Sports Med 2011; 41 (11): 903-923

REVIEW ARTICLE 0112-1642/11/0011-0903/$49.95/0

ª 2011 Adis Data Information BV. All rights reserved.

A Review of Football Injuries on Third

and Fourth Generation Artificial Turfs
Compared with Natural Turf
Sean Williams,1 Patria A. Hume1 and Stephen Kara1,2
1 Sports Performance Research Institute New Zealand (SPRINZ), School of Sport and Recreation,
Auckland University of Technology, Auckland, New Zealand
2 Blues Super 14 Rugby Team, Auckland, New Zealand

Contents
Abstract. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 903
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 904
2. Literature Search Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
2.1 Search Parameters and Criteria . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
2.2 Assessment of Study Quality. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
2.3 Data Extraction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
2.4 Analysis and Interpretation of Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 905
3. Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 908
3.1 Types of Natural and Artificial Surfaces Used by Football Codes . . . . . . . . . . . . . . . . . . . . . . . . . . 908
3.2 Incidence and Nature of Injury as a Result of Playing on Natural Turf or Artificial Turf . . . . . . . . . 908
3.2.1 Ankle Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 909
3.2.2 Knee Injury . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 909
3.2.3 Muscle Strains . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 909
3.2.4 Injury Severity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 909
3.3 Mechanisms and Risk Factors for Injury on Artificial and Natural Turf . . . . . . . . . . . . . . . . . . . . . . . 913
3.3.1 Shoe-Surface Interface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 918
3.3.2 Foot Loading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.3 Impact Attenuation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.4 Physiological Response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.5 Gender. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.6 Age. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.7 Level of Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 919
3.3.8 Training and Matches . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
3.3.9 Weather . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
3.3.10 Changing between Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920
4. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 920

Abstract Football codes (rugby union, soccer, American football) train and play
matches on natural and artificial turfs. A review of injuries on different turfs
was needed to inform practitioners and sporting bodies on turf-related injury
mechanisms and risk factors. Therefore, the aim of this review was to compare
the incidence, nature and mechanisms of injuries sustained on newer generation
904 Williams et al.

artificial turfs and natural turfs. Electronic databases were searched using the
keywords ‘artificial turf’, ‘natural turf’, ‘grass’ and ‘inj*’. Delimitation of 120
articles sourced to those addressing injuries in football codes and those using
third and fourth generation artificial turfs or natural turfs resulted in 11 ex-
perimental papers. These 11 papers provided 20 cohorts that could be assessed
using magnitude-based inferences for injury incidence rate ratio calculations
pertaining to differences between surfaces. Analysis showed that 16 of the 20
cohorts showed trivial effects for overall incidence rate ratios between sur-
faces. There was increased risk of ankle injury playing on artificial turf in eight
cohorts, with incidence rate ratios from 0.7 to 5.2. Evidence concerning risk of
knee injuries on the two surfaces was inconsistent, with incidence rate ratios
from 0.4 to 2.8. Two cohorts showed beneficial inferences over the 90% like-
lihood value for effects of artificial surface on muscle injuries for soccer
players; however, there were also two harmful, four unclear and five trivial
inferences across the three football codes. Inferences relating to injury severity
were inconsistent, with the exception that artificial turf was very likely to have
harmful effects for minor injuries in rugby union training and severe injuries in
young female soccer players. No clear differences between surfaces were evi-
dent in relation to training versus match injuries. Potential mechanisms for
differing injury patterns on artificial turf compared with natural turf include
increased peak torque and rotational stiffness properties of shoe-surface in-
terfaces, decreased impact attenuation properties of surfaces, differing foot
loading patterns and detrimental physiological responses. Changing between
surfaces may be a precursor for injury in soccer. In conclusion, studies have
provided strong evidence for comparable rates of injury between new gen-
eration artificial turfs and natural turfs. An exception is the likely increased
risk of ankle injury on third and fourth generation artificial turfs. Therefore,
ankle injury prevention strategies must be a priority for athletes who play on
artificial turf regularly. Clarification of effects of artificial surfaces on muscle
and knee injuries are required given inconsistencies in incidence rate ratios
depending on the football code, athlete, gender or match versus training.

1. Introduction although injury patterns differ on new generation

In football codes, such as rugby union, soccer
and American football, training and matches are
now being played on both natural and artificial
turf surfaces. First and second generation artifi-
cial surfaces have been associated with an in-
creased injury risk versus natural grass surfaces
across a number of sports.[1-5] The properties of
first and second generation turfs differ distinctly
from recent third and fourth generation sur-
faces.[6] The drawbacks of earlier generation ar-
tificial surfaces, such as hardness and excessive
heat retention, are purported to have been ad-
dressed in newer generation surfaces.[7] A recent
review by Dragoo and Braun[8] concluded that
turfs, the overall injury rate is comparable with
natural turfs. Our review expands on Dragoo and
Braun’s by using magnitude-based inferences to
uniformly analyse studies. We have also analysed
several additional papers that will add to the
knowledge base. Our question is whether third
and fourth generation artificial turfs are asso-
ciated with increased injury risk versus natural
grass surfaces across three football codes. Rugby,
soccer (male and female) and American foot-
ball have dissimilar playing styles and may have
different injury incidence patterns on different
surfaces.
The aim of this review was to compare the
incidence, nature and mechanisms of injuries

ª 2011 Adis Data Information BV. All rights reserved. Sports Med 2011; 41 (11)
Artificial Turf Injury Review
sustained on newer generation artificial turfs and
natural turfs.
2. Literature Search Methodology
Cochrane Collaboration[9] review methodology
(literature search; assessment of study quality; data
collection of study characteristics including par-
ticipants, sport, outcome measures and results;
analysis and interpretation of results; and rec-
ommendations for injury prevention strategies
and further research) was used to evaluate injury
characteristics and risk factors for injury on arti-
ficial turfs compared with natural grass turf.
2.1 Search Parameters and Criteria
Web of Knowledge, Scopus, MEDLINE,
SportDiscus, ProQuest Direct, Google Scholar,
CINAHL and Scirus databases from 1975 to
November 2010 were searched using the key-
words ‘artificial turf’, ‘natural turf’, ‘grass’ and
‘inj*’. Inclusion criteria for the article were provided
as follows: (i) data for injury on natural or artificial
turf in football codes, including rugby union, soccer
and American football; (ii) relevant information for
epidemiology, possible mechanisms and risk factors
of injury on natural or artificial turfs in general; or
(iii) football studies that may provide insight to
possible mechanisms or risk factors associated with
injury on turf. Exclusion criteria were as follows:
(i) unavailable in English and not previously re-
ferred to by other sources; or (ii) not specific to
third or fourth generation artificial turf and did
not add knowledge to the aim of the manuscript.
Additional supportive articles were sought through
article reference lists and a further search using the
key words ‘torque’, ‘stiffness’ and ‘surface*’ to find
papers attending to potential mechanisms. Of the
120 articles sourced, 11 experimental papers pro-
vided data for 20 cohorts that could be assessed
using magnitude-based inferences for injury inci-
dence rate ratio calculations for differences between
new generation artificial and natural turfs.
2.2 Assessment of Study Quality
Methodological limitations were associated
with many of the studies reviewed; namely, a
ª 2011 Adis Data Information BV. All rights reserved.
905
failure to clearly describe the specific type of ar-
tificial turf used, the condition of the natural turf
surface, the characteristics of the cohort, the en-
vironmental conditions, specifications of footwear
used or the p-value associated with the outcome
measure. Variations in injury definitions and se-
verity were encountered, while the data quality
was further reduced by a lack of uniform collec-
tion methods. Many studies in the literature had
made inferences about injury risks based only on
the p-value derived from a null hypothesis test.
This can result in misleading conclusions being
made, depending on the magnitude of the effect
statistic, sample size and error of measurement.[10]
One study in American football[11] was funded by
the artificial turf manufacturer used within the
study and so its conclusions should be treated with
caution. An unsponsored study that also investi-
gated the effects of new generation artificial turf
on injury risk in American football found higher
rates of anterior cruciate ligament injury and
ankle eversion sprains on artificial turf.[12] Un-
fortunately, this study has only been published as
an abstract and so we were unable to include their
study in our data analyses.
2.3 Data Extraction
For studies passing the quality criteria data were
extracted, including participant characteristics, inci-
dence and nature of injuries on natural or artificial
turfs, and main findings (table I). Note that for
analysis of type of injury and risk factors the
number of cohorts was less than the total of 20
cohorts used for the overall analysis, given that
some studies did not report specific injury data.
2.4 Analysis and Interpretation of Results
The outcome measure used to assess each study
was the incidence rate ratio for injuries on artifi-
cial and natural turf, calculated using natural turf
as the reference. Several studies provided more
than one cohort, such as males versus females or
matches versus training information. Clinical in-
ferences regarding the true value of effects were
made in a manner outlined by Batterham and
Hopkins.[10] Where provided, the p-value relating
to the outcome measure (incidence rate ratio) was
Sports Med 2011; 41 (11)
 Table I. Study characteristics, incidence of injuries, incidence rate ratios and 90% confidence intervals, and percentage likelihoods for beneficial, trivial or harmful effects when
ª 2011 Adis Data Information BV. All rights reserved.

906
comparing injury incidence on artificial and natural turfs surfaces for soccer, rugby union and American football
Study No. of subjects or Level of Training Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries natural artificial substantially trivial substantially
turf turf beneficial harmful
Soccer
Ekstrand et al.[13] 613 males Elite Match 21.72 22.37 1.03 0.13 Most likely trivial 0.0 99.9 0.1
Age 25 – 5 y

Soligard et al.[14] ~60 000 players Regional Match 39.70 34.20 0.93 0.15 Very likely trivial 2.4 97.6 0.0
(~one-third were
female)
Age 13–19 y

Steffen et al.[15] 2020 females Regional Match 8.30 8.70 1.05 0.20 Very likely trivial 0.6 95.2 4.2
Age 15 – 1 y

Bjorneboe et al.[16] Males Elite Match 17.00 17.60 1.04 0.16 Very likely trivial 0.1 99.9 0.0
14 teams
No age data

Fuller et al.[17] Males Collegiate Match 23.92 25.43 1.06 0.15 Very likely trivial 0.0 99.2 0.8
2005 season: 52 teams
2006 season: 54 teams
No age data

Fuller et al.[17] Females Collegiate Match 21.79 19.15 0.88 0.13 Likely trivial 7.3 92.7 0.0
2005 season: 64 teams
2006 season: 72 teams
No age data

Ekstrand et al.[18] 492 males Elite Match 21.48 19.60 0.91 0.18 Likely trivial 8.4 91.4 0.2
Age 25 – 5 y

Ekstrand et al.[13] 154 females Elite Match 12.51 14.88 1.19 0.40 Unclear; get 1.6 65.3 33.1
Age 23 – 4 y more data

Fuller et al.[19] Females Collegiate Training 2.79 2.60 0.93 0.15 Very likely trivial 2.6 97.3 0.1
2005 season: 64 teams
2006 season: 72 teams
No age data
Sports Med 2011; 41 (11)

Fuller et al.[19] Males Collegiate Training 3.01 3.34 1.11 0.16 Very likely trivial 0.0 96.9 3.1
2005 season: 52 teams

Williams et al.
2006 season: 54 teams
No age data

Continued next page

 Table I. Contd
ª 2011 Adis Data Information BV. All rights reserved.

Artificial Turf Injury Review

Study No. of subjects or Level of Training Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries natural artificial substantially trivial substantially
turf turf beneficial harmful
Bjorneboe et al.[16] Males Elite Training 1.80 1.90 1.07 0.19 Very likely trivial 0.1 96.5 3.4
14 teams
No age data

Ekstrand et al.[13] 613 males Elite Training 3.47 3.52 1.02 0.16 Very likely trivial 0.2 99.2 0.6
Age 25 – 5 y

Aoki et al.[7] 301 players Unclear Training 4.47 3.80 1.18 0.19 Likely trivial 0.0 83.6 16.4
Age 15 – 2 y
No gender data

Steffen et al.[15] 2020 females Regional Training 1.20 1.20 1.00 0.40 Possibly trivial 13.1 73.8 13.1
Age 15 – 1 y

Ekstrand et al.[18] 290 males Elite Training 2.94 2.42 0.82 0.22 Possibly 34.6 65.2 0.2
Age 25 – 5 y beneficial; use

Ekstrand et al.[13] 154 females Age 23 – 4 y Elite Training 2.79 2.91 1.04 0.44 Possibly trivial 11.3 70.2 18.5

American football
Meyers[11] 465 games Collegiate Match 51.20b 45.70b 0.89 0.03 Most likely trivial 0.0 100.0 0.0
No gender or age data

Meyers and 240 games High school Match 13.90b 15.20b 1.09 0.16 Very likely trivial 0.0 97.5 2.5
Barnhill[20] No gender or age data

Rugby
Fuller et al.[21] 282 males Community Match 26.90 38.20 1.42 0.56 Possibly harmful; 0.4 34.8 64.8
129 backs, age 26 – 4 y don’t use
153 forwards,
age 27 – 6 y
Sports Med 2011; 41 (11)

Fuller et al.[21] 169 males Elite Training 2.30 3.00 1.33 0.54 Possibly harmful; 1.1 45.1 53.8
85 backs, age 25 – 4 y don’t use
84 forwards, age 26 – 8 y
a Incidence rate ratio for injury incidence on artificial and natural turf surfaces, using natural turf as the reference group.
b Incidence given as injuries per ten team games = (n of injuries/n of team games) · 10.
~ indicates approximately.

907
908
used to determine the likelihood that the true
magnitude of the effect was substantial in a bene-
ficial or harmful way. An incidence rate ratio of
0.77 represented a substantial benefit of playing
on artificial turf, while an incidence rate ratio of
1.30 indicated that it was substantially harmful to
play on artificial turf.[22] These thresholds were
chosen as they correspond approximately to the
thresholds for standardized differences in means
(0.20, 0.60, 1.2, 2.0 and 4.0) of the log of time to
injury in the two groups.[23] When investigating
potential mechanisms, the standardized difference
between means for the variable on each surface
was used, with thresholds of –0.2 used to deter-
mine a meaningful difference.[22] Where a study
reported a p-value as ‘p < 0.05’, ‘p = 0.05’ was used
in the analysis. Where a rate ratio was reported with
confidence limits, inferences were calculated using a
spreadsheet for combining independent groups,
with a weighting factor of one for the effect.[24] An
effect was clinically unclear if there was a >25%
likelihood that the true value was beneficial, with
odds of benefit relative to odds of harm (odds ratio)
<66. The effect was otherwise clinically clear: bene-
ficial if the likelihood of benefit was >25%, and
trivial or harmful for other outcomes, depending on
the observed value. The likelihood that an effect
was substantially harmful, trivial or beneficial was
given in plain language terms[10] using the following
scale: 0–0.5%, most unlikely; 0.6–5.0%, very un-
likely; 5.1–25.0%, unlikely; 25.1–75.0%, possible;
75.1–95.0%, likely; 95.1–99.5%, very likely; and
99.6–100%, most likely.[25]
3. Findings
3.1 Types of Natural and Artificial Surfaces
Used by Football Codes
Football codes use two main types of surfaces:
natural (grass) turf or artificial surfaces. The per-
formance of either type of surface is dependent
upon their structural characteristics, response to
physical wear (e.g. shoe-surface interaction and
frequency of use) and environmental exposure
(e.g. sun and rain). Ekstrand and Nigg[1] suggested
that 24% of soccer injuries could be attributed to
unsatisfactory playing surfaces.
ª 2011 Adis Data Information BV. All rights reserved.
Williams et al.
Many species of grass may be used for natural
turf surfaces in sport, all of which possess different
properties pertaining to shoe-surface traction.[26]
Natural turf’s response to wear is determined by the
species of grass, surface compaction and drainage
ability.[27] Properties of natural turf surfaces that
may be connected with injury prevalence include
inappropriate friction characteristics, hardness and
being uneven.[1] Natural turf requires considerable
maintenance throughout the year, including mow-
ing, fertilizing, irrigation, aeration, reseeding and
control against pests, weeds and disease.[6]
The structure of artificial turf has developed
since the initial use of first generation Astroturf
in the late 1960s.[28] Astroturf consisted of a short
grass fibre carpet on top of padding over concrete
and possessed increased stiffness, heat retention
and sliding friction in comparison to natural grass
surfaces.[6] Second generation turfs, developed in
the late 1980s, were characterized by longer, thicker
fibres (22–25 mm), sand fillings and a rubber base
under the turf itself to reduce stiffness.[15] A third
generation of turf with longer fibres (50–60 mm)
and a sand and/or rubber infill was designed for
soccer and is believed to more accurately mimic the
characteristics of natural turf.[29] Fieldturf (a fourth
generation turf) infill consists of a bottom layer of
silica sand, a middle layer that is a combination of
cryogenic rubber and sand and a top layer of rub-
ber.[30] No studies to date have compared injury
risk on third versus fourth generation turfs.
Artificial turfs possess various advantages
over natural turfs, including lower maintenance
costs, greater utility and the ability to be used in
climates where the development of natural turf is
difficult.[18] Use of artificial turf has grown con-
siderably in recent years with its use in matches
approved by the Fédération Internationale de
Football Association (FIFA),[31] as well as by the
International Rugby Union.[6] Artificial turf is
also used in over one-third of the stadiums for the
US National Football League.[32]
3.2 Incidence and Nature of Injury as a Result
of Playing on Natural Turf or Artificial Turf
The incidence rate ratio for injuries on artifi-
cial and natural turf surfaces, using natural turf
Sports Med 2011; 41 (11)
Artificial Turf Injury Review
as the reference, ranged from 0.82 to 1.42, with
the highest rate ratio for rugby (table I). Trivial
effects for the injury rate between surfaces were
shown by 16 of the 20 cohorts, with 13 cohorts
showing trivial percentage likelihoods over 90%.
However, our analysis of Fuller et al.’s study[21]
resulted in the discovery of possibly harmful ef-
fects for rugby union athletes playing matches
and training on artificial turf. Conversely, anal-
ysis of Ekstrand et al.’s study resulted in the dis-
covery of a possibly beneficial effect for elite male
soccer players in training but not matches.[18] Gi-
ven the possible harmful effects for rugby versus
soccer on artificial turf, further research is needed
to investigate the mechanisms that may underlie
these differences in order to reduce the risk of in-
jury when playing rugby on artificial turf.
3.2.1 Ankle Injury
Our analyses showed evidence for an increased
risk of incurring an ankle injury when playing on
artificial turf in 8 of the 14 cohorts, with incidence
rate ratios from 0.71 to 5.20 (see table II). How-
ever, none of the percentage of likelihood cate-
gories had values >95% (very likely harmful).
Evidence of a harmful effect of incurring ankle
injuries whilst playing soccer on artificial turf was
found for elite males in matches and train-
ing,[13,16,18] elite females in training,[13] collegiate
males in training[19] and young females during
matches.[15] A trivial effect was calculated for fe-
male soccer players in training[19] and collegiate
male soccer players during matches.[17] Con-
versely, a beneficial effect was inferred for soccer
matches involving youths[14] and collegiate fe-
males (unlike the trivial effect during training).[17]
A likely harmful effect was calculated for rugby
union match injuries.[21] Unclear effects were cal-
culated for elite female soccer matches[13] and elite
male soccer training.[18] Given the likely increased
risk of incurring an ankle injury on artificial turf,
injury prevention strategies to prevent ankle in-
jury must be a priority for soccer players who
train and play matches on artificial turf regularly.
3.2.2 Knee Injury
Overall, the evidence concerning the risk of
knee injuries on the two surfaces was inconsistent
ª 2011 Adis Data Information BV. All rights reserved.
909
with incidence rate ratios from 0.4 to 2.8 (see
table II). An unclear inference was made across
6[11,13,17,18] of the 16 cohorts analysed. A likely
beneficial inference was calculated for high school
American footballers[20] (the only likelihood over
90% for knee injuries), and a possibly beneficial
effect was calculated for female collegiate soccer
players in training.[19] A likely harmful effect was
inferred for community-level male rugby union
players during matches[21] and young female
soccer players in matches,[15] while a possibly
harmful effect was calculated for elite male soccer
players in matches.[16] Trivial effects were inferred
for male and female collegiate soccer players
during matches,[17] male collegiate soccer players
during training,[19] youth soccer players[14] and
elite male soccer players in training.[13] Given the
inconsistencies in incidence rate ratios depending
on the sport, gender of athlete or match versus
training, more research is required to elucidate the
effect of surface type on knee injuries.
3.2.3 Muscle Strains
Two cohorts[15,18] showed beneficial inferences
over the 90% likelihood value for the effect of
artificial surface on muscle injuries for soccer
players (see table II). A likely beneficial effect was
also calculated for elite male soccer players in
matches,[18] while possibly beneficial effects were
calculated for three other soccer cohorts.[13] High
school American footballers[20] had likely harm-
ful inferences and elite rugby union players had
possibly harmful inferences during training.[21]
Given the six beneficial, two harmful, four un-
clear and five trivial inferences across the three
sports, more research is needed to clarify the ef-
fect of artificial surfaces on muscle strain injury,
particularly given soccer players may have a re-
duced risk of muscle strain injury on artificial sur-
faces compared with American football or rugby.
3.2.4 Injury Severity
Many reviewed studies attempted to quantify
and describe differences between surfaces with
regards to the varying degrees of injury severity
(see table III). However, the range of definitions
used to describe severity made comparisons be-
tween studies difficult. Ekstrand et al.,[13,18] Fuller
Sports Med 2011; 41 (11)
 Table II. Study characteristics, injured body part or injury type, incidence of injuries, incidence rate ratios and 90% confidence intervals, and percentage likelihoods for beneficial,
ª 2011 Adis Data Information BV. All rights reserved.

910
trivial or harmful effects when comparing injury incidence on artificial and natural turfs surfaces for soccer, rugby union and American football
Study No. of subjects or Level of Training or Injured body Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance match part or injury (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries type natural artificial substantially trivial substantially
turf turf beneficial harmful
Soccer
Ekstrand 492 males Elite Match Ankle 2.66 4.83 1.81 0.94 Likely harmful; 0.2 13.5 86.3
et al.[18] Age 25 – 5 y don’t use
Knee 2.66 2.07 0.78 0.49 Unclear; get 48.5 43.6 7.9
more data
Muscle strain 6.16 3.76 0.60 0.25 Likely beneficial; 83.9 16.0 0.1
use
Soligard ~60 000 players Regional Match Ankle 8.40 4.30 0.59 0.20 Likely beneficial; 90.6 9.4 0.0
et al.[14] (~one-third were use
female) Knee 5.60 4.60 0.96 0.32 Likely trivial 13.3 80.3 6.4
Age 13–19 y
Muscle strain 3.00 2.20 0.88 0.43 Unclear; get 31.8 59.8 8.4
more data
Bjorneboe Males Elite Match Ankle 2.20 3.10 1.39 0.55 Possibly harmful; 0.6 38.2 61.3
et al.[16] 14 teams don’t use
No age data Knee 2.00 3.00 1.46 0.59 Possibly harmful; 0.4 30.9 68.7
don’t use
Muscle strain 5.10 4.50 0.88 0.28 Possibly trivial 23.6 74.5 1.8
Fuller Collegiate males Collegiate Match Ankle 4.57 4.59 1.00 0.33 Likely trivial 9.1 81.8 9.1
et al.[17] 2005 season: 52 teams Knee 3.09 3.75 1.21 0.40 Unclear; get 1.1 63.2 35.7
2006 season: 54 teams more data
No age data
Muscle strain 6.47 5.70 0.88 0.25 Likely trivial 21.8 77.0 1.2
Fuller Collegiate females Collegiate Match Ankle 4.21 3.00 0.71 0.28 Possibly 63.4 36.1 0.5
et al.[17] 2005 season: 64 teams beneficial; use
2006 season: 72 teams Knee 4.94 4.86 0.98 0.31 Likely trivial 9.9 83.5 6.6
No age data
Muscle strain 3.17 3.57 1.13 0.42 Unclear; get 4.1 69.5 26.4
more data
Steffen 2020 females Regional Match Ankle 3.00 4.00 1.40 0.50 Possibly harmful; 0.2 35.6 64.2
et al.[15] Age 15 – 1 y don’t use
Sports Med 2011; 41 (11)

Knee 1.10 1.90 1.70 0.80 Likely harmful; 0.3 17.3 82.4
don’t use

Williams et al.
Muscle strain 1.50 0.60 0.40 0.30 Likely beneficial; 93.4 6.3 0.3
use

Continued next page

 Table II. Contd
ª 2011 Adis Data Information BV. All rights reserved.

Artificial Turf Injury Review

Study No. of subjects or Level of Training or Injured body Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance match part or injury (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries type natural artificial substantially trivial substantially
turf turf beneficial harmful
Ekstrand 154 females Elite Match Ankle 2.63 2.89 1.10 0.90 Unclear; get 21.5 43.0 35.6
et al.[13] Age 23 – 4 y more data
Knee 2.30 3.55 1.54 1.26 Unclear; get 6.4 29.1 64.5
more data
Muscle strain 2.96 3.55 1.20 0.89 Unclear; get 14.3 43.4 42.4
more data

Ekstrand 613 males Elite Match Ankle 3.53 4.80 1.36 0.41 Possibly harmful; 0.1 40.1 59.8
et al.[13] Age 25 – 5 y don’t use
Knee 4.34 4.62 1.06 0.30 Likely trivial 3.1 85.2 11.7
Muscle strain 7.44 5.32 0.72 0.17 Possibly 67.7 32.3 0.0
beneficial; use

Ekstrand 492 males Elite Training Ankle 0.33 0.53 1.61 1.30 Unclear; get 5.0 26.7 68.3
et al.[18] Age 25 – 5 y more data
Knee 0.33 0.31 0.95 0.83 Unclear; get 33.0 41.2 25.7
more data
Muscle strain 1.31 0.62 0.48 0.22 Very likely 95.9 4.1 0.0
beneficial; use

Fuller Collegiate males Collegiate Training Ankle 0.58 0.83 1.44 0.41 Possibly harmful; 0.0 27.4 72.6
et al.[19] 2005 season: 52 teams don’t use
2006 season: 54 teams Knee 0.43 0.42 0.99 0.38 Possibly trivial 13.6 74.6 11.8
No age data
Muscle strain 1.16 1.26 1.08 0.24 Likely trivial 0.6 90.9 8.5

Fuller Collegiate females Collegiate Training Ankle 0.45 0.45 1.00 0.40 Possibly trivial 13.5 73.0 13.5
et al.[19] 2005 season: 64 teams Knee 0.54 0.40 0.74 0.31 Possibly 56.3 42.6 1.1
2006 season: 72 teams beneficial; use
No age data
Muscle strain 1.21 1.04 0.86 0.22 Likely trivial 23.1 76.6 0.3

Ekstrand 154 females Elite Training Ankle 0.15 0.76 5.20 13.37 Likely harmful; 3.0 5.6 91.3
et al.[13] Age 23 – 4 y don’t use
Sports Med 2011; 41 (11)

Knee 0.29 0.56 1.90 2.96 Unclear; get 11.2 19.3 69.5
more data
Muscle strain 1.62 1.00 0.62 0.38 Possibly 73.1 25.2 1.7
beneficial; use

Continued next page

911
 Table II. Contd
ª 2011 Adis Data Information BV. All rights reserved.

912
Study No. of subjects or Level of Training or Injured body Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance match part or injury (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries type natural artificial substantially trivial substantially
turf turf beneficial harmful
Ekstrand 613 males Elite Training Ankle 3.24 4.45 1.37 0.43 Possibly harmful; 0.1 38.9 61.0
et al.[13] Age 25 – 5 y don’t use
Knee 3.83 3.99 1.04 0.32 Likely trivial 4.9 84.1 11.0
Muscle strain 1.39 1.13 0.81 0.21 Possibly 37.1 62.8 0.1
beneficial; use

Rugby
Fuller 282 males Community Match Ankle n=1 n=5 3.82 11.19 Likely harmful; 7.2 9.1 83.7
et al.[21] 129 backs don’t use
Age 26 – 4 y Knee n=3 n = 11 2.80 3.62 Likely harmful; 2.4 9.6 88.0
153 forwards don’t use
Age 27 – 6 y
Muscle strain 35.70b 32.70b 0.92 0.48 Unclear; get 28.1 58.8 13.1
more data

Fuller 169 males Elite Training Muscle strain 51.40b 66.70b 1.30 0.36 Possibly harmful; 0.1 49.9 50.0
et al.[21] 85 backs don’t use
Age 25 – 4 y
84 forwards
Age 26 – 8 y

American football
Meyers 240 high school games High school Match Knee 1.00c 0.40c 0.40 0.35 Likely beneficial; 91.3 8.0 0.7
and No gender or age data use
Barnhill[20] Muscle strain 1.10c 2.10c 1.91 1.09 Likely harmful; 0.3 12.0 87.7
don’t use

Meyers[11] 465 collegiate games Collegiate Match Knee 1.30c 1.00c 0.77 0.37 Unclear; get 50.0 47.0 3.0
Sports Med 2011; 41 (11)

were evaluated more data

No gender or age data Muscle strain 7.20c 6.20c 0.86 0.09 Likely trivial 5.1 94.9 0.0

Williams et al.
a Incidence rate ratio for injury incidence on artificial and natural turf surfaces, using natural turf as the reference group.
b Incidence given as proportion (%) of injuries.
c Injury incidence rate = (n of injuries/total n of injuries) · 10.
~ indicates approximately.
Artificial Turf Injury Review
et al.[17,19] and Soligard et al.[14] divided injuries
into four categories of severity according to the
length of absence from matches and training ses-
sions: slight (1–3 days); minor (4–7 days); mod-
erate (8–28 days); and severe (>28 days). Fuller
et al.,[21] in a study of rugby union match injuries,
also used this set of definitions, except for quan-
tifying a slight injury as one that caused 2–3 days
of time loss. Steffen et al.,[15] Fuller et al.[21] and
Bjorneboe et al.[16] used three categories of severity:
minor (1–7 days); moderate (8–21 days); and severe
(>21 days). Meyers[11] and Meyers and Barnhill[20]
used comparable definitions: minor (0–6 days);
moderate (7–21 days); and severe (>22 days).
The difference in incidence of slight injuries
between natural and artificial surfaces was as-
sessed as likely trivial for female collegiate soccer
players during matches,[17] and possibly harmful
for the same group during training.[19] A likely
trivial effect was calculated for elite male soccer
players in training and matches.[13] A likely trivial
inference was also calculated for male collegiate
soccer players during matches,[17] while a possibly
beneficial difference was calculated for this co-
hort during training.[19] A possibly beneficial ef-
fect was calculated for elite male soccer players.[18]
Unclear effects were calculated in five co-
horts.[13,14,18,21] No harmful effects were calcu-
lated in any of the soccer cohorts, suggesting that
artificial turf does not have a negative influence
for slight injury. More studies are required to
clarify the effect of surface for rugby union.
For minor injuries, our analysis showed trivial
results in 6 of the 17 cohorts: namely, collegiate
male soccer matches,[17] collegiate American foot-
ballers[11] and elite male soccer players in training
and matches.[13,16,18] A likely beneficial inference
was made for collegiate female soccer players in
training and matches,[17,19] while a possibly bene-
ficial effect was inferred for young female soccer
players.[15] A very likely harmful effect was calcu-
lated for minor injuries in training within a cohort
of rugby union players,[21] while a likely harmful
effect was found for male collegiate soccer players
participating in training.[19] An unclear inference
was made for rugby union match injuries,[21] high
school American football games,[20] elite female
soccer players in matches and training,[13] elite
ª 2011 Adis Data Information BV. All rights reserved.
913
male soccer matches[18] and youth level soccer
players.[14] The inconsistent results preclude a clear
conclusion with respect to minor injuries.
Evidence concerning moderate injuries was
similarly mixed, with trivial effects calculated
for 7 of the 17 cohorts,[13,15-19] while possibly
harmful inferences were made for 2 cohorts.[19,21]
Conversely, a likely beneficial inference was cal-
culated for elite soccer players,[18] collegiate
American footballers[11] and youth soccer play-
ers,[14] while a possibly beneficial inference was
made for male collegiate soccer players during
matches.[17] An unclear inference was made for
high school American football games,[20] rugby
training,[21] and elite female soccer players in
matches and training.[13] There appears to be no
clear pattern for moderate injury severity.
For severe injury, a very likely harmful infer-
ence was calculated for young female soccer
players.[15] A possibly harmful inference was
made for collegiate males during matches[17] and
elite male soccer players in matches.[16] A likely
trivial effect was calculated for female collegiate
soccer players in training.[19] Likely beneficial
inferences were calculated for both the American
football cohorts[11,20] and elite males in soccer
matches,[13] while clinically unclear inferences were
made in the remaining ten cohorts.[13,14,17-19,21] The
majority of unclear references suggest there is no
clear pattern, although the very likely harmful in-
ference poses questions regarding the role of arti-
ficial turf in the aetiology of severe injuries for
young female soccer players.
Inferences made relating to severities of injury
were inconsistent across cohorts, although certain
findings warrant further investigation; namely,
the very likely harmful effects calculated for
minor injuries in rugby union training[21] and for
severe injuries in young female soccer players.[15]
3.3 Mechanisms and Risk Factors for Injury
on Artificial and Natural Turf
Whiting and Zernicke[33] defined injury me-
chanism as ‘‘the fundamental physical process
responsible for a given action, reaction or result’’
(i.e. the mechanism of injury is the physical action
or cause of injury). Mechanisms of injury are
Sports Med 2011; 41 (11)
 Table III. Study characteristics, injury severity, incidence of injuries, incidence rate ratios and 90% confidence intervals, and percentage likelihoods for beneficial, trivial or harmful
ª 2011 Adis Data Information BV. All rights reserved.

914
effects when comparing injury incidence on artificial and natural turfs surfaces for soccer, rugby union and American football
Study No. of subjects or Level of Training Injury Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match severity (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries (d) natural artificial substantially trivial substantially
turf turf beneficial harmful
Soccer
Fuller Males Collegiate Match Slight (1–3) 7.80 8.34 1.07 0.26 Likely trivial 1.2 89.6 9.2
et al.[17] 2005 season: 52 teams Minor (4–7) 6.91 7.37 1.07 0.28 Likely trivial 1.7 87.9 10.4
2006 season: 54 teams
Moderate 6.19 5.00 0.81 0.25 Possibly 39.1 60.4 0.5
No age data
(8–28) beneficial; use
Severe 2.81 4.17 1.49 0.53 Possibly harmful; 0.1 26.0 73.9
(>28) don’t use

Fuller Females Collegiate Match Slight (1–3) 6.36 6.29 0.99 0.27 Likely trivial 6.0 89.3 4.7
et al.[17] 2005 season: 64 teams Minor (4–7) 6.25 3.86 0.62 0.21 Likely beneficial; 85.2 14.8 0.0
2006 season: 72 teams use
No age data
Moderate 3.92 4.14 1.06 0.36 Likely trivial 5.8 78.3 15.9
(8–28)
Severe 4.75 4.00 0.84 0.28 Unclear; get 33.0 65.5 1.5
(>28) more data

Ekstrand 492 males Elite Match Slight (1–3) 5.83 4.97 0.85 0.34 Unclear; get 33.6 62.8 3.6
et al.[18] Age 25 – 5 y more data
Minor (4–7) 6.66 6.07 0.91 0.34 Possibly trivial 22.3 72.4 5.3
Moderate 6.83 6.35 0.93 0.34 Likely trivial 18.9 75.1 6.0
(8–28)
Severe 2.16 2.21 1.02 0.67 Unclear; get 22.5 51.7 25.8
(>28) more data

Ekstrand 613 males Elite Match Slight (1–3) 6.41 7.34 1.15 0.27 Likely trivial 0.2 80.6 19.2
et al.[13] Age 25 – 5 y Minor (4–7) 6.19 6.13 0.99 0.24 Likely trivial 4.4 92.3 3.3
Sports Med 2011; 41 (11)

Moderate 6.63 6.07 0.92 0.22 Likely trivial 10.6 88.6 0.8
(8–28)

Williams et al.
Severe 2.36 1.97 0.83 0.35 Unclear; get 38.0 58.5 3.5
(>28) more data

Continued next page

 Table III. Contd
ª 2011 Adis Data Information BV. All rights reserved.

Artificial Turf Injury Review

Study No. of subjects or Level of Training Injury Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match severity (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries (d) natural artificial substantially trivial substantially
turf turf beneficial harmful
Ekstrand 154 females Elite Match Slight (1–3) 3.95 4.89 1.24 0.78 Unclear; get 9.8 46.0 44.8
et al.[13] Age 23 – 4 y more data
Minor (4–7) 2.63 4.00 1.52 1.15 Unclear; get 5.4 30.3 66.8
more data
Moderate 2.96 4.44 1.50 1.07 Unclear; get 4.8 31.3 63.9
(8–28) more data
Severe 2.96 1.55 1.52 0.48 Likely beneficial; 78.4 18.3 3.3
(>28) use

Soligard ~60 000 players Regional Match Slight (1–3) 2.70 2.80 1.12 0.51 Unclear; get 8.1 63.0 28.9
et al.[14] (~one-third were more data
female) Minor (4–7) 0.70 1.00 1.50 1.19 Unclear; get 6.5 30.8 62.7
Age 13–19 y more data
Moderate 0.70 0.20 0.28 0.70 Likely beneficial; 84.3 9.4 6.3
(8–28) use
Severe 0.40 0.20 0.49 1.25 Unclear; get 67.2 16.0 16.7
(>28) more data

Fuller Males Collegiate Training Slight (1–3) 1.27 1.03 0.81 0.19 Possibly 39.2 60.8 0.0
et al.[19] 2005 season: 52 teams beneficial; use
No age data Minor (4–7) 0.62 0.97 1.58 0.42 Likely harmful; 0.0 11.3 88.7
don’t use
Moderate 0.59 0.85 1.44 0.41 Possibly harmful; 0.0 27.4 72.6
(8–28) don’t use
Severe 0.52 0.41 0.79 0.31 Unclear; get 45.4 53.1 1.5
(>28) more data

Ekstrand 492 males Elite Training Slight (1–3) 1.09 0.74 0.68 0.32 Possibly 67.4 31.7 0.9
et al.[18] Age 25 – 5 y beneficial; use
Minor (4–7) 0.54 0.64 1.18 0.75 Unclear; get 11.9 48.7 39.4
more data
Sports Med 2011; 41 (11)

Moderate 1.09 0.68 0.63 0.30 Likely beneficial; 76.0 23.5 0.5
(8–28) use
Severe 0.22 0.35 1.61 1.67 Unclear; get 9.0 25.9 65.1
(>28) more data

Continued next page

915
 Table III. Contd
ª 2011 Adis Data Information BV. All rights reserved.

916
Study No. of subjects or Level of Training Injury Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match severity (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries (d) natural artificial substantially trivial substantially
turf turf beneficial harmful
Fuller Females Collegiate Training Slight (1–3) 0.81 1.04 1.29 0.34 Possibly harmful; 0.1 51.8 48.1
et al.[19] 2005 season: 64 teams don’t use
2006 season: 72 teams Minor (4–7) 0.68 0.36 0.53 0.23 Likely beneficial; 92.6 7.4 0.0
No age data use
Moderate 0.54 0.49 0.91 0.35 Possibly trivial 22.9 71.3 5.8
(8–28)
Severe 0.67 0.62 0.92 0.31 Likely trivial 18.8 76.8 4.4
(>28)
Ekstrand 613 males Elite Training Slight (1–3) 1.08 1.16 1.08 0.31 Likely trivial 2.4 83.6 14.0
et al.[13] Age 25 – 5 y Minor (4–7) 0.88 0.80 0.91 0.30 Likely trivial 19.4 77.2 3.3
Moderate 0.98 1.08 1.11 0.33 Likely trivial 2.1 78.9 19.0
(8–28)
Severe 0.49 0.40 0.83 0.37 Unclear; get 38.7 56.7 4.6
(>28) more data
Ekstrand 154 females Elite Training Slight (1–3) 0.88 0.79 0.90 0.73 Unclear; get 36.4 42.8 20.8
et al.[13] Age 23 – 4 y more data
Minor (4–7) 0.88 0.88 1.00 0.80 Unclear; get 27.8 44.5 27.8
more data
Moderate 1.03 0.85 0.83 0.63 Unclear; get 42.9 42.7 14.4
(8–28) more data
Severe 0.00 0.35 NA NA Unclear; get – – –
(>28) more data
Steffen 2020 females Regional Match Minor (1–7) 4.00 2.70 0.70 0.28 Possibly 65.7 33.9 0.4
et al.[15] Age 15 – 1 y beneficial; use
Moderate 2.60 2.80 1.00 0.35 Likely trivial 10.7 78.6 10.7
(8–21)
Severe 1.70 3.30 2.00 0.75 Very likely 0.0 2.6 97.4
(>21) harmful; don’t
use
Bjorneboe Males Elite Match Minor (1–7) 8.50 8.00 0.94 0.22 Likely trivial 7.6 91.3 1.0
Sports Med 2011; 41 (11)

et al.[16] 14 teams Moderate 4.90 4.80 0.99 0.30 Likely trivial 7.9 85.7 6.4
No age data (8–21)

Williams et al.
Severe 3.60 4.80 1.34 0.42 Possibly harmful; 0.1 43.4 56.5
(>21) don’t use

Continued next page

 Table III. Contd
ª 2011 Adis Data Information BV. All rights reserved.

Artificial Turf Injury Review

Study No. of subjects or Level of Training Injury Incidence Incidence 90% Clinical inference Likelihood (%) that true value of the
teams, gender and age performance or match severity (n/1000 h exposure) rate ratioa CI effect statistic is:
(mean – SD or range y) injuries (d) natural artificial substantially trivial substantially
turf turf beneficial harmful
Rugby
Fuller 282 males Community Match Slight (2–3) 2.90 3.70 1.27 1.93 Unclear; get 24.7 26.5 48.8
et al.[21] 129 backs more data
Age 26 – 4 y Minor (4–7) 7.70 8.10 1.05 0.89 Unclear; get 25.2 42.5 32.3
153 forwards more data
Age 27 – 6 y
Moderate 10.60 16.90 1.60 1.02 Possibly harmful; 2.3 26.3 71.4
(8–28) don’t use
Severe 5.80 9.60 1.66 1.50 Unclear; get 6.0 25.1 68.9
(>28) more data

Fuller 169 males Elite Training Minor (1–7) 0.70 1.60 2.11 1.00 Very likely 0.0 4.1 95.9
et al.[21] 85 backs harmful; don’t
Age 25 – 4 y use
84 forwards Moderate 0.70 0.80 1.15 0.78 Unclear; get 14.9 47.5 37.6
Age 26 – 8 y (8–21) more data
Severe 0.60 0.70 1.15 0.85 Unclear; get 16.7 44.9 38.4
(>21) more data

American football
Meyers[11] 465 games Collegiate Match Minor (0–6) 39.90b 38.00b 0.95 0.02 Most likely trivial 0.0 100.0 0.0
No gender or age data Moderate 7.20b 5.00b 0.69 0.09 Likely beneficial; 90.5 9.5 0.0
(7–21) use
Severe 4.10b 2.70b 0.66 0.15 Likely beneficial; 87.0 13.0 0.0
(>22) use

Meyers and 240 games High school Match Minor (0–6) 10.70b 12.10b 1.13 No Unclear; get – – –
Barnhill[20] No gender or age data data more data
Moderate 1.30b 1.90b 1.46 No Unclear; get – – –
Sports Med 2011; 41 (11)

(7–21) data more data

Severe 1.90b 1.10b 0.58 0.31 Likely beneficial; 81.8 17.7 0.5
(>22) use
a Incidence rate ratio for injury incidence on artificial and natural turf surfaces, using natural turf as the reference group.
b Incidence given as injuries per ten team games = (n of injuries/n of team games) · 10.

917
NA = not applicable; ~ indicates approximately; – indicates the effect statistics could not be calculated.
918
usually multifactorial. Information on injury
mechanisms must be considered in a model that
also considers how internal and external risk fac-
tors can modify injury risk.[34] Risk factors are,
therefore, predisposing factors that combined with
the mechanism of injury may make an athlete
more prone to injury. Artificial and natural turf
injuries are the result of many inter-relating fac-
tors, some of which can be isolated to reduce in-
jury risk. However, research-based intervention
strategies to reduce the injury risk are sparse.
3.3.1 Shoe-Surface Interface
The shoe-surface interface has been postulated
as a potential risk factor for differences in injury
patterns between surfaces. Orchard[26] suggested
that measures to reduce shoe-surface traction, such
as ground watering and softening, play during
winter months and player use of boots with shorter
cleats, would reduce the risk of injury. Villwock
et al.[35] highlighted the potential for using shoes
with a more pliable upper, allowing more time for
neuromuscular protective mechanisms and thereby
reducing injury risk. Unfortunately, few studies
have investigated changes in the shoe-surface in-
terface using new generation artificial turfs or have
controlled adequately for confounding variables,
so limited quality evidence exists.
Peak Torque
A strong body of evidence exists to suggest
that a high peak torque between the shoe and
playing surface presents an injury risk to the
lower extremity.[36-39] High frictional forces be-
tween the foot and playing surface result in foot
fixation, which may be responsible for lower ex-
tremity injuries.[40] We calculated a 99.9% like-
lihood that peak torque was substantially higher
on artificial turf from the results of a study by
Villwock et al.[35] Significantly lower torque was
noted in the turf cleat above all other groups. The
turf cleat pattern was described as ‘‘a dense pat-
tern of short elastomeric cleats distributed over
the entire sole.’’ This study used a mobile testing
apparatus with a surrogate leg to collect data, the
aim being to investigate the shoe-surface inter-
action of a number of shoes and surfaces currently
employed in soccer. Similarly, Livesay et al.[41]
ª 2011 Adis Data Information BV. All rights reserved.
Williams et al.
reported higher peak torques on new generation
artificial turf, although insufficient data were
provided to allow us to make a magnitude-based
inference. Highest peak torques were developed
by the grass shoe and third generation turf and the
turf shoe and first generation turf combinations,
with lowest peak torques developed on grass.
This suggests that wearing appropriate footwear
could be an important injury prevention strategy.
Higher speeds are generated in games played on
surfaces with greater traction, a factor that may
be responsible for differences in injury patterns.[26]
While larger peak torque forces on artificial turf
may be implicated in the aetiology of injuries, fur-
ther in vivo human biomechanical analyses are re-
quired to substantiate this observation.
Rotational Stiffness
Rotational stiffness is the rate at which torque
increases with applied rotation at the shoe-
surface interface,[41] and is a more sensitive mea-
sure of the mechanical interaction between the
shoe and surface than peak torque alone. Lower
rotational stiffness indicates a lower rate of
loading upon a joint and may allow more time for
a protective form of neuromuscular control to
stabilize the ankle and knee joints during cutting
manoeuvres, potentially reducing the risk of in-
jury.[35] We inferred a 99.8% likelihood that the
rotational stiffness measured by Villwock et al.[35]
was substantially higher on artificial turf. Vary-
ing cleat patterns did not yield any differences in
rotational stiffness in this study. A similar study
by Livesay et al.[41] suggested that the rotational
stiffness produced on new generation artificial
turf did not differ significantly from that experi-
enced on grass, and both turf shoes and grass
shoes produced a similar initial rotational stiff-
ness on third generation artificial turf and natural
turf. However, the compressive load used to col-
lect the data was limited to a maximum of 511 N,
which is much less than the weight of a typical
soccer player and, therefore, reduces the external
validity of these results. We were unable to make
a magnitude-based inference as insufficient data
were provided. The association between rota-
tional stiffness and rate of loading upon lower
limb joints warrants further investigation.
Sports Med 2011; 41 (11)
Artificial Turf Injury Review
3.3.2 Foot Loading
Foot loading patterns invoked during a cut-
ting manoeuvre performed on artificial and nat-
ural turf by American football players have been
investigated[32] using an in-shoe pressure dis-
tribution-measuring insole. Natural grass pro-
duced higher relative loads on the medial forefoot
and lateral midfoot, whilst on artificial turf, peak
pressures within the central forefoot and lesser
toe regions were higher. Larger lateral forces seen
on artificial turf could indicate a greater degree of
foot inversion[28] and may explain the higher in-
cidence of ankle injuries seen in some of the co-
horts discussed. Increased pressure of the medial
forefoot on natural turf could potentially be
linked to the ‘cleat-catch’ mechanism, which has
been implicated in the aetiology of knee ligament
injuries,[28] but this was not observed in our anal-
ysis of knee injuries on the two surfaces. While
these results are interesting, insufficient data were
provided to allow any calculation of magnitude-
based inferences.
3.3.3 Impact Attenuation
Impact attenuation describes how efficiently
the energy from an impact is absorbed.[42] In-
sufficient impact attenuation is linked to an in-
creased injury risk as a result of overloading in
tissues.[15] Theobald et al.[29] investigated the im-
pact attenuation properties of six third genera-
tion artificial turfs and a grass turf to assess the
risk of incurring a mild traumatic brain injury
after a head impact. A range of impact attenua-
tion properties were reported across different ar-
tificial surfaces, with fall heights ranging from
0.46 m to 0.77 m causing a 10% mild traumatic
brain injury risk. Natural turf was the best per-
former, requiring fall heights exceeding those
achievable during games to reach 10% risk for a
mild traumatic brain injury. However, impact
attenuation performance of natural turf appeared
to be dependent on usage. Impact attenuation of
the third generation turfs tested was independent
of moisture content; hence, risk remained con-
sistent in both dry and wet conditions. These re-
sults led the authors to recommend the use of
third generation turfs. However, the relationship
between surface impact attenuation properties
ª 2011 Adis Data Information BV. All rights reserved.
919
and incidence of injuries resulting from player-to-
surface contact is yet to be investigated.
3.3.4 Physiological Response
Di Michele[43] compared blood lactate con-
centrations and heart rate responses to an incre-
mental running test (which was ended when
blood lactate concentration exceeded 4 mmol/L)
on third generation artificial turf, natural turf
and a treadmill in youth soccer players. Running
speed at the 4 mmol/L threshold was substantially
lower (92.4% likelihood) on artificial turf. Run-
ning on artificial turf at 10 km/h resulted in a
significantly higher heart rate, although the differ-
ence between surfaces was trivial at the 4 mmol/L
threshold. Fatigue has been associated with an in-
creased risk for injury[44] and players have reported
perceived greater physical effort to run on artificial
turf.[45] Therefore, the apparent increased physio-
logical cost of exercising on artificial turf may be a
contributing mechanism to injury.
3.3.5 Gender
A gender difference may exist for minor injuries
sustained at soccer training given that female
soccer players showed a likely beneficial inference
for minor training injuries, whilst the equivalent
male group showed a likely harmful inference.
There were no clear gender effects relating to any
other injury patterns given the lack of comparable
cohorts.
3.3.6 Age
Our review included a number of age groups,
ranging from youths (aged 12–17 years) through
to adult cohorts. The spread of age within single
studies, coupled with the lack of information re-
garding subject age in some studies, made anal-
ysis difficult. Qualitative assessment did not show
any clear patterns related to age.
3.3.7 Level of Performance
It was difficult to determine injury patterns that
may vary by performance level given the small
number of studies: two studies used elite ath-
letes;[18,21] two used junior regional athletes;[7,15]
one used high school athletes;[20] and three used
university/collegiate athletes.[11,17,19] Theoreti-
cally, several factors relating to performance may
Sports Med 2011; 41 (11)
920
predispose to differing injury patterns, such as
player fitness, pitch quality (elite players are more
likely to play on a higher standard of surface),
quality of officiating, incidence of foul play,
differences in postural/joint integrity, muscu-
loskeletal structure and biomechanics of move-
ment. Assessment of the contribution of these
factors needs to be addressed in well controlled
studies.
3.3.8 Training and Matches
Seven studies[13,15-19,21] provided data where
training versus match injuries could be com-
pared. For overall injury incidence, a likely trivial
inference was calculated for match injuries in-
volving elite male soccer players, while a possibly
beneficial inference was made for training in-
juries. For ankle injuries, the risk was likely tri-
vial for collegiate male soccer players in matches
but possibly harmful during training. For female
collegiate soccer players, we calculated a possibly
beneficial effect during matches but a possibly
trivial effect for training. With respect to knee
injuries, a difference was evident for collegiate
females, with a likely trivial effect in matches but
a possibly beneficial effect during training. No
clear differences existed between the match and
training cohorts for muscle strain injuries. In
summary, no clear differences between natural
and artificial surfaces were evident in relation to
training versus match injuries. Potentially, the
inferior impact attenuation properties of artificial
turf[29] could contribute to the formation of micro
damage and result in the development of chronic
injuries. The only study to investigate the devel-
opment of chronic injuries on artificial turf was
by Aoki et al.,[7] who highlighted a higher inci-
dence of lower back pain amongst adolescent
soccer players who trained on artificial turf. Given
that exposure to a certain surface is often much
higher for training as opposed to match play,[15] an
investigation of the long-term risks associated with
training on artificial turf is warranted.
3.3.9 Weather
The potential influence of weather conditions
upon injury risk on the two types of surface has re-
ceived limited attention. Meyers and Barnhill[20]
ª 2011 Adis Data Information BV. All rights reserved.
Williams et al.
reported an increased incidence of injuries during
American high school football games on artificial
turf during temperatures >70F (or >21.1C), an
increase calculated to be likely harmful. This
finding was similar to results reported on earlier
generation artificial turfs.[46] The authors sug-
gested an increased shoe traction at higher turf
temperature as a potential mechanism. These re-
sults are in contrast to Meyers,[11] from which we
inferred a 95.4% likelihood that injury incidence
on artificial turf was substantially lower on hot
days (>70F or >21.1C). These differing results
were postulated to be due to the condition of the
natural turf used in each study, although exact
details regarding the condition of the surfaces
were not provided. During play on wet fields, a
most likely beneficial decrease in injuries oc-
curred on artificial turf.[11] This concurred with
laboratory testing of the surfaces, in which im-
pact attenuation properties of artificial turf were
found to be independent of moisture contrast,
while natural turf varied depending on its
usage.[29] The more consistent nature of artificial
turf may well be beneficial across varying weather
conditions, although these results are yet to be
confirmed in other football codes. How risk of
injury on both forms of surface may change de-
pending on the environmental conditions needs
to be determined.
3.3.10 Changing between Surfaces
Rapid changes between playing on different
surface types may act as a precursor to injury in
soccer,[1] which is a finding that is evident in
American football.[39] Ekstrand and Gillquist[47]
proposed that it took six games for players to
adapt to the new surface, although this related to
older generation turfs. Nevertheless, it appears
that players who change frequently from playing
on one surface to another may be at a greater risk
of injury.
4. Conclusions
 Studies have provided strong evidence overall
for a trivial difference in injury incidence rates
between third and fourth generation artificial
turf compared with natural turf.
Sports Med 2011; 41 (11)
Artificial Turf Injury Review
 Artificial turf increased risk of ankle injuries
for 8 of 14 cohorts.
 Six cohorts showed beneficial effects of artifi-
cial turf on muscle injuries for soccer players
(two were over the 90% likelihood value). No
harmful effects were found.
 Inferences made relating to severities of injury
were inconsistent across cohorts, except artifi-
cial turf had very likely harmful effects for
minor injuries in rugby union training and
severe injuries in young female soccer players.
 No clear differences between natural and
artificial surfaces were evident in relation to
training versus match injuries.
 Inconsistent results from a limited number of
studies made drawing clear conclusions diffi-
cult for analyses by sport, gender, age or
performance level.
 Potential mechanisms for differing injury
patterns on artificial turf include increased
peak torque properties and rotational stiffness
properties of shoe-surface interfaces, differing
foot loading patterns, decreased impact at-
tenuation properties and detrimental physio-
logical responses compared with natural turf.
Research on optimal shoe surface/turf release
factors is required.
 Potential injury prevention strategies include
using boots with shorter cleats and a more
pliable upper on artificial turf, and playing on
one type of surface rather than changing
between surfaces.
 More controlled intervention studies, similar
to the FIFA 11+ scheme,[48] are needed to
provide research-based strategies for injury
prevention as few evidence-based best practice
models for injury risk reduction on natural or
artificial turf exist.
 A proposed benefit of installing artificial turf
is a reduction in ground maintenance costs.
Any money saved could be invested in injury
surveillance to ensure the surface is safe.
 Future research should focus on the long-
term risk of chronic injuries, mechanisms and
risk factors for specific injuries and effective-
ness of injury prevention strategies associat-
ed with match and training on artificial
surfaces.
ª 2011 Adis Data Information BV. All rights reserved.
921
 Future research requires uniformity regarding
injury definitions, statistical reporting and de-
scription of surfaces, subjects and data collec-
tion methods used, as outlined, for example,
by Phillips.[49] Studies should use longitudinal,
prospective cohort designs conducted over
several teams and with one recorder where
possible to ensure high intra-rater reliability.
Definitions of injury and injury severity
should be specific and uniform to allow for
comparison between studies.[49]
Acknowledgements
Auckland University of Technology funded this review.
The authors have no conflicts of interest relevant to the con-
tent of this review. We thank Kelly Sheerin from the Sports
Performance Research Institute New Zealand Running Me-
chanics Clinic for reviewing this manuscript.
There are no competing interests by the authors. The
corresponding author has the right to grant on behalf of all
authors, and does grant on behalf of all authors, an exclusive
license (or nonexclusive for government employees) on a
worldwide basis to the journal editor to permit this article to
be published in the journal.
References
1. Ekstrand J, Nigg BM. Surface-related injuries in soccer.
Sport Med 1989; 8 (1): 56-62
2. Girard O, Eicher F, Fourchet F, et al. Effects of the playing
surface on plantar pressures and potential injuries in tennis.
Brit J Sport Med 2007; 41 (11): 733-8
3. Jamison S, Lee C. The incidence of female injuries on grass
and synthetic playing surfaces. Aust J Sci Med Sport 1989;
21 (2): 15-7
4. Pasanen K, Parkkari J, Rossi L, et al. Artificial playing
surface increases the injury risk in pivoting indoor sports: a
prospective one-season follow-up study in Finnish female
floorball. Brit J Sport Med 2008; 42 (3): 194-7
5. Steele JR, Milburn PD. Effect of different synthetic sport
surfaces on ground reaction forces at landing in netball. Int
J Sport Biomech 1988; 4 (2): 130-45
6. Stiles VH, James IT, Dixon SJ, et al. Natural turf surfaces.
Sport Med 2009; 39 (1): 65-84
7. Aoki H, Kohno T, Fujiya H, et al. Incidence of injury
among adolescent soccer players: a comparative study of
artificial and natural grass turfs. Clin J Sport Med 2010; 20
(1): 1-7
8. Dragoo JL, Braun HJ. The effect of playing surface on in-
jury rate a review of the current literature. Sport Med 2010;
40 (11): 981-90
9. Higgins JPT, Green S, editors. Cochrane handbook for
systematic reviews of interventions 4.2.6 [updated Sep-
tember 2006]. In: The Cochrane Library, issue 4, 2006.
Chichester, UK: John Wiley & Sons, Ltd
Sports Med 2011; 41 (11)
922
10. Batterham AM, Hopkins WG. Making meaningful in-
ferences about magnitudes. IJSPP 2006; 1 (1): 50-7
11. Meyers MC. Incidence, mechanisms, and severity of game-
related college football injuries on FieldTurf versus natural
grass: a 3-year prospective study. Am J Sports Med 2010;
38 (4): 687-97
12. Hershman EB, Powell J, Bergfeld JA, et al. American pro-
fessional football games played on FieldTurf have higher
lower extremity injury rates [abstract no. 692]. Presented at
the 2010 Annual Meeting of the American Academy of
Orthopaedic Surgeons; 2010 Mar 9-13; New Orleans (LA)
13. Ekstrand J, Hagglund M, Fuller CW. Comparison of in-
juries sustained on artificial turf and grass by male and
female elite football players. Scand J Med Sci Sports. Epub
2010 Apr 28
14. Soligard T, Bahr R, Andersen TE. Injury risk on artificial
turf and grass in youth tournament football. Scand J Med
Sci Sports. Epub 2010 Aug 24
15. Steffen K, Andersen TE, Bahr R. Risk of injury on artificial
turf and natural grass in young female football players. Brit
J Sport Med 2007; 41 (1 Suppl.): i33-7
16. Bjorneboe J, Bahr R, Andersen TE. Risk of injury on third-
generation artificial turf in Norwegian professional foot-
ball. Brit J Sport Med 2010; 44 (11): 794-8
17. Fuller CW, Dick RW, Corlette J, et al. Comparison of the
incidence, nature and cause of injuries sustained on grass
and new generation artificial turf by male and female
football players. Part 1: match injuries. Brit J Sport Med
2007; 41 (1 Suppl.): i20-6
18. Ekstrand J, Timpka T, Hagglund M. Risk of injury in elite
football played on artificial turf versus natural grass: a
prospective two-cohort study. Brit J Sport Med 2006; 40
(12): 975-80
19. Fuller CW, Dick RW, Corlette J, et al. Comparison of the
incidence, nature and cause of injuries sustained on grass
and new generation artificial turf by male and female
football players. Part 2: training injuries. Brit J Sport Med
2007; 41 (1 Suppl.): i27-32
20. Meyers MC, Barnhill BS. Incidence, causes, and severity of
high school football injuries on FieldTurf versus natural
grass: a 5-year prospective study. Am J Sports Med 2004;
32 (7): 1626-38
21. Fuller CW, Clarke L, Molloy MG. Risk of injury associated
with rugby union played on artificial turf. J Sport Sci 2010;
28 (5): 563-70
22. Hopkins WG. Linear models and effect magnitudes for re-
search, clinical and practical applications. Sportscience
2010; 14: 49-57
23. Hopkins WG. Progressive statistics updated. Sportscience
2009; 13: 13-4
24. Hopkins WG. A spreadsheet for combining outcomes from
several subject groups. Sportscience 2006; 10: 51-3
25. Hopkins WG. A spreadsheet for deriving a confidence in-
terval, mechanistic inference and clinical inference for a
p-value. Sportscience 2007; 11: 16-20
26. Orchard J. Is there a relationship between ground and cli-
matic conditions and injuries in football? Sport Med 2002;
32 (7): 419-32
27. Lees A, Nolan L. The biomechanics of soccer: a review. J
Sport Sci 1998; 16 (3): 211-34
ª 2011 Adis Data Information BV. All rights reserved.
Williams et al.
28. Wright JM, Webner D. Playing field issues in sports medi-
cine. Curr Sports Med Rep 2010; 9 (3): 129-33
29. Theobald P, Whitelegg L, Nokes LDM, et al. The predicted
risk of head injury from fall-related impacts on to third-
generation artificial turf and grass soccer surfaces: a com-
parative biomechanical analysis. Sports Biomech 2010; 9
(1): 29-37
30. FieldTurf. How FieldTurf synthetic turf works [online].
Available from URL: http://www.fieldturf.us/artificial-
turf-how-fieldturf-works/ [Accessed 2011 Feb 16]
31. FIFA. Quality concept: handbook of requirements for
football turfs. Geneva: Fédération Internationale de
Football Association, 2006
32. Ford KR, Manson NA, Evans BJ, et al. Comparison of in-
shoe foot loading patterns on natural grass and synthetic
turf. J Sci Med Sport 2006; 9 (6): 433-40
33. Whiting WC, Zernicke RF. Biomechanics of musculoskele-
tal injury. Champaign (IL): Human Kinetics, 1998
34. Bahr R, Krosshaug T. Understanding injury mechanisms: a
key component of preventing injuries in sport. Brit J Sport
Med 2005; 39: 324-9
35. Villwock MR, Meyer EG, Powell JW, et al. Football playing
surface and shoe design affect rotational traction. Am
J Sports Med 2009; 37 (3): 518-25
36. Bonstinglu RW, Morehouse CA, Niebel BW. Torques de-
veloped by different types of shoes on various playing
surfaces. Med Sci Sport Exercise 1975; 7 (2): 127-31
37. Lambson RB, Barnhill BS, Higgins RW. Football cleat de-
sign and its effect on anterior cruciate ligament injuries: a
three-year prospective study. Am J Sports Med 1996; 24
(2): 155-9
38. Nigg BM, Yeadon MR. Biomechanical aspects of playing
surfaces. J Sport Sci 1987; 5 (2): 117-45
39. Torg JS, Quedenfeld TC, Landau S. Shoe-surface interface
and its relationship to football knee injuries. Am J Sports
Med 1974; 2 (5): 261-9
40. Olsen O, Myklebust G, Engebretsen L, et al. Relationship
between floor type and risk of ACL injury in team hand-
ball. Scand J Med Sci Sports 2003; 13 (5): 299-304
41. Livesay GA, Reda DR, Nauman EA. Peak torque and ro-
tational stiffness developed at the shoe-surface interface:
the effect of shoe type and playing surface. Am J Sports
Med 2006; 34 (3): 415-22
42. Perez-Soriano P, Llana-Belloch S, Cortell-Tormo JM, et al.
Biomechanical factors to be taken into account to prevent
injuries and improve sporting performance on artificial
turf. J Hum Sport Exerc 2009; 4 (2): 78-92
43. Di Michele R, Di Renzo AM, Ammazzalorso S, et al.
Comparison of physiological responses to an incremental
running test on treadmill, natural grass, and synthetic turf
in young soccer players. J Strength Cond Res 2009; 23 (3):
939-45
44. Small K, McNaughton L, Greig M, et al. The effects of
multidirectional soccer-specific fatigue on markers of
hamstring injury risk. J Sci Med Sport 2010; 13 (1): 120-5
45. Andersson H, Ekblom BR, Krustrup P. Elite football on
artificial turf versus natural grass: movement patterns,
technical standards, and player impressions. J Sport Sci
2008; 26 (2): 113-22
Sports Med 2011; 41 (11)
Artificial Turf Injury Review
46. Orchard JW, Powell JW. Risk of knee and ankle sprains
under various weather conditions in American football.
Med Sci Sport Exercise 2003; 35 (7): 1118-23
47. Ekstrand J, Gillquist J. Soccer injuries and their mechan-
isms: a prospective study. Med Sci Sport Exercise 1983; 15
(3): 267-70
48. FIFA. The 11+ [online]. Available from URL: http://www.
fifa.com/aboutfifa/developing/medical/the11/index.html [Ac-
cessed 2011 Feb 16]
ª 2011 Adis Data Information BV. All rights reserved.
923
49. Phillips LH. Sports injury incidence. Brit J Sport Med 2000;
34 (2): 133-6
Correspondence: Professor Patria Hume, Sports Performance
Research Institute New Zealand (SPRINZ), School of Sport
and Recreation, Auckland University of Technology, Private
Bag 92006, Auckland, New Zealand.
E-mail: patria.hume@aut.ac.nz
Sports Med 2011; 41 (11)