Memantine and Acetylcholinesterase Inhibitor Use in Alzheimer’s Disease Clinical Trials: Potential for Confounding by Indication

Background: Acetylcholinesterase inhibitors (AChEIs) and memantine are commonly prescribed medications for Alzheimer's disease (AD). Their concurrent use in AD randomized clinical trials (RCTs) is generally allowed but their effect in outcome measures is unsettled. Objective: To evaluate whether us e of AChEIs and/or memantine across AD RCTs are associated with different rates of cognitive/functional decline. Methods: We pooled data from 5 RCTs of mild to moderate AD conducted by the Alzheimer's Disease Cooperative Study (ADCS) between 2002-2013. 1,423 participants with MMSE of 14–26 and completion of 12–18 months follow-up visits were analyzed. Trials did not randomize with respect to AChEIs or memantine. We defined 4 groups: AChEI (27%), memantine (16%), AChEIs+memantine (46%), and non-users (11%). Outcome measures were change in ADAS-cog-11, ADCS-ADL, and MMSE from baseline to 18 months. Fisher's exact test, Wilcoxon signed rank, and Spearman's tests were used to identify confounding variables. Mixed model repeated measures were used for adjustments and pairwise tests for comparing change in scores. Results: Age, apolipoprotein E, and initial MMSE were identified as covariates. Memantine and/or AChEIs users had greater impairment at entry than non-users. There was a significant decline on the ADAS-cog-11 in the memantine (estimate –4.2 p < 0.0001) and AChEIs+memantine groups (estimate –3.5 p < 0.0001) than non-users, while there was significantly more decline in MMSE (estimate 2.5 p < 0.0001) and ADCS-ADL in the AChEIs+memantine group (estimate 4.3 p < 0.0001) Conclusion: Memantine monotherapy or combined with AChEIs are associated with more rapid cognitive and functional decline than non-users. We postulated a potential selection bias by indication.