s / Osteoarthritis and Cartilage 24 (2016) S63eS534 S484 Purpose: The primary non-surgical management of patients with knee osteoarthritis (OA) is exercise with emphasis on increasing strength in the lower limb. Despite exercise interventions that focus on muscle strength is recommended as core treatments, important questions on the exact content and dose of the optimal exercise program for knee OA have yet to be answered. For example, it is not clear if interventions that follow the definition of muscle strength training are superior to interventions that do not. The purpose of this systematic review and metaanalysis was to investigate if exercise interventions that follow the American College of Sports Medicine (ACSM) definition of muscle strength training yield different effects on muscle strength, pain, and function than those that do not. Methods: A systematic search was performed in Medline, EMBase, CINAHL, PEDro, Cochrane Central Register of Controlled Trials. Randomised or quasi-randomised controlled trials comparing any type of exercise interventions with no intervention, waiting list, sham, or placebo in patients with knee OA and presenting data on muscle strength outcome and self-reported pain or function was considered eligible. No restriction on date and language in the literature search. ACSM defines strength training as a voluntary contraction against an external resistance, typically performed in especially designed equipment or with free weights. The external load should be above 40% of one repetition maximum (1RM) and the exercise performed in 2e4 sets of 8e12 repetitions. Two authors using predefined selection criteria did independent extraction of articles. Treatment effect sizes are expressed as ‘Standardised Mean Difference’ (SMD), and stratified analyses applied using random effects models (PROSPERO: CRD42014015344). Results: The inclusion criteria yielded 46 studies (5,101 patients). Nineteen trials included strengthening exercise according to the ACSM guidelines while 27 trials did not. Benefits of exercise were replicated with respect to muscle strength, pain, and function (Figure). A statistically significant difference in favour of ACSM exercise with respect to muscle strength gain was found (difference in SMD: 0.43 [95% CI 0.78e 0.07], P 1⁄4 0.02). No differences were found regarding effects on pain (difference in SMD: 0.07 [95% CI 0.42e0.28], P 1⁄4 0.70) and function (difference in SMD: 0.13 [95% CI 0.55e0.30], P 1⁄4 0.56). There was considerable heterogeneity across studies (I2>0.70). Conclusions: The studies that follow the ACSM guidelines for strength training had a superior effect on muscle strength gain compared to those that do not. However, there were no additional benefits regarding pain and function when comparing studies that follow the ACSM guidelines for strength training with studies that do not. 834 IS YOGA BETTER THAN AEROBIC/STRENGTHENING EXERCISES FOR MANAGING KNEE OSTEOARTHRITIS IN OLDER ADULTS? C.K. Cheung y, J. Wyman y, U. Bronas z, T. McCarthy y, J. Switzer y, M. Mathiason y. yUniv. of Minnesota, Minneapolis, MN, USA; z The Univ. of Illinois at Chicago, Chicago, IL, USA Purpose: Osteoarthritis (OA) is the most common type of arthritis among older adults. In addition to using medical treatment for suppressing the symptoms and the inflammation, exercise is often recommended as a first-line intervention approach for OA and yet the optimal type has not been determined. Yoga is a mind-body intervention that is commonly used for promoting flexibility, muscle strength and relaxation. It has been proven to be a safe and feasible intervention for older adults with OA. The purpose of the study was to compare the effects of yoga and aerobic/strengthening exercises on OA symptoms, physical function, mood, spiritual health, fear of falling, and quality of life in community-dwelling older adults with knee OA. Methods: A randomized controlled trial design with three arms was used: Hatha yoga (HY), aerobic/strengthening exercises (AE), and education control. Both HY and AE groups involved 8 weekly 45-minute group classes with additional 2e4 days/week home practice sessions (yoga and aerobic exercise 4 days a week; strengthening exercise 2 days a week on non-consecutive days). Each participant in the HY and AE groups were given a video camera and a log sheet to record their home practices. The HY intervention program was designed specifically for knee OA by a group of certified yoga teachers and reviewed by three yoga masters. The AE intervention program was based on the recommendation by the Arthritis Foundation. Participants in the control group received OA education brochures and weekly phone calls from research staff. Standardized instruments including WOMAC (Pain, stiffness, function), Visual Analog Scale (pain), Short Physical Performance Battery (SPPB) and 50-foot Walk (strength, balance, gait, and speed), Hospital Depression and Anxiety Scale (anxiety/depression), Self-Transcendence Scale (spiritual health), Falls Efficacy Scale-International (fear of falling), and SF-12 (quality of life) were used to collect outcome data at baseline, 4 weeks, and 8 weeks. Class and home practice adherences were assessed weekly using class attendance records, video-recordings and self-reported log sheets. Statistical analysis methods were intention to treat and included ANCOVA, adjusting for baseline to examine between-groups differences, and Pearson correlation. Results: Participants (n 1⁄4 84) were predominantly female (84%), mean age 71.6 ± 8.0 years, range 60e92. Attrition rate was 18%. At baseline, participants in the HY group were significantly younger (p 1⁄4 .028) but had a higher level of pain (p 1⁄4 .037). After 4 weeks of intervention, participants in HY group had a significant improvement in WOMAC total (p 1⁄4 .001), its subscales (pain: p 1⁄4 .021; physical function: p 1⁄4 .001), and falls (p 1⁄4 .004), these effects sustained in 8 weeks. Anxiety in the HY group was significant improved at 4 weeks (p 1⁄4.052) but not at 8 weeks. For physical performance, a subscale of SPPB was significantly improved in HY group in 8 weeks (8” Walk: p 1⁄4 .006). There were no differences in spiritual health and quality of life among the three groups. Out of the 32 participants in the HY group, 20 (63%) participated 75% of classes, 5 (16%) participated 50%e74% of classes, and 7 (19%) participated < 50% of classes. The average videoed home HY practice was 79 ± 54 (0e278) min and 3 ± 1 (0e6) days per week. Out of the 29 participants in the AE group, 16 (55%) participated 75% of classes, 5 (17%) participated 50%e74% of classes, 8 (28%) participated <50% of classes. The average home AE practice was 56 ± 33 (0e126) minutes and 3 ± 1 (0e5) days per week. Correlation between total log and camera times was r 1⁄4.89. No HY or AE related adverse events were reported. Conclusions: Yoga was better in improving OA symptoms perception, gait, anxiety, and fear of falling compared to the current recommended aerobic/strengthening exercises. The mind-body intervention may have added benefits to older adults with OA. Although majority of participants attended most of classes offered for both HA and AE groups, home practice adherence remains an issue. The average home practice was below the prescribed weekly dose and there was a wide range in minutes practiced. Strategies that promote home yoga practice in older adults with OA should be developed to optimize its therapeutic benefits. Log sheet may be a valid source for collecting adherence data. A main limitation of the study was the convenient and homogenous sample. Using multiple sites with a larger sample size and diverse participants will increase the generalizability of the findings. Including physiological and biomechanical data to determine the mechanism of how yoga improves OA symptoms will be valuable in future research.