You are on page 1of 8

AJSM PreView, published on March 26, 2010 as doi:10.

1177/0363546509359065

Footwear Traction and Lower Extremity Joint Loading


John W. Wannop,* Jay T. Worobets, PhD, and Darren J. Stefanyshyn, PhD From the University of Calgary, Human Performance Lab, Faculty of Kinesiology, Calgary Alberta, Canada
Background: Traction is influenced by the sole architecture and playing surface, with increases in traction potentially leading to injury. The mechanism as to how or why increased traction could lead to injury remains unknown. Purpose: This study was undertaken to determine how shoes of different sole designs and traction influence knee and ankle joint moments. Study Design: Controlled laboratory study. Methods: Traction testing was performed on 2 shoes of varying sole designs (tread vs smooth) using a robotic testing machine. All testing was conducted on a 60-cm 3 90-cm piece of sample track surface. Kinematic and kinetic data were then collected on 13 recreational athletes performing running V-cuts in the 2 different shoe conditions. Five trials per condition were collected with reflective markers placed on the right shank and shoe of each participant. Kinematic and kinetic data were collected using an 8high-speed camera system and force plate. Results: The coefficient of translational traction and the peak moment of rotation were both significantly higher in the tread shoe compared with the smooth shoe (1.00 vs 0.87 and 23.87 Nm vs 16.12 Nm, respectively). The high-traction shoe had significantly higher peak ankle external rotation moments (89.58 Nm vs 80.17 Nm), peak knee external rotation moments (36.23 Nm vs 32.02 Nm), peak knee adduction moments (224.0 Nm vs 186.8 Nm), and knee adduction angular impulse (2.10 Nms vs 1.83 Nms) compared with the low-traction shoe. Conclusion: Increased shoe traction significantly increased ankle and knee joint moments during a V-cut. Despite the significant difference in traction, no difference in performance was observed. These changes could have an effect on ankle and knee joint injury. Clinical Relevance: Shoes with decreased traction could be used in sports to reduce the joint moments in the knee and ankle and potentially reduce injury without a loss in performance. Keywords: biomechanics; traction; knee; ankle; moments

The knee and ankle are 2 of the most frequently injured joints in the body, with most of these injuries occurring without player-to-player contact.14,16,17 It has been reported that 78% of all anterior cruciate ligament (ACL) injuries are noncontact, and the majority occur when landing from a jump, while cutting, or during sudden deceleration.18 Myklebust et al16,17 reported that the highest number of ACL injuries were recorded in pivoting sports during plant-and-cut high-speed movements and it has also been shown that the
*Address correspondence to John W. Wannop, University of Calgary, Human Performance Lab, Faculty of Kinesiology, 2500 University Drive NW, Calgary Alberta, Canada T2N 1N4 (e-mail: bwannop@kin.ucalgary.ca). One or more authors has declared a potential conflict of interest: Financial support was received from the International Society of Biomechanics and adidas International donated the footwear used in testing.
The American Journal of Sports Medicine, Vol. XX, No. X DOI: 10.1177/0363546509359065 2010 The Author(s)

ankle accounts for between 15% and 30% of all reported injuries during sideways movements.25 Some studies have concluded that traction between the shoe and surface is one of the major factors affecting lower extremity noncontact injuries.11,31 Traction is a property of both the shoe and surface, and is usually divided into 2 categories: linear translational traction and rotational traction. Linear translational traction is dependent on both the force normal and horizontal to the surface and is usually described as a coefficient defined by the ratio of horizontal to vertical force. Rotational traction is generally described using the moment of rotation with respect to the center of pressure,18 which refers to rotation of the foot around a point of contact on the shoe sole.6 Increases in traction may allow for sharper cutting angles, which can enhance performance, but may also produce greater stress to supporting structures of joints.12,26 Bramwell et al4 and Luethi and Nigg13 showed higher injury rates on football and tennis surfaces with high coefficients of
1

Wannop et al

The American Journal of Sports Medicine

Figure 1. The 2 shoes used in the study: A, adidas Response 12 CPT (smooth shoe); and B, adidas Response52(A) (tread shoe). traction. Also, two-thirds of all noncontact soccer injuries may be due to excessive shoe-surface traction.5 Thus, the goal should be to determine what range of traction coefficients provides high enough traction for maximum performance yet low enough traction to prevent injury.6 For athletic performance, a required amount of traction is needed to perform the various sport movements. Higher available traction may lead to higher absolute peak force development but does not necessarily constitute a larger impulse or shorter stance time.28 Therefore the increased traction of the shoe-surface interface may not alter or increase performance, but it may increase the peak resultant joint moments and loading on the joints. Although these previous studies answered some questions regarding footwear traction and injury, many questions still remain. It has been shown that traction is influenced by the sole architecture and playing surface,26 and that increases in traction may lead to injury,11,31 but the mechanism as to how or why increased traction leads to injury remains unknown. It is generally believed that increases in joint loading may lead to joint injury.8,20,21,27 More specifically, joint loading is defined by peak joint moments, which represent the maximal torque and twisting loading on the joint, and joint angular impulse, which represents the cumulative loading experienced by the joint throughout the stance phase calculated by multiplying the load by the length of time it is applied.27 While resultant joint moments calculated from inverse dynamics cannot determine the exact nature of the loading on the actual joint structures, it has been used as a valid predictor of the total load distribution across a joint.9 As well, a 2009 study by Mizuno et al15 showed that increases in resultant joint moments placed on the knee resulted in direct increases in ACL strain. The question remains as to whether there is a functional relationship between traction and joint loading. Therefore, the purpose of this study was to determine how shoes of different traction influence loading in the knee and ankle joints. It is hypothesized that as the traction of the shoe increases, increases in resultant joint moments will also result, with performance remaining constant.

Figure 2. Setup on the robotic testing machine used for translational and rotational traction testing.

METHODS Traction Testing


Two adidas shoe models (adidas, Herzogenaurach, Germany) of varying sole properties were tested on a 6 degrees of freedom robotic testing machine to determine the translational and rotational frictional properties between the shoe and surface. The 2 shoe types were adidas Response 12 CPT (smooth) and adidas Response52(A) (tread). All aspects of both shoes were identical with the exception of the shoe sole. The smooth shoe was composed of 2 different rubbers, and had a completely smooth sole, whereas the tread shoe had rubber grooves of varying directions and small rubber studs (Figure 1). A sample track surface measuring 60 3 90 cm was bolted to the movable platform of the robotic testing machine. A prosthetic foot was used to simulate a physiological foot, and was fitted in a left size-9 shoe of each condition. The foot was then attached to the upper portion of the robotic testing machine and the platform was angled at 20 of plantar flexion to simulate the orientation of the foot during a cutting movement (Figure 2). A triaxial load cell was used to measure forces and moments in the anteroposterior, mediolateral, and inferosuperior directions during the testing procedure. For translational traction, a normal load of 750 N was applied to the shoe. After the normal load had been reached, the platform moved anteriorly to the shoe at a speed of 100 mm/sec. For rotational traction, a normal load of 750 N was applied to the shoe and the platform was internally rotated at a speed of 75 deg/s. For both traction tests, force data were collected at a sampling frequency of 1000 Hz during the duration of the movement. The data were filtered using a low-pass fourth order Butterworth filter with a cutoff frequency of 50 Hz, and were analyzed using MATLAB (MathWorks, Natick, Massachusetts). The coefficient of translational traction was calculated throughout the movement and the maximum coefficient was taken for each trial. For rotational traction,

Vol. XX, No. X, XXXX

Footwear Traction and Joint Loading 3

the moment of rotation was measured and recorded and the maximum moment during the movement duration was taken for each trial. Ten trials were performed on each shoe condition and the mean value of these 10 trials was used to determine the traction properties of each shoe.

Motion Analysis
Data were collected on 13 recreational athletes (11 male, 2 female) from the Kinesiology Faculty at our university. The body weight of the participants ranged from 617 N to 982 N with an average of 806 6 124 N. Before the study, all volunteers were required to read and sign a consent form approved by the university ethics committee. To be included in the study, all volunteers must have been involved in recreational sport, free from recent lower extremity injury, and properly fit the size-9 shoe. Each individual performed a running V-cut with the 2 different shoe conditions on a 30-m laboratory runway. The V-cuts were performed by running forward and performing a cut at a 45 angle (Figure 3). During each V-cut, 3-dimensional (3D) force data were collected using a force platform (Kistler AG, Winterthur, Switzerland) mounted in the center of the runway floor. The sample surface that was used during traction testing was securely adhered to the force platform. Force data were collected at 2400 Hz, and participants were required to land and perform the cut with their right foot in the center of the force platform. The participants were instructed to perform each movement at their maximum speed, which was monitored using photocells placed 1.9 m apart. They were given enough practice trials before testing to ensure proper movement technique and to determine each individuals maximum speed. Five accepted trials were required, with a trial being accepted if the individuals were within 5% of their maximum previously recorded speed in each shoe condition. Three-dimensional kinematics of the lower limb were collected for the right leg of each participant during each trial. The shoe and shank were defined by attaching retroreflective markers, measuring 19 mm in diameter, to each segment using medical adhesive. Three markers per segment were used, attached to the following locations: head of the fibula, upper tibial crest, distal lateral lower leg, posterior shoe heel, distal shoe heel, and lateral side of the shoe below the lateral malleolus. Eight high-speed digital video cameras (Motion Analysis Corp, Santa Rosa, California) at a sampling frequency of 240 Hz were used to capture the motion of the markers. The system was calibrated to an accuracy defined by a 3D residual below 0.6 mm. A standing neutral trial was captured using the video system to determine the 3D coordinates of the ankle and knee joint center. The participant was asked to stand with the feet hip-width apart, with the knee and hip fully extended. To determine the ankle joint center, additional markers were placed on the medial and lateral malleoli. For the knee joint center, additional markers were placed on the lateral knee and at the center of the patella. The kinematic and kinetic data were imported into Kintrak 6.3 (Motion Analysis Corp) for analysis, and filtered at cutoff frequencies of 24 Hz and 100 Hz, respectively, using

Figure 3. Diagram of the cutting movement performed on the sample track surface adhered to the force platform. a fourth-order low-pass Butterworth filter. The analyzed variables for each shoe condition were peak internal resultant joint moments and peak angular impulses in all 3 planes using a segment coordinate inverse dynamics approach (Figure 4).

Statistical Analysis
For traction testing, the 10 trials were compared between the 2 shoe conditions. For moment data, average values for all 5 trials per condition were compared. All data were compared between shoe conditions using a paired t test at the 95% level of confidence. All statistical analysis was performed using SPSS software version 12.0 (SPSS Inc, Chicago, Illinois).

RESULTS Robotic Testing


From linear translational traction testing, the coefficient of linear translational traction was found to be significantly higher in the tread shoe compared with the smooth shoe (1.00 6 0.03 vs 0.87 6 0.02; P \ .001). From rotational traction testing, the peak moment of rotation was significantly higher in the tread shoe compared with the smooth shoe (23.89 6 0.55 Nm vs 16.12 6 0.16 Nm; P \ .001) (Figure 5). Because of the difference in traction, the smooth shoe will be referred to as the low-traction shoe and the tread shoe as the high-traction shoe.

Ankle Joint Moments


All values of peak ankle joint moments can be seen in Figure 6 and ankle angular impulse can be seen in Figure 7.

Wannop et al

The American Journal of Sports Medicine

Figure 4. Ankle and knee joint moment sample curves. The circled region represents the peak joint moment values compared between conditions. The area under the curves represents the joint angular impulse. Curves are mean and standard deviation of 1 representative study participant.

Figure 5. Translational traction coefficients and peak moment of rotation of each shoe from the robotic traction testing. Results are the mean of 10 trials with standard deviation. The high-traction shoe had significantly higher peak external rotation moments when compared with the low-traction shoe (89.6 6 10.1 Nm vs 80.2 6 10.1 Nm; P 5 .003). The peak ankle eversion moment, plantar flexion moments, and total ankle angular impulse (Figure 7) were similar between the 2 conditions with no statistically significant differences seen. adduction angular impulse (2.10 6 1.1 Nms compared with 1.83 6 0.93 Nms) were significantly higher in the high-traction shoe compared with the low-traction shoe (P 5 .041 and P 5 .01). No other significant differences were seen between conditions.

Performance Knee Joint Moments


All values of peak knee joint moments can be seen in Figure 8 and knee angular impulse can be seen in Figure 7. The high-traction shoe had significantly higher peak external rotation moments when compared with the low-traction shoe (36.2 6 7.9 Nm vs 32.0 6 7.7 Nm; P 5 .043). As well, the knee adduction moment (224.0 6 57.7 Nm compared with 186.8 6 47.3 Nm) and knee Running speed and stance time were similar between conditions, with no significant differences seen (Figure 9).

DISCUSSION
Over the past several years, many studies have investigated the interaction between the shoe and surface.1,3,7,11,18,26,29-32 These studies have attempted to

Vol. XX, No. X, XXXX

Footwear Traction and Joint Loading 5

Figure 6. Values of peak ankle joint moments compared between the high-traction and low-traction shoes. Values are the mean with standard deviations of all participants.

Figure 7. Values of ankle and knee joint angular impulses compared between the high-traction and low-traction shoes. Values are the mean with standard deviations of all participants. *Indicates a significant difference (P \ .05).

Figure 8. Values of peak knee joint moments compared between the high-traction and low-traction shoes. Values are the mean with standard deviations of all participants. determine the relationship between traction and injury, as it has been speculated that high traction between the shoe and surface leads to noncontact injuries at the knee and ankle.4,11,13,23,31 Although these studies examined how different shoes can affect injury rates, in some cases they failed to actually measure the traction of the shoes being tested.10,16,17 As well, studies that concluded that high traction may affect injury made no mention or speculation as to why or how traction could affect injury. Although resultant joint moments calculated from inverse dynamics cannot determine the exact nature of the loading on the actual joint structures, they can be used as an indicator of what is occurring in the joint.9 The loading that actually reaches the joint structures is dependent on many things such as the complex interaction of the load, magnitude, and frequency of muscular forces, as well as the timing of these muscular forces and neuromuscular control. Without muscle activation studies, it cannot be determined how much loading is transferred to the ligament structures, but it is believed that increases in values of resultant joint moments calculated from inverse dynamics would reflect increases in values of joint loading.9 No studies have inspected how changes in traction of the shoe-surface interface will affect resultant joint

Wannop et al

The American Journal of Sports Medicine

Figure 9. Performance measurements for each condition consisted of running speed and stance time. Values are the mean with standard deviations for all participants.

loading acting on the joints of the body. This study sought to determine how shoes of varying traction properties influenced the resultant loading in the ankle and knee joints.

Loading in Transverse and Frontal Plane


It has been speculated that cutting, stopping, landing, and rotating are high-risk maneuvers for ankle sprain,24 and the movement was selected in this study to incorporate most of those aspects. The ankle joint peak loading in the transverse plane increased by approximately 12% in the high-traction shoe. It has been shown that 90% of all ligamentous injuries to the ankle are caused by internal rotation1,25; when shoe traction was increased, the loading in the ankle joint increased, possibly bringing the joint closer to injury. If this high internal rotation moment caused by the movement is not compensated by the high external rotation moment created by the muscles surrounding the ankle joint, injury could result. In the knee joint, the plant-and-cut movement has been proposed as a possible mechanism of ACL tear due to sudden deceleration and rapid twisting of the ligament.2,16,17 As traction increased, peak knee joint moments in the transverse plane increased by approximately 13% and the peak adduction moments in the frontal plane increased by 20%. Previous studies have shown that loading of 35 to 80 Nm in the transverse plane and 125 to 210 Nm in the frontal plane can cause damage and rupture of knee ligaments.19 Although these values were obtained from a static ligament failure test performed on cadavers, which may limit the comparison, no dynamic ligament failure studies have been performed for obvious ethical reasons. The average loading in the low-traction shoe kept these knee loads below the rupture zone (32 Nm), while the high-traction shoe increased these knee loads into the rupture zone (35 Nm). In the frontal plane, both shoes produced average joint moments that were in the zone at which rupture could occur; however, the high-traction shoe raised these

moments far above the previously reported rupture zone (224 Nm). There is no way to know how much of the increase in resultant joint loading was passed to the ligament structures, and perhaps implementation of a muscle model could help determine dynamically the loading of the ligament structures. It is safe to assume, however, that as the resultant joint moments increase, the likelihood of joint injury also increases, and if neuromuscular perturbations occur altering the timing or magnitude of the muscle forces, more of this loading could be transferred to the ligament structures, possibly resulting in injury. As well, the angular impulse of the adduction moment increased by 15% in the high-traction shoe, putting further strain on the knee joint, which may well lead to chronic injury. These increases in knee adduction load and angular impulse occurred right at the plant-and-cut portion of the movement where the knee would be most vulnerable and may enhance the risk for ACL injury. The results of this study may be considered to be in contrast to the results of Kaila.10 He found the joint loading between different soccer shoe conditions to not be significantly different. However, the major limitation in that study is that the traction of the shoes and surface were never measured. If the traction properties of the shoesurface interface were similar between conditions, then the loading may be similar as well. Also, the study participants did not perform the movement at maximum effort, and the only information on participant speed was that it was between 5.5 m/s and 6.0 m/s. Perhaps the fact that the participants were not performing at their maximum effort affected the results. The present study determined that if 2 shoes have significantly different traction, the frontal and transverse plane loading at the ankle and knee joint between the 2 conditions will also be significantly different.

Performance
In sports, each athlete strives to achieve maximum athletic performance with the traction of the shoe and surface having a large effect on performance. In each condition, the maximum speed of movement was determined during the practice trials. Within each condition, the speed of each trial was required to be within 5% of the maximum speed attained during the practice trials; however, between conditions, the speed was not restrained. When the speed of motion was examined as well as the stance time between the 2 conditions, there were no significant differences present. This indicates that although the traction of the shoe types was quite dissimilar, the athletes were still able to perform identically across conditions. Therefore, if an athlete were to take a shoe that has less traction, his or her risk of injury would be decreased because of the decrease in joint loading, but the performance would remain constant. This may not be true for all movements, shoes, and traction coefficients as there must be a point when the traction of the shoe-surface interface becomes too low, hindering the performance of

Vol. XX, No. X, XXXX

Footwear Traction and Joint Loading 7

the athlete, and future work should attempt to determine this point.

Traction
From the traction testing results, it appears that the higher linear translational traction correlated to higher peak moments of rotation (resistance to rotation). Shorten et al22 found similar results previously using soccer shoes. As linear traction increased, rotational traction seemed to also increase proportionally. It is generally believed that high rotational traction is what places an athlete at risk of injury; however, from the present study, this cannot be verified. It cannot be determined if the differences in joint loading were due to the linear traction, rotational traction, or some combination of the 2. Further research is needed, such as testing a shoe-surface combination in which 2 shoes have the same linear traction properties but different rotational traction properties, or vice versa, to determine which component of traction is more foretelling of injury risk. When studying footwear traction, all aspects with regard to the surface and shoe that could affect traction should be considered such as temperature and surface conditions (eg, grass length, moisture).32

CONCLUSION
This study has shown that as traction at the shoe-surface interface increases, so does the resultant joint loading in both the ankle and knee. The loading in the ankle joint increased 12% in the transverse plane, whereas the loading in the knee joint increased 13% and 20% in the transverse and frontal planes. The increased traction was enough to increase the knee joint loading into the previously reported rupture zone, and with the addition of muscular perturbations, the joint may be more susceptible to injury. At this time, it is unclear as to whether the linear or rotational traction of the shoe caused these increases in joint loading, or if the relationship between traction and loading is linear. Future studies will examine those aspects.

ACKNOWLEDGMENT
The authors thank the Markin-Flanigan Scholarship and the International Society of Biomechanics for financial support, as well as adidas International for the donation of the footwear used in testing.

REFERENCES
1. Andreasson G, Lindenberger U, Renstrom P, Peterson L. Torque developed at simulated sliding between sport shoes and an artificial turf. Am J Sports Med. 1986;14(3):225-230. 2. Baker BE. Prevention of ligament injuries to the knee. Exerc Sport Sci Rev. 1990;18:291-305. 3. Bowers KD JR, Martin RB. Cleat-surface friction on new and old Astroturf. Med Sci Sports. 1975;7(2):132-135.

4. Bramwell ST, Requa RK, Garrick JG. High school football injuries: a pilot comparison of playing surfaces. Med Sci Sports Exerc. 1972;4:166-169. 5. Ekstrand J, Nigg BM. Surface-related injuries in soccer. Sports Med. 1989;8:56-62. 6. Frederick EC. Kinematically mediated effects of sports shoe design: a review. J Sports Sci. 1986;4:169-184. 7. Heidt RS Jr, Dormer SG, Cawley PW, Scranton PE Jr, Losse G, Howard M. Differences in friction and torsional resistance in athletic shoe-turf surface interfaces. Am J Sports Med. 1996;24(6): 834-842. 8. Hewett TE, Myer GD, Ford KR, et al. Biomechanical measures of neuromuscular control and valgus loading of the knee predict anterior cruciate ligament injury risk in female athletes: a prospective study. Am J Sports Med. 2005;33(4):492-501. 9. Hurwitz DE, Sumner DR, Andriacchi TP, Sugar DA. Dynamic knee loads during gait predict proximal tibial bone distribution. J Biomech. 1998;31(5):423-430. 10. Kaila R. Influence of modern studded and bladed soccer boots and sidestep cutting on knee joint loading during match play conditions. Am J Sports Med. 2007;35(9):1528-1536. 11. Lambson RB, Barnhill BS, Higgins RW. Football cleat design and its effect on anterior cruciate ligament injuries: a three-year prospective study. Am J Sports Med. 1996;24(2):155-159. 12. Levy IM, Skovron ML, Agel J. Living with artificial grass: a knowledge update, part 1: basic science. Am J Sports Med. 1990;18:406-412. 13. Luethi S, Nigg BM. The influence of different shoe constructions on discomfort and pain in tennis. In: Winter DA, Norman RW, Wells RP, Hayes KC, Patla AE. eds. Biomechanics IX-B. Champaign, IL: Human Kinetics; 1985:149-153. 14. Malone TR, Hardaker WT, Garrett WE, et al. Relationship of gender to ACL injuries in intercollegiate basketball players. J South Orthop Assoc. 1992;2:36-39. 15. Mizuno K, Andrish JT, van den Bogert AJ, McLean SG. Gender dimorphic ACL strain in response to combined dynamic 3D knee joint loading: implications for ACL injury risk. Knee. 2009;16(6):432-440. 16. Myklebust G, Maehlum S, Engebretsen L, Strand T, Solheim E. Registration of cruciate ligament injuries in Norwegian top level team handball: a prospective study covering two seasons. Scand J Med Sci Sports. 1997;7:289-292. 17. Myklebust G, Maehlum S, Holm I, Bahr R. A prospective cohort study of anterior cruciate ligament injuries in elite Norwegian team handball. Scand J Med Sci Sports. 1998;8:149-153. 18. Nigg BM, Yeadon MR. Biomechanical aspects of playing surfaces. J Sports Sci. 1987;5:117-145. 19. Seering WP, Piziali RL, Nagel DA, Schurman DJ. The function of the primary ligaments of the knee in varus-valgus and axial rotation. J Biomech. 1980;13:785-794. 20. Sharma L, Hurwitz DE, Thonar EJ, et al. Knee adduction moment, serum hyaluronan level, and disease severity in medial tibiofemoral osteoarthritis. Arthritis Rheum. 1998;41(7):1233-1240. 21. Shin CS, Chaudhari AM, Andriacchi TP. The effect of isolated valgus moments on ACL strain during single-leg landing: a simulation study. J Biomech. 2009;42(3):280-285. 22. Shorten MR, Hudson B, Himmelsbach J. Shoe-surface traction of conventional and in-filled synthetic turf football surfaces. In: Milburn P, Wilson B, Yanai T, eds. Proceedings XIX International Congress of Biomechanics. Dunedin, NZ: University of Otago; 2003. 23. Skovron ML, Levy IM, Agel J. Living with artificial grass: a knowledge update, part 2: epidemiology. Am J Sports Med. 1990;18:510-513. 24. Stacoff A, Steger J, Stussi E. The control of the rearfoot in lateral movements in sports. Sportverletz Sportschaden. 1993;7:22-29. 25. Stacoff A, Steger J, Stussi E, Reinschmidt C. Lateral stability in sideward cutting movements. Med Sci Sports Exerc. 1996;28:350-358. 26. Stanitski CL, McMaster JH, Ferguson RJ. Synthetic turf and grass: a comparative study. J Sports Med. 1974;2:22-26. 27. Stefanyshyn DJ, Stergiou P, Lun VMY, Meeuwisse WH, Worobets JT. Knee angular impulse as a predictor of patellofemoral pain in runners. Am J Sports Med. 2006;34(11):1844-1851.

Wannop et al

The American Journal of Sports Medicine

28. Stucke H, Baudzus W, Baumann W. On friction characteristics of playing surfaces. In: Frederick EC, ed. Sport Shoes and Playing Surfaces. Champaign, IL: Human Kinetics Publishers; 1984: 87-97. 29. Torg JS, Quedenfeld TC. Effect of shoe type and cleat length on incidence and severity of knee injuries among high school football players. Res Q. 1971;42:203-211.

30. Torg JS, Quedenfeld TC. Knee and ankle injuries traced to shoes and cleats. Phys Sportsmed. 1973;1:39-43. 31. Torg JS, Quedenfeld TC, Landau S. The shoe-surface interface and its relationship to football knee injuries. J Sports Med. 1974;2:261-269. 32. Torg JS, Stilwell G, Rogers K. The effect of ambient temperature on the shoe-surface interface release coefficient. Am J Sports Med. 1996;24(1):79-82.

For reprints and permission queries, please visit SAGEs Web site at http://www.sagepub.com/journalsPermissions.nav