ABSTRACT
OBJECTIVES: To examine the longitudinal association between physical performance and risk of dementia in individuals aged 90 and older without dementia. DESIGN: Population-based longitudinal study. SETTINGS: The 90+ Study, Laguna Woods, California. PARTICIPANTS: Men n = 176 and women n = 402 without dementia from The 90+ Study (n = 578, mean age 93.3). At baseline, 54% of participants were cognitively normal, and 46% had cognitive impairment, no dementia. MEASUREMENTS: Physical performance measures (4-m walk, 5 chair stands, handgrip, standing balance) were scored from 0 (unable to perform) to 4 (best performance). The outcome was dementia, diagnosed according to the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, criteria. Hazard ratios (HRs) for dementia in relation to baseline physical performance were estimated using Cox regression after adjustment for potential confounders. HRs and P-values for the overall Wald chi-square are reported to show the magnitude of each physical performance measure and the strength of the association between each measure and incident dementia. RESULTS: Poor physical performance in most measures was associated with greater risk of incident dementia over a mean follow-up of 2.6 years (range 0.6-9.0 years). After controlling for potential confounders, standing balance had the strongest association with incident dementia (HRs = 1.9-2.5, overall P = .02), followed by 4-m walk (HRs = 1.1-1.8, overall P = .04) and handgrip (HRs = 1.0-2.0, overall P = .03). The association with five chair stands was not significant. In a subanalysis limited to cognitively normal participants, HRs were attenuated, but most remained in the same direction. CONCLUSION: Poor physical performance is associated with risk of developing dementia over an average 2.6-year follow-up in the oldest-old, indicating that poor physical performance may be an early sign of late-age dementia.
