Part 1 of the present study of football injuries¹ confirmed that there were no major differences between the incidence, severity, nature or cause of match injuries sustained on new generation artificial turf and grass for either male or female players. Although the new generation of artificial turf surfaces have not yet been widely adopted for matches, professional and non-professional clubs use them to provide year-round, all-weather training surfaces. It is important, therefore, to compare the epidemiology of injuries sustained during training activities on artificial turf of this type and grass to understand fully the risks associated with the use of artificial turf. Comparative epidemiological data for training injuries sustained on artificial turf and grass in football are limited^2,3 and the available information is limited to male elite professional players. Although Árnason et al² reported that the incidence of injuries sustained by Icelandic footballers playing on older types of artificial turf was twice that recorded on grass surfaces (p<0.01), their results did not differentiate between match and training exposures so it was not possible to determine the contribution of training activities to the increased risk. Ekstrand et al³ reported that there was no significant difference between the overall incidence of training injuries in elite male European footballers on the new generation of synthetic infill artificial turf and grass pitches, although the incidence of lower leg strains on artificial turf was half the level recorded on grass (p<0.01).

The aim of the present study was to compare the incidence, nature, severity and cause of training injuries sustained on new generation synthetic infill artificial turf and grass in male and female players to determine whether the risk of injury during football training is greater on artificial turf than on grass surfaces. A similar comparative study for match injuries has been reported separately in this supplement.¹

METHOD

As our objective in this study was to identify whether footballers were subjected to a greater risk of injury while training on new generation artificial turf than on grass, as for the study of match injuries, the sample size needed to be sufficiently large to identify statistically significant results.⁴

Sample size calculation

For this calculation, the incidence of training injuries on grass in male and female players was estimated to be 3 injuries/1000 player hours; this value was based on the data presented in the review of football injuries by Junge and Dvorak.⁵ An increase in the incidence of training injuries on artificial turf of one-third (33%) compared with the level experienced on grass was again accepted as a significant size effect for the same reasons as those discussed in Part 1 for match injuries.¹ Because for this study there was a much larger number of teams that trained on grass compared with the number training on artificial turf, we could increase the statistical power of the study by using a 4:1 ratio of grass to artificial turf player exposures. The minimum sample sizes required for a study to identify an effect of this size with a 95% significance level and 90% power⁴ were calculated to be 45 927 player hours for the cohorts training on artificial turf and 183 708 player hours for the control cohorts training on grass.

Implementation

This prospective two-cohort study evaluated men’s and women’s football training injuries collected by the National Collegiate Athletic Association (NCAA) Injury Surveillance System (ISS)^1,6 over a two-season period (2005 and 2006). The study included all organised practice sessions from the preseason, in-season and postseason periods (August through December). The sample of American college and university football teams (2005 season: men 52 teams, women 64 teams; 2006 season: men 54 teams, women 72 teams) represented approximately 7% of all NCAA schools sponsoring football over the period. Eighteen men’s and 18 women’s teams used artificial turf as their home pitch and 44 men’s and 56 women’s teams used grass as their home pitch. All squad players at each team were included in the study. We defined training injury as:


 “any physical complaint sustained by a player during a football training session that prevented the player from taking a full part in training or match play activities for one or more days beyond the day of injury.”

We did not include absences from competition and/or training caused by medical conditions not related to football. The severity of individual training injuries was defined by the number of days the player missed from training and/or competition or, where the injury persisted beyond the end of the season, as a “season ending” injury. Injury severities were grouped as minimal (1–3 days), mild (4–7 days), moderate (8–28 days) and severe (>28 days plus season ending injuries). The total player training exposures over the two seasons, for each cohort on each playing surface, were calculated on the basis of the sum of all team training sessions on the surface, the number of players attending and the average estimated length of time of the training sessions (2 h). Athletic trainers who worked with the teams and who were qualified health professionals recorded every training injury according to the specified requirements of the NCAA-ISS⁶: these procedures provide details of the training surface (grass/synthetic infill artificial turf) and the location, type, diagnosis, severity and cause (acute/gradual onset; contact/non-contact) of each injury. The definitions and procedures used in this study were consistent with the international consensus statement on injury definitions and procedures for epidemiological studies of football injuries.⁷ An NCAA research review board approved the data collection procedures with regard to the protection of human subjects.

Data analysis

Incidences of training injuries were reported as the number of injuries/1000 player training hours together with 95% CI.⁴ We considered differences between the incidence of training injuries on artificial turf and grass to be significant if the 95% CI of the incidence ratio (equivalent to the risk ratio) did not include the value of 1.0 and the p value of the two-sided z test for the comparison of rates⁴ was <0.05. Differences in the mean and median severity (days) of training injuries for different groups of players were assumed to be significant if the 95% CI did not overlap. Differences in the distributions of training injuries as a function of the location and type of injury were considered significant if the p value of the two-sided z test for the comparison of proportions⁴ was <0.05.

RESULTS

Table 1 presents the numbers of team training and player training sessions, the average number of players attending training sessions and the training exposures on artificial turf and grass for men and women in each of the two seasons. The total number of training injuries recorded for men was 818 (artificial turf: 189; grass: 629) and for women 774 (artificial turf: 122; grass: 652).

Incidence and severity of injury

The overall incidence of injury on artificial turf and grass did not differ significantly for either male or female players (table 2). Incidences of mild and moderate injuries were significantly higher on artificial turf than on grass for men but for women the incidence of mild injuries was significantly lower on artificial turf (see table 2 for p values). For injuries that were not season ending, the mean or median severity of injuries sustained on artificial turf and grass for either male or female players did not differ significantly (table 3).

The most common season ending injury for men on artificial turf was an ankle ligament tear, which was significantly more common on artificial turf than on grass (31% (4), 95% CI 6% to 56% and 5% (4), 95% CI 0 to 11%, respectively; p<0.01), whereas on grass the most common injury was a knee ligament tear (11% (8), 95% CI 4% to 18% and 23% (3), 95% CI 2 to 46%, respectively; p = 0.22). For women, the most common season ending injury on both playing surfaces was a knee ligament tear (artificial turf: 30% (3), 95% CI 2% to 58%; grass: 23% (18), 95% CI 13% to 32%; p = 0.60).

Nature of injury

We did not find any significant differences in the incidence of injury as a function of general body location (table 4). The lower limb was the most common location of injuries sustained on artificial turf and grass in both male and female players. For men, the ankle and thigh were the most commonly injured lower limb structures on artificial turf and grass; this was the same situation for women training on artificial turf but on grass the thigh and knee were injured more often (table 4). Men sustained significantly more ankle (p = 0.04) and foot (p = 0.04) injuries on artificial turf than on grass. Muscle/tendon injuries were the most common type of training injury on artificial turf and grass for women and for men on grass but joint (non-bone)/ligament/cartilage injuries were more common on artificial turf for men (table 5). Joint (non-bone)/ligament/cartilage injuries for men was the only type of training injury for which the incidence of injury was significantly higher on artificial turf than on grass (p = 0.04).

Cross-tabulations of the location and type of injury (table 6) showed that joint (non-bone)/ligament/cartilage injuries to the lower limb was the most common category of injury in men on both surfaces and in women on grass; however, muscle/tendon injuries to the lower limb were more common in women on artificial turf. In addition, the proportion of lower limb, joint (non-bone)/ligament/cartilage injuries was significantly lower on artificial turf than on grass in women (p = 0.02).

Table 7 presents the incidences of the most common training injury diagnoses in male and female players. For men, the three most common injuries on grass were ankle lateral ligament complex, hamstring and adductor tears while on artificial turf they were ankle lateral ligament complex, quadriceps and adductor tears; for women the three most common injuries on grass were ankle lateral ligament complex, quadriceps and hamstring tears and on artificial turf ankle lateral ligament complex and hamstring tears and concussion. We specifically investigated the overall incidence of lower limb strains on artificial turf and grass in order to review Ekstrand and colleagues’³ conclusions on these injuries. We did not find any surface related significant differences in the incidence of these injuries in men (artificial turf: incidence 0.50, 95% CI 0.34 to 0.72; grass: incidence 0.45, 95% CI 0.36 to 0.55; incidence ratio 1.11, 95% CI 0.73 to 1.70; p = 0.62) or women (artificial turf: incidence 0.34, 95% CI 0.21 to 0.55; grass: incidence 0.52, 95% CI 0.44 to 0.62; incidence ratio 0.65, 95% CI 0.39 to 1.10; p = 0.11).

Injury causation

The incidences of acute, gradual onset, contact or non-contact training injuries on artificial turf and grass did not differ significantly for either male or female players (table 8). Only one concussion injury (female player) sustained on artificial turf (total concussions: men 4; women 10) was caused by player-to-surface contact; on the other hand, 18% (4/22) of concussion injuries sustained on grass by men and 10% (3/30) by women were caused by player-to-surface contact.

DISCUSSION

Although a range of manufacturers supplied the artificial turf training surfaces included in this study, all the surfaces incorporated synthetic infill materials, such as rubber crumb. The power of the study was improved because it was possible to increase the total player training exposure times on grass in the study to approximately four times the levels available for artificial turf.⁴ The training exposures recorded in this study, for both male and female players, exceeded the number calculated to identify, with statistical significance, a relative risk of 1.33 for the overall incidence of training injuries on artificial turf compared with grass. These exposure levels were also sufficient to identify, with statistical significance, a relative risk of 2.00 (the standard adopted by the UK Industrial Injuries Advisory Council⁸) for subcategories of injury in which the incidence of injury was at least 0.4 injuries/1000 player hours on grass (ie, ∼13% of the estimated overall incidence of training injuries).

The incidence of injury recorded in the present study on grass for men (3.01) was consistent with values reported previously (2–7) but the value obtained for female players (2.79) was lower (4–7).⁵ The incidence of training injuries recorded for male players was similar to but higher than that reported in the comparative study among elite professional footballers³ for both artificial turf (3.34 vs 2.42, respectively) and grass (3.01 vs 2.94, respectively); however, this previous study³ excluded gradual onset/overuse injuries from the calculations of incidence. There are no equivalent published data with which to compare the incidence of training injuries for women on artificial turf surfaces. We did not find any significant differences between the overall incidence or the mean and median severity of injuries sustained on artificial turf and grass in either male or female players, which is consistent with Ekstrand and colleagues’³ conclusions in male players. However, the incidences of mild and moderate injuries were significantly higher on artificial turf than on grass in men although the incidence of mild injuries was lower on artificial turf in women.

As with match injuries, the most common location for training injuries was the lower limb, especially the ankle, knee and thigh. Although joint (non-bone)/ligament/cartilage injuries remained common injuries, muscle/tendon injuries formed a much higher proportion of the training injury burden than they did for match injuries. In terms of the combined location/type of injury, joint (non-bone)/ligament/cartilage injuries to the lower limb were the most common training injuries in men on artificial turf and grass surfaces and for women on grass; however, women were more likely to sustain a lower limb muscle/tendon injury on artificial turf. Ankle lateral ligament complex tears were the most common training injury on artificial turf and grass in both men and women, which was a similar situation to that observed with match injuries. Ekstrand et al³ suggested that the incidence of lower limb muscle strains for men was significantly (p<0.01) lower on artificial turf than grass: this effect was specifically investigated in this study for both male and female players. A small increase was observed in the incidence of injury for men (incidence ratio 1.11) and a reduction in the incidence for women (incidence ratio 0.65) on artificial turf but the differences were not significant in either case.

There was no indication of a higher incidence of non-contact or gradual onset (chronic) injuries on artificial turf than on grass in either male or female players; however, the incidences of gradual onset injuries recorded in this study for men were much lower than those reported by Ekstrand et al.³ The lower incidence of these injuries in the present study may reflect the characteristics of the sample population, as the playing/training season is much shorter in the American non-professional collegiate (∼5 months) than in the European professional (∼9 months) season, and training and playing intensity may also be higher in the European professional leagues. Gradual onset training injuries were no more common on artificial turf than they were on grass for either men (artificial turf: 11.7%; grass: 12.3%) or women (artificial turf: 19.6%; grass: 19.0%); however, they did represent a significantly (<0.01) higher proportion of the injury burden than was the case for match injuries for men and women on artificial turf and grass (<5%).¹ In terms of injury causation, although “contact with the playing surface” was responsible for a higher proportion of training injuries among female (artificial turf: 15.4%; grass: 11.5%) than in male (artificial turf: 8.4%; grass: 8.6%) players, there were no differences between artificial turf and grass for either men or women. These trends were similar to those observed for match injuries on artificial turf (men: 6.6%; women: 11.2%) and grass (men: 7.8%; women: 15.5%).¹ Although the proportions of non-contact training injuries sustained on artificial turf and grass did not differ significantly in men (artificial turf: 48.2%; grass: 51.8%) or women (artificial turf: 48.5%; grass: 55.9%), they represented a significantly (p<0.01) higher proportion of injuries than was the case for match injuries for both men (artificial turf: 27.3%; grass: 31.1%) and women (artificial turf: 25.4%; grass: 23.5%).

This study consisted of a large, uniform sample population of male and female college and university students, who trained on a regular basis throughout the season. This sample population, therefore, provided an ideal setting to compare the risks of training injuries on new generation artificial turf and grass. The protocols used were consistent with the international consensus statement on injury definitions and procedures for studies of injuries and football⁷ and with the procedures used by Ekstrand et al³ in a similar comparative study of training injuries sustained on new generation artificial turf and grass by elite professional footballers. The statistical power of the present study allowed a more detailed analysis and comparison of injury subcategories than has been possible previously. This study confirms that there are generally no differences between the overall incidence, severity, nature or cause of training injuries sustained on artificial turf and grass in male or female players. These results taken in conjunction with the results reported separately for match injuries¹ and the results presented previously by Ekstrand et al³ and Fuller⁹ suggest that there is no difference in the overall risk of injury to male or female footballers on new generation artificial turf surfaces compared with grass.