Objective: To identify clinical predictors of cognitive decline in Alzheimer's disease.
Design: A cohort of patients was followed up longitudinally and the likelihood of arriving at two cognitive end points was assessed using the Cox proportional hazards model and eight explanatory variables.
Setting: Subjects were chosen from patients examined for memory loss at two medical centers affiliated with the University of Southern California, Los Angeles.
Patients: The sample included 135 patients who met National Institute for Neurological and Communicative Disorders and Stroke-Alzheimer's Disease and Related Disorders Association criteria for probable or definite Alzheimer's disease, had initial Mini-Mental State Examination (MMSE) scores of 14 or greater, and had been seen on at least two occasions.
Main outcome measures: The time to reach either of two end points, ie, MMSE score of 8 and a decline of six points on the MMSE, was assessed.
Results: After controlling for initial severity of dementia (eg, by dividing the sample into mild and moderate dementia subgroups or by using the individually defined end point of a six-point decline on the MMSE), the presence at baseline of extrapyramidal signs (risk-hazard ratio, 10.34; 95% confidence interval, 2.76 to 38.68; P = .0005), agitation (risk-hazard ratio, 2.98; 95% confidence interval, 1.35 to 6.61; P = .007), and hallucinations (risk-hazard ratio, 3.85; 95% confidence interval, 1.35 to 11; P = .01) predicted a shorter time to reach an end point.
Conclusions: After controlling for initial severity of dementia, the presence of extrapyramidal signs and behavioral symptoms (agitation and hallucinations) significantly predict faster cognitive decline. These findings may reflect the effects of neuroleptic medication, the presence of underlying diffuse Lewy body disease, or alterations in biogenic amine systems.