Cummings_2008_Alzheimers.Dement_4_49

BACKGROUND: Behavioral abnormalities and neuropsychiatric symptoms are common manifestations of Alzheimer's disease (AD). Many clinical trials of cholinesterase inhibitors (ChE-Is) and memantine have included behavioral measures as primary or secondary outcomes, and most have observed behavioral benefits in conjunction with treatment. The purpose of this review was to determine the frequency of positive behavioral outcomes in AD clinical trials and clinical reports, to determine the symptoms most responsive to antidementia agents, and to explore factors that correlate with negative outcomes in clinical trials of antidementia agents with regard to behavioral measures.
METHODS: We performed a computerized search of randomized clinical trials and open-label studies of ChE-Is and memantine for AD including a behavioral outcome measure. Studies involving 10 or more patients using therapeutic doses of the target agents and including a behavioral measure as a primary or secondary outcome were included in this review.
RESULTS: One hundred fifty-seven peer-reviewed articles and 68 publicly presented abstracts were identified in the literature search. Subsequent review established that 15 publications arising from 13 randomized, double-blind, placebo-controlled AD trials met the review inclusion criteria. Positive outcomes on behavioral measures were reported in 8 of 15 publications as a primary or secondary outcome. In most blinded AD clinical trials, behavioral measures were secondary outcomes. Mood symptoms and apathy have most commonly responded to ChE-Is, whereas memantine has been associated with a reduction in irritability and agitation. However, there is substantial variability among trials in terms of behavioral outcomes. Studies that assessed patients with more severe dementia, included patients with less severe behavioral disturbances at baseline, involved institutionalized patients, or were international in scope tended to have negative outcomes. In institutionalized patients there is commonly an improvement in the placebo group that confounds the observation of any drug-placebo difference.
CONCLUSIONS: Antidementia agents have been associated with beneficial behavioral outcomes in many randomized clinical trials and open-label studies. Most studies are not designed to test the psychotropic properties of antidementia drugs. Trials with negative behavioral outcomes are most likely to involve patients who are institutionalized and have few behavioral disturbances at baseline. Clinical trials designed to assess behavioral effects of antidementia agents should anticipate these factors.