Pharmacological treatment of neuropsychiatric symptoms of dementia: a review of the evidence.

Study Design And Objective: Systemic review of double-blind, placebo-controlled, randomized controlled trials (RCTs) and meta-analyses of medication efficacy in the treatment of patients diagnosed with dementia and experiencing neuropsychiatric symptoms(hallucinations, delusions, agitation, aggression, combativeness, wandering).

Exclusion Criteria: Studies were excluded if they reported only depression, if the medication was not available no longer used in the United States, or duplicated another study already included.

Data Sources: Medline of English articles between 1966 and June 2004, Cochrane Database of Systematic Reviews, and a manual search by the authors for other relevant articles.

Outcomes: Diverse outcome measures ranging from global benefit to behavioral rating scales. Some of the 29 reports listed several instruments; in total, 24 rating scales were used. Statistical outcome was described and some, but not all, noted clinical impression. Adverse outcomes were listed.

Results: The results were clustered in groups: conventional antipsychotics, atypicals, antidepressants, cholinesterase inhibitors, mood stabilizers, and others. Treatment duration ranged from 17 days to 16 weeks. Types of dementia and levels of severity varied. The authors reported little benefit and some evidence for harm for typical (or conventional) agents. In contrast, some RCTs of atypical antipsychotics reported "modest" benefit, with olanzapine and risperidone leading others. Although trials reported minimal side effects at low doses, authors acknowledged an increased risk for stroke. No studies adequately compared benefit of typical with atypical agents. With the possible exception of citalopram, antidepressant agents did not reduce agitation, but did improve depression. Cholinesterase inhibiting agents demonstrated significant efficacy toward behavior, while memantine had mixed results. Valproate did not prove to be efficacious, and results for carbamazepine were conflicting.