A Retrospective Study of Cholinesterase Inhibitors for Alzheimer's Disease: The Effect of Cerebrovascular Disease on Patient Outcomes and the Impact of Biases on the Results

Introduction: Dementia may be caused by Alzheimer's disease (AD), cerebrovascular disease (CVD), or a combination of both. When CVD is associated with dementia, survival is thought to be reduced. It is unclear whether treatment with cholinesterase inhibitors (ChEIs), which has been found to improve cognitive symptoms and global function in AD patients, has similar benefits in vascular forms of dementia.;Methods: A retrospective cohort study was undertaken using the Regie de l'Assurance Maladie du Quebec (RAMQ) databases to examine the time to NHP or death for AD patients aged 66+, with or without CVD, treated with ChEIs between July 1, 2000, and June 30, 2003. Because ChEIs are approved only for AD in Canada, a ChEI prescription was used as a surrogate for an AD diagnosis. Concomitant CVD was identified on the basis of a lifetime diagnosis of stroke or endarterectomy, or a diagnosis of transient ischemic attack within the six months prior to the index date. Separate analyses were performed for patients with persistent ChEI use and those who discontinued ChEI therapy. Seven Cox proportional hazard regression models which varied in the definition of the index date (start of follow-up) and the duration of follow-up were used to evaluate the impact of immortal time bias.;Results: 4,428 patients met inclusion criteria for AD with CVD; 13,512 were classified as having AD alone. For the composite endpoint of NHP or death, 1,000-day survival rates were lower among AD patients with versus without CVD (p<0.01), but absolute differences were very small (84% vs. 86% with continuous ChEI use; 77% vs. 78% with discontinuous ChEI therapy). Of the secondary endpoints, time to death was shorter for patients with versus without CVD, but time to NHP did not differ between groups. In the primary, unbiased analysis, no association was found between ChEI treatment type and death or NHP. However, after introduction of immortal time bias, a strong differential effect was observed.;Limitations: Results may have been affected by selection (misclassification) bias, between-group differences in smoking and body mass index (information on which was not available in the RAMQ databases), and duration of ChEI therapy.;Objectives: The present study was designed to determine whether co-existing CVD is associated with survival or time to nursing home placement (NHP) among AD patients treated with ChEIs. Findings of poorer outcomes in patients with versus without CVD might argue against the use of ChEIs for AD patients in whom CVD co-exists. The objective of a second analysis was to assess for the first time in patients with AD the potential impact of immortal time (and follow-up) bias on risk for these outcomes.;Conclusions: Associations between co-existing CVD and time to NHP or death appeared to be of little clinical relevance among AD patients treated with ChEIs. The lack of difference between AD patients with and without CVD suggests that CVD should not be used as a reason to deny AD patients access to ChEI treatment. Properly accounting for unexposed person-time in the analysis eliminates biased estimates of drug efficacy.;Keywords: cholinesterase inhibitors; dementia; Alzheimer's disease; cerebrovascular disease; administrative database...