doi:10.1136/bjsm.37.3.233
Br. J. Sports Med. 2003;37;233-238
C Woods, R Hawkins, M Hulse and A Hodson
football: an analysis of ankle sprains
Programme: an audit of injuries in professional
The Football Association Medical Research
http://bjsm.bmjjournals.com/cgi/content/full/37/3/233
Updated information and services can be found at:
These include:
References
http://bjsm.bmjjournals.com/cgi/ ... 3/233#otherarticles
6 online articles that cite this article can be accessed at:
http://bjsm.bmjjournals.com/cgi/content/full/37/3/233#BIBL
This article cites 41 articles, 13 of which can be accessed free at:
Rapid responses
http://bjsm.bmjjournals.com/cgi/eletter-submit/37/3/233
You can respond to this article at:
service
Email alerting
top right corner of the article
Receive free email alerts when new articles cite this article - sign up in the box at the
Topic collections
• Sports Medicine (1329 articles)
• Physiotherapy (148 articles)
• Injury (833 articles)
Articles on similar topics can be found in the following collections
Notes
http://www.bmjjournals.com/cgi/reprintform
To order reprints of this article go to:
http://www.bmjjournals.com/subscriptions/
To subscribe to British Journal of Sports Medicine go to:
Downloaded from bjsm.bmjjournals.com on 22 October 2006
ORIGINAL ARTICLE
The Football Association Medical Research Programme:
an audit of injuries in professional football: an analysis
of ankle sprains
C Woods, R Hawkins, M Hulse, A Hodson
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Br J Sports Med 2003;37:233–238
Aim: To conduct a detailed analysis of ankle sprains sustained in English professional football over two
competitive seasons.
Methods: Club medical staff at 91 professional football clubs annotated player injuries. A specific
injury audit questionnaire was used together with a weekly form that documented each club’s current
injury status.
Results: Completed injury records for the two competitive seasons were obtained from 87% and 76%
of the participating clubs. Ankle ligament sprains accounted for 11% of the total injuries over the two
seasons, with over three quarters (77%) of sprains involving the lateral ligament complex. A total of
12 138 days and 2033 matches were missed because of ankle sprains. More sprains were caused by
contact mechanisms than non-contact mechanisms (59% v 39%) except in goalkeepers who sustained
more non-contact sprains (21% v 79%, p<0.01). Ankle sprains were most often observed during tackles
(54%). More ankle sprains were sustained in matches than in training (66% v 33%), with nearly half
(48%) observed during the last third of each half of matches. A total of 44% of sprains occurred during
the first three months of the season. A high number of players (32%) who sustained ankle sprains
were wearing some form of external support. The recurrence rate for ankle sprains was 9% (see methodology
for definition of reinjury).
Conclusion: Ankle ligament sprains are common in football usually involving the lateral ligament complex.
The high rate of occurrence and recurrence indicates that prevention is of paramount importance.
Ankle sprains (especially those involving the lateral
ligament complex) have often been reported as the
most common injuries in sport.1–6 It has been suggested
that such injuries are usually sustained in sports involving
running,2 cutting,2 jumping,2 7 and contact with other
players,8 9 and this partly explains the high incidence of ankle
sprains in football.10–12 Ankle sprains in this population have
been reported to have a high recurrence rate.11 13–15
The findings of the initial Football Association Audit of
Injuries study were consistent with these findings.16 Over two
seasons, the authors observed that 17% of all injuries were to
the ankle, the same figure being reported by Ekstrand and
Gillquist.11 Ekstrand and Tropp13 found that ankle sprains
comprised 19% of all injuries. Sandelin et al17 observed that
75% of ankle injuries were ligament sprains (mostly lateral
ligament complex), whereas Hawkins et al16 reported this figure
to be 67% (80% being to the lateral ligament complex).
Hawkins et al16 found that a total of 76% of ligament sprains
that recurred during the same season were to the ankle. Given
the high incidence of ankle sprains, the authors suggested
that prevention and rehabilitation of ligament sprains
warranted further investigation.
As a follow up to the initial study, the aim of this study was
to undertake a detailed analysis of the data on ankle sprains.
Information on incidence, time lost, mechanism of injury, use
of external support, and timing of ankle sprains could help to
suggest the best methods of preventing and rehabilitating
such injuries.
METHODS
Player injuries were prospectively reported from July 1997
through to the end of May 1999 inclusive. A total of 91 of the
92 football clubs from the English football leagues (Premier
and Football League) committed themselves to the project.
Injuries were recorded by club physiotherapists and/or doctors
on a specific player injury audit questionnaire designed for
this study. Injury audit questionnaires for players who had
returned to full training/competition during a particular week
were returned weekly together, with a form indicating which
players had been absent and the number of days and competitive
matches each had missed that week. Before the study,
medical staff from clubs attended a briefing day and were
issued with guidance notes on how to complete the questionnaires.
Only professional players with a squad number were
involved in the study. Participants were asked to complete a
consent form, and each club provided details of their squad at
the beginning of each season. Table 1 presents the information
obtained. New players who joined the club were included, and
players leaving clubs were omitted from the study if they did
not stay within one of the four English leagues.
A recordable injury was defined as one sustained during
training or competition and which prevented the injured
player from participating in normal training or competition
for more than 48 hours (not including the day of the injury).
Injuries unrelated to football were not included, nor was any
absence resulting from illness. Injuries acquired during international
duty were included because details of such injuries
were generally reported back to club medical staff. The severity
of each injury was defined as slight, minor, moderate, or
major depending on whether the player was absent from
training or competition for two to three days, four to seven
days, one to four weeks, or more than four weeks, respectively.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Abbreviations: ATFL, anterior talofibular ligament; CFL, calcaneofibular
ligament
See end of article for
authors’ affiliations
. . . . . . . . . . . . . . . . . . . . . . .
Correspondence to:
Caroline Woods, Lilleshall
National Sports Centre, Nr
Newport, Shropshire
TF10 9AT, UK;
The FA.com
caroline.woods@TheFA.com
Accepted 23 August 2002
. . . . . . . . . . . . . . . . . . . . . . .
233
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006
Reinjury was defined as an injury of the same nature and
location involving the same player in the same season. The
dominant foot was defined as the predominant foot used for
kicking a ball.
Data were analysed using SPSS (Chicago, Illinois, USA).
Descriptive and comparative data are presented. The c2
significance test was used to investigate differences, and
significance was accepted at p<0.05 level. All players agreed to
participate in the study, and there were no drop outs during
the study period.
RESULTS
Of the 91 clubs starting the study, completed injury records for
the entirety of the 1997/1998 and 1998/1999 competitive seasons
were attained from 87% and 76% respectively. During the
study period, 1011 ankle injuries were documented, comprising
17% of the 6030 total number of injuries sustained over the
two seasons
Table 2 displays the nature of all ankle injuries. Ankle ligament
injuries (sprains) accounted for 11% of the total injuries
sustained over the two seasons. There was no significant
difference between the incidence of dominant and nondominant
ankle sprains based on expected values (56% v
42%). No significant differences in the incidence of ankle
sprains between Premier, 1st, 2nd, and 3rd divisions were
observed.
Table 3 shows the medical classification of ankle sprains.
Most involved injury to some portion of the lateral ligament
complex, that is the anterior talofibular, calcaneofibular, and
posterior talofibular ligaments (77%).
Table 4 shows the diagnostic investigations performed on
ankle sprains. Only six players underwent some form of
surgery, and 19 players had injections.
One third of ankle sprains were sustained during training
and two thirds during matches; there was no significant
difference between the observed and expected incidence of
ankle sprains based on the percentage of total match and
training injuries reported. Player to player contact was
responsible for 59% of injuries, and 39% were non-contact
injuries. Tackling (36%) and being tackled (18%) were the
most common mechanisms of sustaining an ankle sprain.
Figure 1 displays the non-contact mechanisms of ankle
sprains: 77% of non-contact sprains were caused during landing,
twisting and turning, and running. Ankle sprains in goalkeepers
were the result of significantly more non-contact
mechanisms of injury than contact mechanisms (79% v 21%,
p<0.01). The most common mechanisms of injury for this
position were landing (36%), twisting/turning (21%), and diving
(10%).
The total number of days that players were absent over the
two seasons was 12 138, and a total of 2033 matches were
missed. A total of 83% of the ankle sprains required players to
miss one month or less.
Figure 2 shows the timing of match injuries. A total of 48%
of injuries were sustained during the last third of the first and
second halves of the match. There was no significant
difference between the number of ankle sprains sustained in
the first and second halves of matches. There was no
significant difference between the timing of contact and noncontact
ankle sprains during matches or training.
Figure 3 shows the number of ankle sprains during each
month of the season. During the first three months of the season,
44% of ankle injuries were sustained (p<0.01).
Table 1 Division, playing position, and age
distribution of the cohort at the beginning of the study
No %
Division
Premier 618 26
1st 712 30
2nd 550 23
3rd 496 21
Total* 2376 100
Playing position
Goalkeeper 223 9
Defender 817 34
Midfielder 739 31
Forward 597 25
Total* 2376 99
Age distribution
17–22 970 41
23–28 817 34
29–34 508 21
35+ 81 3
Total* 2376 99
*Percentage totals may be subject to rounding errors associated with
individual components.
Table 2 Nature of ankle injuries
Nature No %
Sprain and rupture 677 67
Tissue bruising 79 8
Tendonitis and paratendonitis 65 6
Inflammatory synovitis 31 3
Fracture 25 3
Capsular tear 21 2
Strain 21 2
Other* 74 7
Not specified 18 2
Total 1011 100
*Other includes periostitis, dislocation, chondral lesion, muscular
contusion, tendon rupture, cut, overuse, and bursitis.
Table 3 Medical classification of ankle ligament
injuries
Name of ligament No %
Anterior talofibular 493 73
Medial 97 14
Unspecified 28 4
Anterior tibiofibular 23 3
Calcaneofibular 14 2
Posterior talofibular 13 2
Other* 5 1
Missing 4 1
Total 677 100
*Other includes interosseous membrane and posterior tibiofibular
ligament.
Table 4 Diagnostic investigation of ankle sprains
Nature No %
x Ray 59 9
MRI 12 2
x Ra+MRI 3 0.4
Ultrasound 1 0.1
Arthroscopy 1 0.1
x Ray+ultrasound 1 0.1
None 600 89
Total* 77 101
*Percentage totals may be subject to rounding errors associated with
individual components.
234 Woods, Hawkins, Hulse, et al
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006
Table 5 shows the number of players wearing external support
to the ankle. In 32% of injuries, players had been wearing
some form of external support.
The reinjury rate for ankle sprains was 9%, whereas the
average reinjury rate for all injuries was 7%. Although not significant,
there were more non-contact reinjuries than initial
injuries (47% v 39%). The average number of training days
missed and the average number of matches missed per ankle
sprain for reinjuries and initial injuries did not differ
significantly (19 days and four matches v 18 days and three
matches).
DISCUSSION
Of all the injuries sustained over the two seasons, ankle injuries
were responsible for 11%. This figure is lower than most
other studies, with figures of 15%,18 22%,12 and 32%19 being
reported. The differences in injury definition and
methodology18 20 21 makes comparison between studies difficult
and may help explain differences in the results. For
example, some studies record injury rate per 1000 hours.
However, in this study, the exposure of players to training and
matches was not measured, therefore injury rate could only be
reported in absolute terms. Also, we did not include any injuries
where players missed training for less than 48 hours,
whereas other studies have used the definition that an injury
is any incident that causes a player to miss the next scheduled
game or practice.11–13 On consultation with doctors and physiotherapists
working in professional football, it was felt that the
definition used in the present study was more appropriate. It
should also be noted that the results of this study are based on
the diagnoses of individual club medical personnel, which
may vary from practitioner to practitioner.
We found that a sprain was, by far, the most common type
of injury to the ankle (67%). Ankle sprains most often
involved the lateral ligament complex (77%). Lewin12 also
found the lateral ligament to be the most commonly injured
structure (67%). This may be because of the relative shortness
of the medial malleolus and the natural tendency for the ankle
to go into inversion rather than eversion.5 We observed
involvement of the anterior talofibular ligament (ATFL) in
73% of cases.Other authors have also found the ATFL to be the
most commonly sprained ligament,6 with Sitler et al22
reporting that 66% of the ligamentous injuries of the ankle
were to the ATFL. A possible reason for the high incidence of
injury to the ATFL could be that it has a lower load to failure
than the calcaneofibular ligament (CFL).2 Clanton and
Porter23 quoted values of 138 N and 345 N for the ATFL and
CFL respectively. Secondly, in plantarflexion, the ATFL is relatively
taut, whereas the CFL is relatively loose; in dorsiflexion,
the converse is true.23 This would fit with the common mechanism
of injury to the lateral ligament, which typically
involves the foot and ankle just at the moment of loading with
a plantarflexion and inversion force.23–27
Injuries to the medial or deltoid complex accounted for only
14% of ankle sprains. Clanton and Porter23 stated that medial
ligament complex injuries occur in 10% of all ankle sprains;
however, their review of ankle sprains included many different
Figure 1 Mechanism of non-contact
ankle sprains.
Figure 2 Timing of ankle sprains sustained during match play.
Figure 3 Month in which injury occurred: ankle sprains and all
injuries.
Table 5 Type of external support
worn by players who sustained ankle
sprains
No %
No support 336 50
Taping 167 25
Joint support 46 7
Missing 128 19
Total* 677 101
*Percentage totals may be subject to rounding
errors associated with individual components.
Ankle sprains in professional football 235
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006
sports. It is hardly surprising that the incidence of medial
ligament complex sprains in our study was higher than 10%
given that the demands of soccer include kicking with the
inside of the foot and ankle as well as receiving tackles to this
area.
This study shows that the anterior and posterior tibiofibular
ligament and interosseous membrane were injured in 4% of
sprains. These structures generally constitute the syndesmosis
of the ankle making this value comparable to that of Renström
and Konradsen,27 who reported a 3% incidence for isolated
syndesmosis injuries.
In our study, 11% (77) of ankle sprains were diagnostically
investigated, mostly by x ray examinations (59). According to
the Ottawa strategy for ankle injuries,28 radiographs should be
taken if there is bone tenderness at the tip or posterior aspect
of the lateral malleolus, at the tip or posterior aspect of the
medial malleolus, at the navicular tuberosity or base of the
fifth metatarsal, or if the patient is unable to weight bear
immediately after the injury and at the initial examination.
This system can then be used to reduce the use of radiographs.
A low number of players (6) had surgery for their ankle
sprains. This may be because functional non-operative
treatment is the accepted choice for grade I and grade II ankle
sprains.27 In the case of grade III sprains, the treatment is less
clear—that is,whether to immobilise in a cast, to operate, or to
allow early controlled mobilisation. Kuwada9 stated that,
when conservative measures have been exhausted and the
patient is not satisfied with his or her condition, surgical correction
is a reliable and viable treatment.
Our results show that more ankle injuries were sustained to
the dominant side than the non-dominant side, although the
difference was not significant based on expected incidence.
Other studies have shown significant differences.11 29 30 It could
be expected that most sprains would be to the dominant side,
as the main mechanisms of injury discussed previously generally
involve the dominant leg.
More contact than non-contact mechanisms of injury were
observed (59% v 39%). Árnason et al18 also found contact ankle
sprains to be more common than non-contact (69% v 31%).
Similarly they found that “tacklings”, which presumably
includes tackling and being tackled, to be the major
mechanism of injury (62%); in comparison, we observed this
value to be 54%. Non-contact mechanisms were most
commonly landing, twisting and turning, and running. The
only positional variation in mechanism of injury was that
goalkeepers sustained significantly more non-contact injuries
(namely twisting and turning, landing, and diving). This
would correlate with the functional profile of a goalkeeper as
they are regularly performing these activities as part of their
positional requirements. The mechanism of injury is vital from
the point of view of functional rehabilitation programmes and
in devising strategies for the prevention of reinjury. It has been
suggested that athletes be trained and rehabilitated in potential
positions of injury.26 31 If this principle is applied to
football, activities involving jumping, landing, cutting and
turning, and running could be performed during late stage
rehabilitation and preventive protocols to maximise ankle stability
during such manoeuvres. Contact positions of injury can
also be used, but as this generally involves tackling, it may be
more difficult to simulate and control safely. Laskowski et al32
stated that sport specific training is crucial in regaining proprioception
to “hard wire the proprioceptive pathways and
solidify a neuromuscular engram specific to these activities.”
According to Hawkins et al,16 the impact of an injury on a
club can be considered in relation to its severity and the
number of potential competitivematches missed.We observed
that 12 138 days and 2033 matches were missed because of
ankle sprains, which equates to an average of 18 days and
three games missed per sprain. Ekstrand and Gillquist30
reported that players were absent from practice on average for
four weeks after an ankle sprain, but the number of players in
their study was much smaller than in the present one. In this
study, 83% of ankle sprains had a rehabilitation period of less
than one month. This suggests thatmost ankle sprains are not
severe, and it is the incidence rather than severity of ankle
sprains that makes them problematic injuries. It also suggests
that the rehabilitation period was rather short, which may
explain the higher than expected reinjury rate for ankle
sprains compared with total injuries (9% v 7%), as the injury
may not have had enough time to heal completely. Houglum33
stressed the importance of understanding the phases and
timing of healing for appropriate, efficient, and effective rehabilitation.
There is no uniform consensus on how long
injured ligaments take to reach normal tensile strength;
figures range from 16 weeks to 40–50 weeks for a return to
85–95% of normal tensile strength.33 With periods of rehabilitation
being much shorter than the duration of ligament
healing, players may have returned to full function without
full tensile strength of the ligament. Applying stress to collagen
in the maturation phase helps to organise the collagen
fibres, enhancing the strength of the scar.33 This may present a
case for continuing treatment of the ligament during the
maturation and remodelling stage even when the player has
returned to full training. This would ensure that the ligament
regains as much strength and organisation as possible.
Konradsen et al34 monitored changes in ankle eversion
strength and sensorimotor control functions after acute ankle
inversion injury. They found that 12 weeks after the injury, an
increased error in accuracy of ankle position was still present
compared with the healthy ankle. It took six weeks for
normalisation of eversion strength. These findings justify continued
proprioceptive and strength type training even after
players have returned to play. Tropp et al35 recommended wobble
board training after return to play to prevent reinjury. This
training may also help to avoid the development of chronic
ankle joint instability (especially functional instability), ankle
instability being common among athletes.2 4 35
More injuries were sustained during matches than in training
(66% v 33%). Árnason et al18 also reported a higher injury
rate for matches, but the difference was much greater (4.4 v
0.1 per 1000 hours equating to 98% v 2%). This correlates with
the increased number of contact mechanisms, as more contact
injuries would be expected during games.18 Nearly half (48%)
of ankle sprains sustained during games occurred during the
last one third of each half. This pattern was observed by
Hawkins et al16 for all injuries, with the authors citing Gleeson
et al36 who suggested that the risk of ligamentous injury may
be increased by increases in electromechanical delay and
anterior tibiofemoral displacement. This emphasises the
importance of endurance training in ankle rehabilitation to
avoid fatigue at the end of each half. Itmay also present a case
for preventive training programmes when players are more
fatigued—that is, at the end of training sessions.However, this
requires further research, as other studies have found ankle
injuries to be evenly distributed throughout games.11 37
The timing of injuries throughout the season is also important;
44% of ankle injuries were sustained during the first
three months of the season, considerably more than expected.
The importance of structured neuromuscular coordination
and proprioceptive training during the closed season and preseason
months is emphasised, as the number of ankle sprains
peak in August and September. In their systematic review on
the prevention of ankle sprains, Thacker et al20 emphasised the
importance of conditioning of the ankle before the competitive
season and during the course of the season, with emphasis
on ankle strength and proprioception. According to Gauffin
et al38 postural sway and the pattern for postural correction
were improved by wobble board training.
Ankle sprains are commonly known as recurrent injuries,
with 56%,19 75%,39 and 69%18 of sprains involving players with
a previous history of ankle sprain. The problem with comparing
these data with our own is that this study only recorded
236 Woods, Hawkins, Hulse, et al
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006
injuries over two seasons and therefore the past medical history
of the players is not known—that is, if they sustained an
ankle sprain in previous years. Also, the studies cited above
have not recorded how they defined and measured previous
injury. Of the 677 injuries recorded over the two seasons in
this study, 57 were reinjuries, (9%). Although not significant,
it was found that those players sustaining recurrent injuries
missed on average more training days and matches than those
with first time injuries (18 v 19 days, three v four matches).
Missing four matches instead of three may not be significant
in terms of statistics, but in terms of football, it is crucial that
players, especially “first choice” ones, miss as few matches
possible.
More non-contact mechanisms were responsible for reinjuries
than initial injuries (47% v 39%). Nielsen and Yde19
described a characteristic pattern of major trauma causing the
initial injury, with minor trauma (for instance during
running) being responsible for the reinjury. Ekstrand and
Gillquist39 reported that many major injuries were preceded by
minor injuries; they suggested that this may be due to impairment
of timing and neuromuscular coordination. Árnason et
al37 suggested that reinjuries were caused by lack of preventive
measures and inadequate rehabilitation. Controlled rehabilitation
and strict adherence to directions for resumption of
play should therefore be insisted upon. It may also help to
have preinjury or normative measures of ankle strength and
proprioception as a component of player functional profiles.
The objective measures could then be used to help decide
when the player is fully fit. Waddington and Shepherd26
suggested measuring postural sway as a prediction of injury
risk. Athletes in the “higher injury incidence zone” would
then carry out a specifically designed functional training programme
to potentially reduce the risk of ankle injury.
Our study showed that 32% of players were wearing some
form of ankle support when they sustained an injury. This
appears remarkably high given that this is often considered to
be a form of prevention of ankle sprains.14 29 31 40 The question
must be posed as to why so many players were wearing an
ankle support. Perhaps it was for prophylactic reasons to prevent
initial injury or because of mechanical and or functional
instability from a previous injury. The high number of injuries
in taped ankles may be explained if the players involved had a
history of ankle sprain, because the risk of reinjuring a previously
sprained ankle is high. Some players are keen to return
to training without reaching full fitness and may request to
have their ankle strapped in the hope that this will provide
extra support and protection from reinjury. This may also help
to explain the high number of players sustaining injury even
with an ankle support. This study did not record how many
players were wearing an ankle support and who did not sustain
an injury. This, along with more detail on the ankle supports
used (for example the method of application, skill of
applicant, and the type of joint support used), would be
required to draw further conclusions. A discussion on the
effectiveness of joint support for the ankle joint as a preventive
tool in football is beyond the scope of this paper, although it is
an issue that undoubtedly requires further investigation.
As the lateral ligament ankle sprain is so common in football,
prevention of initial and recurrent injuries is of
paramount importance. Methods of preventing contact ankle
sprains have previously been suggested. These include rules to
control and minimise unnecessary or hazardous contact with
other players and appropriate officiating to ensure compliance
with event rules.20 These may in practice be very difficult to
implement, and so more practical interventions such as the
education of coaches and players to minimise contact in training
sessions and the wearing of an ankle guard component of
shin guards are recommended. None of these factors have
been subject to rigorous scientific review, but common sense
suggests that they would be useful in the prevention of such
injuries. Ekstrand and Gillquist39 recommended that coaches
emphasise injury prevention and that athletes be taught basic
principles of injury prevention. Other suggestions for the prevention
of ankle sprains include adequate maintenance of
pitches and training surfaces.39 This is a plausible suggestion
because it has been reported that one of the risk factors for
ankle injury is an uneven surface.5 Complete rehabilitation
and preseason ankle conditioning (involving functional
stimulus to both proprioceptive and muscular control systems
closely related to the action that overloads the system in the
first instance) have already been suggested. The use of external
support in the prevention of ankle sprains has yet to be
validated. However, both taping and braces have been shown
to prevent ankle sprains in football players.14 29 41 The design
selected for some of these studies may form a basis for questioning
the validity of the results.7
As a component of long term planning of athlete
development, Bayli42 emphasised the importance of mastering
eye-foot coordination and balance at an early age (6–10 years).
If such fundamentals are not mastered early in an athlete’s
career, his or her ability to move to a higher level of sporting
achievement will be limited. This so called “window of opportunity”
could also be used as a long term injury prevention
strategy by educating coaches to introduce proprioceptive and
coordination activities at this early age.
Ankle sprains (especially those involving the lateral
ligament) are common injuries in football. It is the frequency
and risk of reinjury rather than severity (time missed) that
makes these injuries problematic. Emphasis is therefore on
prevention through the use of functional profiles (including
normative and preinjury measures of ankle stability),
adequate rehabilitation, preseason conditioning of the ankle,
and education of coaches and players.
ACKNOWLEDGEMENTS
We acknowledge the financial support given by The Professional
Footballers’ Association together with the support of The League
Managers Association, The Premier League, and The Football League,
and the commitment of the medical practitioners working at professional
football clubs in England and Wales. We also gratefully
acknowledge the contributions made by the members of the Project
Consultative Committee Working Group, namely Mr R Myles Gibson
(Chairman), Dr C Cowie, Dr M Waller, Mr G Lewin, and Mr A Jones.
. . . . . . . . . . . . . . . . . . . . .
Authors’ affiliations
C Woods, R Hawkins, M Hulse, A Hodson, The Football Association,
Medical and Exercise Department, Lilleshall National Sports Centre,
Shropshire, UK
REFERENCES
1 Robbins S, Waked E, Rappel R. Ankle taping improves proprioception
before and after exercise in young men. Br J Sports Med
1995;29:242–7.
2 Barrett J, Bilisko T. The role of shoes in the prevention of ankle sprains.
Sports Med 1995;20:277–80.
3 Ogilvie-Harris DJ, Gilbart M. Treatment modalities for soft tissue injuries
of the ankle: a critical review. Clin J Sport Med 1995;5:175–86.
4 Karlsson J, Swärd L, Andréasson GO. The effect of taping on ankle
stability. Sports Med 1993;16:210–15.
5 Garrick JG. The frequency of injury, mechanism of injury, and
epidemiology of ankle sprains. Am J Sports Med 1977;5:241–2.
Take home message
Ankle sprains are common in football and usually involve
the lateral ligament. Their frequent occurrence and
recurrence indicates that preventive strategies such as
functional profiles (including normative and preinjury
measures of ankle stability), effective rehabilitation,
preseason conditioning of the ankle, and education of
coaches and players are of paramount importance.
Ankle sprains in professional football 237
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006
6 Orteza LC, Vogelbach WD, Denegar CR. The effect of molded and
unmolded orthotics on balance and pain while jogging following
inversion ankle sprain. Journal of Athletic Training 1992;27:80–4.
7 Callaghan MJ. Role of ankle taping and bracing in the athlete. Br J
Sports Med 1997;31:102–8.
8 Garrick JG, Requa RK. The epidemiology of foot and ankle injuries in
sports. Clinics in Podiatric and Medical Surgery 1989;6:629–37.
9 Kuwada GT. Current concepts in the diagnosis and treatment of ankle
sprains. Clinics in Podiatric Medicine and Surgery 1995;12:653–65.
10 Pardon ET. Lower extremities are site of most common soccer injuries.
Physician and Sports Medicine 1977;6:43–8.
11 Ekstrand J, Gillquist J. Soccer injuries and their mechanisms: a
prospective study. Med Sci Sports Exerc 1983;15:367–70.
12 Lewin G. The incidence of injury in an English professional soccer club
during one competitive season. Physiotherapy 1989;75:601–5.
13 Ekstrand J, Tropp H. The incidence of ankle sprains in soccer. Foot
Ankle Int 1990;11:41–4.
14 Tropp H, Askling C, Gillquist J. Prevention of ankle sprains. Am J Sports
Med 1985;13:259–62.
15 Putukian M, Knowles WK, Swere S, et al. Injuries in indoor soccer: The
Lake Placid Dawn to Dark Soccer Tournament. Am J Sports Med
1996;24:317–22.
16 Hawkins RD, Hulse MA, Wilkinson C, et al. The association football
medical research programme: an audit of injuries in professional football.
Br J Sports Med 2000;34:0–4.
17 Sandelin J, Santavirta S, Kiviluoto O. Acute soccer injuries in Finland in
1980. Br J Sports Med 1985;19:30–3.
18 Arnason Á, Gudmundsson Á, Dahl HA, et al. Soccer injuries in Iceland.
Scand J Med Sci Sports 1996;6:40–5.
19 Nielsen AB, Yde, J. Epidemiology and traumatology of injuries in
soccer. Am J Sports Med 1989;17:803–7.
20 Thacker SB, Stroup DF, Branche CM, et al. The prevention of ankle
sprains in sports. A systematic review of the literature. Am J Sports Med
1999;27:753–60.
21 Lüthje P, Nurmi I, Kataja M, et al. Epidemiology and traumatology of
injuries in elite soccer: a prospective study in Finland. Scand J Med Sci
Sports 1996;6:180–5.
22 Sitler M, Ryan J, Wheeler B, et al. The efficacy of a semirigid ankle
stabilizer to reduce acute ankle injuries in basketball. Am J Sports Med
1994;22:454–61.
23 Clanton TO, Porter DA. Primary care of foot and ankle injuries in the
athlete. Clin Sports Med 1997;16:435–66.
24 Robbins S, Waked E. Factors associated with ankle injuries. Sports Med
1998;25:63–72.
25 Liu SH, Jason WJ. Lateral ankle sprains and instability problems. Foot
and Ankle Injuries 1994;13:793–809.
26 Waddington GS, Shepherd RB. Ankle injury in sports: role of motor
control systems and implications for prevention and rehabilitation.
Physical Therapy Review 1996;1:79–87.
27 Renström AFH, Konradsen L. Ankle ligament injuries. Br J Sports Med
1997;31:11–20.
28 Stiell IG, McKnight RD, Greenberg GH, et al. Implementation of the
Ottawa ankle rules. JAMA 1994;271:827–32.
29 Surve I, Schwellnus MP, Noakes T, et al. A fivefold reduction in the
incidence of recurrent ankle sprains in soccer players using the
sport-stirrup orthosis. Am J Sports Med 1994;22:601–6.
30 Ekstrand J, Gillquist J. The frequency of muscle tightness and injuries in
soccer players. Am J Sports Med 1982;10:75–8.
31 Safran MR, Zachazewski JE, Benedetti RS, et al. Lateral ankle sprains: a
comprehensive review. Part 2: treatment and rehabilitation with an
emphasis on the athlete. Med Sci Sports Exerc 1999;31:S438–47.
32 Laskowski ER, Newcomer-Aney K, Smith J. Refining rehabilitation with
proprioception training. Physician and Sports Medicine
1997;25:89–102.
33 Houglum PA. Soft tissue healing and its impact on rehabilitation. Journal
of Sport Rehabilitation 1992;1:19–39.
34 Konradsen L, Olesen S, Hansen HM. Ankle sensorimotor control and
eversion strength after acute ankle inversion injuries. Am J Sports Med
1998;26:72–7.
35 Tropp H, Odenrick P, Gillquist J. Stabilometry recordings in functional
and mechanical instability of the ankle joint. Int J Sports Med
1985;6:180–2.
36 Gleeson NP, Reily T, Mercer T, et al. Influence of acute endurance
activity on leg neuromuscular and musculoskeletal performance. Med Sci
Sports Exerc 1998;30:596–608.
37 Árnason Á, Jóhannsson E, Gudmundsson Á, et al. Strains, sprains and
contusions in Icelandic elite soccer players. Med Sci Sports Exerc
1994;26:s17.
38 Gauffin H, Tropp H, Odenrick P. Effect of ankle disk training on postural
control in patients with functional instability of the ankle joint. Int J Sports
Med 1988;9:141–4.
39 Ekstrand J, Gillquist J. The avoidability of soccer injuries. Int J Sports
Med 1983;2:120–8.
40 Hume PA, Gerrard DF. Effectiveness of external ankle support. Sports
Med 1998;25:285–312.
41 Ekstrand J, Gillquist J, Liljedahl S. Prevention of soccer injuries. Am J
Sports Med 1983;11:116–20.
42 Bayli I. Long-term planning of athlete development: the training to train
phase. Faster Higher Stronger 1998;1:8–11.
238 Woods, Hawkins, Hulse, et al
www.bjsportmed.com
Downloaded from bjsm.bmjjournals.com on 22 October 2006

doi:10.1136/bjsm.35.4.212
Br. J. Sports Med. 2001;35;212-213
J Orchard
professional football
The use of local anaesthetic injections in
http://bjsm.bmjjournals.com/cgi/content/full/35/4/212
Updated information and services can be found at:
These include:
References
http://bjsm.bmjjournals.com/cgi/ ... 4/212#otherarticles
3 online articles that cite this article can be accessed at:
Rapid responses
http://bjsm.bmjjournals.com/cgi/eletter-submit/35/4/212
You can respond to this article at:
service
Email alerting
top right corner of the article
Receive free email alerts when new articles cite this article - sign up in the box at the
Topic collections
• Sports Medicine (1329 articles)
• Other Anesthesia (168 articles)
• Pain (271 articles)
Articles on similar topics can be found in the following collections
Notes
http://www.bmjjournals.com/cgi/reprintform
To order reprints of this article go to:
http://www.bmjjournals.com/subscriptions/
To subscribe to British Journal of Sports Medicine go to:
Downloaded from bjsm.bmjjournals.com on 22 October 2006
Leader
The use of local anaesthetic injections in professional football
Local anaesthetic use in professional football is one of the
greatest taboos in sports medicine. The subject is not covered
comprehensively in any sports medicine textbook or
review article. Most of the publications citing local anaesthetic
use are legal cases from the National Football
League (NFL) in which the procedure has been connected
with a career ending injury.1 2 There are some candid
accounts of how commonplace local anaesthetic use is in
the NFL,3 4 and anecdotal evidence suggests that the situation
is no different in professional football competitions
elsewhere in the world. This includes professional rugby
union, in which the practice is officially banned. I have previously
published my personal statistics of local anaesthetic
injections over a four year period with a professional Australian
football team,5 6 but cannot find any similar
documentation from other doctors in the sports medicine
literature. The attitude of most professional football
doctors may be that this practice is a necessary evil that can
be performed on certain occasions, but not mentioned or
justified in public.
I contend that we end this hypocrisy. Either sports physicians
should cease using local anaesthetics in professional
football and recommend that the practice be universally
banned, or we should study the procedure, speak and write
freely about it, and produce guidelines for its rational use.
The effects of officially banning local anaesthetic use, as
the International Rugby Board has done, would include an
increase in injury prevalence, which is measured by players
missing games. This is because players who could take the
field with a painkilling injection would be forced onto the
sidelines by team physicians adhering to the law. Anecdotal
evidence is that the ban in rugby union has not eliminated
the practice from occurring altogether. Most of the desire
to take the field with injury arises from the players
themselves, and if regular doctors refuse to use local anaesthetics,
some players find an alternate doctor who is willing
to do it in secret. For a suitable injury in a critical match, a
player may be able convince his doctor to administer the
injection even if it was banned, with both parties officially
denying it. Suspending players for testing positive to a local
anaesthetic in a drug test is not practical, as these drugs are
regularly needed in professional football during the procedure
of suturing lacerations. A ban on local anaesthetic use
is neither realistic, nor in my opinion necessary, so I will
proceed with the argument in favour of the judicious use of
local an

CLINICAL IMPLICATIONS: The normal values will help both in
diagnosis and follow up of patients with chronic lung diseases.
DISCLOSURE: Nisar Rao, None.
Pulmonary Function: Pulmonary Function
Testing
12:30 PM - 2:00 PM
UTILIZING THE FEV1/FEV6 AS A SURROGATE FOR FEV1/FVC
TO DETECT OBSTRUCTION IN AN INNER-CITY HOSPITAL
POPULATION
Roberto B. Dos Remedios MD* Ria Gripaldo MD Harlem Hospital, New
York, NY
PURPOSE: Recent literature sites the use of Forced Expiratory
Volume in 1 second/Forced Expiratory Volume in 6 seconds (FEV1/
FEV6)as a surrogate for FEV1/Forced Vital Capacity(FVC) in diagnosing
obstruction, with an average sensitivity of 94%. The aforementioned
studies were predominantly on white patient populations. It is the
objective of this group to evaluate the same indices in an inner city black
patient population.
METHODS: A total of 380 studies were obtained that met acceptability
criteria as defined by the American Thoracic Society. Prediction
equations for FEV1, FEV1/FVC and FEV1/FEV6 as established by
Hankinson et al, were used for Predicted Normals as well as Lower Limits
of Normal (LLN). Obstruction was diagnosed by ratios below the
predicted LLN. Sensitivity, specificity, positive and negative predictive
values (PPV and NPV) were calculated for the FEV1/FEV6 vs. the
FEV1/FVC.
RESULTS: The FEV1/FEV6 showed a sensitivity of 92% and a
specificity of 97% compared to the FEV1/FVC. It had a PPV of 93% and
its NPV was 97%. On examination of the discordant results (There were
8 false positives and 9 false negatives), majority fell close to the lower
limits of normal, and when obstruction was stratified for severity according
to the ATS guidelines they fell under mild obstruction.
CONCLUSION: Our population shows good correlation between the
FEV1/FEV6 and the FEV1/FVC in identifying obstruction. It becomes
apparent however that the discordant results identified may pose a
problem to subgroups of patients with mild obstruction or ratios close to
the lower limit of normal, such as smokers 45 years old and above for
whom office spirometry which relies on FEV1/FEV6 is a recommended
screening tool.
CLINICAL IMPLICATIONS: Although the utility of FEV1/FEV6 as
a surrogate for FEV1/FVC to identify obstruction in patients who might
have difficulty meeting standard end of test criteria is supported by our
data, the utility of FEV1/FEV6 as a screening tool where the primary
population will be at or near normal, needs to be evaluated further.
DISCLOSURE: Roberto Dos Remedios, None.
INTERPRETABILITY OF SPIROMETRY IN COMMUNITYBASED
STUDIES
Robert A. Cohen MD, FCCP David Angulo MD* Diana Hackbarth RN,
PhD., Jeanine Solinski MPP Rhonda Williams MES Christa McKee Lori
Younker Yuanhe Li MPH Kristen Donnelly MPH John H. Stroger, Jr.
Hospital of Cook County, Chicago, IL
PURPOSE: To determine factors that influence interpretability rates
of screening spirometries obtained in community settings. Widespread
spirometric testing has been advocated to diagnose early chronic obstructive
pulmonary disease. There is little information on the efficacy of mass
screening spirometry performed in community settings.
METHODS: Two cohorts were studied: 1) Cohort P were 1819
subjects presenting to a national pharmacy chain for a “free lung test” to
rule out COPD. They were tested by 55 volunteer respiratory care
practitioners who had received four hours of training. 2) Cohort S who
were 350 subjects presenting to local schools to rule out asthma. They
were tested by 7 lay research assistants who had completed 16 hours of a
NIOSH spirometry. The spirometry studies were read by independent
pulmonologists according to ATS 1994 acceptability and reproducibility
criteria.
RESULTS: 36% of Cohort P were male as were 22% of Cohort S.
Cohort P had a mean age of 58 compared to 31 in Cohort S. 70% of the
studies in Cohort A were interpretable and 90% of studies in Cohort B
(p  0.001). Heavier individuals were more likely to have interpretable
spirometries, BMI of 29 kg/m2 versus 27 kg/m2 (p  .001). Binary logistic
regression analysis showed subjects tested by personnel who had received
16 hours training and who averaged  50 studies per person had the
highest predictive power for interpretability vs personnel who had 4 hours
of training and averaged  12 studies per person. In Cohort P, female
gender and higher BMI were associated with interpretability whereas age
and smoking status were not significant. In Cohort S, male gender was the
only predictor of interpretability.
CONCLUSION: Large scale screening can be performed successfully
in community settings with interpretability rates of greater than 70%. The
number hours of training and the number of studies performed per
technician appear to be more important than patient characteristics.
CLINICAL IMPLICATIONS: Large scale spirometric screening
programs may provide useful data even when performed in non-clinical
settings.
DISCLOSURE: David Angulo, None.
REFERENCE VALUES FOR PULMONARY FUNCTION OF RETIRED
PROFESSIONAL AMERICAN FOOTBALL PLAYERS
Janette D. Foster MD, MBA* Larisa V. Buyantseva MD, MS Erik B.
Lehman MS Zhengmin Qian MD, PhD Robert L. Vender MD Arthur J.
Roberts MD Rebecca Bascom MD, MPH Pennsylvania State University
College of Medicine, Hershey, PA
PURPOSE: To establish reference values for pulmonary function
(PFTs) of retired professional American football players (rPAFP). We
hypothesized: 1) rPAFP derived prediction equations differ from the
NHANESIII equations in these large men 2) No lung function difference
exists between Caucasian-Americans (CA) and African-American (AA)
men of similar body builds and adult physical activity history and
socioeconomic status.
METHODS: Cross sectional study of 549 volunteers recruited to ten
health screenings conducted by the Living Heart Foundation. Inclusion
criteria for reference values were current or former professional American
football play, good quality spirometry, lifetime nonsmokers, absence of
self-reported asthma or other chronic lung disease. Simple linear regression
derived prediction equations for forced vital capacity (FVC, liters)
and forced expiratory volume one second (FEV1, liters).
RESULTS: 311 participants (191 CA and 120 AA) met inclusion
criteria. Mean demographic characteristics were age 50.7  13.0 years
(range 23-85), height 73.3  2.6 inches (range 64-81), weight 241  44
pounds (range 155-393), and body mass index 31.4 4.7 (range 22.3-
50.5).rPAFP prediction equations correlated with NHANESIII equations
for FVC (r0.951) and for FEV1 (r0.955), but paired t-test showed
statistically significant differences for %predicted FVC (9712.6
NHANESIII vs 10013 rPAFP, p0.0001) and %predicted FEV1
(9511 NHANESIII vs 10012 rPAFP, p0.0001). FVC differed by
race (4.610.78 AA vs 5.351.00 CA (p0.05)) as did FEV1 (3.580.60
AA vs 4.050.75 CA (p0.05)). The rPAFP prediction equations are:
FEV1(AA)  (-0.01891)* age  (0.1055)* height  (-3.2197) FVC (AA)
 -0.01696)* age  (0.1433)* height  (-5.0254) FEV1 (CA) 
(-0.03254)* age  (0.1151)* height  (-2.6857) FVC (CA)  (-0.03902)*
age  (0.1539)* height  (-3.8886).
CONCLUSION: rPAFP derived prediction equations differ from the
NHANESIII equations in these large men although the large residual
error suggests a need to additional rPAFP modeling. Lung function
differences between highly athletic Caucasian-American and African-
American men are similar in magnitude to results from general population
samples.
CLINICAL IMPLICATIONS: These data support the use of population
specific prediction equations of pulmonary function testing, and for
CA and AA specific prediction equations even among men who during
their early adult life shared similar body habitus, physical activity history,
and socioeconomic factors.
DISCLOSURE: Janette Foster, None.
Wednesday, October 25, 2006
Pulmonary Function: Exercise and
Rehabilitation, continued
CHEST / 130 / 4 / OCTOBER, 2006 SUPPLEMENT 247S
POSTER PRESENTATIONS

J. Am. Coll. Cardiol. 2003;41;280-284
William B. Abernethy, III, Joseph K. Choo and Adolph M. Hutter, Jr
Echocardiographic characteristics of professional football players
This information is current as of October 22, 2006
http://content.onlinejacc.org/cgi/content/full/41/2/280
located on the World Wide Web at:
The online version of this article, along with updated information and services, is
Downloaded from content.onlinejacc.org by on October 22, 2006
Cardiac Imaging
Echocardiographic Characteristics
of Professional Football Players
William B. Abernethy, III, MD, Joseph K. Choo, MD, Adolph M. Hutter, JR, MD
Boston, Massachusetts
OBJECTIVES We examined the echocardiographic characteristics of highly trained American football
players.
BACKGROUND Intense physical training is associated with morphologic and physiologic cardiac changes
often referred to as the “athlete’s heart.” Echocardiographic features peculiar to elite football
players have not been described.
METHODS We studied cardiac morphology and function as assessed by rest and stress echocardiography in
156 asymptomatic National Football League players. Resting and stress ejection fraction (EF),
wall thickness, and diastolic left ventricular internal diameter (LVID) were measured. Left
ventricular (LV) mass was calculated, as was relative wall thickness (RWT) defined as septal and
posterior wall thickness divided by LVID. Control data were obtained from published studies.
RESULTS The mean LVID (53
0.5 mm) and maximal wall thickness (11.2
0.2 mm) were increased
over normal reported control subjects. There was a correlation between LVID and body
weight (p  0.01) and body surface area (BSA) (p  0.01). The average LVID indexed to
BSA was 23
2 mm/M2. There was also a correlation between maximal wall thickness and
body weight (p  0.01) and BSA (p  0.01). The average wall thickness indexed to BSA was
5.05
0.88 mm/M2. Of the players, 23% had evidence of LV hypertrophy. Two players had
an increased septal-to-posterior-wall-thickness ratio (
1.3), although no player had an
outflow gradient. The RWT for the players was 0.424 (
0.1). The mean resting EF was 58%
(
4.4%), and every player undergoing exercise testing had an appropriate hyperdynamic
response in cardiac function.
CONCLUSIONS Both wall thickness and LVID of elite American football players are increased and correlate
with body size. There is a high RWT, reflecting an emphasis on strength training. The LV
EF was normal and not supranormal, as is sometimes believed. Regardless of the resting EF,
all players had hyperdynamic cardiac responses with exercise. (J Am Coll Cardiol 2003;41:
280–4) © 2003 by the American College of Cardiology Foundation
Intense physical training is associated with characteristic
changes in cardiac function and morphology that have been
termed the “athlete’s heart” (1). Physiologic alterations from
training include an increased stroke volume and decreased
heart rate (2), while morphologic changes include increased
left ventricular (LV) cavity dimension, wall thickness, and mass
(1,3,4). The extent of morphologic changes varies between
sports (5,6), and complicates the differentiation between these
normal physiologic alterations and a cardiomyopathy. There is
also a misperception among some that the LV ejection fraction
(EF) of a well-trained athlete should be supranormal. Although
the exercise-induced cardiac changes in athletes participating
in various sports have been well described, changes
associated with American football have been less well studied.
Accordingly, we reviewed the echocardiographic features of
156 professional American football players.
METHODS
Between 1996 and 1999, 1,282 apparently healthy collegiate
football players underwent a routine physical exam and
electrocardiography as part of their evaluation for selection
into the National Football League. No athlete was excluded
for any reason. Each of the athletes had performed at an
exceptional level and was being assessed for professional
employment. Steroid use among the participants was not
known. All players had similar training methods emphasizing
efforts to improve power and speed rather than endurance,
although there was some variability based on the
player’s position. The blood pressure in all players was
consistently or predominantly 140/90 mm Hg. All subjects
were asymptomatic. The 156 athletes undergoing
echocardiography had a clinical suspicion of possible heart
disease based on mild abnormalities on history, clinical
evaluation (often a family history of hypertension), or the
electrocardiogram. In other respects they were representative
of the elite football players entering the National
Football League. Many players with electrocardiographic
abnormalities did not have an echocardiogram. The players
studied underwent M-mode, Doppler, and twodimensional
echocardiography at rest and with exertion
using commercially available equipment (Hewlett-Packard,
Andover, Massachusetts). Wall thickness, left ventricular
internal diameter (LVID) at end-diastole, resting and exertional
EF, left ventricular mass (LVM), and relative wall
From the Cardiac Unit, Massachusetts General Hospital, Boston, Massachusetts.
Manuscript received September 20, 2001; revised manuscript received July 8, 2002,
accepted August 20, 2002.
Journal of the American College of Cardiology Vol. 41, No. 2, 2003
© 2003 by the American College of Cardiology Foundation ISSN 0735-1097/03/$30.00
Published by Elsevier Science Inc. PII S0735-1097(02)02633-5
Downloaded from content.onlinejacc.org by on October 22, 2006
thickness (RWT) were assessed. Interventricular septal
(IVST) and posterior wall thickness (PWT) were considered
mildly abnormal if 11 mm and markedly abnormal if
13 mm. Septal-to-PWT ratio was considered abnormal if
1.3 (7). Left ventricular cavity dimension was considered
mildly abnormal if 55 mm and clearly abnormal if 60
mm. Relative wall thickness was calculated (PWT  IVST/
LVID) and expressed as a fraction (4). Left ventricular mass
was calculated from end-diastolic wall thickness and cavity
dimension using the Penn-cube formula: LVM  1.04
([LVID  PWT  IVST]3  LVID)  13.6 g (8). Left
ventricular hypertrophy was considered present when mass
index exceeded 116 g/M2 (9). Left ventricular EF was
measured by biplane Simpson’s rule with measurements
taken from endocardial contour measured in the apical and
two-chamber views (7). Stress echocardiograms were performed
utilizing two-stage stress with a modified Naughton
protocol achieving a maximum heart rate of 85%
predicted. The images were qualitatively evaluated visually
for wall motion abnormalities. The relationships
between wall thickness, LVID, body weight, and body
surface area (BSA) were determined by calculation of
Pearson’s correlation collation with statistical significance
determined by two-tailed t testing. Data are presented as

SE with a p value 0.05 considered significant.
Interobserver reproducibility assessment of the measurements
was not available.
RESULTS
The subjects had a mean age of 22 years. The height ranged
from 172 to 201 cm (69 to 79 in.) with a mean 185 cm (73
in.), and their weight ranged from 85 to 157 kg (187 to 347
lbs), with a mean 105 kg (231 lbs). The BSA ranged from
1.95 to 2.84 M2, with a mean 2.29 M2.
The mean LVID was 53 mm (
0.5 mm). While over
one-third of the players (43 of 156) had a mildly enlarged
LVID of 55 mm, only 10 players (6%) had markedly
enlarged ventricles (LVID 60 mm) (Fig. 1). There was a
correlation between LVID and body weight (Pearson correlation
of 0.468, p  0.01) as well as between LVID and
BSA (Pearson correlation of 0.479, p  0.01). The average
LVID indexed to BSA was 23
2 mm/M2.
The mean value for maximal wall thickness was 11.2 cm
(
0.2 cm) (Figs. 2 and 3). There was a correlation between
wall thickness and body weight (Pearson correlation of
0.490, p  0.01) as well as between wall thickness and BSA
(Pearson correlation of 0.457, p  0.01). The average wall
thickness indexed to BSA was 5.05
0.88 mm/M2.
Seventeen athletes (11%) had a wall thickness 13 mm, and
two players had a wall thickness 15 mm. Both of these
players with markedly thickened walls weighed more than
118 kg (260 lbs) and had LVIDs 55 mm. One of the
athletes had increased wall thickness with a relatively small
LV cavity (wall thickness of 15 mm, LVID of 40 mm).
There was no asymmetry in wall thickness or resting outflow
gradient in that athlete, although this was suggestive of
hypertrophic nonobstructive cardiomyopathy. The mean
septal-to-PWT ratio was 1.02. Only two players had a
markedly increased ratio (
1.3), and neither of these players
had a murmur or outflow tract gradient.
The mean LVM was 236 g (Fig. 4). Only 3% of players
had an LVM 350 g. The mean LVM/BSA ratio was 103
g/M2 (
20.9 g/M2). Left ventricular hypertrophy was
present in 23% of the athletes (26 of the 113 subjects for
which sufficient information was available to allow calculation
of the LVM). The RWT was 0.424. Thirty-six athletes
Abbreviations and Acronyms
BSA  body surface area
EF  ejection fraction
IVST  interventricular septal wall thickness
LV  left ventricular
LVID  left ventricular internal diameter
LVM  left ventricular mass
PWT  posterior wall thickness
RWT  relative wall thickness
Figure 1. Distribution plot of left ventricular diastolic diameter. Six percent of the players had a left ventricular internal diameter (LVID) at end-diastole
6.0 cm. There was a correlation between LVID and body weight (p  0.01) as well as body surface area (p  0.01).
JACC Vol. 41, No. 2, 2003 Abernethy III et al. 281
January 15, 2003:280–4 Echocardiographic Characteristics of Football Players
Downloaded from content.onlinejacc.org by on October 22, 2006
(23%) had an RWT 0.44 (considered the upper limit of
normal of control subjects).
The mean resting EF was 58% (
4.4%). A total of 39%
of players had an EF of 50% to 55%, although no player had
an EF 50% (Fig. 5). All subjects had an appropriate
hyperdynamic EF with exertion to an average EF with stress
of 76% (
14%).
While trivial regurgitant valvular abnormalities were
common, only five players had moderate regurgitation (four
with pulmonic insufficiency and one with tricuspid insufficiency).
Two players had mitral valve prolapse, one had mild
pulmonic stenosis, and one had a bicuspid aortic valve. No
valvular abnormalities were found that would prompt limitation
of participation in professional football.
DISCUSSION
Physiologic hypertrophy is a common feature of the “athlete’s
heart.” The extent of change in cardiac dimensions
varies between athletes and training methods, and adds to
the clinical dilemma when attempting to distinguish athlete’s
heart from pathologic heart disease (1,4,9,10). Football
remains very popular in the U.S., and cardiac adaptations
peculiar to training for this sport have not been well
described. We intended to assess the cardiac structural
changes commonly seen among athletes intensely training
and competing in American football.
Although our study did not include cardiac morphology
on normal controls, a recent meta-analysis summarized such
findings among the control subjects of previous reports (4).
Cardiac structure in 78 original studies comparing athletes
to control subjects included the findings of over 800
controls, and reported an LVID of 49.6 mm, wall thickness
of 8.8 mm, and LVM of 174 g (4). The cardiac morphologies
found in our study of football players are similar to
average values reported among athletes pursuing a variety of
other sports (4,6,11). While both the wall thickness and LV
cavity size were increased, the increased LVM among the
football players was primarily due to an enlarged LV cavity.
This is an important point, as highly trained competitive
athletes (particularly rowers and cyclists) without apparent
heart disease can develop markedly thickened ventricular
walls that may resemble hypertrophic cardiomyopathy
(1,5,12,13). Previous studies have demonstrated that a LV
wall thickness of 13 mm was very uncommon among
highly trained athletes, and the upper limit to which the
ventricle wall thickens with training is 16 mm (11). In large
part, our findings on elite American football players have
been similar. Although nearly a one-quarter of the subjects
would be classified as having evidence of LV hypertrophy,
only a small minority had extreme levels of hypertrophy. In
Figure 2. Maximal wall thickness. The mean value was 11.2 mm (
0.2 mm). Six percent of players had a wall thickness 14 mm.
Figure 3. Maximal wall thickness related to player weight. There was a correlation between wall thickness and body weight (p  0.01) as well as body
surface area (p  0.01).
282 Abernethy III et al. JACC Vol. 41, No. 2, 2003
Echocardiographic Characteristics of Football Players January 15, 2003:280–4
Downloaded from content.onlinejacc.org by on October 22, 2006
our study both LVID and wall thickness correlated with
body size of the athlete. Six percent of the football players in
our series had a wall thickness in the borderline range of
13 mm, but no player had a thickness 16 mm, and no
player was found to have a dynamic outflow tract obstruction.
Morganroth et al. (5) first suggested two different morphologic
types of athlete’s heart: a strength-trained heart
and an endurance-trained heart. The pressure overload
states accompanying strength training result in predominantly
increased LV wall thickness with little change in LV
chamber size. The sustained episodes of high cardiac output
and volume overload that occur in endurance training would
preferentially increase LV chamber size with less effect on
wall thickness. Such divergent cardiac adaptations between
static and more dynamic sports were recently confirmed in a
large meta-analysis (4). The football players in our series
had cardiac changes more similar to other athletes that have
emphasized strength training, with a relatively larger increase
in LV wall thickness and a slight increase in LVID
(4). This is not surprising, as most players emphasize weight
lifting during their training, an activity that causes transient
large increases in blood pressure (14,15). This training
favors more ventricular wall hypertrophy, as reflected in the
high RWT among the football players. The RWT of 0.424
among the football players in this series is more than the
control patients (0.356) and endurance-trained athletes
(0.389) reported in past studies, although not as great as
values of RWT reported for classic strength trained athletes
(0.442) (4).
Although auscultation of a murmur prompted obtaining
the echocardiogram in some of the subjects, most players
were found to have only trivial regurgitant valves. There
were two cases of mitral valve prolapse, one of bicuspid
aortic valve, and one mild pulmonic stenosis, but no lesions
were found that would prompt surgical intervention or
limitation of activities. A physiologic flow murmur is a
well-recognized phenomenon in highly trained athletes.
Left ventricular systolic function is commonly assessed by
EF (7). There is a misperception among some that the EF
of a well-trained athlete should be supranormal. However,
studies suggest this is not the case. The largest study of
athletes has shown LV systolic function as assessed by EF or
fractional fiber shortening is similar to sedentary control
subjects (4). We found similar results, with a large percentage
(nearly 40%) of the football players having low normal
values between 50% and 55%. All subjects who underwent
stress echocardiography had an appropriate hyperdynamic
response in systolic function.
In summary, elite American football players in the National
Football League have cardiac dimensions similar to
other well-trained athletes. There is relatively more hypertrophy
of the LV wall as opposed to larger LV chamber size,
although both are increased. Both wall thickness and LV
chamber size correlate with body size. Resting EFs were
normal and not hyperdynamic, but all increased with exercise.
Figure 4. Distribution of left ventricular mass. Three percent of players had a left ventricular mass 350 g.
Figure 5. Resting ejection fraction. The mean resting ejection fraction was 58% (
4.4%).
JACC Vol. 41, No. 2, 2003 Abernethy III et al. 283
January 15, 2003:280–4 Echocardiographic Characteristics of Football Players
Downloaded from content.onlinejacc.org by on October 22, 2006
Acknowledgment
The authors thank Eugene V. Pomerantsev, MD, PhD, for
his advice and for the performance of the statistical analyses.
Reprint requests and correspondence: Dr. Adolph M. Hutter,
Jr., ACC 467 Massachusetts General Hospital, 14 Parkman
Street, Boston, Massachusetts 02114. E-mail: hutter.adolph@mgh.
harvard.edu.
REFERENCES
1. Maron B. Structural features of the athlete heart as defined by
echocardiography. J Am Coll Cardiol 1986;7:190–203.
2. Blomqvist CG, Saltin B. Cardiovascular adaptations to physical
training. Annu Rev Physiol 1983;45:169–89.
3. Spirito P, Pelliccia A, Proschan MA, et al. Morphology of the
“athlete’s heart” assessed by echocardiography in 947 elite athletes
representing 27 sports. Am J Cardiol 1994;74:802–6.
4. Pluim BM, Zwinderman AH, van der Laarse A, van der Wall EE.
The athlete’s heart: a meta-analysis of cardiac structure and function.
Circulation 2000;101:336–44.
5. Morganroth J, Maron BJ, Henry WL, Epstein SE. Comparative left
ventricular dimensions in trained athletes. Ann Intern Med 1975;82:
521–4.
6. Pelliccia A, Culasso F, Di Paolo FM, Maron BJ. Physiologic left
ventricular cavity dilatation in elite athletes. Ann Intern Med 1999;
130:23–31.
7. Weyman A. Principles and Practice of Echocardiography. 2nd ed.
Philadelphia, PA: Lea & Febiger, 1994.
8. Devereux RB, Reichek N. Echocardiographic determination of left
ventricular mass in man: anatomic validation of the method. Circulation
1977;55:613–8.
9. Palmieri V, Dahlof B, DeQuattro V, et al. Reliability of echocardiographic
assessment of left ventricular structure and function: the
PRESERVE study. J Am Coll Cardiol 1999;34:1625–32.
10. Maron BJ, Pelliccia A, Spirito P. Cardiac disease in young trained
athletes: insights into methods for distinguishing athlete’s heart from
structural heart disease, with particular emphasis on hypertrophic
cardiomyopathy. Circulation 1995;91:1596–601.
11. Pelliccia A, Maron BJ, Spataro A, Proschan MA, Spirito P. The upper
limit of physiologic cardiac hypertrophy in highly trained elite athletes.
N Engl J Med 1991;324:295–301.
12. Gilbert CA, Nutter DO, Felner JM, Perkins JV, Heymsfield SB,
Schlant RC. Echocardiographic study of cardiac dimensions and
function in the endurance-trained athlete. Am J Cardiol 1977;40:528–
33.
13. Shapiro LM, Kleinbenne A, McKenna WJ. The distribution of left
ventricular hypertrophy in hypertrophic cardiomyopathy: comparison
to athletes and hypertensives. Eur Heart J 1985;6:967–74.
14. MacDougall JD, Tuxen D, Sale DG, Moroz JR, Sutton JR. Arterial
blood pressure response to heavy resistance exercise. J Appl Physiol
1985;58:785–90.
15. MacDougall JD, McKelvie RS, Moroz DE, Sale DG, McCartney N,
Buick F. Factors affecting blood pressure during heavy weight lifting
and static contractions. J Appl Physiol 1992;73:1590–7.
284 Abernethy III et al. JACC Vol. 41, No. 2, 2003
Echocardiographic Characteristics of Football Players January 15, 2003:280–4
Downloaded from content.onlinejacc.org by on October 22, 2006
J. Am. Coll. Cardiol. 2003;41;280-284
William B. Abernethy, III, Joseph K. Choo and Adolph M. Hutter, Jr
Echocardiographic characteristics of professional football players
This information is current as of October 22, 2006
& Services
Updated Information
http://content.onlinejacc.org/cgi/content/full/41/2/280
including high-resolution figures, can be found at:
References
http://content.onlinejacc.org/cgi/content/full/41/2/280#BIBL
free at:
This article cites 14 articles, 10 of which you can access for
Citations
icles
http://content.onlinejacc.org/cgi/content/full/41/2/280#otherart
This article has been cited by 6 HighWire-hosted articles:
Rights & Permissions
http://content.onlinejacc.org/misc/permissions.dtl
tables) or in its entirety can be found online at:
Information about reproducing this article in parts (figures,
Reprints
http://content.onlinejacc.org/misc/reprints.dtl
Information about ordering reprints can be found online:
Downloaded from content.onlinejacc.org by on October 22, 2006

Br. J. Sports Med. 1998;32;326-332
RD Hawkins and CW Fuller
football
injuries and incidents at three levels of professional
An examination of the frequency and severity of
http://bjsm.bmjjournals.com/cgi/content/abstract/32/4/326
Updated information and services can be found at:
These include:
References
http://bjsm.bmjjournals.com/cgi/ ... 4/326#otherarticles
30 online articles that cite this article can be accessed at:
Rapid responses
http://bjsm.bmjjournals.com/cgi/eletter-submit/32/4/326
You can respond to this article at:
service
Email alerting
top right corner of the article
Receive free email alerts when new articles cite this article - sign up in the box at the
Notes
http://www.bmjjournals.com/cgi/reprintform
To order reprints of this article go to:
http://www.bmjjournals.com/subscriptions/
To subscribe to British Journal of Sports Medicine go to:
Downloaded from bjsm.bmjjournals.com on 22 October 2006
An examination of the frequency and severity of
injuries and incidents at three levels of
professional football
Richard D Hawkins, Colin W Fuller
Abstract
Objective—To assess the risk of injury to
professional footballers during European
international and English Premier and
First Division league matches.
Methods—Videotaped recordings of 29,
49, and 93 matches from the 1996 European
Championship, 1996/1997 English
Premier season and 1994 to 1997 English
First Division seasons respectively were
analysed. During each match, several relevant
variables, including the number of
fouls, injuries, time of incident, player
identity, and injury mechanism, were
recorded.
Results—Significantly more free kicks
were awarded during international
matches than during league matches;
however, there were no significant differences
between the numbers of free kicks
awarded over the three First Division seasons
assessed. Between 1.7 and 3.0% of
fouls resulted in a player requiring treatment
for injury, but only 15–28% of all
injuries resulted from foul play. In all
“non-foul” situations, in which injury
resulted, at least 60% still involved player
to player contact. No significant differences
in injury frequency were observed
between playing positions or match
halves.
Conclusions—The results equate to a total
of 808 players per season from the estimated
2600 players in the four English
professional football leagues sustaining a
match injury that caused them to miss at
least one game. The large number of
underlying “non-injury” incidents is
identified as the reason for this level of
injury rather than a higher ratio of
“injury” to “non-injury” incidents in
professional football compared with other
occupations.
(Br J Sports Med 1998;32:326–332)
Keywords: football; injury; risk factors
It is not possible or usually acceptable to make
all sports injury-free as the inherent risks associated
with sport often provide the main
attraction for participants and spectators.1
However, there is a legal duty for employers of
professional sportspeople in the United Kingdom
to assess the risks associated with the
sport and reduce these risks to an acceptable
level.2 To make meaningful decisions about the
level of risk in professional sport, it is important
to collect relevant data so that the risks can be
quantified and compared with other activities.
To comply with UK health and safety legislation,
the risk management process should
aim to identify factors that could lead to injury.
As the severity of the outcome of any incident
is often dependent on chance, the best way to
assess the overall levels of risk in any activity is
to identify the underlying causes and measure
their frequency of occurrence and the severity
of injuries resulting from these causes. This
stage should then be followed by the development
and implementation of control strategies
to reduce the magnitude of the risks or
eliminate the causes completely. As most injuries
in professional football are claimed to
occur during competitive games,3 4 a major step
in the overall risk management process should
therefore be quantification of the risks to players
during competitive matches. An assessment
of the level of this risk at the highest level of
competition was undertaken during the 1994
World Cup Finals.5 The aim of the current
work was to extend the assessment process to a
further three levels of professional competition
and to compare these results with those
obtained previously from the 1994 World Cup.
Methods
Three levels of professional football were
assessed, namely European international, English
Premier League, and English First Division.
Assessments were made from videotaped
recordings of 171 televised matches over a
period of three years. These matches comprised
29 of the 31 matches played during the
European Championship Finals staged in England
during June 1996, 49 matches played in
the English Premier League during the 1996/
1997 season, and 93 matches played in the
English First Division during the 1994/1995
(26), 1995/1996 (27) and 1996/1997 (40) seasons.
The criteria used for analysis were the same
as those used previously for the 1994 World
Cup Finals.5 They included: the use of refereeing
decisions on free kicks for the specification
of incident category—that is, foul and non-foul
play; players receiving treatment on or off the
pitch for the specification of injury; and three
levels of incident severity: “moderate” (the
player received treatment and subsequently
missed at least the next match), “minor” (the
player received treatment but did not miss the
next match through injury), and “non-injury”
(the referee deemed a foul to have been
committed but no player required treatment).
326 Br J Sports Med 1998;32:326–332
Centre for Hazard and
Risk Management,
Loughborough
University,
Loughborough, Leics
LE11 3TU, United
Kingdom
R D Hawkins
C W Fuller
Correspondence to:
Dr C W Fuller.
Accepted for publication
27 July 1998
Downloaded from bjsm.bmjjournals.com on 22 October 2006
During the European Championship, reports
available through national television,
radio, and press provided detailed information
which enabled the severity of players’ injuries
to be determined. For matches in the English
Premier and First Division leagues such
detailed information was not available on a
reliable basis to categorise injury severity so it
was not possible to identify moderate and
minor injuries in this way; therefore for most of
these matches the two levels of injury severity
were combined. However, comparable information
was available for some matches from
club physiotherapists involved in a complementary
study,6 and this information was used
to provide representative data on the severity of
players’ injuries.
The numbers of injuries recorded at each
competitive level were analysed and assessed in
three formats5: “accident ratio” (AR), the ratio
of the number of injuries at each level of injury
severity; “injury incidence rate” (IIR), the percentage
likelihood of an individual player
receiving treatment for an injury (moderate or
minor) during a match; and “injury frequency
rate” (IFR), the number of injuries (moderate
or minor) per 100 000 hours played.
Data were analysed using the Statistical
Package for Social Scientists (SPSS Inc,
Chicago, Illinois, USA). Student’s t test for
correlated means was used to examine differences
in results obtained between match
halves. A one way analysis of variance for
correlated means was used to investigate
differences between match quarters, the Tukey
post-hoc test being used to identify where significant
differences occurred. A two way analysis
of variance with repeated measures on one
factor was used to identify any differences
between each competitive level of play, the
Scheffe S post-hoc test being used to identify
where significant differences occurred. Results
are reported as means, and, where differences
in results are documented, statistical significance
was met at a minimum level of p<0.05.
Results
Table 1 shows the numbers of games analysed,
free kicks awarded for foul play, and player
treatments for injury arising from foul and
non-foul incidents for each season and/or level
of competition. The average numbers of free
kicks per game were found to be: World Cup,5
29; European Championship, 35; Premier
League, 20; First Division, 22 (table 1).
Significantly more free kicks were awarded on
average per game during the World Cup and
European Championship than during the Premier
League and First Division matches, and
the 1996 European Championship also produced
significantly more free kicks per game
than the 1994 World Cup. There were no
significant differences between the numbers of
free kicks awarded in each of the three First
Division seasons analysed. Between 1.7 and
3.0% of fouls committed actually led to a
player requiring treatment, with no significant
differences being observed between the four
levels of competition (table 1). In addition, no
significant correlation was found between the
number of free kicks awarded and the number
of injuries in matches at any level of competition.
Figure 1 shows the mean number of free
kicks awarded for foul play in each quarter of
play, and figs 2, 3, and 4 show the mean
number of player injuries resulting from all,
Table 1 Number of games analysed, free kicks awarded and injuries observed
Level of competition Year/ season
Games
analysed
Foul play free
kicks
Player treatments
Foul Non-foul
World Cup5 1994 44 1272 33 81
European Championship 1996 29 1011 22 58
Premier League 1996/97 49 998 17 100
First Division 1994/95 26 571 14 75
First Division 1995/96 27 604 22 54
First Division 1996/97 40 873 25 77
Figure 1 Mean number of free kicks awarded for foul play in each quarter of play.
10
8
6
4
2
0
World Cup
European Championship
Premier League
First Division
Period of play (minutes)
Mean number of free kicks per game
0–25 25–45+ 45–70 70–90+
Figure 2 Mean number of all injuries in each quarter of play.
1.0
0.8
0.6
0.4
0.2
0
World Cup
European Championship
Premier League
First Division
Period of play (minutes)
Mean number of injuries per game
0–25 25–45+ 45–70 70–90+
Table 2 Accident ratios: all injuries
Incident
severity
Level of competition
World
Cup5
European
Championship
Premier
League
First
Division
Moderate 1 1 1 1
Minor 5.7 5.7 5.3* 4.5
Non-injury 73 82 53 41
*See text for the calculation of this value.
Frequency of injuries and incidents in professional football 327
Downloaded from bjsm.bmjjournals.com on 22 October 2006
foul, and non-foul incidents respectively in
each quarter of play at each level of competition.
For consistency of reporting, free kicks
and injuries occurring during the extra time
periods played in five of the European
Championship matches are not included in the
values shown in figs 1–4.
Based on the typical 1:4:4:2 (goalkeeper,
defender, midfielder, forward) playing system
used by teams in these competitions, no
significant differences were found in the
number of incidents at any level of competition
as a function of playing position. Similarly, the
ratios of injuries arising from foul and non-foul
incidents showed no significant differences
between playing positions.
ACCIDENT RATIOS
The approach developed by Bird and Germain7
and Heinrich8 of using AR triangles to report
and analyse industrial accidents was used previously
to illustrate the ratio of injuries and
underlying non-injury incidents during the
1994World Cup.5 This approach was also used
here to show the ratio of results for moderate,
minor, and non-injury incidents. For the English
Premier and First Division leagues, where
it was not possible to distinguish between
moderate and minor injuries from media
reports, injury data were provided by physiotherapists
for teams involved in eight Premier
and 21 First Division matches analysed.
All injuries
A total of 12 moderate injuries, 68 minor injuries,
and 989 non-injury incidents were recorded
during the 1996 European Championship
(table 1). Therefore, for every moderate
injury that prevented participation for a
minimum of one game, there were 5.7 minor
injuries and 82.4 non-injury incidents which
could potentially have led to an injury (table 2).
A total of 117 moderate plus minor injuries and
981 non-injury incidents were recorded during
the 1996/97 Premier League matches; for the
1994–1997 First Division league matches, 267
moderate plus minor injuries and 1987 noninjury
incidents were recorded (table 1). For
the 29 League matches for which injury data
were provided by club physiotherapists, there
were no moderate injuries and six minor
injuries for the eight Premier League matches
and four moderate and 18 minor injuries for
the 21 First Division matches. Therefore, by
using this ratio of moderate to minor injuries of
1:4.5, it was calculated that the overall AR for
the First Division was for every moderate
injury there were 4.5 minor injuries and 40.9
non-injury incidents (table 2). As no moderate
injuries were reported in the eight Premier
matches, it is not possible to derive a direct
ratio of moderate to minor injuries; however,
the average ratio (1:5.3) observed for theWorld
Cup, European Championship, and First Division
has been used to calculate a comparable
AR for the Premier League of one moderate
injury for every 5.3 minor injuries and 52.8
non-injury incidents (table 2). These results
can be further subdivided into two categories:
those injuries resulting from foul and those
from non-foul incidents.
Injuries resulting from foul situations
In the European Championship matches, only
one incident resulted in a moderate injury, with
a further 21 minor injuries incurred from the
1011 fouls recorded. During the tournament
three players were treated in the final game of
their country, one of whom was substituted. It
was assumed in the absence of other information
that these players would have been fit if
any further games had been played and
therefore these injuries were classed as minor.
From the 998 fouls recorded during the 49
Premier League matches, 17 resulted in
moderate or minor injuries. From the 2048
Figure 3 Mean number of injuries resulting from foul play in each quarter of play.
1.0
0.8
0.6
0.4
0.2
0
World Cup
European Championship
Premier League
First Division
Period of play (minutes)
Mean number of injuries per game
0–25 25–45+ 45–70 70–90+
Figure 4 Mean number of injuries resulting from non-foul situations in each quarter of
play.
1.0
0.8
0.6
0.4
0.2
0
World Cup
European Championship
Premier League
First Division
Period of play (minutes)
Mean number of injuries per game
0–25 25–45+ 45–70 70–90+
Table 3 Accident ratios: foul and non-foul situations
Incident severity
Level of competition/incident category
World Cup5 European Championship Premier League First Division
Foul Non-foul Foul Non-foul Foul Non-foul Foul Non-foul
All injuries 1 2.5 1 2.6 1 5.9 1 3.4
Non-injury 38 — 45 — 58 — 33 —
328 Hawkins, Fuller
Downloaded from bjsm.bmjjournals.com on 22 October 2006
fouls recorded during the 93 First Division
league matches, there were 61 incidents which
resulted in moderate or minor injuries. Taking
the number of moderate and minor injuries
together produces ARs for all injury to
non-injury incidents of 1:45, 1:58, and 1:33 for
the European Championship, Premier League
and First Division respectively (table 3).
Injuries resulting from non-foul situations
Eleven injuries were judged to be at least moderate
during the European Championship, with
a further 47 minor injuries. Fifteen of the injuries
categorised as minor occurred in the final
game of the country. During the Premier
League matches, 100 injuries were judged to be
at least moderate or minor, and 206 were
judged to be at least moderate or minor during
the First Division league matches. By normalising
these results on the number of injuries
resulting from foul situations at each competitive
level, ARs of 2.6, 5.9, and 3.4 were
obtained for all injuries resulting from non-foul
incidents during the European Championship,
Premier League, and First Division matches
respectively (table 3). The number and nature
of the underlying causes of non-foul injuries
have not been identified, and therefore these
ratio figures are missing from table 3.
INJURY INCIDENCE RATES
The IIR has been defined previously5 as the
ratio of the sum of all moderate and minor
injuries recorded to the number of players taking
part in all the matches analysed, expressed
as a percentage. The calculations assumed that
22 players were present at all times during each
game. These figures therefore represent the
likelihood that any individual player will receive
medical treatment as the result of an injury
during a match at each level of competition.
There were no significant differences in IIRs
for the World Cup5 (11.8%), European Championship
(12.5%), Premier League (10.9%),
and First Division (13.0%) (table 4). There
were also no significant differences in the IIR
values as a function of match halves or playing
position (fig 5) for the three levels of competition.
INJURY FREQUENCY RATES
The IFR has been defined previously5 as the
sum of all moderate and minor injuries
recorded during the total player playing time
normalised to 100 000 playing hours. These
values are therefore consistent with criteria
used in other working environments for assessing
levels of risk. The calculations assumed that
there were 100 minutes in each standard game,
130 minutes in games where extra time was
played (except in the European Championship
final when 105 minutes were played as a result
of the golden goal ruling), and that 22 players
were involved in each match at all times. This
gives total equivalent playing hours of 1109
hours for the 29 European Championship
matches, 1797 hours for the 49 Premier
League matches, and 3410 hours for the 93
First Division matches assessed. Table 5 shows
the IFR values for injuries received in foul,
non-foul, and all categories for the European
Championship, Premier League, and First
Division matches. The overall IFR values in
professional football were not found to differ
significantly between the four levels of competition,
but IFRs were found to be significantly
higher in non-foul than in foul situations at all
levels of competition.
Table 6 summarises the impact of player to
player contact in non-foul situations, and fig 6
shows the mechanisms leading to injury
through player to player contact in non-foul
situations.
Table 4 Injury incidence rates (IIR)
Level of competition
World
Cup5
European
Championship
Premier
League
First
Division
Matches 44 29 49 93
Players 968 638 1078 2046
Injuries 114 80 117 267
IIR (%) 11.8 12.5 10.9 13.0
Table 5 Injury frequency rates per 100 000 playing hours
for all injuries
Incident
category
Injury frequency rate
World
Cup5
European
Championship
Premier
League
First
Division
Foul 1992 1984 946 1789
Non-foul 4888 5230 5565 6041
All 6880 7214 6511 7830
Figure 5 Injury incidence rates as a function of playing position.
20
15
10
5
0
World Cup
European Championship
Premier League
First Division
Playing position
Injury incidence rate (%)
Goalkeeper Defender Midfielder Attacker
Figure 6 Mechanisms of player to player contact injuries in non-foul situations.
0.6
0.4
0.2
0
World Cup
European Championship
Premier League
First Division
Injury mechanism
Mean number of injuries per game
Tackled Heading Tackling Collision Violent
conduct
Frequency of injuries and incidents in professional football 329
Downloaded from bjsm.bmjjournals.com on 22 October 2006
Discussion
This study assesses injury frequency and severity
during competitive games at three levels of
professional football, namely European international,
English Premier, and English First
Division, and complements our earlier work
covering international football at World level.5
The results and conclusions from this work
have therefore been compared with the earlier
results to present a more complete assessment
of incidents and injuries within professional
football.
The higher number of free kicks awarded on
average during World Cup and European
Championship games may be attributable to the
specific guidance provided to all referees before
these international competitions with the aim of
achieving consistency in decision making.
The ratio values for moderate to minor injuries
shown in table 2 were used to calculate
moderate IFR values from the overall values
shown in table 5: 1027, 1077, 1033, and 1424
per 100 000 playing hours for the World Cup,
European Championship, Premier League,
and First Division respectively. These compare
with previous results of 2384 for a Scottish
Premier League club,9 1690 for Swedish senior
amateur league teams,10 and 940 for international
youth (9 to 19 years) teams.11 The
present results therefore do not support previous
reports that injury levels increase with
increasing levels of competition,4 12–14 and,
within the highest levels of professional football,
the levels of risk to players appear to be
similar at the present time.
The percentage of injuries resulting from foul
situations compared with the total number of
injuries recorded (table 1) were calculated to be
29, 28, 15, and 23% for the World Cup,5 European
Championship, Premier League, and First
Division respectively. These results are very
similar to those reported previously3 12 15 where
fouls were found to be responsible for between
19 and 28% of match injuries. Over 60% of the
injuries resulting from foul situations were found
to be sustained while the injured player was
being tackled, the free kick being awarded in
favour of the injured player in over 97% of cases
at all competitive levels.
The main primary causation factor for injuries
resulting from non-foul situations was still
found to be player to player contact (table 6).
The range of results obtained for the four levels
of competition (49–76%) is consistent with
the work of Sandelin et al4 (52%) and Hoff and
Martin16 (66%). The apparently lower proportion
of injuries from this cause in the World
Cup is thought to be because more injuries
were defined in the non-contact and nonestablished
categories as a consequence of the
adverse temperature and humidity conditions
in many of those match