Content uploaded by Anders H Engebretsen
Author content
All content in this area was uploaded by Anders H Engebretsen
Content may be subject to copyright.
Although differences in study design and injury definitions
make a direct comparison between studies difficult,
13
the
incidence of injuries among adult male soccer players on
the elite level has been estimated to range between 25 and
35 per 1000 game-hours.
5,14,19,34
Thus, the injury risk is
considerable and high compared with most other team
sports.
19
Studies from the professional leagues in Europe
(Norway, Sweden, Iceland, Britain, Fédération
Internationale de Football Association [FIFA], and Union
of European Football Associations [UEFA]) agree that
injuries to the lower extremities constitute the biggest
problem.
3,5,14,18,19,33,34
The 4 dominating injury types in soc-
cer are sprains to the ankle and knee and strains to the
hamstring and groin. These account for more than 50% of
all injuries, and prevention programs for soccer should
therefore target these.
Prevention of Injuries Among
Male Soccer Players
A Prospective, Randomized Intervention Study Targeting
Players With Previous Injuries or Reduced Function
Anders H. Engebretsen,
*†
Grethe Myklebust,
†
PT, PhD, Ingar Holme,
†
PhD,
Lars Engebretsen,
†‡
MD PhD, and Roald Bahr,
†
MD, PhD
From
†
Oslo Sports Trauma Research Center, Norwegian School of Sports Sciences,
Oslo Norway, and the
‡
Department of Orthopaedic Surgery, Ulleval University Hospital,
University of Oslo, Oslo, Norway
Background: This study was conducted to investigate whether the most common injuries in soccer could be prevented, and to
determine if a simple questionnaire could identify players at increased risk.
Hypothesis: Introduction of targeted exercise programs to male soccer players with a history of previous injury or reduced func-
tion in the ankle, knee, hamstring, or groin will prevent injuries.
Study Design: Randomized controlled trial; Level of evidence, 2.
Methods: A total of 508 players representing 31 teams were included in the study. A questionnaire indicating previous injury
and/or reduced function as inclusion criteria was used to divide the players into high-risk (HR) (76%) and low-risk (LR) groups.
The HR players were randomized individually into an HR intervention group or HR control group.
Results: A total of 505 injuries were reported, sustained by 56% of the players. The total injury incidence was a mean of 3.2
(95% confidence interval [CI], 2.5-3.9) in the LR control group, 5.3 (95% CI, 4.6-6.0) in the HR control group (P = .0001 vs the
LR control group), and 4.9 (95% CI, 4.3-5.6) in the HR intervention group (P = .50 vs the HR control group). For the main out-
come measure, the sum of injuries to the ankle, knee, hamstring, and groin, there was also a significantly lower injury risk in the
LR control group compared with the 2 other groups, but no difference between the HR intervention group and the HR control
group. Compliance with the training programs in the HR intervention group was poor, with only 27.5% in the ankle group, 29.2%
in the knee group, 21.1% in the hamstring group, and 19.4% in the groin defined as having carried out the minimum recom-
mended training volume.
Conclusion: The players with a significantly increased risk of injury were able to be identified through the use of a questionnaire,
but player compliance with the training programs prescribed was low and any effect of the intervention on injury risk could not
be detected.
Keywords: football; injury prevention; ankle injuries; knee injuries; hamstring injuries; groin injuries; risk factors; randomized con-
trolled trial
1
*Address correspondence to Anders H. Engebretsen, Oslo Sports Trauma
Research Center, Norwegian School of Sports Science, PO Box 4014 Ullevål
Stadion, N-0806 Oslo, Norway (e-mail: Anders.Engebretsen@nih.no).
No potential conflict of interest declared.
The American Journal of Sports Medicine, Vol. X, No. X
DOI: 10.1177/0363546508314432
© 2008 American Orthopaedic Society for Sports Medicine
AJSM PreView, published on April 3, 2008 as doi:10.1177/0363546508314432
Copyright 2008 by the American Orthopaedic Society for Sports Medicine.
2 Engebretsen et al The American Journal of Sports Medicine
As early as 1983, Ekstrand et al
10
showed a significant
reduction in the overall number of soccer injuries through
a 7-part prevention program. The rate of the most common
types of soccer injuries, sprains and strains to ankles and
knees, was reduced significantly. However, in more than 20
years, only 9 more injury prevention studies have been
published in soccer, and only 5 of them among men at the
senior level.
4,7,8,31,32
Tropp et al
32
showed that a balance
training program or the use of orthoses resulted in signifi-
cantly fewer ankle sprains than for a control group. Later,
orthoses and proprioceptive training were proven useful to
prevent ankle and knee injuries, respectively.
8,31
Finally,
Askling et al
7
and Arnason et al
4
have recently observed a
reduction in hamstring strains among male players
through eccentric strength-training programs.
Although the incidence and pattern (injury type, local-
ization, and severity) of injuries in soccer have been
described in detail,
9,12,24,29
much less is known about their
risk factors. Therefore, we do not know which players should
be targeted, for instance through specific training programs.
The risk of injury seems to be influenced by age,
11,17,24,29
sex,
17,25
and level of play.
12,24
In addition, a history of previous
injury was shown to be a significant risk factor for ankle
sprains in soccer as early as 1985.
32
This was recently con-
firmed by Arnason et al
6
in the largest cohort study from elite
soccer to date, in which the main risk factors for the 4 main
injury types were previous injury and age. We therefore
hypothesized that players with a history of previous injury or
symptoms indicating reduced function would represent a
group with an increased injury risk, who should be targeted
with specific prevention programs addressing their reported
deficits. A previous injury could compromise joint function
through reduced mechanical instability or neuromuscular
control, or muscle function through scar tissue formation,
reduced strength, or more subtle changes in the length-
tension relationship.
One aim of this study was therefore to examine whether
we could identify players with an increased risk of injury
using a questionnaire focusing on history of previous
injury and joint/muscle function. We also wanted to exam-
ine if exercise programs targeting players with an
increased risk of injury could prevent the 4 most common
injury types in soccer, ankle and knee sprains and ham-
string and groin strains.
METHODS
Teams playing in the Norwegian 1st, 2nd, or the top of the 3rd
division that were geographically located in the proximity of
Oslo (n = 35 teams, 769 players) were invited to participate in
the study. The 8 3rd division teams included either won their
league or finished as first runners-up the previous season,
resulting in a relatively homogeneous group of teams, even if
they competed in 3 different divisions.
Three of the teams (n = 60 players) declined the invita-
tion to participate, 177 players did not show up for testing,
3 players did not speak Norwegian and could therefore not
complete the questionnaire, and 4 players were excluded
for other reasons (Figure 1). Hence, 244 of the players
invited could not be included. In addition, 1 team (n = 17
players) was later excluded because the physiotherapist
did not instruct the intervention group players nor record
Invited to the study
(n = 769) (35 teams)
Players participating in the
study (n = 525) (32 teams)
Declined invitation
(n = 244)
High-risk group
(n = 401)
Individual randomization
within each team
Low-risk group
(n = 124)
High-risk criteria
-Injury previous 12 months
- Reduced function
High-risk intervention group
(HR intervention) (n = 193)
High-risk control group
(HR control) (n = 195)
Low-risk control group
(LR control) (n = 120)
Excluded (n = 17)
n = 4
n = 13
Figure 1. Flow chart showing movement of patients through the study.
Vol. X, No. X, XXXX Injury Prevention in Soccer 3
injuries, resulting in a final sample of 508 players repre-
senting 31 teams.
The teams were tested during the preseason (January
through March 2004) at the Norwegian School of Sports
Sciences. The players were asked to fill out a questionnaire in
5 parts. The first section consisted of general information
(date of birth, team, field position, and player experience). The
second through fifth sections included information on the
ankle, knee, hamstring, and groin, respectively. Each of these
sections covered the history of previous injuries (severity,
nature, and number of months since the most recent injury,
use of protective gear such as tape or brace, and if the injury
had caused the player to miss matches), and a function score
for each region. The questionnaires used to assess function
were the Foot and Ankle Outcome Score (FAOS) and Knee
Osteoarthritis Outcome Score (KOOS) score, which were
translated to Norwegian.
27,28
For the hamstring and groin,
we developed similar function scores, Hamstring Outcome
Score (HaOS) and Groin Outcome Score (GrOS), based on
the same principles as FAOS and KOOS, only specific to these
regions and their typical symptoms (see Appendix, available
in the online version of this article at http://ajsm.sagepub
.com/cgi/ content/full/X/X/X/DC1/).
Based on the questionnaire, the 508 players were
divided into 2 groups (Figure 1), a high-risk (HR) and a
low-risk (LR) group. The criteria for classifying a player as
having an assumed increased risk of injury were a history
of an acute injury to the ankle, knee, hamstring or groin
during the previous 12 months or a reduced function with
an average score of less than 80% for any of the body parts
mentioned. A player fulfilling any of the inclusion criteria
for any of the 4 body parts was assigned to the HR group.
The players in the HR group were randomized individu-
ally, but stratified within each team, into 2 groups, the HR
intervention group and the HR control group (Figure 1). In
this way, each team would normally have players from all
3 groups (HR intervention, HR control, and LR control).
The players in the HR intervention group were only
included on the basis of the inclusion criteria they fulfilled,
meaning that they only received a training program for the
body part(s) to which they were assumed to have an
increased risk of injury. In a situation in which a player
Figure 2. Example of an ankle exercise. The exercise was
prescribed to be performed with a straight leg and with a
gradual progression in difficulty (see Table 1).
Figure 3. Example of a knee exercise. The exercise was
prescribed to be performed in the knee-over-toe position and a
flexed knee, with gradual progression in difficulty (see Table 2).
Figure 4. Groin exercise. A variety of different exercises were
prescribed for strengthening of the groin muscles. In this
example, the player is pushing the legs together for 15 sec-
onds while keeping a ball between the knees (see Table 3).
4 Engebretsen et al The American Journal of Sports Medicine
ended up with 4 programs, the team physical therapist
was asked to merge the programs into 1 continuous pro-
gram. However, even if a player fulfilled the inclusion cri-
teria for 1 body part on only 1 side, he was asked to
perform the prevention exercises for both legs.
The players were asked to complete the ankle, knee, and
groin training programs (Tables 1 through 3, Figures 2
through 4) 3 times a week for 10 weeks during the presea-
son, in separate training sessions done in addition to the
regular team training. For the hamstring program (Table 4,
Figure 5), a 10-week progression was prescribed.
22
The inter-
vention players were also asked to perform the exercises
once per week for the rest of the season as maintenance. The
programs were meant to progress in difficulty, to challenge
the players as their performance improved. The players
were also asked to report all exercises they performed on a
form, checking a box if they had carried out the preventive
training that day. The form covered all 10 weeks for com-
pliance assessment.
Most of the teams from the 1st and 2nd divisions already
had a physical therapist working with the team. When
there was no physical therapist attached to the team, we
provided them with one. Each physical therapist was
rewarded with a stipend (10 000 NOK, or approximately
1900 USD). In addition to reporting injuries throughout
the preseason and season, the physical therapist was
TABLE 1
The Ankle Exercise Program
32,a
Weeks 1-2
Balance board Both legs on the board, with arms crossed. Attempt to stand still and maintain the balance.
Similar exercise, but now performed on 1 leg.
Both legs on the board, bouncing a ball alternately with both hands, standing as still as possible during the exercise.
Both legs on the board, throwing the ball and catching it.
Balance pad One leg on the pad, maintaining balance for 30 seconds on alternating legs.
Jumping exercise—from outside the pad, landing on alternating legs.
Weeks 3-5
Balance board Ball juggling performed while standing on 1 leg.
Balance pad Bouncing the ball around the pad while standing on 1 leg.
Calf raise while standing on both legs on the pad.
Weeks 6-10
Balance board Soccer-specific exercises, juggling the ball while standing on 1 leg, also combining both the balance board and
balance pad, placing the pad on top of the board.
Balance pad Closing the eyes while standing on 1 leg, and other exercises including landing on 1 or 2 legs while jumping
from a box/stairs.
a
All exercises were prescribed to be performed with a straight leg (no knee flexion) (Figure 2) and with a gradual progression in difficulty.
The players were instructed to switch between the balance board and pad, and, as they became more proficient, to include ball-based exer-
cises while keeping their balance.
TABLE 2
The Knee Exercise Program
8,23,a
Weeks 1-2
Balance board Both legs on the board, with arms crossed, always keeping the knee-over-toe position.
Similar exercise, but now performed on 1 leg.
Both legs on the board, bouncing a ball alternately with both hands, standing as still as possible during the exercise.
Both legs on the board, throwing the ball and catching it.
Balance pad One leg on the pad, maintaining balance for 30 seconds on alternating legs.
Walk onto the pad, stopping and keeping the balance.
Jumping exercise—from outside the pad, landing on alternating legs.
Weeks 3-5
Balance board Ball juggling performed while standing on 1 leg.
Two-legged squats, with knee-over-toe position.
Balance pad Bouncing the ball around the pad while standing on 1 leg.
Weeks 6-10
Balance board Soccer-specific exercises, juggling the ball while standing on 1 leg, also combining both the balance board and
balance pad, placing the pad on top of the board.
Balance pad Closing the eyes while standing on 1 leg, and other exercises including landing on 1 or 2 legs while jumping
from a box/stairs.
One-legged squats, and balance exercises while closing the eyes.
Floor exercise One-legged jumping on 1 foot in an imaginary zig-zag course.
a
All exercises were prescribed to be performed with the knee-over-toe position and a flexed knee (Figure 3) with gradual progression in
difficulty. As with the ankle program, the players were instructed to switch between the balance board and pad, and include ball-based exer-
cises as they progressed.
Vol. X, No. X, XXXX Injury Prevention in Soccer 5
responsible for instructing all the players who were ran-
domized into the HR intervention group in their training
programs. Each player was given a folder describing the
exercises he was asked to do, as well as any necessary
equipment such as balance mats and balance boards.
An injury was defined as any physical complaint sus-
tained by a player that resulted from a soccer match or soc-
cer training, resulting in a player being unable to take a full
part in future soccer training or match play (“time-loss”
injury). Acute injuries were defined as injuries with a sud-
den onset associated with a known trauma, whereas over-
use injuries were those with a gradual onset without any
known trauma. Injuries were classified into 3 severity cate-
gories according to the time it took until the player was fully
fit to take part in all types of organized soccer play: minor
(1-7 days), moderate (8-28 days), and major (>28 days).
Match exposure was defined as play between teams from
different clubs, while training exposure was defined as
team-based and individual physical activities under the
control or guidance of the team coaching or fitness staff
aimed at maintaining or improving soccer skills or physi-
cal condition. All injuries were categorized by the authors
based on the injury reports from each physiotherapist.
The main outcome measure was the sum of the risk for an
ankle sprain, knee sprain, groin strain, or hamstring strain.
Statistical Methods
Exposure was calculated in hours as the sum of all indi-
vidual exposures recorded during training and match play
TABLE 3
The Groin Exercises
16,a
Warm-up Keeping a ball between the extended legs, pushing the legs together for 15 s, while lying on the ground. Repeated 10×.
Similar exercise, only difference having the knees flexed and the ball between the knees.
Transverse abdominal Lie facing the ground, only resting on the forearms and toes in a straight position, contracting the abdominal
muscles muscles, “forcing the umbilicus inwards.” Performed in 20 s, repeated 5×.
Sideways jumping Knee-over-toe position while jumping sideways with arms resting on the hips.
Sliding Wearing only socks, slide a leg alternately away and towards the other that is bearing the weight. The exercise
can be performed both sideways and diagonally for 30-60 s before switching legs.
Diagonal walking Exercise described by Holmich et al
16
performed 5 × 15 s on each leg.
a
The players were instructed to perform the exercise 3 times a week for approximately 15 minutes. A ball was needed for some of the exer-
cises (Figure 4), and the exercises could be performed without warming up.
TABLE 4
The Hamstring Exercise Program
a
No. of Training No. of
Week Sessions Per Week Repetitions
115 + 5
226 + 6
333 × 6-8
433 × 8-10
5-10 3 12 + 10 + 8
a
The Nordic hamstring exercise is performed standing on the
knees on a soft foundation, slowly lowering the body toward the
ground using the hamstrings while the feet are held by a partner
(Figure 5). Progression is achieved by increasing the initial speed,
and eventually having a partner push forward.
Reproduced with permission from Mjolsnes R, Arnason A, Osthagen T,
Raastad T, Bahr R. A 10-week randomized trial comparing eccentric
vs. concentric hamstring strength training in well-trained soccer
players. Scand J Med Sci Sports. 2004;14:311-317.
Figure 5. The hamstring exercise program. (A) start position.
The player stands on his knees on a soft foundation with the
feet being held by a partner. (B) slowly lowering the body
toward the ground using the hamstrings while focusing on
keeping the body straight (see Table 4).
6 Engebretsen et al The American Journal of Sports Medicine
during the season. The injury rate was compared between
the HR control group and the HR intervention group, and
the HR control group and the LR control group, respec-
tively, using a z test, reporting 95% confidence intervals
(CIs) based on the Poisson model. Chi-square tests were
used to compare the proportion of injured players between
the HR intervention group and the HR control group, and
between the HR control group and the LR control group,
respectively. Otherwise, results are presented as the
means with standard deviations.
RESULTS
Screening and Randomization
Of the 508 players included in the study, 388 (76%) were
assumed to have an increased risk for 1 or more injury
types based on their history of previous injury and/or func-
tion scores. Of these, 195 players were randomized to the
HR control group and 193 players to the HR intervention
group (Figure 1). In the intervention group, 2 players were
asked to perform all of the 4 training programs; 22 players,
3 programs; 62 players, 2 programs; and 107 players, 1 pro-
gram. Of the 305 training programs prescribed, 102 were
for the ankle, 65 for the knee, 76 for the hamstring, and 62
for the groin.
Player Exposure
The total exposure to match play and training was 108 111
player hours (Table 5), and there was no difference in
mean player exposure between the HR intervention group
(217 ± 94 hours), the HR control group (210 ± 103 hours),
and the LR control group (211 ± 88 hours).
Injuries
A total of 505 injuries were reported (Table 5), sustained
by 283 (56%) of the 508 players included in the study. In
the LR control group, there were 82 injuries, while there
were 216 injuries in the HR control group and 207 injuries
in the HR intervention group. There was no difference in
the incidence of injuries between the HR intervention
group and the HR control group (relative risk [RR], 0.94;
95% CI, 0.77-1.13), while the incidence was lower in the
LR control group than both other groups (RR, 0.65 vs
the HR intervention group; 95% CI, 0.51-0.85; RR, 0.61 vs
the HR control group; 95% CI, 0.48-0.79).
During the season, 45.8% of the players in the LR control
group (55 of 120 players) sustained 1 or more injuries, com-
pared with 58.5% in the HR control group (114 of 195 players;
P = .029 vs the LR control group; χ
2
test) and 59.1% in the HR
intervention group (114 of 193 players; P = .90 vs the HR con-
trol group).
There was no difference in injury severity among the 3
groups (Table 6).
Intervention Outcome: Intention-to-Treat Analysis
For the main outcome measure, the sum of injuries to the
ankle, knee, hamstrings, and groin, the total incidence was
2.3 injuries per 1000 playing hours (95% CI, 2.1-2.6). The
corresponding figures were 1.3 (95% CI, 0.9-1.8) for the LR
control group, 2.8 (95% CI, 2.3-3.3) in the HR control
group, and 2.6 (95% CI, 2.1-3.0) in the HR intervention
group. There was a significantly lower injury risk in the LR
TABLE 5
Match, Training, and Total Exposure, Number of Injuries, and Injury Incidence
(With 95% Confidence Intervals) for the 3 Groups Throughout the Season
a
Training Match Total
Exposure Incidence Exposure Incidence Exposure Incidence
(h) Injuries (per 1000 h) (h) Injuries (per 1000 h) (hours) Injuries (per 1000 h)
HR intervention 34 422 100 2.9 (2.3-3.5) 7434 93 12.5 (10.0-15.1) 41 856 207 4.9 (4.3-5.6)
(n = 193)
HR control 33 757 103 3.1 (2.5-3.6) 7156 100 14.0 (11.2-16.7) 40 913 216 5.3 (4.6-6.0)
(n = 195)
LR control 20 925 40 1.9 (1.3-2.5) 4417 34 7.7 (5.1-10.3) 25 342 82 3.2 (2.5-3.9)
(n = 120)
Total 89 103 243 2.7 (2.4-3.1) 19 008 227 11.9 (10.4-13.5) 108 111 505 4.7 (4.3-5.1)
a
HR, high risk; LR, low risk.
TABLE 6
Injury Type and Injury Severity
(Based on Time Loss) in the 3 Groups
a
HR Intervention HR Control LR Control
Group Group Group
(n = 193) (n = 195) (n = 120)
Injury type
Acute (%) 143 (41) 153 (43) 57 (16)
Overuse (%) 62 (42) 61 (41) 25 (17)
Other (%) 2 (50) 2 (50) 0 (0)
Time loss
1-3 days (%) 37 (7) 54 (11) 24 (5)
4-7 days (%) 47 (9) 42 (8) 13 (3)
1-4 weeks (%) 81 (16) 66 (13) 27 (5)
>4 weeks (%) 30 (6) 40 (8) 12 (2)
Not specified (%) 12 (2) 14 (3) 6 (1)
a
Percentages are shown within each group. HR, high risk; LR,
low risk.
Vol. X, No. X, XXXX Injury Prevention in Soccer 7
control group compared with the 2 other groups (RR, 0.49;
95% CI, 0.33-0.71 vs the HR control group; RR, 0.53; 95%
CI, 0.36-0.77 vs the HR intervention group). However, no
difference was seen between the HR intervention group
and the HR control group (RR, 0.93; 95% CI, 0.71-1.21).
When the players in the HR intervention and HR control
groups with increased risk of injury were compared, we
found no significant differences in the risk of injury to the
body part in question between the 2 groups for any cate-
gory (ankle: RR, 0.64; 95% CI, 0.32-1.29; knee: RR, 0.96;
95% CI, 0.35-2.64; hamstrings: RR, 1.55; 95% CI, 0.83-2.90;
and groin: RR, 1.18; 95% CI, 0.55-2.54) (Table 7).
Compliance With the Training Program
and Per-Protocol Analysis
Compliance with the training programs in the HR inter-
vention group was poor, with only 27.5% (28 players) in the
ankle group and 29.2% (19 players) in the knee group hav-
ing completed 30 or more training sessions. For the ham-
string and groin exercises, compliance was even less, with
only 21.1% (12 players) and 19.4% (16 players) completing
20 or more training sessions, respectively. Hence, the com-
pliant (more than 30 exercises for ankle and knee, and
more than 20 training sessions for hamstring and groin)
groups are small. As many as 15.7% (16 players) reported
not having done any ankle exercises; 11.8% (12 players), 1
to 9 exercise sessions; and 24.5% (25 players), 10 to 19 ses-
sions; while 20.6% (21 players) reported having carried out
20 or more sessions, but less than the target number of 30.
The corresponding figures for knee exercises were 23.1%
(0 exercise sessions reported), 9.2% (1-9 sessions), 13.8% (10-
19 sessions), and 24.6% (20-29 sessions). For hamstring exer-
cises, the figures were 63.2% (0 exercise sessions reported),
7.9% (1-9 sessions), and 7.9% (10-19 sessions); and for groin
exercises, 67.7% (0 exercise sessions reported), 4.8% (1-9 ses-
sions), and 8.1% (10-19 sessions).
In a per-protocol analysis on ankle injuries, the inci-
dence of ankle injuries in the compliant group, who sus-
tained 3 injuries (2 of 28 injured players), was 0.5 (95% CI,
–0.1 to 1.0) injuries per 1000 hours, compared with 0.9
(95% CI, 0.5-1.3) injuries per 1000 hours among players
with an increased risk of ankle injuries in the HR control
group (RR = 0.51; 95% CI, 0.2-1.7). Similarly, we could not
detect any difference in the risk of knee injury between
players in the HR intervention group who were compliant
with the knee program (0.2 [95% CI, –0.2 to 0.7] injuries
per 1000 hours) and the HR players in the HR control
group (0.5 [95% CI, 0.2-0.9] injuries per 1000 hours; RR =
0.46; 95% CI, 0.1-3.7). In the same way, no difference was
observed in the incidence of hamstring (RR = 0.94; 95% CI,
0.3-3.2) and groin injuries (RR = 1.6; 95% CI, 0.5-5.6)
between players in the HR intervention group who were
compliant with the respective training programs and the
HR control group.
DISCUSSION
The main finding of this study was that, although we were
able to identify players with an increased injury risk through
a comprehensive questionnaire, there was no effect of the tar-
geted intervention on injury risk. The most likely explanation
for this is the low compliance with the exercise programs.
With such low compliance in the intervention group (ranging
from 20% to 30% for the different exercise programs), no effect
could be expected on injury rate.
In contrast to most previous intervention studies, we
chose to randomize players individually to the intervention
or control group. We relied on the team physical therapists
to instruct the players in the intervention program. However,
to avoid contamination, the players were asked to do the exer-
cises outside the regular team training sessions—before or
after training or at home. The low overall compliance in
the intervention group indicates that significant contami-
nation between groups is unlikely to have occurred. As
seen in previous studies, the main challenge is getting
players in the intervention group to follow preventive
training programs, not keeping other players from train-
ing.
23
However, a potentially bigger risk of contamination
is the fact that 19 of the 31 teams did team-based preven-
tive exercises similar to ours regularly throughout the pre-
season, and 16 of these reported good training regimens. We
could not, of course, keep the teams from carrying out their
normal preventive exercises. Although these team-based
exercises were done by players in both groups, the fact that
players from the control group trained with exercises similar
to our intervention exercises does, of course, reduce the poten-
tial of showing a positive preventive effect of our intervention
TABLE 7
Intention-to-Treat Analysis for Ankle, Knee, Hamstring, and Groin Injuries
a
HR Intervention Group HR Control Group
Players at Injury Players at Injury P Value
Increased Injured Incidence Increased Injured Incidence (Control vs
Risk Injuries Players (95% CI) Risk Injuries Players (95% CI) Intervention)
Ankle 102 13 10 (10%) 0.6 (0.3-0.9) 107 20 14 (13%) 0.9 (0.5-1.3) .21
Knee 65 7 6 (9%) 0.5 (0.1-0.9) 66 8 7 (11%) 0.5 (0.2-0.9) .93
Hamstring 85 23 17 (20%) 1.5 (0.9-2.0) 76 17 14 (18%) 0.9 (0.5-1.4) .17
Groin 62 11 10 (16%) 0.9 (0.4-1.4) 98 16 13 (13%) 0.7 (0.4-1.1) .67
a
The number of players with an increased risk of injury to the different body parts (ankle, knee, hamstring and groin) within the two high
risk groups and the number of injuries that occurred to the same body part among these players. HR, high risk; LR, low risk; CI, confidence
interval.
8 Engebretsen et al The American Journal of Sports Medicine
and represents a limitation in this study. Moreover, exer-
cises carried out by each player on his own are probably
not as effective as when they are carried out under quali-
fied supervision,
30
not just because of a lower compliance,
but also because the quality of the exercises performed can
not be ensured in the same way. Potentially negative fac-
tors such as initial muscle soreness or eventual boredom
could possibly be overcome more effectively in a group
training session with a qualified instructor.
Because the compliance was low, the statistical power is
also too low to assess the effect of the training programs in
the subgroups that did follow the training protocol (ie,
through the per-protocol analyses). The 4 programs used
were selected either because there was evidence from pre-
vious prevention studies to indicate that they are effective
or because they have been shown to be effective as reha-
bilitation exercises after injury. To prevent ankle and knee
injuries, various forms of balance-training exercises have
been shown to be effective in other study popula-
tions.
8,10,15,20,21,26,31,32
Strain injuries to the hamstrings have
been effectively prevented through eccentric strength train-
ing, such as that used in the present study.
4,7,16,22
A program of
strength training and core stability exercises has been shown
to be highly effective in the treatment of long-standing groin
pain in a population consisting mainly of soccer players.
16
This program formed the basis for the present program, but
because we thought it would be unrealistic to implement the
entire groin program, we prescribed an abbreviated 10- to
15-minute session to increase compliance. In other words,
each of the program components were based on studies indi-
cating their effect in prevention or rehabilitation of the 4
main injury types. However, previous injury as a risk factor
is not fully understood; it may be that ankles and knees are
not fully restored structurally or functionally. Although the
injury prevention literature supports several different exer-
cises, there is limited evidence that reinjuries can be prevented
through the same exercises. We do not know which exercises
should be chosen to prevent reinjuries and which have poten-
tial for primary prevention of the same injury types.
It is possible that the compliant players may have bene-
fited from the programs if they had carried out more ses-
sions. We know that a certain minimum of exercise must
be performed before an effect may be expected.
23
For the
purposes of data analysis, we suggested that at least 30
exercises (20 for hamstring and groin) needed to be carried
out. However, this number is arbitrary, as there is no evi-
dence on the dose-effect relationship for any exercise pro-
gram to prevent injuries.
Although the intervention was ineffective, this study
demonstrates that the players who had the most to gain from
preventive exercises could be identified. The risk of injury was
approximately twice as high among athletes with a history of
previous injury and/or reported reduced function. This identi-
fication was achieved through the use of a simple question-
naire only, and the addition of more elaborate functional tests
did not increase the predictive value of the screening (data not
shown). The rationale for the approach used, employing a self-
completed questionnaire, provided the potential for expand-
ing the range of application of the athlete screening process.
The questionnaire represents a cost-effective means of
player screening, which could also be done using Web-
based solutions. In this way, teams and players with no
medical staff could do a self-test in the preseason to find
out whether they have an increased risk of injuries.
The increased risk associated with a history of previous
injury and reduced function also has other implications. One is
that preventing the first injury should be a high priority, to
keep players from entering the vicious cycle of repeated
injuries to the same body part. This cannot be achieved only at
the team level; more research is needed and effective injury
prevention may also involve changes to the rules of the game
and more specific training of referees.
1,2
The most likely expla-
nation for previous injury being such a consistent risk factor
for reinjuries is that the joints or muscles in question are not
fully restored structurally and/or functionally. Based on this, it
seems reasonable to suggest that one thing teams can do, even
at lower levels of play, is to focus on improving rehabilitation
after injury and implementing adequate return-to-play guide-
lines. Players with reduced function after previous injury
should undergo a structured rehabilitation program until full
function is established. However, it remains to be proven
whether this would reduce injury risk significantly. The pres-
ent study also shows that this cannot be left to the players
themselves; adequate supervision is necessary.
One limitation of this study is the difference in physical
therapist contact between the HR intervention players and
the other groups. To instruct in the intervention exercises, the
physical therapist became well acquainted with each of the
intervention players, and not always to the same extent
with the other players on the team. Thus, there was a
potential for a bias in injury reporting, as the same physi-
cal therapist also was responsible for reporting injuries.
CONCLUSION
We were able to identify the players with an increased risk
of injury through a questionnaire on previous injuries and
joint and muscle function only. However, the introduction
of individual specific preventive training programs did not
affect the injury risk in this intervention, most likely due
to a low compliance with the training programs prescribed.
ACKNOWLEDGMENT
The Oslo Sports Trauma Research Center has been estab-
lished through generous grants from the Eastern Norway
Regional Health Authority, the Royal Norwegian Ministry
of Culture, the Norwegian Olympic Committee and
Confederation of Sport, and Norsk Tipping AS. We thank
all personnel for making this study possible.
REFERENCES
1. Andersen TE, Arnason A, Engebretsen L, Bahr R. Mechanisms of
head injuries in elite football. Br J Sports Med. 2004;38:690-696.
2. Andersen TE, Floerenes TW, Arnason A, Bahr R. Video analysis of the
mechanisms for ankle injuries in football. Am J Sports Med.
2004;32:69S-79S.
3. Andersen TE, Tenga A, Engebretsen L, Bahr R. Video analysis of
injuries and incidents in Norwegian professional football. Br J Sports
Med. 2004;38:626-631.
Vol. X, No. X, XXXX Injury Prevention in Soccer 9
4. Arnason A, Andersen TE, Holme I, Engebretsen L, Bahr R. Prevention
of hamstring strains in elite soccer: an intervention study. Scand J
Med Sci Sports. 2007 Mar 12 [Epub ahead of print].
5. Arnason A, Gudmundsson A, Dahl HA, Johannsson E. Soccer injuries
in Iceland. Scand J Med Sci Sports. 1996;6:40-45.
6. Arnason A, Sigurdsson SB, Gudmundsson A, Holme I, Engebretsen L,
Bahr R. Risk factors for injuries in football. Am J Sports Med. 2004;
32:5S-16S.
7. Askling C, Karlsson J, Thorstensson A. Hamstring injury occurrence
in elite soccer players after preseason strength training with eccentric
overload. Scand J Med Sci Sports. 2003;13:244-250.
8. Caraffa A, Cerulli G, Projetti M, Aisa G, Rizzo A. Prevention of anterior cru-
ciate ligament injuries in soccer. A prospective controlled study of propri-
oceptive training. Knee Surg Sports Traumatol Arthrosc. 1996;4:19-21.
9. Dvorak J, Junge A. Football injuries and physical symptoms: a review
of the literature. Am J Sports Med. 2000;28:S3-S9.
10. Ekstrand J, Gillquist J, Liljedahl SO. Prevention of soccer injuries: supervi-
sion by doctor and physiotherapist. Am J Sports Med. 1983;11:116-120.
11. Ekstrand J, Gillquist J, Moller M, Oberg B, Liljedahl SO. Incidence of
soccer injuries and their relation to training and team success. Am J
Sports Med. 1983;11:63-67.
12. Ekstrand J, Tropp H. The incidence of ankle sprains in soccer. Foot
Ankle. 1990;11:41-44.
13. Fuller CW, Ekstrand J, Junge A, et al. Consensus statement on injury
definitions and data collection procedures in studies of football (soc-
cer) injuries. Clin J Sport Med. 2006;16:97-106.
14. Hawkins RD, Fuller CW. A prospective epidemiological study of
injuries in four English professional football clubs. Br J Sports Med.
1999;33:196-203.
15. Heidt RS Jr, Sweeterman LM, Carlonas RL, Traub JA, Tekulve FX.
Avoidance of soccer injuries with preseason conditioning. Am J
Sports Med. 2000;28:659-662.
16. Holmich P, Uhrskou P, Ulnits L et al. Effectiveness of active physical
training as treatment for long-standing adductor-related groin pain in
athletes: randomised trial. Lancet. 1999;353:439-443.
17. Inklaar H. Soccer injuries. I: Incidence and severity. Sports Med. 1994;
18:55-73.
18. Juma AH. Outline of sport injuries in the V World Youth Championship
for FIFA Cup in Saudi Arabia. Indian J Med Sci. 1998;52:433-437.
19. Junge A, Dvorak J, Graf-Baumann T, Peterson L. Football injuries dur-
ing FIFA tournaments and the Olympic Games, 1998-2001: develop-
ment and implementation of an injury-reporting system. Am J Sports
Med. 2004;32:80S-89S.
20. Junge A, Rosch D, Peterson L, Graf-Baumann T, Dvorak J. Prevention
of soccer injuries: a prospective intervention study in youth amateur
players. Am J Sports Med. 2002;30:652-659.
21. Mandelbaum BR, Silvers HJ, Watanabe DS, et al. Effectiveness of a
neuromuscular and proprioceptive training program in preventing
anterior cruciate ligament injuries in female athletes: 2-year follow-up.
Am J Sports Med. 2005;33:1003-1010.
22. Mjolsnes R, Arnason A, Osthagen T, Raastad T, Bahr R. A 10-week
randomized trial comparing eccentric vs. concentric hamstring
strength training in well-trained soccer players. Scand J Med Sci
Sports. 2004;14:311-317.
23. Myklebust G, Engebretsen L, Braekken IH, Skjolberg A, Olsen OE,
Bahr R. Prevention of anterior cruciate ligament injuries in female
team handball players: a prospective intervention study over three
seasons. Clin J Sport Med. 2003;13:71-78.
24. Nielsen AB, Yde J. Epidemiology and traumatology of injuries in soc-
cer. Am J Sports Med. 1989;17:803-807.
25. Nilsson S, Roaas A. Soccer injuries in adolescents. Am J Sports Med.
1978;6:358-361.
26. Olsen OE, Myklebust G, Engebretsen L, Holme I, Bahr R. Exercises to
prevent lower limb injuries in youth sports: cluster randomised con-
trolled trial. BMJ. 2005;330:449.
27. Roos EM, Brandsson S, Karlsson J. Validation of the foot and ankle
outcome score for ankle ligament reconstruction. Foot Ankle Int.
2001;22:788-794.
28. Roos EM, Roos HP, Lohmander LS, Ekdahl C, Beynnon BD. Knee
Injury and Osteoarthritis Outcome Score (KOOS)—development of a
self-administered outcome measure. J Orthop Sports Phys Ther.
1998;28:88-96.
29. Schmidt-Olsen S, Jorgensen U, Kaalund S, Sorensen J. Injuries
among young soccer players. Am J Sports Med. 1991;19:
273-275.
30. Soderman K, Werner S, Pietila T, Engstrom, Alfredson H. Balance
board training: prevention of traumatic injuries of the lower
extremities in female soccer players? A prospective randomized
intervention study. Knee Surg Sports Traumatol Arthrosc. 2000;8:
356-363.
31. Surve I, Schwellnus MP, Noakes T, Lombard C. A fivefold reduction in
the incidence of recurrent ankle sprains in soccer players using the
Sport-Stirrup orthosis. Am J Sports Med. 1994;22:601-606.
32. Tropp H, Askling C, Gillquist J. Prevention of ankle sprains. Am J
Sports Med. 1985;13:259-262.
33. Walden M, Hagglund M, Ekstrand J. Injuries in Swedish elite
football—a prospective study on injury definitions, risk for injury
and injury pattern during 2001. Scand J Med Sci Sports.
2005;15:118-125.
34. Walden M, Hagglund M, Ekstrand J. UEFA Champions League study:
a prospective study of injuries in professional football during the
2001-2002 season. Br J Sports Med. 2005;39:542-546.
