Introduction: Two phase 3 studies evaluated the efficacy and safety of rosiglitazone (RSG), a type 2 diabetes treatment, in an extended release (RSG XR) form as adjunctive therapy to ongoing acetylcholine esterase inhibitor (AChEI) treatment in AD (REFLECT-2, adjunctive to donepezil; REFLECT-3, to any AChEI). An open-label extension study (REFLECT-4) assessed RSG XR long-term safety. Methods: In these two double-blind, placebo-controlled studies, subjects with mild-to-moderate probable AD were randomized within 2 apolipoprotein E (APOE) allelic strata (APOE 4- positive, APOE 4-negative) to once daily placebo, 2 mg RSG XR, or 8 mg RSG XR for 48 weeks (REFLECT-2, N=1,496; REFLECT-3, N=1,485). Co-primary efficacy endpoints were change from baseline in Alzheimers Disease Assessment Scale-Cognitive subscale (ADAS-Cog) and Clinical Dementia Rating scale - Sum of Boxes (CDR-SB) scores at week 48. Three populations were analyzed: APOE4-negative, all subjects except APOE 4 homozygotes, and the full intent- to-treat population. Results: No statistically or clinically relevant differences between treatment groups were observed on the a priori primary endpoints in REFLECT-2 or REFLECT-3. Edema was the most frequent adverse event with RSG in each study (14% and 19%, respectively, at 8 mg RSG XR). Conclusions: No evidence of statistically or clinically significant efficacy in cognition or global function was detected for 2 mg or 8 mg RSG XR as adjunctive therapy to ongoing AChEIs. There was no evidence of an interaction between treatment and APOE status. Safety and tolerability of RSG XR was consistent with the known profile of rosiglitazone.