Background: Neuroleptics are commonly used to treat behavioral disorders associated with dementia. However, their safety and efficacy have not been well established in these patients.
Method: A meta-analysis of randomized, controlled (either placebo or active drug), double-blind trials published since 1966 (N = 16; 499 treated, 112 active controls, and 123 placebo) was conducted. Data were collected on proportion of patients with clinically significant improvement, significant side effects, and dropout rates.
Results: Pooled mean percentages of patients who improved (95% CI): all neuroleptics, 64% (54% to 74%); low potency, 63% (54% to 72%); moderate potency, 70% (56% to 85%); moderate-high potency, 62% (49% to 75%); and high potency, 69% (49% to 90%). Thus, no differences in efficacy existed between different potencies of neuroleptics. Therapeutic effect (neuroleptic minus placebo) was only 26% (14% to 38%). Treatment-emergent side effects were more common for neuroleptics vs. placebo (mean difference = 25%, 13% to 37%), but pooled mean dropout rates were not different (mean difference = 4%, -7% to 14%). Neither weighting by clinical trial quality (3 raters; weighted agreement, 83% to 92%) nor exclusion of poor quality trials changed the results.
Conclusion: Neuroleptics have small but significant efficacy over placebo in this population, and the efficacy rate is equivalent to the side effect rate. Comparing different neuroleptics shows they have similar efficacy, side effects, and dropout rates. Further study to determine more specific drug-responsive behaviors is needed to maximize benefits of these drugs.