Association between aluminum in drinking water and incident Alzheimer’s disease in the Canadian Study of Health and Aging cohort

•Epidemiological evidence linking aluminum in drinking water and Alzheimer’s disease is inconsistent.•This study represents the largest prospective evaluation of aluminum in drinking water and Alzheimer’s disease conducted to date.•Exposure to aluminum in drinking water was not clearly associated with increased risk of Alzheimer’s disease.•The positive linear trend observed in the sub-cohort genotyped for apolipoprotein E warrants further investigation.•Further studies would benefit from consideration of multiple aluminum chemical species in risk analysis. Epidemiological evidence linking aluminum in drinking water and Alzheimer’s disease (AD) has been inconsistent, with previous studies often limited by small sample sizes. The present study addresses this issue using data from the Canadian Study of Health and Aging (CSHA), a prospective cohort of 10,263 subjects followed-up from 1991-1992 through 2001–2002. Participants’ residential histories were linked to municipal drinking water sources in 35 Canadian municipalities to obtain ecologic pH, aluminum, fluoride, iron and silica concentrations in drinking water. Cox proportional hazards models were used to examine associations between aluminum and incident AD [Hazard Ratios (HRs), 95% confidence intervals (CIs)], adjusting for age, gender, history of stroke, education, and high blood pressure. A total of 240 incident AD cases were identified during follow-up of 3, 638 subjects derived from the CSHA cohort with complete data on all covariates. With categorical aluminum measurements, there was an increasing, but not statistically significant, exposure-response relationship (HR = 1.34, 95% CI 0.88−2.04, in the highest aluminum exposure category; p = 0.13 for linear trend). Similar results were observed using continuous aluminum measurements (HR=1.21, 95% CI 0.97−1.52, at the interquartile range of 333.8 μg/L; p = 0.09 for linear trend). In a subsample genotyped for ApoE-ε4, there was some evidence of an association between aluminum and AD (p = 0.03 for linear trend). Although a clear association between aluminum in drinking water and AD was not found, the linear trend observed in ApoE-ε4 subsample warrants further examination.