This study examines the role of aluminium in the etiology of Alzheimer's disease (AD). Brains taken at autopsy (n = 50) from patients with a history of long-term hemodialysis (HD) and intake of aluminium (Al)-containing drugs were examined by light microscopy. Using our modified silver stain we have been able to demonstrate and clearly discriminate between AD changes and dialysis-associated encephalopathy (DAE) on paraffin sections; evaluation was done with a 3-point scale. DAE morphology is characterized by lysosome-derived intracytoplasmic, Al-containing, pathognomonic, argyrophilic inclusions in choroid plexus epithelia, cortical glia and neurons. A statistically significant difference was found between the amounts of drug-related Al ingested and the degree of DAE-related morphological change (P < 0.001). On the other hand no apparent microscopical increase in AD morphology was found. No AD changes were seen whatsoever in patients under the age of 60, despite a history of long-term HD with ingestion of "pure" Al up to 2.5 kg. Patients over 60 years of age occasionally presented with sparse deposits of beta A4 amyloid (beta A4) and/or a low incidence of AD-type neurofibrillary tangles (NFT). In accordance with CERAD criteria these were identified as normal, age-related phenomena (P < 0.001 for beta A4; P < 0.001 for NFT). Rare, isolated cases from a group of 127 long-term hemodialyzed patients have been reported previously, who presented with intermingled, clearly distinguishable lesions of both age-related AD morphology and DAE changes. Comparison of AD morphology with an age-matched control group was not statistically significant (P > 0.6 for beta A4, P > 0.7 for NFT). In our experience, Al does not cause an increase in AD morphology, at least not in terms of bioavailable Al in drugs or as a result of long-term HD.