measured with a Harpenden skinfold caliper. two analyses were carried out.05). including 5 min of cycling on a Monarck bike. The takeoff was from behind a line on the floor.05. 61 vs. 27). and. The body mass index (BMI) was calculated as body weight (kg) divided by height (m) squared. 54% was play. 28 of the subjects (16 elite. To assess the longitudinal relationship between strength and soccer ability (elite or nonelite).2 (10) Values are given in h/wk with % in parentheses. and 28 of these were also tested a fourth time. During the study some boys stopped playing soccer and were then excluded from subsequent tests. 2) In the second analysis. P ⬍ 0. Furthermore. Significance was accepted at P ⱕ 0. P ⬍ 0. All GEE analyses were carried out with the Statistical Package for Interactive Data Analysis (10). body size dimensions. 67.01 yr.4* 5 11. etc. Ref. The advantages of using this method are that all available longitudinal data are used to estimate the regression coefficients and that the method is suitable for designs with unequally spaced time intervals. a longitudinal linear regression technique that is extensively described elsewhere (26.4 (nonelite) yr with a significant difference (1. The age of the subjects was assessed to the nearest 0.5* 4.9 yr.).7* (16) 7. . The boys were all encouraged to the highest effort by the investigators. and IGF-I concentration. 27 vs. No supplemental weight training was used.9 (19) 10. and the sensitivity for testosterone was 0.1 cm. Grip strength was measured with a hand dynamometer in subjects while seated and for the dominant arm only.

5 45 11. The elite players were slightly older than the nonelite players (0. i. mm Testes volume.1 ⫾ 1. however.6 ⫾ 5. Serum testosterone concentrations. The results from the GEE analysis regarding the longitudinal relationship between being an elite player/ nonelite player and strength development are presented in Table 4. n.1 43.2 47. suprailiac).9 42 13. Differences in development between the two groups were analyzed by MANOVA for repeated measures. The results of these analyses (Table 3) showed that elite players had higher values (P ⬍ 0. sum of 4 skinfolds (biceps.1 ⫾ 15. the elite players were significantly taller (P ⬍ 0.5 ⫾ 0.6 ⫾ 6. of subjects.6 ⫾ 0.5-yr intervals starting when the subjects had a mean age of 11.1 ⫾ 15. yr n Height.97 17. no. significant differences in development between the groups according to Tanner stages were present (MANOVA for repeated measures).05) and had lower values for skinfold measurements (P ⬍ 0.05). P ⬍ 0.3 47 12.8 ⫾ 2. n.e. of subjects.9 49 12.4 ⫾ 6. the positive relationships between all strength parameters and being an elite player remained highly significant. Insulin-like growth factor I levels in elite and nonelite young male soccer players. Values are means ⫾ SE. When the relationships between elite/nonelite player and strength were adjusted for height.5 ⫾ 6. only the relationships with back muscles and handgrip remained significant (P ⬍ 0. body mass.9 43 12.3 12. 2. however.6 27.3 18. n.3 ⫾ 10.6 27.4 12 Values are means ⫾ SD.2 ⫾ 7.7 ⫾ 2.4 ⫾ 7.9 ⫾ 1.7 43 154.9 5. Fig.4 ⫾ 2.0 ⫾ 7. No significant differences between the groups in BMI or IGF-I (Fig.5 17.1143 STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS Table 2.3 43. Difference (adjusted for age) between groups is significant for 4 test rounds (P ⫽ 0.0 ⫾ 3.7 ⫾ 7.8 ⫾ 0. In univariate analysis a significant positive relationship was found between all strength parameters and being an elite player.05).7 47 150.5 ⫾ 1. Dotted line indicates that the last test round included a reduced number of players (n ⫽ 28). no.0 17. Measurements were taken at 0. BMI.4 9.7 17.6 16 166.6 37.9 ⫾ 6.4 12 160.7 ⫾ 14.6 ⫾ 4. no significant elite/nonelitetime interaction for any of the strength parameters. No Fig.8 41.7 50 147.076). a small decrease in regression coefficients was observed.2 ⫾ 8. When adjustment was made for age. kg/m2 Skinfold.1 33.4 43 12.3 ⫾ 18.3 3. The increase in strength over time was. and sum of skinfolds.4 ⫾ 0.1 ⫾ 0.8 19. ml n 2 3 4 Elite Nonelite Elite Nonelite Elite Nonelite Elite Nonelite 11. When adjustment was made for serum testosterone and IGF-I. with an adjustment for age.6 ⫾ 0.5 44 160 ⫾ 8.5 16 13. In general.9 ⫾ 0. RESULTS The characteristics of the subjects are presented in Table 2.9 40. 1. of subjects.1 29.05) compared with nonelite players for all strength parameters throughout the measurement period.3 ⫾ 6.9 17.1 36.5 9. The elite players had greater testicular volume than did the nonelite players and higher serum testosterone concentration (Fig.0 53. P ⬍ 0. and skinfold n is the same as for age. 2) were present.0 ⫾ 7. Characteristics of subjects Test Round 1 Age.3 46. as well as for numbers of years of training in organized soccer.6 6. Adjustment for body dimensions also led to a decrease in regression coefficients.5 28.5 48 152.4 yr.015) with a tendency for significance present for 3 test rounds (P ⫽ 0.3 ⫾ 5.7 39. kg BMI.05).5 ⫾ 0.9 ⫾ 1.5-yr intervals.7 ⫾ 7. a more marked decrease was found for the relationship with leg extension by using both legs.7 yr (nonelite).1 ⫾ 2.5 44 155. body mass index.05). .1 18. Values are means ⫾ SE. subscapularis.4 5. the regression coefficients for elite/nonelite players decreased slightly.1 ⫾ 6.5 (SD) (elite) and 11. Each test round was carried out at 0. no.6 ⫾ 7.3 ⫾ 5.9 ⫾ 0.8 ⫾ 3. similar in both groups.3 ⫾ 8. weight.1 ⫾ 6.7 ⫾ 6.6 ⫾ 3.5 ⫾ 2. cm Body mass. triceps. No significant difference between elite and nonelite players was present.9 ⫾ 0. 1.. BMI.0 ⫾ 5. For height. skinfold.8 7. Figures 3–6 show the development in strength parameters for elite and nonelite players.3 35.

Isometric strength for abdominal and back muscles measured in standing elite and nonelite young male soccer players. This relationship was independent of testosterone and IGF-I. Broad jump performance in elite and nonelite young male soccer players. P ⫽ 0. Values are means ⫾ SE.011. 3. no. 6.e. no. P ⫽ 0.001. 5.7 yr (nonelite).002. DISCUSSION The results of the GEE analysis regarding the longitudinal relationship between strength and serum testosterone are presented in Table 5. P ⫽ 0. of subjects for elite players/nonelite players in each test round. Difference between the groups is significant: abdominal. n. Values are means ⫾ SE. no. n. of subjects. Values are means ⫾ SE. and back. i. Isometric strength for leg extensors measured in seated elite (E) and nonelite (N-E) young male soccer players.002. Increases in muscle strength with age in young boys cannot simply be explained by growth. i. and IGF-I showed more or less the same picture as for the longitudinal relationships between being an elite/nonelite player and strength. because it has been shown in both prepubertal and pubertal boys that strength increases more rapidly than does height (6). no. n. n.004. It Fig. subjects for elite players/nonelite players in each test round. P ⫽ 0.01. Adjustment for age. body dimensions.1144 STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS Fig..5 (SD) (elite) and 11.. The adjustment for body weight. indicating that the greater strength was not solely due the level of serum testosterone. height. of subjects.6 ⫾ 0. Values are means ⫾ SE.5-yr intervals starting when the subjects had a mean age of 11. a strong positive relationship between being an elite player and the level of strength was observed. a decrease in regression coefficients. . Difference between the groups is significant.9 ⫾ 0. Differences between the groups are significant. indicating that testosterone is important for development of muscle strength in young boys. P ⫽ 0. In addition. P ⫽ 0. 4. none of the relationships between serum testosterone and strength parameters was significant.002. The present study showed that both elite and nonelite players through a 2-yr period had an increase in strength that was related to the levels of serum testosterone. and nondominant leg. dominant leg. and sum of skinfolds led to a dramatic decrease in regression coefficients. Fig. in general. Differences between the groups are significant: both legs. Fig. Univariate analysis showed a significant positive relationship between development in all strength parameters and serum testosterone concentration. Isometric handgrip strength (dominant hand) in elite and nonelite young male soccer players.e. Values are means ⫾ SE. Measurements were taken at 0. P ⫽ 0.

3 ⫺47.1 0. carried out at 0.4 ⫺62.5–42. indicating that these factors also play a role in the development of strength.002 0.4–243.1 ⫺71.001 0.4 to ⫺45. This relationship is dependent on anatomic dimensions and skinfold thickness.3 95% CI ⫺67.5-yr intervals on isometric strength and broad jump.8 95% CI ⫺44 to ⫺0.4 to ⫺2.057 ␤ ⫺22.5 ⫺17.762 0.2 ⫺651.000 0.000 0.502 0. who showed a Table 4.97 2.0 ⫺264.5 to ⫺21.4 ⫺66. age per se has a positive influence on muscular strength.6 95% CI ⫺29.7 ⫺158.7 ⫺100.8–1. Furthermore. From experimental data and from the recognition that testosterone has a prominent anabolic effect. insulin-like growth factor I. adjusted for confounding factors Leg Extension Univariate Analysis Abdominal Muscles Back Muscles Handgrip Dominant leg Elite/nonelite P 0.002 ␤ ⫺35.7–117 0. Results of MANOVA for 3 test rounds.2–166.4 to ⫺5.003 0.9 ⫺145.3 ⫺266.4 ⫺365.9 ⫺11.000 0. which is in agreement with Asmussen and Heebøll-Nielsen (3).1 ⫺43 to ⫺7.000 0.8 ⫺413.052 0.1 95% CI ⫺50.6 ⫺181.004 ␤ ⫺29.014 ␤ ⫺11.9 ⫺47. weight.9–3.001 ⬍0.4 ⫺61.226 ⫺165.1 0.006 0.011 0.8 ⫺15.000 0.2 to ⫺175.2 ⫺39.1 ⫺41.1 0. especially in tests that require a high degree of neuromuscular coordination. IGF-I.001 ⫺242. and 95% confidence intervals (CI) obtained from generalized estimating equations analysis. in the present study the percentile difference between the sum of the strength of each leg and both legs decreased with age [from 39 to 15% (elite) and from 34 to 22% (nonelite)].2 Adjusted for age P 0.4 0.003 ⫺178.8 ⫺177. stature.014 ⫺305.000 ⫺13.7 ⫺94.3 ⫺63.2 95% CI 29.3 22.91 0.2 Adjusted for dimensions: Height and body mass P 0.7 ⫺353. time by team.428 ⫺79. longitudinal regression coefficients (␤).191 0. except for leg extension with both legs.3 to ⫺41.038 0.8 ⫺298.2 354. .000 0.1 ⫺12.008 ⫺152.8 0.004 0.594 0.9 Adjusted for IGF-I P 0.2 to ⫺22.999 0.7 0. body mass ⫹ 4 skinfolds P 0.9 ⫺15.7 0.9 ⫺112.9–6.000 ⫺11.7 ⫺67.5 95% CI ⫺24.029 ⫺93. besides dimensions. it has been suggested that testosterone is responsible for the in- crease in strength in male individuals at puberty (5).3 ⫺19 to ⫺7.9 to ⫺1.9 0.1 to ⫺11.7 0.1 ⫺54.5 0. and time of measurement Handgrip Abdomen Back Both legs Nondominant leg Dominant leg Broad jump Team Time by Team Time of Measurement 0. 28) that MVC for leg extension with both legs are less than the sum of MVC for each leg.133 ␤ ⫺6.8 0.7 ⫺122.7 ⫺69.001 ⫺9.5 to ⫺1.7 ⫺550 to ⫺61.000 0.8 ⫺478.5–0.6 0.6 Broad Jump 0.4–151.001 ⬍0. Longitudinal relationships between being an elite or a nonelite player and development of isometric strength/broad jump. multiple ANOVA for repeated measures.4 Nondominant leg Both legs 0.048 0.073 ⫺180 ⫺376.4 to ⫺11.7 to ⫺69.5–102.316 ⫺44.8 0.000 0.8 to ⫺10.032 ⫺249.2 to ⫺20.2 0. Asmussen and Heebøll-Nielsen (2) suggested that.058 ⫺6.5 to ⫺49.763 ␤ ⫺1.001 ⬍0.1145 STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS Table 3.3 ⫺67.3–20.3 to ⫺124.002 0.6 ⫺155.4 ⫺275.000 0.696 ⬍0. We found that the relationship between strength development and changes in serum testosterone was independent of age.9 Values are P values.507 ⫺1.8 to ⫺9.001 ⬍0. is more likely to be due to an interrelationship between several factors such as age.9 ⫺7.8 to ⫺13. indicating a limit in neural output.214 0. MANOVA.8 to ⫺25.1 to ⫺59.9 to ⫺38. and maturation (endocrine and neurological).0 to ⫺27.8 ⫺46 ⫺185.8 to ⫺3.000 0.5–20. testing the effects of team (elite/nonelite).4 to ⫺29.001 ⬍0.000 ␤ ⫺44.371 0.4 ⫺433.046 ⫺121.1 to ⫺18.000 0.002 0. In the present study this is confirmed by a significant positive relationship between development in all strength parameters and serum testosterone concentration.6 ⫺46.8 ⫺301.9 ⫺57.010 0.8 ⫺30.001 ⬍0.3 ⫺72.6 Adjusted for serum testosterone P 0.6 to ⫺15.4 ⫺25.45 0.4 to ⫺20.4 Adjusted for years of training P 0.000 0.9 ⫺132. muscle size.1 0.6 ⫺34.8 ⫺101.005 0.4 ⫺66. It has been shown for both children and adults (14.7 to ⫺4.001 Values are P values adjusted for age (age as covariate).6 0.4 to ⫺8.719 ⫺54.17 ⫺81.5 ⫺89 to ⫺35.3 ⫺49 ⫺239.7 ⫺241.4 ⫺34.6 95% CI ⫺59.159 ⫺4.004 0.1 Height.000 0.4 to ⫺120.4–3.2 ⫺31.2 to ⫺26.001 ⫺9.3 ⫺228.

7 0.2 0. which only revealed a tendency for independence (P ⫽ 0.2–12.8–33.011 3. abdominal muscles.8 0. In the present study no difference in BMI was found between the two groups.1 0.1 0.790 0.000 8.1 0. maturation.3 0. Leatt et al.3–2.5–1. had almost the same strength values as did Danish schoolboys. the GEE analysis was carried out. Longitudinal relationships. between serum testosterone levels and development of isometric strength/broad jump.1–49.7–2.4 0.3–51 0.000 1.0 ⫺9.5 0.0 0.3 0.055 13 ⫺0.991 0.3 ⫺2.3 22.6 ⫺30.000 2.2 0.000 11.000 6.6–13.1 0. who were examined in 1981 (11).7 3.6–151.4–1 0.2–5.1–1.8 0.000 46.8 0.3–1.7 20.572 5.1–19. The elite players in the present study were also stronger compared with the nonelite players when a correction was made for the small difference in age between the two groups.000 60 37.6 0.9 22. it may have been caused by qualitative changes in the muscles such as a lower ratio of connective tissue to muscle tissue so that the same mass of musculature may be brought to produce more tension in the elite players.001 5.165 2.2–4.758 ⫺5. Alternatively.834 ⫺1.7 2.1 0.7–14. Ref.7–70.9–72.9 0.9–82.g.000 99.484 5.9–14.1 2.000 6.4 0.8 5.466 0.9–7. indicating that the development in strength was related to factors associated with being an elite player Both legs 0.012 4. Leatt et al. 7).2 1. It seems likely that the development of leg muscle strength in particular would give an advantage to the elite soccer player.9 independent of testosterone concentration. possibly caused by muscle hypertrophy as a response to training.000 11 7.1 0. adjusted for confounding factors Leg Extension Univariate Analysis Testosterone P ␤ 95% CI Adjusted for age P ␤ 95% CI Adjusted for dimensions Height and body mass P ␤ 95% CI Height.008 86. It is also plausible that part of the difference in strength may result from a better mastery of the neuromuscular system in the elite players caused by the training regimen that the elite players were exposed to from an early age.9–5.1 ⫺14.1 0.9 0. body mass ⫹ 4 skinfolds P ␤ 95% CI Adjusted for IGF-I P ␤ 95% CI Abdominal Muscles Back Muscles Handgrip Dominant leg Nondominant leg 0.9–9.058.029 17.029 0.4 0.2–15.000 46.7–76. It cannot. while not as strong as the elite players.1 50. The nonelite players.3 gradual decrease of this difference in male subjects aged 15–35 yr. The fact that the elite players were initially stronger compared with the nonelite players could partly explain why the increase in strength for the elite players during the test period did not lead to an increase in the difference between the two groups because strength is known to increase more from lower initial levels (e. respectively).. which indicate a larger lean body mass in the elite players.8 ⫺42. but the elite players had less subcutaneous fat evaluated from skinfold measurements.4 ⫺22.4–9.2 ⫺12. The relationship between strength development and elite/ nonelite was independent of serum testosterone and IGF-I.726 ⫺0. or competition level. (17) showed a greater isokinetic and explosive strength in Canadian national soccer players who were under 18 yr old compared with the national players who were under 16 yr old.7–147. This indicates that the development in strength is related to some extent to a hypertrophy of the muscles.6 Broad Jump 0.7–23.3 ⫺0.073 and P ⫽ 0. so the increase could be due to growth. The difference in strength between elite players and nonelite players appears not to be due to the difference in height between the two groups because the difference was independent of dimensions.2 0.9 ⫺2. When the sum of skinfolds was included with dimensions in the GEE analysis of relationships between elite/nonelite and strength development.2 7.0 0.604 ⫺0.5 ⫺1. except for leg extension with both legs and broad jump. The elite players were taller and more mature compared with the nonelite players.8–31. Thus dependency of age when examining extension with two legs (with some degree of neuromuscular coordination) could be explained by a requirement for a neuromuscular maturation possibly related to age.1–13.3 1.9 ⫺19.3 ⫺0.4 0.1–9.2 0.1–2 0.9 1. aged 11 yr. for both elite and nonelite players as one group.1 0.4 0.1146 STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS Table 5. The reason for this increase in strength may be due to a greater relative increase in muscle mass of the elite players and thus a larger cross-sectional area of the muscles.000 56.4 3.7 ⫺3.9 4. how- .5–2.8 0.6 36.7–5.9 0.636 9.9 0.442 14.000 8.575 0.6 0.829 0. found that elite players had more lean body mass compared with normal subjects. and broad jump but not for back muscles and handgrip strength. To examine this relationship.3–26. a dependency was found for leg extensions.000 2.

. 10. Sale. P. stage of puberty.. and W. J. 14. Metab. height velocity. 8: 371–380. A. 3. W. Beillot. and S. Zachmann. 5. E. no. J.. D. Influence of sex. D. Oseid and K. 1–44. development of isometric strength and performance in broad jump was related to changes in serum testosterone concentrations but also influenced by body size. 1147 11. V. and Team Danmark. accepted in final form 26 April 1999. edited by S. Denmark (E-mail: L1Hansen@aki. K. E. R. Rochcongar. Copenhagen. Gebski. J. biochemical. Address for reprint requests and other correspondence: L. 1972. Main. testicular size. 34: 319–326. In: Physiology of Sports. Longitudinal data analysis for discrete and continuous outcomes.-H. P. Muscle strength of boys and girls. Tanner. W. Sports Med. In: Communications From the Testing and Observation Institute. G. 1986. J. Round. Kaplan. Pediatr. p. 1994. Shepard. 1981 compared to 1956. and D.. Kemper.. London: Spon. Champaign. vol. J. Clin. J.. Isokinetic investigation of knee extensors and knee flexors in young French soccer players. and K. Champaign. Comparison of motor unit activation during unilateral and bilateral leg extension. and N. 17: 199–211. The Encyclopedia of Sports Medicine. E. L. Dalgaard. Res. 1989. Carmel. T. 1959. weight. Relation between the longitudinal development of lipoprotein levels and lifestyle parameters during adolescence and young adulthood. Training in elite young athletes [the training of young athletes (TOYA) study]: injuries. Sacco. 18. Twisk.. 1996. 1989. Tanner. 2. D. P. J. endocrinologic. V. 1984. and S. Universitetsparken 13. H. McNeil. Appl. 51: 170–179. 9: 161–171. Br. Sports Sci. H. and stages of puberty. August. flexibility and isometric strength. Soc. B. Children and Exercise XIII. T. and P. .dk). 1974. K. and H. be excluded that the differences between the elite and nonelite players are due to an early selection of boys with higher strength for the elite group. and K. M. R. 1956. R. Asmussen. 1986. K. R. A. 27. van Mechelen. E. 16. Physiol. edited by O.Carlsen. 39: 1–37. Falkner and J. Testicular volume during adolescence. Metab. p. Hall. Denmark: Danish Natl. New South Wales. 1996. Asmussen. M.. Rowland. 22. E. III. Champaign. G. J. Liang. J. Carron. In summary. 3–13. neurologic. Heebøll-Nielsen. Hormonal changes in puberty. A dimensional analysis of physical performance and growth in boys. DK-2100. Ann. A. A. Mero. Clin. Strength development in boys from 10 through 16 years. V. 6: 246–256. Furthermore. 30. Age. Marshall. Jaakkola. Hansen. A. Heebøll-Nielsen. Kings. Appl. The biological aspects of puberty. In: Communications From the Testing and Observation Institute. H. and K.. Gisolfi and D. (Editor). Resistance training during pre. 1992. In: The Child and the Adolescent Athlete. Molbech. Dept. 1990. mechanisms. K. W. edited by T. J. Helms. Maffulli. 78: 744–752. FSH. IL: Human Kinetics Books. and weight-training. of Human Physiology. Mellenbergh. J. In: Human Growth. K. Budliger. Methods for evaluation of muscle strength. adolescent. Kemper. 25.ku. indicating that being an elite player per se affected the development of strength. V. Blimkie. and J. Bang. User’s Guide (3rd ed. 2. Prader. C.. Med 28: 123–126. Komi. In: Perspectives in Exercise and Sports Medicine.. Haflinger. Dis. and J. Dassonville. IL: Blackwell Scientific. Copenhagen. IL: Human Kinetics. and B. REFERENCES 1. P. P. A longitudinal study of changes in physical performance of 10–15-year-old girls and boys. and D. L. 1976. Plyley. F. 17. Cardiorespiratory and metabolic responses to exercise: maturation and growth. 1988.. weight velocity. Sports Sci. and R. Developmental Exercise Physiology. J. Serum insulin-like growth factor-1 in 1030 healthy children. Leatt. Vandervoort. V. Chicago. Kulin. 1982. Biometrics 42: 121–130. This study was supported by grants from the Danish Association of Soccer. Leung. K. Reilly. K. 20. J. SPIDA User Manual. H. A cross-sectional survey of upper and lower limb strength in boys and girls during childhood and adolescence. UK: Blackwell Scientific.. E. Puberty. 21. Morvan. 9: 448–450. Lunn. Cooper. R. G. 1990. Physiol. 1996. Copenhagen. J. and W. 1992. Zeger. C. Sports Sci. J. N. Endocrinol. G. 26.. S. Sport Sci. J. Heebøll-Nielsen. R. New York: Plenum. D. Lamb. Denmark: Danish Natl. A. 3. Rasmussen.STRENGTH AND DEVELOPMENT IN YOUNG SOCCER PLAYERS ever. Blimkie. M. Physiol. testicular size and bone age with male pubertal development. Child Dev. 6. 41–67. G. Acta 29: 61–72.. O. and body mass index. Skakkebæk. K. B. Paediatr. Twisk. C. Clinical longitudinal standards for height.. the Danish Association of Sports. Oxford. Correlations of plasma testosterone. W. Can. 17: 264–279. 17: 75–82. Bar-Or. 1991.. Klausen. 12.and sex-associated variation in strength during childhood: anthropometric. 1994. Monogr. 6. P. Tanner. 13. 56: 46–51. J. 5: 165– 175. 1988. Rev.). Bailey. 5. Child. Strength. A. Hum. Scand. 1961. Komi. LH. Relationships between muscle fibre characteristis and physical performance capacity in trained athletic boys. Parker. 9. Assoc.-Y.. p. Hertel. 29. 24. Mu ¨ ller. Physical performence and growth in children.. Schibye. Epidemiol. A. trainability. Assoc. G. Heebøll-Nielsen. edited by C.and early puberty: efficacy. 23. Appl. Klausen. SPSS. 1992. Jan. of Infantile Paralysis. Endocrinol. J. J. Juul. Whitehouse. 4. Institute of Exercise and Sport Sciences. and J. Received 9 April 1998.. Isometric muscle strength of adult men and women. N. 7. Int. 28. Helv. 4: 37–43. Specific muscular development in under-18 soccer players. Med. Mu ¨ ller. J.. no. and M. Biol. Epidemiol. Kind. 1996. Grumbach. Strength and Power in Sport. Jørgensen. M. Moroz. J. Australia: Macquarie Univ. 19. J. R.. and physical activity correlates. morphologic. and K. edited by F. and J. M. Community Health 50: 505–511. A. sex. IL: SPPS. 11. M. and P. 1974. Ann. van Mechelen. 7: 593–603. Relation between the longitudinal development of lipoprotein levels and biological parameters during adolescence and young adulthood in Amsterdam. age and intelligence. ver. IN: Benchmark. J. 1955. and adults: relation to age. H. Mellenbergh. L. Asmussen. 15.. Asmussen. 8. and persistence. Heebøll-Nielsen. Sp. Jørn Mu¨ller for doing the measurements of testicular volume and the Tanner stage evaluations and further for the valuable discussions regarding the interpretations of the maturational data. Arch. vol. J. genetic. indicating that testosterone is important for development of strength in young boys. P. 1996. and D. P. 1987. SPSS-X. elite players were stronger than nonelite players independent of testosterone concentration also with correction for body size. M. M. p. The authors gratefully acknowledge Dr. C. Jones. H. 113–122. for Infantile Paralysis. J.