Treatment success in Alzheimer disease (AD) trials is generally based on benefits over placebo-treated controls. Consequently, variation in rates of decline among placebo-treated patients could impact outcomes from AD trials. In the present analyses, individual patient data [baseline Mini-Mental State Examination (MMSE): 10 to 26] were pooled from randomized, placebo-controlled studies of donepezil for AD conducted during the 1990s, and grouped by initiation year-group 1: 1990 to 1994; group 2: 1996 to 1999. Changes in MMSE and Alzheimer's Disease Assessment Scale-cognitive subscale (ADAS-cog) were compared between groups 1 and 2 for placebo, and then between donepezil and placebo. Data were available from 3403 patients in 13 trials. Group 2 (post-1995) included patients with lower baseline MMSE scores, older patients, fewer males, more comorbidity, and more concomitant medications. MMSE decline by week 24 was significantly greater among group 1 (pre-1995) placebo patients versus group 2; a similar trend was observed with the ADAS-cog. Nevertheless, donepezil-mediated treatment effects were consistent over the decade of enrollment. These analyses suggest that patients are showing slower rates of cognitive decline in more recent trials compared with older trials, although having more comorbidities. This finding may have important potential implications for future clinical trial design.