QUADRICEPS MUSCLE ATROPHY frequently follows

1211 Bilateral Kinematic and Kinetic Analysis of the Squat Exercise After Anterior Cruciate Ligament Reconstruction George J. Salem, PhD, Ruben Salinas, DPT, F. Victor Harding, MA ABSTRACT. Salem GJ, Salinas R, Harding V. Bilateral kinematic and kinetic analysis of the squat exercise after anterior cruciate ligament reconstruction. Arch Phys Med Rehabil 2003;84:1211-6. Objectives: To characterize the bilateral lower-extremity kinematics and kinetics associated with squatting exercise after anterior cruciate ligament (ACL) reconstruction. Design: We evaluated bilaterally sagittal plane kinematics and kinetics of the ankle, knee, and hip joints during submaximal squatting exercise in rehabilitating patients after ACL reconstruction. Comparisons were performed between involved and noninvolved limbs, and regression models were created to examine the relations between the bilateral kinetic differences and time postsurgery. Setting: A motion analysis laboratory. Participants: Eight adults (27.9 6.8y) with unilateral ACL reconstruction (postsurgical time, 30 12wk). Interventions: Not applicable. Main Outcome Measures: Sagittal plane ankle, knee, and hip peak net moments of force, maximum joint excursion angles, and peak vertical ground reaction forces. Results: Peak vertical ground reaction forces did not differ between limbs. The peak knee extensor moment generated during the submaximal squatting exercise was 25.5% greater in the noninvolved limb than in the involved limb (P.003). The peak ankle plantarflexor moment did not differ between limbs (P.85); however, there was a trend toward a greater hip extensor moment in the involved limb (P.06). The ratio of the peak hip extensor moment to the peak knee extensor moment was 46.5% greater in the involved limb (P.02). Only the peak dorsiflexion angle differed between limbs (P.02). None of the linear models examining the relations between differences in the involved limb and noninvolved limb kinetics, and postsurgical time, were statistically significant. Conclusions: Patients performing the squat exercise, within 1 year of ACL reconstructive surgery, used 2 strategies for generating the joint torques required to perform the movement: (1) in the noninvolved limb, patients used a strategy that equally distributed the muscular effort between the hip and From the Musculoskeletal Biomechanics Research Laboratory, Department of Biokinesiology and Physical Therapy, University of Southern California, Los Angeles, CA. Supported by the California Physical Therapy Fund (grant no. 99-14). Clinical support was provided by Fortanasce and Associates, Physical Therapy. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors(s) or upon any organization with which the author(s) is/are associated. Reprint requests to George J. Salem, PhD, Dept of Biokinesiology and Physical Therapy, University of Southern California, 1540 E Alcazar St, CHP-155, Los Angeles, CA 90089, e-mail: gsalem@hsc.usc.edu. 0003-9993/03/8408-7733$30.00/0 doi:10.1016/s0003-9993(03)00034-0 knee extensors, and (2) in the involved limb, patients used a strategy that increased the muscular effort at the hip and reduced the effort at the knee. These intra- and interlimb motor-programming alterations (ie, substitution strategies) could potentially slow or limit rehabilitation, and induce strength and performance deficits. Key Words: Anterior cruciate ligament; Biomechanics; Joints; Rehabilitation. 2003 by the American Congress of Rehabilitation Medicine and the American Academy of Physical Medicine and Rehabilitation QUADRICEPS MUSCLE ATROPHY frequently follows anterior cruciate ligament (ACL) injury and repair. 1-3 Because this muscle is a fundamental extensor and stabilizer of the knee joint, restoration of quadriceps muscle mass and performance is paramount to successful ACL rehabilitation. 3-5 Despite this muscle s importance, however, there is currently no consensus regarding the optimal therapeutic prescription for restoring its function after ACL surgical repair. 6-8 Therapeutic exercise selection has evolved from the classic model (eg, limited weight bearing, immobilization, single-joint exercises) 1,9 to the current model emphasizing early weight bearing, functional progression, and multiple-joint exercises. 10-12 The current preference for exercises performed by using simultaneous movements at several joints is based on the assumption that these exercises more closely approximate functional activities, such as stair climbing and rising from a chair, because of the interplay between synergistic muscle activity and multiplanar force application. 3,13,14 Additionally, it is believed that the strain that some single-joint exercises place on the maturing graft may adversely affect the long-term stability of the reconstructed knee. 13,15,16 Biomechanical, electromyographic, and strain gauge studies have been used to compare and contrast single- and multiple-joint exercises 3,13,15,17-19 ; in response, some authors 10,12 have recommended the exclusive use of multiple-joint exercise and the exclusion of traditional single-joint exercise. The squat exercise is a classic multiple-joint exercise that has become an integral part of most lower-extremity strengthening and postoperative ACL rehabilitation programs. 10,12 During squatting activities, extensor moments produced at the ankle, knee, and hip joints must be coordinated to appropriately control the descent and ascent of the body s center of mass. The multiple-joint characteristics of these movements, however, may allow patients to use intralimb substitution patterns that shift the effort from a targeted muscle group (eg, knee extensors) to another muscle group (eg, hip extensors). Moreover, when these exercises are performed bilaterally, patients may shift the effort from the targeted limb (ie, involved limb) to the contralateral limb. These substitution patterns may ultimately limit the effectiveness of the intervention exercise preventing the correction of weakness in the involved limb.

1212 ACL REHABILITATION: SQUATTING MECHANICS, Salem Lower-extremity, postsurgical, bilateral strength differences are a concern for rehabilitation professionals because retrospective studies suggest that side-to-side deficits (10% 25%) in quadriceps cross-sectional area and peak torque may persist for up to 49 months after surgery. 2,20 Additionally, strength training studies using multiple-joint exercise, but testing isolated quadriceps function alone, have found no statistically significant increases in quadriceps strength during single-joint strength measures, even though increases in strength were identified by using multiple-joint strength measures. 6,14 These findings suggest that the exclusive use of multiple-joint exercises, for the purpose of reversing isolated quadriceps atrophy and disfunction, could be problematic. During the squat, the knee and hip extensor moments are coupled because of the effects of trunk flexion on the position of the line of gravity with respect to the 2 joint centers. A more flexed trunk posture moves the line of gravity closer to the knee and away from the hip decreasing the internal knee extensor moment and increasing the internal hip extensor moment. 21 Conversely, a more extended hip posture will shift the muscular effort away from the hip extensors and toward the knee extensors. Importantly, a continuum of knee-hip contribution is available for patients performing this multiple-joint task, and successful squatting motions can be achieved by using either a knee-dominant or hip-dominant strategy. We hypothesized that insults to the knee as significant as an ACL rupture and reconstruction would alter motor performance and the relative contributions of the knee and hip musculature during the squat exercise. Specifically, we believed that our postoperative subjects would adopt a movement strategy that increased the hip extensor moments of force and decreased the knee extensor moments of force in their injured limb during performance of the maneuver. To characterize these relative contributions, we report the bilateral lower-extremity kinematics and kinetics of postoperative ACL reconstruction patients during performance of the back squat exercise. METHODS Participants Seven men and 1 women who had undergone ACL reconstructive surgery within the previous year participated in this study. They were recruited through physician referrals and from physical therapy clinics in the greater Los Angeles area. The participants average age was 27.9 6.8 years, average body mass was 93.3 18.2kg, and average time since surgery was 30 12 weeks. Scheduling of the biomechanical testing sessions was based on considerations of subject availability and time postsurgery. The participants were asked, With which leg do you kick a ball? All participants responded that they kicked a ball with their right leg; thus, we classified all participants as right-leg dominant. Four subjects had dominant limb (right) ACL reconstructions, and 4 subjects had nondominant limb ACL reconstructions. Inclusion criteria for participants were (1) bone patellar ligament or bone-autograft reconstruction, (2) no other ligamentous injury, (3) no meniscal injury, (4) full pain-free range of motion, and (5) absence of joint pain or swelling. A clinical examination was performed by a licensed physical therapist to ensure that all subjects were free of knee joint pain and swelling. The University of Southern California s Institutional Review Board approved all procedures and all participants provided their consent before testing. Instrumentation Bilateral tracking markers were taped to the skin above bony landmarks of the lower extremities (anterior superior iliac spine, posterior superior iliac spine, greater trochanter, lateral thigh, lateral femoral condyle, lateral shank, lateral malleolus, fifth and second metatarsals). Three-dimensional coordinate data were then collected by means of a 6-camera motion analysis system (Vicon a ) at a 60-Hz sampling rate. Ground reaction forces were recorded from 2 force platforms b at a 600-Hz sampling rate. Experimental Procedures Testing, anthropometric measures (body weight, height, leg length; ankle, knee, hip widths) were obtained from each participant. The participants then warmed up for 5 minutes on a stationary bicycle and practiced the testing procedures before data collection began. During the biomechanical analysis, efforts were made to minimize alterations to the participant s normal (ie, previously utilized) training techniques, including foot position, movement speed, and joint kinematics. These efforts were employed for 3 reasons: (1) we were interested in quantifying the kinematics and kinetics associated with the performance of their adopted, well-practiced, rehabilitation exercise patterns; (2) we were concerned that alterations that participants might make, in response to more rigid performance instructions, would be transitory and not reflect longer-term movement strategies; and (3) the participants had received extensive previous instruction from physical therapists and were experienced in performing the exercise all subjects had performed the exercise 2 to 3 times weekly, for a duration of at least 6 weeks, and all subjects continued to perform the exercise throughout their postsurgical rehabilitation period. Participants, while standing with 1 foot on each of 2 force platforms, were instructed to assume a stance width consistent with that used during their previous rehabilitation training. They were then instructed to descend to a level to where your posterior thighs are parallel to the floor and then ascend. Participants performed 3 sets of 10 repetitions of the back squat exercise using a resistance weight of 35% body weight. The subjects self-selected their movement pace when performing the repetitions; however, a 3-minute rest period between sets was instituted to limit the potential effects of muscle fatigue. Data processing software (Bodybuilder a ) was then used to calculate maximum joint excursion angles in the sagittal plane and peak net moments of force produced at the ankle, knee, and hip, during the back squat exercise (fig 1). All moments were normalized to the subject s body mass. Using this data, a ratio of the hip-to-knee peak net extensor moments of force (hip: knee ratio) was also created by dividing the peak hip net extensor moment by the peak knee net extensor moment. The hip:knee ratio is a measure of the relative targeting of the knee and hip extensor musculature during the back squat exercise. A hip:knee ratio greater than 1 indicates that an exercise selectively loads the hip extensors (induces a greater peak net joint moment); whereas a hip:knee ratio less than 1 indicates that the exercise selectively loads the knee extensors. Similar hip and knee moment comparisons have been used to contrast squatting techniques in athletic populations. 22 This instrumentation and data processing have previously been used in our laboratory to assess resistance-exercise performance with high reliability (Cronbach.98). Statistical Analysis All statistical analyses were performed by using SPSS, version 7.5, c for Windows. The maximum joint excursion angles

ACL REHABILITATION: SQUATTING MECHANICS, Salem 1213 Table 1: Involved Limb and Noninvolved Limb Values of Maximum Joint Excursion Angles for the Ankle, Knee, and Hip (N 8) Measurement P* Noninvolved Involved Ankle dorsiflexion.02 34.12 6.64 31.80 6.55 Knee flexion.12 109.55 11.47 106.98 10.41 Hip flexion.62 95.64 7.68 96.95 10.36 NOTE. All values are maximum standard deviation (SD) joint excursion angles in degrees, averaged over 3 squatting repetitions, across 3 exercise sets, for each participant. *P value for t test the average maximum joint excursion angles differed between noninvolved and involved limbs. P values were adjusted for the multiple comparisons by using Bonferroni adjustments. involved limb (table 1). No other statistically significant differences, in lower-extremity peak joint excursion angles, were identified between involved and noninvolved limbs (P.05). Fig 1. Free-body diagram illustrating the direction of the internal (muscular) moments (M) and joint excursion angles (A) for the ankle (a), knee (k), and hip (h) joints. Abbreviations: ASIS, anterior superior iliac spine; PSIS, posterior superior iliac spine. and peak moments of force at the hip, knee, and ankle joints were averaged across 3 squatting sets and used for statistical analysis. Repeated-measures t tests were used to determine if statistically significant differences in the mean values of the maximum joint excursions and peak moments of force existed between the noninvolved and involved limbs. Statistical analyses were adjusted for the multiple comparisons by using Bonferroni adjustments. A secondary analysis was also conducted to determine if the differences between kinetics of the involved and noninvolved limbs varied by postsurgical time. We created between-limb difference variables (defined as the peak net moment of the involved limb ankle, knee, or hip less peak net moment of the noninvolved limb ankle, knee, or hip). We then used multiple regression models to fit these ankle, knee, and hip difference variables to the postsurgical time (weeks postsurgery). Adjusted regression models were controlled for participant age and height. Finally, in order to assess the repeatability of participant performance across the 3 squatting sets, we calculated reliability coefficients (Cronbach ) for hip: knee ratio (our primary endpoint) for both the involved and noninvolved limbs. RESULTS Kinematics Average peak ankle dorsiflexion angles assumed during the squat were 7.3% greater in the noninvolved limb than in the Kinetics Peak vertical ground reaction forces averaged 757 173N and did not differ between limbs (P.28). Representative joint moment curves for the knee and hip are presented in figure 2. Across all participants, the average peak hip net extensor moment generated in the involved limb did not statistically significantly differ from that generated in the noninvolved limb; however, there was a trend toward an increased peak hip moment in the involved limb (P.06). The peak knee extensor moment generated during the squat was 25.5% greater in the noninvolved limb than in the involved limb (P.003; table 2). The peak ankle plantarflexor moment did not differ between limbs (P.85). The hip:knee ratio of the involved limb was 46.5% greater than the hip:knee ratio of the noninvolved limb (P.02; fig 3). None of the models examining the relations between differences in kinetics of the involved limb and noninvolved limb and postsurgical time was statistically significant (table 3). Between-set hip:knee ratio repeatability was high for both the involved (.985) and noninvolved (.968) limbs. DISCUSSION Our findings show 2 distinct, limb-specific strategies for generating the joint torques required to perform the squat exercise in our sample of patients less than 1-year after ACL reconstruction. In the noninvolved limb, subjects used a strategy that equally distributed the peak moments between the hip and knee; thus, the hip:knee ratio of the noninvolved limb approached unity. Conversely, these same participants used a more hip dominant strategy in their involved limb the hip: knee ratio was 1.7 in the involved limb and 46.5% greater than Table 2: Involved Limb and Noninvolved Limb Values of Average Peak Moments of Force for the Ankle, Knee, and Hip (N 8) Measurement P* Noninvolved Involved Ankle plantarflexor.85 0.79 0.07 0.80 0.16 Knee extensor.003 1.28 0.28 1.02 0.31 Hip extensor.06 1.12 0.38 1.67 0.68 NOTE. All values are net peak SD sagittal plane moments of force, averaged over 3 squatting repetitions, across 3 exercise sets, and normalized to body mass (Nm/kg). *Bonferonni-adjusted P value for t test that the average maximum peak moments of force differed between noninvolved and involved limbs.

1214 ACL REHABILITATION: SQUATTING MECHANICS, Salem Fig 2. Representative net moment/time curves for a single participant for (A) knee extensors and (B) hip extensors. Data are from 3 successive repetitions within a single set; vertical bars identify the time of maximum knee excursion angle; the ascending (ASC) and descending (DES) phases of each repetition are shown. the hip:knee ratio of the noninvolved limb. The greater hip: knee ratio generated by these rehabilitating participants in their involved limb suggests that their motor-control strategy reduced their knee extensors peak effort by increasing the peak effort of their hip extensors. These findings are supported by the knee extensor moment analysis that identified a 25.5% greater peak knee extensor moment in the noninvolved limb. These results suggest that bilateral motor-performance differences may be dramatically influenced by ACL injury, repair, and rehabilitation. Further, the substitution strategies used during bilateral functional rehabilitative exercises could potentially slow or limit rehabilitation, induce strength and performance deficits, and increase the risk of future injury. Indeed, statistically significant bilateral strength deficits in ACL rehabilitating patients may exist up to 49 months after surgery. 2,21 These previous reports highlighted the potential limitations associated with ACL rehabilitation protocols; however, they did not experimentally address the potential mechanisms contributing to these limitations. The present investigation identified a potential mechanism by demonstrating the substitution patterns adopted by rehabilitating patients during the performance of a bilateral multiple-joint exercise. The present study also suggests that these strategies may be used even during submaximal exercise with moderate resistance weight (35% body weight). Thus, these kinetic-pattern changes are likely not driven by deficits in peak quadriceps strength alone. Furthermore, sagittal plane peak net moment hip:knee ratio differences were not accompanied by differences in the maximum excursion angles or peak vertical ground reaction forces between limbs. These findings may be explained in part by temporal pattern differences in the reaction forces, maximum excursion angles, and peak net moments (ie, they do not all occur at the same time); however, nonsagittal plane excursions could also influence both intra- and interlimb kinetic patterns by altering the position of the joint centers. These potential nonsagittal plane excursions and moments are likely to be greatest at the Table 3: Linear Relations Between the Difference* of the Peak Joint Moments Generated in the Involved and Noninvolved Limbs During the Squat and Weeks Postsurgery Model (dependent variable) Model R P Plantarflexor difference.67.44 Knee extensor difference.44.81 Hip extensor difference.57.63 Fig 3. The average peak hip extensor moment to peak knee extensor moment ratio. *Hip:knee ratio (HKR) was 46.5% greater in the involved limb (P.02). *Between-limb difference variables were defined as the peak involved limb ankle, knee, or hip moment minus peak noninvolved limb ankle, knee, or hip moment. Models adjusted for age and height.

ACL REHABILITATION: SQUATTING MECHANICS, Salem 1215 hip joint; consequently, we believe that clinicians, in addition to monitoring sagittal plane joint excursions, should also give attention to transverse and frontal plane movements of the hip joint. The current preference for multiple-joint functional exercise is based on the assumption that these exercises more closely approximate functional activities, and, thus, train task-appropriate motor programs. 3,13,14 The present findings, however, highlight the potential limitations of exclusively using bilateral- and multiple-joint exercises to rehabilitate patients after ACL reconstruction. For example, the bilateral kinetic differences identified in the present study did not vary with rehabilitation time during the 1-year postsurgical period none of the models that we used to examine the relations between differences in the involved and noninvolved limb kinetics, and postsurgical time, were statistically significant (.44 P.88). These findings suggest that even with explicit instruction, careful observation, and numerous rehabilitative training sessions, patients may continue to use substitution patterns to unload involved structures. Because these kinetic patterns occurred without concomitant differences in knee or hip maximum excursions, without bilateral differences in ground reaction forces, and with moderate resistance, clinicians may find it difficult to identify potential joint substitution patterns by qualitatively examining movement form (eg, relative joint excursions) alone or by using gross measures of lower-extremity loading (eg, weight scales, balance boards). We believe it may be prudent for clinicians to monitor bilateral hip and knee strength changes frequently throughout the rehabilitation period, and to use both multiple-joint exercises and single-joint isolation exercises to target specific lower-extremity muscle groups. Ultimately, randomized controlled trials should be performed to examine the relative efficacy of various multiplejoint exercise programs, single-joint exercise programs, and combination programs on the restoration of lower-extremity muscle mass, bilateral kinetic symmetry, and physical function after ACL reconstruction. Limitations The present investigation did not permit an assessment of patient technique adaptations after 1 year because all of our patients were less than 1 year postsurgery. Because substitution patterns could be transitory, future investigations should include longitudinal assessments of patients for longer than 1 year. Our small sample size may have limited our ability to find statistically significant differences in our outcome measures. The observed power for the bilateral knee moment, hip moment, and hip:knee ratio differences were all above 0.9; whereas, the observed power for the ankle moment difference was only.07. Sample size calculations, however, revealed that approximately 204 subjects would be required to find statistical significance between the bilateral peak ankle moments. The present study only examined a single multiple-joint exercise, the squat, and did not quantify the kinetics associated with other multiple-joint exercises such as the lunge and leg press. Future investigations should examine additional multiple-joint therapeutic activities. Finally, it would be informative to have a systematic characterization of the effects that various resistance weights, foot positions, and protocol instructions have on the kinetic patterns associated with squatting in rehabilitating ACL patients. CONCLUSION Patients performing the squat exercise, within a 1 year after ACL reconstructive surgery, used 2 distinct limb-specific strategies for generating the joint torques required to perform the movement. In the noninvolved limb, patients used a strategy that equally distributed the muscular effort between the hip and knee extensors. In the involved limb, patients used a strategy that increased the muscular effort at the hip and reduced the effort at the knee. These intra- and interlimb motor-programming alterations, which occurred without concomitant differences in maximum joint excursions and peak vertical ground reaction forces, could potentially slow or limit rehabilitation, and induce strength and performance deficits. References 1. Johnson RJ, Beynnon BD, Nichols CE, Renstrom PA. Current concepts review: the treatment of injuries of the anterior cruciate ligament. J Bone Joint Surg Am 1992;74:140-51. 2. LoPresti C, Kirkendall DT, Street GM, Dudley AW. Quadriceps insufficiency following repair of the anterior cruciate ligament. J Orthop Sports Phys Ther 1988;9:245-9. 3. Wilk KE, Escamilla RF, Fleisig GS, Barrentine SW, Andrews JR, Boyd ML. A comparison of tibiofemoral joint forces and electromyographic activity during open and closed chain exercises. Am J Sports Med 1996;24:518-27. 4. Arms SW, Pope MH, Johnson RF, Fischer RA, Arvidsson I, Eriksson E. The biomechanics of anterior cruciate ligament rehabilitation and reconstruction. Am J Sports Med 1984;12:8-18. 5. Snyder-Mackler L, Delitto A, Bailey SL, Stralka SW. Strength of the quadriceps femoris muscle and functional recovery after reconstruction of the anterior cruciate ligament: a prospective, randomized clinical trial of electrical stimulation. J Bone Joint Surg Am 1995;77:1166-73. 6. Augustsson J, Esko A, Thomee R, Svantesson U. Weight training of the thigh muscles using closed vs. open kinetic chain exercises: a comparison of performance enhancement. J Orthop Sports Phys Ther 1998;27:3-8. 7. Beynnon BD, Fleming BC, Johnson RJ, Nichols CE, Renstrom PA, Pope MH. Anterior cruciate ligament strain behavior during rehabilitation exercises in vivo. Am J Sports Med 1995;23:24-34. 8. Fitzgerald GK. Open versus closed kinetic chain exercise: issues in rehabilitation after anterior cruciate ligament reconstructive surgery. Phys Ther 1997;77:1747-54. 9. Fu FH, Woo SL, Irrgang JJ. Current concepts for rehabilitation following anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther 1992;15:270-8. 10. Bynum EB, Barrack RL, Alexander AH. Open versus closed chain kinetic exercises after anterior cruciate ligament reconstruction. Am J Sports Med 1995;23:401-6. 11. Shelbourne KD, Nitz P. Accelerated rehabilitation after anterior cruciate ligament reconstruction. Am J Sports Med 1990;18:292-9. 12. Shelbourne KD, Klootwyk TE, DeCarlo MS. Update on accelerated rehabilitation after anterior cruciate ligament reconstruction. J Orthop Sports Phys Ther 1992;15:303-7. 13. Lutz GE, Palmitier RA, An KN, Chao EY. Comparison of tibiofemoral joint forces during open-kinetic-chain and closedkinetic-chain exercises. J Bone Joint Surg Am 1993;75:32-9. 14. Worrell TW, Borchert B, Erner K, Fritz J, Leerar P. Effect of lateral step-up exercise protocol on quadriceps and lower extremity performance. J Orthop Sports Phys Ther 1993;6:646-53. 15. Henning CE, Lynch MA, Glick KR. An in vivo strain gage study of elongation of the anterior cruciate ligament. Am J Sports Med 1985;13:22-6. 16. Stuart MJ, Meglan DA, Lutz GE, Growney ES, An KN. Comparison of intersegmental tibiofemoral joint forces and muscle activity during various closed kinetic chain exercises. Am J Sports Med 1996;24:792-9. 17. Escamilla RF, Fleisig GS, Zheng N, Barrentine SW, Wilk KE, Andrews JR. Biomechanics of the knee during closed kinetic chain and open kinetic chain exercises. Med Sci Sports Exerc 1998;30:556-69.

1216 ACL REHABILITATION: SQUATTING MECHANICS, Salem 18. Gryzlo SM, Patek RM, Pink M, Perry J. Electromyographic analysis of knee rehabilitation exercises. J Orthop Sports Phys Ther 1994;20:36-43. 19. Signorile JF, Weber B, Roll B, Caruso JF, Lowensteyn I, Perry AC. An electromyographical comparison of the squat and knee extension exercises. J Strength Cond Res 1994;8:178-83. 20. Arangio GA, Chen C, Kalady M, Reed JF. Thigh muscle size and strength after anterior cruciate ligament reconstruction and rehabilitation. J Orthop Sports Phys Ther 1997;26:238-43. 21. Escamilla RF, Flesig GS, Lowry TM, Barrentine SW, Andrews JR. Three-dimensional biomechanical analysis of the squat during varying stance widths. Med Sci Sports Exerc 2001;33:984-98. 22. Wretenberg P, Feng Y, Arborelius UP. High- and low-bar squatting techniques during weight-training. Med Sci Sports Exerc 1996;28:218-24. Suppliers a. Vicon Motion Systems, 14 Minns Business Park, West Way, Oxford OX2 0JB, UK. b. Advanced Mechanical Technology Inc, 176 Waltham St, Watertown, MA 02472. c. SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.