s / Osteoarthritis and Cartilage 24 (2016) S63eS534 S92 143 BALANCE RECOVERY DURING INDUCED FALLS IS IMPAIRED IN PEOPLE WITH KNEE OSTEOARTHRITIS: IMPLICATIONS FOR FALLS PREVENTION P. Levinger y, H. Nagano y, C. Downie y, A. Hayes y, K. Sanders z, F. Cicuttini x, R. Begg y. yVictoria Univ., Melbourne, Australia; zAustralian Catholic Univ., Melbourne, Australia; xMonash Univ., Melbourne, Australia Purpose: Knee osteoarthritis (OA) is a major risk factor for falls with high falls rates reported; 50% in people with knee OA who are over the age of 60. Increased neuromuscular deficits are common in people with knee OAwhichmay further impact on postural control mechanisms and the ability to defend against a fall. Despite the high prevalence of falls in people with knee OA, the mechanism of falling in this group is unclear. This study investigated the biomechanical response of the lower limb joints during a forward induced fall under different task conditions in people with knee OA. Moreover, the balance recovery step responses (single vs multi steps) were also investigated to determine how dynamic postural control would be affected due to the presence of knee OA while undertaking the three task conditions. Methods: Twenty four participants with knee OA (68.6± 6.2 years) and fifteen asymptomatic healthy controls (72.4± 4.8 years) participated in the study. Forward fallwas induced by releasing participants froma static forward leaning position. Participants were required to recover balance during three conditions: normal, physical dual task (obstacle clearance) and cognitive dual task (counting backwards). Spatiotemporal parameters, lower limb joint kinematics and kinetics of the recovery limb were compared between the two groups and across the three task conditions using two way full factorial Multiple Analyses of Covariate (MANCOVA) with Bonferroni-adjusted post-hoc tests with age as a covariate. Results: TheOAgroupdemonstrated significantly slower spatio-temporal characteristics and reduced hip and knee flexion angles, joint moments/ powers and lower knee and ankle negative works. Cognitive dual task resulted in reduced centre of mass velocity and step length (p1⁄4 0.03) compared to the physical dual task condition. Reduced knee (p1⁄4 0.02) and hip powers (p1⁄4 0.03) were demonstrated in the OA group during the dual task conditions especially obstacle clearance where additional cognitive demand was required. The number of participants who responded with multi steps was similar for both groups during normal balance recovery task (60%; p>0.05). Although not significant, with increased task complexity the number of multi-step responders increased in the OA group compared to the control (75% for OA vs 60% for control during the obstacle task and 83% for OA vs 60% for control during the cognitive task). Conclusions: When simulating a forward fall, participants with knee OA demonstrated difficulty in absorbing the impact and slowing down the forward momentum of the body during a recovery step. Differences in biomechanical response of the hip and knee were identified which could provide useful and important information regarding the strategies used by people with knee OA to recover balance and avoid a potential fall. Poor dynamic postural control was also demonstrated when required to simultaneously recover balance and attend to a secondary task such as obstacle clearance and cognitive task. As task complexity increased (i.e. use of physical/cognitive dual task), poor recovery strategies (multiple steps) were used more often in the knee OA group. Hence, balance impairment may be further exacerbated by impaired attention dynamics in people with knee OA. 144 TIBIOFEMORAL CONTACT FORCES 2e3 YEARS FOLLOWING ANTERIOR CRUCUATE LIGAMENT RECONSTRUCTION: A COMPARISON WITH HEALTHY INDIVIDUALS D.J. Saxby y, A.L. Bryant z, L. Modenese y, P. Gerus x, B. Killen y, J. Konrath y, K. Fortin z, T.V. Wrigley z, K.L. Bennell z, F.M. Cicuttini k, C. Vertullo y, J.A. Feller ¶, T. Whitehead#, P. Gallie yy, D.G. Lloyd y. yCtr. for Musculoskeletal Res., Menzies Hlth.Inst. Queensland, Griffith Univ., Gold Coast, Australia; zCtr. for Hlth., Exercise and Sports Med., Univ. of Melbourne, Melbourne, Australia; x Lab. of Human Motion, Ed. and Hlth., Univ. of Nice Sophia-Antipolis, Nice, France; kDept. of Epidemiology and Preventive Med., Monash Univ., Melbourne, Australia; Coll. of Sci., Hlth.and Engineering, La Trobe Univ., Melbourne, Australia; OrthoSport Victoria, Melbourne, Australia; yy Pacific Private Clinic, Gold Coast, Australia Purpose: Due to the risk of osteoarthritis development in those with anterior cruciate ligament (ACL) rupture and reconstruction (ACLR), we need to better understand in vivo articular contact loading in ACLR knees. To date, few studies have explored in vivo tibiofemoral contact forces in ACLR knees, and of those studies the examination has focused on affected and unaffected limbs, considered only females and/or been confined to small sample sizes. As a result, it remains unclear whether the ACLR tibiofemoral contact forces are comparable with healthy individuals, or whether they are peculiar in some way. The purpose of this study was to compare in vivo tibiofemoral contact forces in a large sample of ACLR individuals and healthy controls during normal walking and sporting gait tasks. Methods: This cross-sectional study was conducted at Griffith University’s Centre for Musculoskeletal Research (CMR) and University of Melbourne’s Centre for Health, Exercise and Sports Medicine (CHESM) (ethics approvals CMR: PES/36/10/HREC, CHESM: 0932864.3). Both ACLRs (68% male, n1⁄4 107, 29.7 ± 6.5 years, 78.1± 14.4 kg) and healthy controls (58% male, 27.5 ± 5.4 years, 67.8± 14.0 kg) participated. ACLRs were tested 2e3 years after a single combined hamstrings autograft reconstruction (2.51± 0.44 years). All participants underwent comprehensive 3-dimensional gait analysis wherein they walked (1.42 ± 0.20 m.s-1), ran (4.46± 0.46 m.s-1) and ran with a 45 diagonal sidestep (3.66± 0.60 m.s-1). Gait tasks were performed shod (Dunlop Volleys), and surface electromyography (EMG) was performed on the eight major muscles (rectus femoris, vastii, gastrocnemii, hamstrings, tensor fasciae latae) on the reconstructed (ACLRs) or randomized (controls) knee. From the gait analysis data, joint kinematics and kinetics were determined using OpenSim, and an EMG-driven neuromusculoskeletal model was then developed for each subject to determine their in vivo tibiofemoral contact forces during the different gait tasks. Analyses of variance were performed to determine if main and interaction effects existed for group (ACLR and controls) and gait task (walk, run, sidestep). If significant effects were found, ttests with Bonferroni corrections were applied to assess specific differences. Results: ACLRs and controls were of similar age, sex, tested limb-side and height, and had similar spatiotemporal parameters (i.e. velocity, cadence, etc...) during the different gait tasks. Compared to controls, ACLRs had a significantly smaller peak flexion angle, excursion and moment during running (angle: 41.4 ± 6 vs 44± 5.9 , excursion: 32.4 ± 5.4 vs 34.4 ± 4.6 , and moment: 0.25± 0.08 Nm.kg-1 vs 0.29 ± 0.06 Nm.kg-1) and sidestepping (angle: 50.6± 6.9 vs 53.4 ± 7.3 , excursion: 39.9 ± 6.7 vs 42.5 ± 7.1 , and moment: 0.29 ± 0.1 Nm.kg-1 vs 0.33 ± 0.1 Nm.kg-1) (all p