RT Journal Article SR Electronic T1 Further thoughts on the aluminum-Alzheimer's disease link. JF Journal of Epidemiology and Community Health JO J Epidemiol Community Health FD BMJ Publishing Group Ltd SP 401 OP 403 DO 10.1136/jech.50.4.401 VO 50 IS 4 A1 Forbes, W F A1 McLachlan, D R YR 1996 UL http://jech.bmj.com/content/50/4/401.abstract AB STUDY OBJECTIVE AND METHOD: The results of studies on aluminum (Al) and Alzheimer's disease (AD) from groups in Newcastle, UK and Ontario, Canada were compared in order to explain why the former were unable to detect a link while the latter could, and to suggest alternative ways of examining the data. RESULTS: The Al concentrations in the Newcastle study were relatively small compared with the Ontario ones. When Al concentrations > 250 micrograms/l were used, the RRs were greater for AD than for other forms of dementia, and the RRs were lower for those under 75 years and greater at ages 85 years and over than at ages above 75 years. The relationship between dementia and Al concentrations was U or J shaped--there was a minimum at an Al concentration of 100 micrograms/l. Other constituents or properties such as silicic acid, fluoride, turbidity, iron, and pH all have an effect on the relationship. CONCLUSIONS: Analyses of the type reported from Newcastle can yield further information if they are extended to include multivariate analyses that take account of other water constituents which can affect the relationship between Al water concentrations and AD are carried out. The relationship between Al and dementia may be U or J shaped rather than linear. With regard to AD, the group aged less than 65 years is not the best one in which to explore a relationship. Lastly, it may be that a link with AD is most meaningful at relatively high Al water concentrations.