Journal of Strength and Conditioning Research, 2004, 18(1), 121–128
q 2004 National Strength & Conditioning Association
CHANGES IN EXERCISE PERFORMANCE AND
HORMONAL CONCENTRATIONS OVER A BIG TEN
SOCCER SEASON IN STARTERS AND NONSTARTERS
WILLIAM J. KRAEMER,1 DUNCAN N. FRENCH,1 NIGEL J. PAXTON,2 KEIJO HA¨KKINEN,3
JEFF S. VOLEK,1 WAYNE J. SEBASTIANELLI,2 MARGOT PUTUKIAN,2 ROBERT U. NEWTON,5
MARTYN R. RUBIN,1 ANA L. GO´MEZ,1 JASON D. VESCOVI,1 NICHOLAS A. RATAMESS,1
STEVEN J. FLECK,4 J. MICHAEL LYNCH,2 AND HOWARD G. KNUTTGEN2
1
Human Performance Laboratory, Department of Kinesiology, The University of Connecticut, Storrs, Connecticut
06269; 2Center for Sports Medicine/Laboratory for Sports Medicine, Pennsylvania State University, University
Park, Pennsylvania 16802; 3Department of Biology of Physical Activity, The University of Jyva¨skyla¨, Jyva¨skyla¨,
Finland; 4Department of Sport Science, Colorado College, Colorado Springs, Colorado 80903; 5School of
Biomedical and Exercise and Sport Science, Edith Cowan University, Joondalup, Australia.
ABSTRACT. Kraemer, W.J., D.N. French, N. Paxton, K. Ha¨kkinen, J.S. Volek, W.J. Sebastianelli, M. Putukian, R.U. Newton,
M.R. Rubin, A.L. Go´mez, J. Vescovi, N.A. Ratamess, S.J. Fleck,
J.M. Lynch, and H.G. Knuttgen. Changes in exercise performance and hormonal concentrations over a Big Ten soccer season in starters and nonstarters. J. Strength Cond. Res. 18(1):
121–128. 2004.—As a consequence of the physiological demands
experienced during a competitive soccer season, the antagonistic
relationship between anabolic and catabolic processes can affect
performance. Twenty-five male collegiate soccer players were
studied throughout a season (11 weeks) to investigate the effects
of long-term training and competition. Subjects were grouped as
starters (S; n 5 11) and nonstarters (NS; n 5 14). Measures of
physical performance, body composition, and hormonal concentrations (testosterone [T] and cortisol [C]) were assessed preseason (T1) and 5 times throughout the season (T2–T6). Starters
and NS participated in 83.06% and 16.95% of total game time,
respectively. Nonstarters had a significant increase (11.6%) in
body fat at T6 compared to T1. Isokinetic strength of the knee
extensors (1.05 rad·sec21) significantly decreased in both S
(212%) and NS (210%; p # 0.05) at T6. Significant decrements
in sprint speed (14.3%) and vertical jump (213.8%) were found
at T5 in S only. Though within normal ranges (10.4–41.6
nmol·L21), concentrations of T at T1 were low for both groups,
but increased significantly by T6. Concentrations of C were elevated in both groups, with concentrations at the high end of
the normal range (normal range 138–635 nmol·L21) at T1 and
T4 in NS and T4 in S, with both groups remaining elevated at
T6. Data indicate that players entering the season with low circulating concentrations of T and elevated levels of C can experience reductions in performance during a season, with performance decrements exacerbated in starters over nonstarters. Soccer players should therefore have a planned program of conditioning that does not result in an acute overtraining
phenomenon prior to preseason (e.g., young players trying to get
in shape quickly in the 6 to 8 weeks in the summer prior to
reporting for preseason camp). The detrimental effects of inappropriate training do not appear to be unloaded during the season and catabolic activities can predominate.
KEY WORDS. testosterone, cortisol, overtraining, conditioning,
training, football
INTRODUCTION
S
occer has been characterized as a high intensity sport that combines intermittent bouts of
anaerobic and aerobic exercise (6, 12). During
the course of a competitive soccer season,
player’s bodies are continuously subjected to
a variety of psychological and physical stresses from both
practice and competition (4, 6). As a consequence of the
physiological demands placed on players, conditioning
programs for soccer require that aerobic capacity,
strength, power, speed, and speed endurance be developed as fundamental components of physical conditioning. Consequently, the goal of soccer practice and physical
conditioning is to provide a stimulus for soccer-specific
adaptations that will result in improved athletic performance (7). The maintenance or improvement in performance standards is not, however, solely determined by
appropriate conditioning. The ability of bodily systems
(e.g., neuromuscular system, endocrine system) to recover
and regenerate following composite stresses including
strenuous activity, psychological stress of practice and
competition, etc., can also influence physical performance. Of particular importance to force development is
the manner in which muscles respond and remodel following exercise stressors, (i.e., practice, conditioning, or
competition). When a player is both training and competing, the dynamic homeostatic balance created between
anabolic (building) and catabolic (breakdown) processes
within the muscle can ultimately influence muscular
force characteristics and, therefore, affect the quality of a
player’s performance (8).
The effects of soccer training on immune function (22),
muscular strength (19, 20), the hormonal responses to a
single game (17), and long-term training (3–5) have previously been reported. Yet few data exist to track hormonal responses over the course of a competitive season.
Indeed, biological and quantitative assessments are now
a common practice used for the evaluation of physiological responses to many sport competitions and training
programs. A wide variety of biochemical, hematological,
and physiological markers have been used for the longterm monitoring of athletes (9). In particular, cortisol and
testosterone have been identified as reliable markers of
training stress (2, 21, 10, 11, 18). Reflective of this study,
Carli et. al (4) have demonstrated that during the course
of a competitive soccer season basal levels of both cortisol
and testosterone can become significantly elevated. Based
on such data, we hypothesized that increases in testosterone and cortisol would occur over the course of the sea121
122
KRAEMER, FRENCH, PAXTON
ET AL.
son. However, the potentially greater competitive responsibilities being placed on the actual starters in a sport,
and the differential physiological and performance effects
related to starter (S) or nonstarter (NS) status have not
been clarified in prior soccer research. Thus, it was not
apparent how hormonal changes would temporally track
in S and NS over the course of a National Collegiate Athletic Association (NCAA) Division I Big Ten Conference
season.
While the use of biochemical markers for long-term
monitoring of exercise stress is widespread, the relationship between hormonal concentrations and physical performance standards during the course of a competitive
season remains unclear. If the physical demands of practice, conditioning, and competition are too great, it might
be hypothesized that catabolic activities will predominate. If, however, the body is able to successfully cope
with the demands, anabolic metabolism can help to maintain or improve performance over the course of the season
(9). Nevertheless many factors impact this delicate metabolic balance, including conditioning activities, practice
schedules, academic demands, psychological stressors,
and competition, each contributing to the overall physiological status of a player. The purpose of this study was
to investigate the physiological and performance changes
that take place over a Big Ten season in college soccer
players. In other words, we wanted to determine how
these players adapted to the effects of conditioning, practice, and high level competition over an entire season. We
were specifically interested in determining whether differences existed between S and NS on a collegiate soccer
team.
METHODS
Experimental Approach to the Problem
We utilized a collegiate soccer team in the United States
of America who were as a team ranked in the top 10.
Their competition included the national champion and
top contenders in NCAA Division I soccer. We monitored
the team throughout an 11-week competitive season consisting of 19 games using a longitudinal study design.
Measures of physical performance (isokinetic and isometric strength, sprint speed, vertical jump), body composition (body density, percent body fat), and hormonal concentrations (testosterone, cortisol) were assessed 6 times
during the course of the season. Baseline testing was performed 1 week before the first competitive game (T1), 4
assessments were made during the season (i.e., weeks 3,
7, 8, and 9; T2-T5), and a further assessment was made
1 week following the end of the competitive period (T6) to
determine if any dramatic recovery occurred. In all cases,
a minimum of 7 days separated test administration.
Throughout tracking players continued to participate in
regular team practices (i.e., speed, agility, and speed endurance conditioning; simulated game play) and an inseason strength and conditioning program performed
twice per week. All workouts were supervised by team
coaches. The in-season strength-training program targeted the major muscle groups (i.e., legs, back, chest) and
consisted of varied workouts with exercises focusing on
muscular power development (e.g., jump squats, back
squats, power cleans, bench throws) using loads of up to
75–85% of 1 repetition maximum (1RM). It should be noted that prior to T1 players completed a preseason heavy
strength development phase, with training loads at 90–
95% 1RM, but endurance run/sprint interval training was
based on individual preferences which we hypothesize led
to a potential acute overtraining stressor due to the desire
by each player to quickly gain cardiovascular fitness for
the start of the preseason practice. Between T1 and T6
the mean weekly amount of combined training was ;10
h·wk21. Subjects did not adjust their diets or lifestyles
significantly during the course of the season.
Subjects
NCAA Division I male soccer players at Pennsylvania
State University volunteered to participate in the 11week tracking study designed to examine physiological
responses to a competitive Big Ten soccer season. Prior
to commencing the study, players underwent physical examination by the team physician, and each was cleared
of any medical or endocrine disorders that might confound or limit their ability to participate fully in the investigation. The study was approved by the institutional
review board for use of human subjects, and all players
gave their written informed consent to participate. At the
conclusion of study, the 25 subjects fit the criteria of the
study and completed the season without confounding injury. Those subjects were assigned to 1 of 2 groups, S (n
5 11) or NS (n 5 14) based on the amount of game time
each played during the season. Starters and NS participated in 83.06% and 16.95% of total game time, respectively. None of the subjects were injured to the extent to
limit practice or playing time or confound the study by
injury. The physical characteristics (mean 6 SD) of the S
and NS, respectively, were: age, 19.91 6 0.9 and 18.71 6
0.9 years; height, 177.68 6 5.6 and 179.48 6 6.9 cm; body
mass, 73.43 6 6.7 and 71.59 6 6.6 kg; body fat, 9.43 6
2.2 and 8.29 6 1.7%.
Physical Performance Testing
All subjects had prior familiarization with all of the testing protocols to insure a stable baseline. Upon reporting
to the laboratory, subjects performed a standardized
warm-up. Submaximal cycling was carried out at 60 W
against a load of 0.5 kg for 5 minutes, followed by light
stretching. Concentric strength of the knee flexors and
extensors was determined at 1.05 rad·sec21 and 5.25
rad·sec21 using a Cybex isokinetic dynamometer (Lumex
Corp., Ronkonkoma, NY). Each player was given a series
of submaximal practice trials prior to each testing velocity. Subjects were seated in an upright position and secured around the chest and thigh. The test limb (dominant leg) was positioned and secured to the dynamometers lever arm according to the manufacturer’s guidelines,
just proximal to the medial malleolus. All positions were
noted and documented to assure reliability of testing conditions. Upon instruction, subjects performed 3 continuous maximal effort contractions (extension/flexion)
throughout the complete range of motion. Similar encouragement was given to each player by the investigator
team. The highest torque produced during each of the assessments was recorded and the data was used to represent the maximal isokinetic strength for that velocity. In
addition, angle-specific torque was evaluated at 608 knee
extension and knee flexion, and 458 knee extension and
knee flexion at each velocity. Tests of different contractile
velocities were assigned in a balanced, random order, and
2- to 3-minutes rest separated trials.
PHYSIOLOGICAL RESPONSES
Isometric strength of the knee extensors was also assessed for the dominant limb of the players. Again, the
Cybex isokinetic dynamometer was used for this testing
protocol. The positioning of the test limb was carried out
in an identical manner to that previously discussed. Assessments of isometric strength were performed at a constant position of 458 of knee flexion. Upon command, subjects were instructed to kick out maximally against the
lever arm in an attempt to exert maximal force as quickly
as possible. An investigator terminated each trial when a
positive rise in force tracing ceased to be observed. Each
trial had an average duration of between 4 and 6 seconds
using a rapid force production effort protocol for the test.
Three trials with a minimum of 2 minutes rest between
were performed with the highest of the 3 recordings obtained and defined as the player’s maximal isometric
force.
Maximal vertical jump height was measured using a
Vertec vertical jump measurement device (Sports Imports, Columbus, OH). Prior to testing, each subject’s
standing vertical reach was determined. Care was taken
to make sure the standing reach was accurately determined with regard to limb stretch. Subjects were then
given 3 trials to jump for maximum height, with 2 to 3
minutes rest separating trials. The highest jump of the 3
trials was recorded. Determination of maximal vertical
jump height was calculated by subtracting the standing
reach from the jump height.
Sprint speed was assessed over 18.3 meters (20 yards)
and 36.7 meters (40 yards) on a synthetic grass (AstroTurf) surface which was used for all testing. Sprint times
were recorded manually in the field using handheld digital stopwatches with the same timers used for each test.
Timers were assigned to a specific set of subjects and
were used for each test to assure test-to-test reliability.
The timers were placed at both the 18.3 meter (20 yard)
and 36.7 meter (40 yard) marks so that both times could
be recorded during a single trial. Subjects performed 3
trials, with the fastest time recorded serving as the measure of maximal speed.
Body Composition
Body composition was determined from 7 skinfolds taken
using a Lange skinfold caliper (Country Technology, Gays
Mills, WI). The 7 sites used were the triceps, subscapula,
suprailiac, thigh, midaxilla, abdomen, and chest. Measures were consistently recorded on the right side of the
subject’s body. Bone density was calculated using the
methods of Jackson-Pollock (13) and percent fat determined using the Siri equation (25). Three skinfold measures were taken at each site and a variance of less than
5% was used as an acceptable criterion for measurement.
The same experienced technician performed all measurements throughout the study period. Body mass was measured to the nearest 0.5 kg using an electronic scale
(American Business Equipment Company, Inc., New Holland, PA).
Hormonal Analyses
A resting blood sample of 10 ml was obtained from the
antecubital vein in the arm following a 12-hour overnight
fast. Samples were collected at the same time of the
morning (0800–1000 hours) for each player’s visit in order
to control for circadian variances. Whole blood was allowed to clot at room temperature, following which it was
TO A
COMPETITIVE SOCCER SEASON 123
FIGURE 1. Changes in percent body fat during the course of
an 11-week competitive soccer season. Values are means 6 SE
and represent the percent body fat recorded for nonstarters
(white) and starters (black). * p # 0.05 from T1.
centrifuged at 3200 rpm (48 C) for 20 minutes. The resultant serum was removed and stored at 2808 C until
subsequent analysis. Serum concentrations of testosterone and cortisol were determined in duplicate using a 125I
solid phase radioimmunoassay (Diagnostic Systems Laboratories, Inc., Webster, TX). Immunoreactivity values
were determined using a gamma counter and on-line data
reduction system (Model 1272, Pharmacia LKB Nuclear,
Inc., Gaithersburg, MD). The intra- and interassay variances for testosterone were less than 3.45% and 7.02%,
respectively. The intra- and interassay variances for cortisol were less than 4.44% and 5.66%, respectively.
Statistical Analyses
A one-tailed t-test was used to compare active game time
in order to determine starters from nonstarters, which
proved to be significantly different (i.e., again, S and NS
participated in 83.06% and 16.95% of total game time, respectively). A two way analysis of variance (ANOVA) with
repeated measures was used to determine the pairwise differences between starters and nonstarters where appropriate. When a significant F value was achieved, appropriate Tukey posthoc tests procedures were used to locate
the difference between the means. Test-retest reliabilities
for the experimental tests demonstrated intraclass correlations of R $ 0.95. Linear regression analyses were performed to examine the bivariate relationships between
hormonal and performance variables. In all cases, the level
of significance was set at p # 0.05.
RESULTS
Measures of body composition indicated that a change in
percent body fat occurred only in nonstarters during the
season (Figure 1). A significant increase of 1.6% was
found at T6 compared to preseason levels (T1). This increase in body fat was identified as having occurred during the latter half of the season, with no significant
changes found between T1 and midseason (T3). No significant changes in percent body fat were observed at any
time during the season for starters.
In both groups, isokinetic leg strength decreased over
time when measured at 1.05 rad·sec21 (Table 1). NS had a
decreased peak torque (;10%) of the knee extensors at T6
(204.73 6 7.40 Nm) that was significantly smaller when
compared to the corresponding T1 assessment (227.40 6
10.60 Nm). S also demonstrated a decrement during the
124
KRAEMER, FRENCH, PAXTON
ET AL.
TABLE 1. Changes in isokinetic knee extension and flexion
strength during the course of an 11-week competitive soccer season.
Torque (Nm)†
Group
Test
NS
S
NS
S
NS
S
T1
T1
T3
T3
T6
T6
Extension
227.40
240.31
213.12
214.11
204.73
211.48
6
6
6
6
6
6
Flexion
10.60
11.12
11.03
9.08*
7.40*
8.43*
126.67
152.04
119.16
136.44
131.55
139.80
6
6
6
6
6
6
6.65
9.27**
7.23
8.21**
4.78
6.85
† Values are means 6 SE and represent isokinetic strength
recorded at 1.05 rad·sec21 for nonstarters (NS) and starters (S).
* p # 0.05 from T1.
** p # 0.05 from corresponding NS group.
FIGURE 2. Changes in 20-yard sprint times during the course
of an 11-week competitive soccer season. NS 5 nonstarters; S
5 starters. Values are means 6 SE and represent the sprint
times recorded for NS (white) and S (black). * p # 0.05 from
T1.
knee extension, with significant reductions in peak torque
found at both T3 (;11%) and T6 (;12%) compared to T1
(214.11 6 9.08 Nm and 211.48 6 8.43 Nm vs. 240.31 6
11.12 Nm, respectively). No changes in peak torque of the
knee flexors were found in either of the groups during the
tracking period; however, there were significant differences between NS and S at T1 (NS, 126.67 6 6.65 Nm; S,
152.04 6 9.27 Nm) and T3 (NS, 119.16 6 7.23 Nm; S,
136.44 6 8.21 Nm). Comparisons of angle-specific isokinetic strength at 608 of knee extension were significantly
decreased (;19%) in the nonstarters at T6 (154.04 6 6.47
Nm vs. 191.17 6 11.61 Nm) (Table 2). In S, significant
decreases in angle-specific peak torque of the knee flexors
were observed at T3 for both 458 (17%; 116.66 6 7.03 Nm
vs. 140.35 6 8.60 Nm) and 608 (;9%; 103.06 6 6.58 vs.
113.36 6 8.53). Players classified as S also demonstrated
significantly greater angle-specific isokinetic torque for
knee flexion at 458 than NS at T1.
No significant changes in isokinetic leg strength at
5.25 rad·sec21 for knee extension or knee flexion were
found in either group. Furthermore, there were no significant differences between the 2 groups at any point
during the tracking period. Both groups did, however,
show significant increases during the isokinetic knee flexion at 608 at T6 compared to T3 (71.96 6 4.49 and 74.25
6 6.70 vs. 55.44 6 5.00 and 56.82 6 5.94 for NS and S
respectively; Table 3).
Measures of isometric strength showed no significant
changes during the season for either group. Furthermore,
no significant differences were observed in isometric
strength measures between the 2 experimental groups
(NS and S).
Nonstarters showed no change across time for the
18.3-meter (20-yard) sprint. A significant difference in
sprint speed was found in the S group, with the times at
T5 significantly longer than T1 (Figure 2). Neither group
demonstrated any changes in times for the 36.7-meter
TABLE 2. Changes in angle-specific isokinetic knee strength during the course of an 11-week competitive soccer season.
Angle-specific torque at 1.05 rad·sec 21 (Nm)†
Group
Test
NS
S
NS
S
NS
S
T1
T1
T3
T3
T6
T6
Ext—608
191.17
198.83
172.20
197.64
154.04
171.90
6
6
6
6
6
6
11.61
7.71
9.93
8.78
6.47*
10.89
Ext—458
132.50
145.79
133.38
148.68
119.31
121.85
6
6
6
6
6
6
8.42
7.02
7.50
9.24
7.83
10.44
Flex—608
113.36
133.00
103.46
103.06
117.69
117.76
6
6
6
6
6
6
8.53
8.91
6.66
6.58*
3.91
6.73
Flex—458
116.66
140.35
112.77
116.81
130.75
130.83
6
6
6
6
6
6
7.03
8.60**
7.63
7.84*
4.78
7.40
† Values are means 6 SE and represent isokinetic strength recorded at 1.05 rad·sec21 for nonstarters (NS) and starters (S).
* p # 0.05 from T1.
** p # 0.05 from corresponding NS group.
TABLE 3. Changes in angle specific isokinetic knee strength during the course of an 11-week competitive soccer season.
Angle-specific torque at 5.25 rad·sec 21 (Nm)†
Group
Test
NS
S
NS
S
NS
S
T1
T1
T3
T3
T6
T6
Ext—608
87.42
92.48
79.72
83.40
76.64
81.60
6
6
6
6
6
6
7.32
5.61
6.46
3.93
4.67
5.48
Ext—458
77.52
87.31
73.33
83.25
71.83
76.16
6
6
6
6
6
6
8.06
6.08
6.35
5.13
5.03
4.90
Flex—608
65.37
71.53
55.44
56.82
71.96
74.25
6
6
6
6
6
6
5.39
6.52
5.00
5.94
4.49*
6.70*
Flex—458
67.22
77.79
60.26
67.70
68.24
69.36
6
6
6
6
6
6
5.30
7.67
4.77
5.19
3.46
4.39
† Values are means 6 SE and represent isokinetic strength recorded at 5.25 rad·sec21 for nonstarters (NS) and starters (S).
* p # 0.05 from T1.
PHYSIOLOGICAL RESPONSES
FIGURE 3. Changes in vertical jump height during the course
of an 11-week competitive soccer season. NS 5 nonstarters; S
5 starters. Values are means 6 SE and represent the height
jumped for NS (white) and S (black). * p # 0.05 from T1 and
T3; $ p # 0.05 difference to corresponding NS group.
FIGURE 4. Changes in resting concentrations of serum
testosterone during the course of an 11-week competitive
soccer season. NS 5 nonstarters; S 5 starters. Values are
means 6 SE and represent the circulating concentrations for
NS (white) and S (black). * p # 0.05 from T2; ** p # 0.05 from
T1. $ p # 0.05 difference to corresponding S group.
(40-yard) sprint, nor were there any differences between
groups.
Significant changes in vertical jump performance
were observed only in the starters group. A significant
decrease in jump height was found at T5 when compared
to the baseline scores recorded at T1 (Figure 3). No changes were observed at any time for the NS. At T5, NS also
had a significantly higher vertical jump height when compared to S.
In both the NS and S groups, changes in circulating
concentrations of plasma testosterone were found during
the season. Significant increases in testosterone concentrations were observed for both groups (Figure 4). At T6,
the NS demonstrated significantly higher (;23%) concentrations compared to the values recorded at T2 (18.2
nmol/L vs. 13.95 nmol/L). In a similar manner, testosterone concentrations for the S group at T6 were significantly higher (;29%) when compared to baseline values
recoded at T1 (17.2 nmol/L vs. 12.25 nmol/L). Differences
between groups were only recorded at T3, with the NS
having significantly higher concentrations of circulating
testosterone compared to the S group. When correlations
between testosterone levels and performance characteristics were carried out, a significant relationship was
found in the NS between testosterone and vertical jump
TO A
COMPETITIVE SOCCER SEASON 125
FIGURE 5. Changes in resting concentrations of serum
cortisol during the course of an 11-week competitive soccer
season. NS 5 nonstarters; S 5 starters. Values are means 6
SE and represent the circulating concentrations for NS (white)
and S (black). * p # 0.05 from T2, T3, T5; $ p # 0.05
difference to corresponding S group.
FIGURE 6. Changes in the testosterone to cortisol ratio (T/C
ratio) during the course of an 11-week competitive soccer
season. NS 5 nonstarters; S 5 starters. Values are means 6
SE and represent the T/C ratio for NS (white) and S (black).
* p # 0.05 from T1 and T4.
performance at T5 (r 5 0.77). In those players categorized
as S, significant correlations were found between testosterone concentrations and knee flexion at 1.05 rad·sec 21
for T6 (r 5 0.58), knee flexion at 5.25 rad·sec 21 for T1 (r
5 0.55) and T6 (r 5 0.61), and peak isometric torque at
T1 (r 5 0.64) and T3 (r 5 0.71).
At T1, plasma cortisol concentrations were significantly higher in NS compared to S. This concentration
was above normal ranges (138–635 nmol·L21). There were
no significant changes in plasma cortisol concentrations
during the course of the season, with the exception of S,
who experienced a significant increase in cortisol concentrations at T4 only (Figure 5). The only significant correlations between cortisol concentrations and physical
performance were found in the NS, with vertical jump at
T2 (r 5 20.64) and T6 (r 5 20.59), 20-yard sprint at T2
(r 5 20.78), 40-yard sprint at T2 (r 5 20.57), and 1.05
rad·sec21 for knee extension at T1 (r 5 20.58) having a
relationship to cortisol concentrations. No correlations between cortisol concentrations and physical performance
were found in the players designated as S. The testosterone/cortisol ratio (T/C ratio) was found to change during
the season in the NS. Significant elevations in the T/C
ratio were observed at T6 when compared to the values
calculated for T1 and T4 (Figure 6). Again, no significant
changes were demonstrated during the season in those
126
KRAEMER, FRENCH, PAXTON
ET AL.
players recognized as S. Starters did, however, show that
the serum T/C ratio had a significant correlation to the
vertical jump performances assessed at T5 (r 5 0.65).
Nonstarters were found to have a significant correlation
between the T/C ratio and both the 20-yard sprint at T2
(r 5 0.56) and the peak isometric torque recorded at T1
(r 5 0.61).
DISCUSSION
Temporal responses to training and competition may mediate general performance decrements during the course
of the 11-week soccer season. Both S and NS experienced
reductions in exercise performance that were highlighted
in the latter stages of the competitive period (i.e., T6; near
playoffs and tournaments). Though more pronounced in
the starters, performance reductions were observed in all
players, indicating that performance adaptations may be
independent of total match play. Starters were found to
have significant reductions in both sprint speed (14.3%)
and vertical jump height (213.8%) compared to preseason
values (T1). Decreases in the peak isokinetic torque (1.05
rad·sec21) of the thigh musculature were found in both S
(212%) and NS (210%) at the end of the season. Nonstarters were also observed to have a significant increase
in body fat of 1.6%, a change not reflected in the S.
In association with the changes in exercise performance, changes in circulating concentrations of testosterone and cortisol were also tracked throughout the season.
Though within the normal range of 10.4–41.6 nmol·L21
reported in the literature (23), initial serum testosterone
concentrations for all players (NS, 14.9 nmol·L21; S, 12.25
nmol·L21) were considered somewhat low following preseason conditioning (T1). These concentrations were
found to increase significantly by T6 to 18.2 nmol·L 21 and
17.2 nmol·L21 for NS and S, respectively. In contrasting
fashion to the trends observed for testosterone, initial cortisol concentrations were considered elevated (548.49
nmol·L21) in S compared to the normal resting range of
138–635 nmol·L21, and above normal ranges in NS
(664.89 nmol·L21). Furthermore, circulating concentrations of cortisol remained elevated throughout the 11week season without any significant reductions in circulating concentrations. Since both testosterone and cortisol
have been identified as reliable markers of training stress
(1, 10, 11, 18, 18, 21), these data indicate that for much
of the season (T1-T5) catabolic processes (i.e., elevated
cortisol, reduced testosterone) may have predominated.
This assumption is supported by the low testosterone/cortisol ratio, previously referred to by Ha¨kkinen et al. (10,
11), which suggests catabolic metabolism was foremost
throughout the majority of the season. Interestingly the
catabolic environment that was apparently already initiated in the preseason was not obviated over the course of
the season.
Bosco et al. (3) have previously reported basal concentrations of both testosterone and cortisol to be maintained
within normal recognized ranges for a group of professional male soccer players. Celani and Grandi (5) examined the testosterone responses of trained nonprofessional soccer players over 12 weeks, and after an initial 30day rest period these investigators also observed that basal concentrations of testosterone were within normal
ranges. The present data indicate basal concentrations of
testosterone were within normal ranges, however, circulating concentrations were considered low for the predom-
inance of the season with S demonstrating lower values
at the beginning of the preseason, which may be due to
acute overtraining related to endurance overtraining.
Only at T6 did basal concentrations of testosterone significantly increase, with S and NS responding similarly.
This recovery appears to reflect a dramatic reduction in
total stress related to the season. Serum cortisol concentrations were found to remain elevated throughout the
entire season, with S having basal levels at the higher
end of normal ranges and NS above normal values (138–
635 nmol·L21). Carli et al. (4) indicate that serum testosterone and cortisol increase during a soccer season, however, the current study suggests that concentrations of
these training stress markers following off-season and
preseason conditioning may affect both the anabolic and
catabolic responses to the competitive season. These
changes may be exclusive to the training approach of collegiate soccer, in which players try to get into shape
quickly during the summer break prior to a very rigorous
preseason practice schedule in August. In addition, if
such high concentrations are created prior to preseason
training camp, our data indicates that no abatement of
this adrenal-cortical stress occurs in the circulating concentrations over the season. Both S and NS are exposed
to the same conditioning and soccer training stress and
competition is only one part of the total stress equation,
which is obviously greater in S. However, it appears that
practice and conditioning stress may also be important to
the physiological recovery as NS were not devoid of hypopituitary-adrenal-testicular maladaptations over the season. The dramatic increase in testosterone observed in
both S and NS at T-6 represents a potential rebound of
physiological function with stress reduction.
As corroborated in studies by Kraemer et al. (14) and
Aldercrentz et al. (1), sprint running increases circulating
concentrations of cortisol and decrease concentrations of
plasma testosterone. Categorized as high-intensity intermittent exercise, soccer has been identified as a sport that
has a prevalence of repeated bouts of maximal effort
sprints (6). Florini (8) previously reported the antagonistic relationship between anabolic and catabolic hormones,
and indicates that a decrease in plasma testosterone coupled with the catabolic state that accompanies elevated
cortisol levels can result in reduced physical performance
standards. Elevated cortisol concentrations lead to increased binding at the glucocorticoid receptor, and have
been shown to result in reduced protein synthesis and
concomitant losses in muscular force and functional performance (13). An elevation in testosterone concentrations can assist in balancing these catabolic effects; however decreased testosterone concentrations do not provide
adequate or optimal anabolic stimulation to offset the catabolic effects consequent to chronically elevated cortisol
levels (e.g., protein catabolism, immune suppression, in
attempt to preserve glycogen concentrations in muscle).
This was the case observed in this tracking study over
the season.
An increased catabolic environment within the muscles may have contributed to the observed diminished
knee extensor strength. Discussed previously, a catabolic
state can lead to diminished force production due to losses
of contractile proteins or neural transmitters typically
stimulated by testosterone interactions and such mechanisms would result in potential decreases in isokinetic
strength (8). This has been supported in work by Ha¨kki-
PHYSIOLOGICAL RESPONSES
nen et al. (10, 11) who have shown significant decreases
in testosterone with heavy weight lifting, and correlative
decreases between maximal strength and power and the
testosterone/cortisol ratio. The decreased ability to produce force may manifest itself as concomitant performance reductions in speed, vertical jump height, and
strength. The data from our investigation indicates that
S had low levels of circulating testosterone and what may
be considered elevated levels of cortisol. This may have
created a state that may be due to the greater amount of
physical stress evoked during the significantly higher
amount of game time. The earlier onset of decrements in
knee extension strength in S (at T3) vs. NS (at T6) may
reflect this dramatic influence of the catabolic dominance.
Increased muscle catabolism may also be a possible explanation as to why S experienced significantly greater
peak torque for knee flexion at 1.05 rad·sec 21 at T1 and
T3 compared to NS, but then demonstrated no difference
in knee flexion strength at T6. The fact that neither group
showed increases in isokinetic or isometric strength during the season with soccer training is consistent with the
¨ berg et al. (19). However, with the use of an
findings of O
in-season conditioning program, one might have expected
strength gains but sport practice and competition apparently overrode any gains from anabolic exercise in the
weight room and points to the importance of sport coaching decisions related to practice stress and total work
planned for different phases (e.g., amount of scrimmage
time in practice) of the season. It is not readily apparent
why elevated cortisol concentrations and decreased testosterone concentrations (i.e., low T/C ratio) would only
affect slow-velocity high-force contractions and not highvelocity low-force contractions. Kuzon et al. (16) have indicated that soccer is not only a high intensity exercise,
but also provides a significant oxidative stimulus that
may have a greater impact on Type I fibers. Since slow
velocity muscle actions have a greater contribution of
Type I motor units, any negative effects on Type I (slow
twitch fibers) would be reflected in slow velocity isokinetic testing and help explain our results (15). For this
reason, overuse or overtraining of endurance-related soccer practice drills and conditioning activities may be more
detrimental to Type I fibers and the slow twitch motor
units that contribute to slow velocity force production.
As the players in the present study entered the competitive soccer season with low values for testosterone
and what can be considered elevated levels of circulating
cortisol, it appears that preseason training played a critical role in determining the metabolic status of the players as they entered the competitive period. Data indicate
that players entering the season in apparent catabolic
state experienced reductions in performance throughout
the season, with performance decrements exacerbated in
S compared to NS. It is plausible that intensive training
prior to the start of the season, combined with continued
high intensity training during both practice and competition, contributed to chronically elevated cortisol concentrations and suppressed concentrations of testosterone.
Such physiological conditions may have contributed to
the subsequent decrements in muscle force production,
supporting the findings of previous research (8). From
these data, we suggest that monitoring players prior to
preseason conditioning during the summer could be a
necessary next step in gaining a more complete understanding of the changes that occur during the course of a
TO A
COMPETITIVE SOCCER SEASON 127
competitive soccer season. As most sport seasons are now
extending to year-round training, a full year of study may
be needed to fully document changes in collegiate soccer
players to a competitive season. Furthermore, greater attention may need to be paid to the exact structure of the
cardiovascular endurance training in relationship to
strength-related conditioning during the summer.
PRACTICAL APPLICATIONS
During the course of a competitive season, collegiate soccer players are exposed to a variety of physical and psychological stresses from practice, conditioning, and competition. The ability of the players to recover following
such activities can ultimately affect the quality of performance for ensuing physical activity. This study has indicated that the nature in which players enter a season
can have significant effects on physical performance tests
over the season. Without appropriate rest and recovery,
intensive preseason conditioning can be found to decrease
circulating levels of testosterone while simultaneously increasing concentrations of the catabolic hormone cortisol.
With the continued high intensity stress experienced
throughout the season, the consequence of entering the
competitive period in this nature is reflected as a chronic
catabolic environment for the neuromuscular system.
Such a catabolic physiological status results in significant
losses in muscular force that are manifested as decreases
in speed, vertical jump height, and strength. The latter
stages of a season may well be regarded as critical in determining the fortunes of a team, as it is during this period that play-offs and tournaments are contested. Present data indicate that both S and NS are susceptible to
decrements in performance capabilities if entering the
season with catabolic processes predominating. Assessment of the circulating concentrations of testosterone and
cortisol represent a possible means of monitoring the
‘stress’ of training and competition, therefore providing a
means of avoiding the performance decrements associated with the rigors of practice, conditioning, and competition. Sport practice schedules may need to be more carefully considered by head coaches who may not appreciate
the impact of their practice decisions on their players’
physiological status. Strength and conditioning professionals need to monitor such changes and help to educate
sport coaches as to the potentially negative physiological
impact of high scrimmage and total work practices over
the entire season. Such interfaces with strength and conditioning professionals and sport coaches will be the ‘‘new
frontier’’ for strength coaches as part of the total performance and sports medicine teams.
REFERENCES
1.
2.
3.
4.
ALDERCRENTZ, H., M. HARKONEN, K. KUOPPASALMI, H. NAVERI,
I. HUHTANIEMI, H. TIKKANEN, K. REMES, A. DESSYPRIS, AND J.
KARVONEN, Effect of training on plasma anabolic and catabolic
steroid hormones and their response during physical training.
Int. Sports Med. 7:27–28. 1986.
BANFI, G., M. MARINELLI, G.S. ROI, AND V. AGAPE. Usefulness
of free testosterone/cortisol ratio during a season of elite speed
skating athletes. Int. J. Sports Med. 14:373–379. 1993.
BOSCO, C., J. TIHANYI, AND A. VIRU. Relationships between
field fitness test and basal serum testosterone and cortisol levels in soccer players. Clin. Physiol. 16:317–322. 1996.
CARLI, G., C.L. DI PRISCO, G. MARTELLI, AND A. VITI. Hormonal
changes in soccer players during an agonistic season. J. Sports
Med. Phys. Fitness 22:489–494. 1982.
128
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
KRAEMER, FRENCH, PAXTON
ET AL.
CELANI, M.F., AND M. GRANDI. The pituitary-testicular axis in
non-professional soccer players. Exp. Clin. Endocrinol. 94:244–
252. 1989.
EKBLOM, B. Applied physiology of soccer. Sports Med. 3:50–60.
1986.
FILAIRE, E., X. BERNAIN, M. SAGNOL, AND G. LAC. Preliminary
results on mood state, salivary testosterone:cortisol ratio and
team performance in a professional soccer team. Eur. J. Appl.
Physiol. 86:179–184. 2001.
FLORINI, J.R. Hormonal control of muscle growth. Muscle Nerve
10:577–598. 1987.
FRY, A.C., W.J. KRAEMER, M.H. STONE, B.J. WARREN, J.T.
KEARNEY, C.M. MARESH, C.A. WESEMAN, AND S.J. FLECK. Endocrine and performance responses to high volume training
and amino acid supplementation in elite junior weightlifters.
Int. J. Sport Nutr. 3:306–322. 1993.
HA¨KKINEN, K., A. PAKARINEN, M. ALEN, H. KAUHANEN, AND
P.V. KOMI. Relationships between training volume, physical
performance capacity, and serum hormone concentrations during prolonged training in elite weight lifters. Int. J. Sports Med.
8:61–65. 1987.
HA¨KKINEN, K., A. PAKARINEN, M. ALEN, AND P.V. KOMI. Serum
hormones during prolonged training of neuromuscular performance. Eur. J. Appl. Physiol. Occup. Physiol. 53:287–293. 1985.
KIRKENDALL, D.T. The applied sport science of soccer. Phys.
Sportsmed. 13:53–59. 1985.
KONAGAYA, M., P.A. BERNARD, AND S.R. MAX. Blockade of glucocorticoid receptor binding and inhibition of dexamethasoneinduced muscle atrophy in the rat by RU38486, a potent glucocorticoid antagonist. Endocrinology 119:375–380. 1986.
KRAEMER, W.J., S.J. FLECK, R. CALLISTER, M. SHEALY, G.A.
DUDLEY, C.M. MARESH, L. MARCHITELLI, C. CRUTHIRDS, T.
MURRAY, AND J.E. FALKEL. Training responses of plasma betaendorphin, adrenocorticotropin, and cortisol. Med. Sci. Sports
Exerc. 21:146–153. 1989.
KRAEMER, W.J., J.F. PATTON, S.E. GORDON, E.A. HARMAN,
M.R. DESCHENES, K. REYNOLDS, R.U. NEWTON, N.T. TRIPLETT,
AND J.E. DZIADOS. Compatibility of high-intensity strength and
endurance training on hormonal and skeletal muscle adaptations. J. Appl. Physiol. 78:976–989. 1995.
KUZON, W.M., JR., J.D. ROSENBLATT, S.C. HUEBEL, P. LEATT,
M.J. PLYLEY, N.H. MCKEE, AND I. JACOBS. Skeletal muscle fi-
17.
18.
19.
20.
21.
22.
23.
ber type, fiber size, and capillary supply in elite soccer players.
Int. J. Sports Med. 11:99–102. 1990.
LUPO, C., L. BALDI, M. BONIFAZI, L. LODI, G. MARTELLI, A. VITI,
AND G. CARLI. Androgen levels following a football match. Eur.
J. Appl. Physiol. Occup. Physiol. 54:494–496. 1985.
MUJIKA, I., J.C. CHATARD, S. PADILLA, C.Y. GUEZENNEC, AND
A. GEYSSANT. Hormonal responses to training and its tapering
off in competitive swimmers: relationships with performance.
Eur. J. Appl. Physiol. Occup. Physiol. 74:361–366. 1996.
OBERG, B., M. MOLLER, J. GILLQUIST, AND J. EKSTRAND. Isokinetic torque levels for knee extensors and knee flexors in
soccer players. Int. J. Sports Med. 7:50–53. 1986.
OBERG, B.E., M.H. MOLLER, J. EKSTRAND, AND J. GILLQUIST.
Exercises for knee flexors and extensors in uninjured soccer
players: effects of two different programs. Int. J. Sports Med.
6:151–154. 1985.
PASSELERGUE, P., AND G. LAC. Saliva cortisol, testosterone and
T/C ratio variations during a wrestling competition and during
the post-competitive recovery period. Int. J. Sports Med. 20:
109–113. 1999.
REBELO, A.N., J.R. CANDEIAS, M.M. FRAGA, J.A. DUARTE, J.M.
SOARES, C. MAGALHAES, AND J.A. TORRINHA. The impact of soccer training on the immune system. J. Sports Med. Phys. Fitness 38:258–261. 1998.
YOUNG, D.S. Implementation of SI units for clinical laboratory
data. Am. J. Clin. Nutr. 67:166–181. 1998.
Acknowledgements
We would like to thank Head Coach Barry Gorman and
especially the Pennsylvania State University soccer team for
their enthusiastic participation in this initial study while one of
the authors (WJK) was the Director of Research in the Center
for Sports Medicine. To his credit, Coach Gorman, used the
findings of this work as the basis for substantive alterations in
his subsequent season’s practice schedule, work, and scrimmage
demands, which helped to reduce the stress levels observed in
this initial study and demonstrated that sport science can be
used successfully by sport coaches to improve coaching practices
and influence success on the field. We also thank all of the
laboratory staff for their efforts and help in the collection of data
and laboratory work in this study.
Address correspondence to Professor Wiliiam J. Kraemer,
[email protected].