Rosiglitazone and pioglitazone for the treatment of Alzheimer's disease.

Objective: To review the literature on the efficacy and safety of rosiglitazone and pioglitazone for the treatment of Alzheimer's disease (AD).

Data Sources: Literature was accessed through MEDLINE (1948-August 2011 week 2) and EMBASE (1980-2011 week 32) using the search terms rosiglita-zone, pioglitazone, and Alzheimer's disease. Results were limited to studies conducted in humans and published in English.

Study Selection And Data Extraction: Clinical trials evaluating the efficacy and safety of rosiglitazone or pioglitazone in patients with AD were critically evaluated.

Data Synthesis: The mechanism for development of AD has been linked to both inflammation and decreased insulin sensitivity. Because of this, rosiglitazone and pioglitazone have been evaluated as potential treatments for AD because of their insulin-sensitizing and antiinflammatory effects. Five clinical trials were evaluated (3 assessing rosiglitazone, 2 assessing pioglitazone); 1 trial evaluating rosiglitazone demonstrated a beneficial effect on cognition in patients with probable AD. However, the largest randomized, double-blind, placebo-controlled trials conducted to date failed to demonstrate a difference between rosiglitazone and placebo when assessing primary endpoints. Two small trials evaluating pioglitazone produced conflicting results regarding efficacy in AD; numerous limitations make results difficult to interpret. The safety of these agents was also evaluated in these trials; edema was seen more commonly in patients receiving rosiglitazone or pioglitazone than in those receiving placebo; however, each drug was generally well tolerated.

Conclusions: Results from clinical trials and current safety data suggest that rosiglitazone should not be used for the treatment of AD. Application of results from trials evaluating pioglitazone in the treatment of AD is limited because of major trial limitations; therefore, it should not be recommended at this time. Although these drugs are not commonly used in the treatment of AD, further pharmacoepidemiologic studies are warranted before their use can be recommended.