PREDICTORS OF PHYSICAL ACTIVITY IN TWO COHORTS WITH KNEE OSTEOARTHRITIS. Dina L. Jones*1; James J. Irrgang2; Andrea M. Kriska3; Stanley Wearden4; Lawrence S. Crossett5; C. K. Kwoh6 1. Department of Orthopaedics and Physical Therapy, West Virginia University, Morgantown, WV; 2. Department of Physical Therapy, University of Pittsburgh, Pittsburgh, PA; 3. Department of Epidemiology, University of Pittsburgh, Pittsburgh, PA; 4. Department of Orthopaedics, West Virginia University, Morgantown, WV; 5. Department of Orthopaedics, University of Pittsburgh, Pittsburgh, PA; 6. Division of Rheumatology and Clinical Immunology, University of Pittsburgh, Pittsburgh, PA PURPOSE: The purpose of this study was to determine which sociodemographic and clinical factors significantly predict physical activity in two cohorts with knee osteoarthritis (OA), one undergoing primary total knee arthroplasty (TKA cohort) and one not undergoing TKA (non-TKA cohort). BACKGROUNDS/SIGNIFICANCE: The prevalence of inactivity is high in individuals with knee OA. Predictors of physical activity have been identified in the general population and in individuals with arthritis, but not specifically in those with knee OA or undergoing primary TKA. SUBJECTS: The TKA cohort contained 74 patients, aged 45 years or older, who were scheduled for elective primary TKA due to OA. The non-TKA cohort contained 66 patients, aged 45 years or older, with knee OA and no prior or planned TKA. METHODS AND MATERIALS: Sociodemographic factors, clinical factors, and physical activity were assessed by structured interviews and self-administered questionnaires. Sociodemographic factors included gender, age, race/ethnicity, marital status, income, and education. Clinical factors included the number of knees with OA, knee OA severity (Index of Severity for Knee Disease), OA severity in other joints (Rapid Assessment of Disease Activity in Rheumatology), knee pain (Numerical Pain Rating Scale, Visual Analogue Scale for pain, SF-36 Bodily Pain scale, and WOMAC pain scale), stiffness (WOMAC stiffness scale), history of prior knee injury or fracture, history of prior knee surgery, comorbidities (Charlson Index), body mass index (BMI), and mental (SF-36 Mental Composite Score) and physical functioning (WOMAC physical function scale and SF-36 Physical Composite Score). Leisure activity, occupational activity, and instrumental activities of daily living (IADLs) during the past month were assessed using standardized physical activity questionnaires. ANALYSES: Correlational analyses were conducted to determine if the sociodemographic and clinical factors are significantly associated with total physical activity (leisure and occupational activity combined) and IADLs in each cohort. Multiple linear regression was performed to identify predictors of total physical activity and IADLs in each cohort. RESULTS: There were 23 males and 51 females in the TKA cohort with a mean ± SD age of 65.6 ± 9.9 years. The non-TKA cohort contained 14 males and 52 females with a mean age of 68.4 ± 9.8 years. The TKA cohort reported a median of 21.8 (interquartile range, 7.7-45.1) MET-hours of total physical activity and 37.2 (interquartile range, 19.5-57.0) MET-hours of IADLs per week. Total physical activity was associated with gender (r = -0.27, p = 0.02), knee OA severity (r = -0.31, p < 0.01), and comorbidities (r = -0.25, p = 0.03) in the TKA cohort. The best set of predictors of total physical activity in the TKA cohort were knee OA severity (estimate = -2.87, p = 0.01) and comorbidities (estimate = -5.35, p = 0.05) (R2 = 0.15). The non-TKA cohort reported a median of 38.7 (interquartile range, 25.9-68.4) MET-hours of total physical activity and 52.5 (interquartile range, 32.6-88.7) MET-hours of IADLs per week. Total physical activity in the non-TKA cohort was associated with gender (r = -0.27, p = 0.03), knee OA severity (r = -0.27, p = 0.03), and BMI (r = -0.33, p < 0.01). The best set of predictors of total physical activity in the non-TKA cohort were gender (estimate = 13.82, p = 0.02), BMI (estimate = -3.17, p < 0.01), and BMI squared (estimate = 0.13, p = 0.02) (R2 = 0.25). The relationship between total physical activity and BMI was curvilinear in the non-TKA cohort. In both cohorts, IADLs were only associated with BMI (TKA cohort: r = -0.30, p = 0.02; non-TKA cohort: r = -0.32, p < 0.01). CONCLUSIONS: Individuals undergoing primary TKA who reported more severe knee OA and comorbidities had lower levels of total physical activity. In individuals not undergoing TKA, total physical activity was lower in females than in males, and decreased with increasing BMI. Due to the low levels of total physical activity in individuals with high BMI, the rate of decrease in total physical activity was lower in those with higher BMI than lower BMI. IADLs were lower in individuals in both cohorts with higher BMI. Physical therapists can use this information to identify patients with knee OA or those undergoing primary TKA who are more likely to be less active. Physical therapists, in their unique roles as patient educators and physical activity specialists, have the opportunity to promote participation in regular physical activity in these two groups of patients. A better understanding of the predictors of physical activity could lead to the development of tailored interventions aimed at increasing activity levels in particular subgroups of people with arthritis. FUNDING SOURCE: American College of Rheumatology Research and Education Foundation. KEYWORDS: Knee, Osteoarthritis, Physical activity, Total knee arthroplasty Copyright 2009 by the American Physical Therapy Association. Requests for reprints should be directed to the corresponding author of the article. Educators, students, and other academic customers may receive permission to reprint copyrighted material from Physical Therapy (ISSN 1538-6724) by contacting the Copyright Clearance Center Inc, 222 Rosewood Dr, Danvers, MA 01923. Other types of customers who want permission to reprint should contact the APTA Editorial Office, Attn: Physical Therapy.