Motor signs predict poor outcomes in Alzheimer disease.

OBJECTIVE To examine whether the presence of motor signs has predictive value for important outcomes in Alzheimer disease (AD). METHODS A total of 533 patients with AD at early stages (mean Folstein Mini-Mental State Examination [MMSE] 21/30 at entry) were recruited and followed semiannually for up to 13.1 years (mean 3) in five University-based AD centers in the United States and European Union. Four outcomes, assessed every 6 months, were used in Cox models: cognitive endpoint (Columbia Mini-Mental State Examination < or = 20/57 [ approximately MMSE < or = 10/30]), functional endpoint (Blessed Dementia Rating Scale > or = 10), institutionalization equivalent index, and death. Using a standardized portion of the Unified PD Rating Scale (administered every 6 months for a total of 3,149 visit-assessments, average 5.9 per patient), the presence of motor signs, as well as of individual motor sign domains, was examined as time-dependent predictor. The models controlled for cohort, recruitment center, sex, age, education, a comorbidity index, and baseline cognitive and functional performance. RESULTS A total of 39% of the patients reached the cognitive, 41% the functional, 54% the institutionalization, and 47% the mortality endpoint. Motor signs were noted for 14% of patients at baseline and for 45% at any evaluation. Their presence was associated with increased risk for cognitive decline (RR, 1.72; 95% CI, 1.24 to 2.38), functional decline (1.80 [1.33 to 2.45]), institutionalization (1.68 [1.26 to 2.25]), and death (1.38 [1.05 to 1.82]). Tremor was associated with increased risk for reaching the cognitive and bradykinesia for reaching the functional endpoints. Postural-gait abnormalities carried increased risk for institutionalization and mortality. Faster rates of motor sign accumulation were associated with increased risk for all outcomes. CONCLUSIONS Motor signs predict cognitive and functional decline, institutionalization, and mortality in Alzheimer disease. Different motor sign domains predict different outcomes.