Polypharmacy, Gait Performance, and Falls in Community-Dwelling Older Adults. Results from the Gait and Brain Study

BACKGROUND AND OBJECTIVES Polypharmacy, defined as the use of five or more medications, has been repeatedly linked to fall incidence, and recently it was cross-sectionally associated with gait disturbances. Our objectives were to evaluate cross-sectional and longitudinal associations between polypharmacy and gait performance in a well-established clinic-based cohort study. We also assessed whether gait impairments could mediate associations between number of medications and fall incidence. DESIGN Prospective cohort of community-dwelling older adults, with 5 years of follow-up. SETTING Geriatric clinics in an academic hospital in London, ON, Canada. PARTICIPANTS Community-dwelling older adults aged 65 and older (n = 249; 76.6 +/- 8.6 y; 63% women). MEASUREMENTS Number of medications, quantitative spatiotemporal gait parameters, and fall incidence during follow-up. RESULTS The number of medications was cross-sectionally associated with poor gait performance (slow gait, speed p < .001; higher variability, p < .001; and higher stride, p < .001; step, p = .013, and double support times, p < .001). Prospectively, the number of medications was associated with overall gait decline (odds ratio = 1.23; 95% confidence interval [CI] = 1.13-1.33; p < .001), faster gait decline (hazard ratio = 4.62; 95%CI = 1.82-11.73; p < .001), and higher falls incidence (p = .006). These associations remained true after adjusting for age, sex, and accounting for "confounding by indication bias" by using a comorbidity propensity score adjustment. Each additional medication taken, significantly increased gait decline risk by 12% to 16% and fall incidence risk by 5% to 7%. Mediation analyses revealed that gait impairments in stride length, step length, and step width mediated the strength of the association between medications and fall incidence. CONCLUSION Polypharmacy was cross-sectionally associated with poor gait performance and longitudinally associated with gait decline and fall incidence. Despite our use of propensity matching, confounding by indication could have influenced the results. Quantitative spatial gait parameters performance mediated the strength of the association between medications and falls, suggesting a role of gait disturbances in the medication-related falls pathway.