CURRENT REVIEW * ACTUALITES

Would decreased aluminum ingestionreduce the incidence of Alzheimer's disease?

Donald R. Crapper McLachlan,*t MD, FRCPC; Theo P. Kruck,t PhD; Walter J. Lukiw,t MSc;Sivarama S. Krishnan,4 PhD

Although the cause of Alzheimer's disease (AD) remains unknown there is mountingevidence that implicates aluminum as a toxic environmental factor of considerableimportance. Four independent lines of evidence - laboratory studies of the effects ofintracerebral aluminum on the cognitive and memory performance of animals,biochemical studies, epidemiologic studies and the slowing of the progress of the diseasewith the use of an agent that removes aluminum from the body - now support theconcept that aluminum is one of the pathogenic factors in AD. The evidence warrantsserious consideration of reducing human exposure to aluminum. We hypothesize that apublic health effort to restrict human ingestion of aluminum would reduce the incidenceof this common chronic illness in the elderly.

Meme si l'on ne connait toujours pas la cause de la maladie d'Alzheimer (MA), de plusen plus d'indices semblent demontrer que l'aluminium est un facteur environnementaltoxique tres important. Quatre niveaux de preuve independants - etudes de laboratoireau sujet des effets de l'aluminium intracerebral sur la cognition et la memoire chez lesanimaux, etudes biochimiques, etudes epidemiologiques et ralentissement du progres dela maladie lorsqu'on utilise un agent qui elimine l'aluminium du corps - appuientmaintenant le concept selon lequel l'aluminium est un des facteurs pathogenes de laMA. Ces preuves justifient d'envisager serieusement de reduire l'exposition des etreshumains a l'aluminium. Nous posons l'hypothese suivante: une campagne d'hygienepublique visant a limiter l'ingestion d'aluminium par les etres humains reduiraitl'incidence de cette maladie chronique repandue chez les personnes agees.

T he cause of Alzheimer's disease (AD) remainsunknown, but both genetic and environmen-tal factors appear to be important. Mutation

on chromosome 211,2 may account for a smallnumber of families with familial AD; recently, twofamilies of patients with AD were found to have apoint mutation in the amyloid precursor protein, agene on chromosome 21.3 This mutation causes anamino acid substitution of isoleucine for valine inthe transmembrane domain of the ,B-amyloid precur-sor protein at a site just two amino acids from the

3-amyloid peptide. It is still uncertain whether thispoint mutation "causes" AD or whether the muta-tion is closely linked to the yet-to-be-discovered generesponsible for AD on chromosome 21. For mostfamilies, however, linkage to chromosome 21 has notbeen established, and the genetic form of the diseaseappears to involve mutations on more than onechromosome.

The much more common, sporadic form of ADappears to be related to unidentified environmentalfactors. Even in genetically identical twins nongenet-

From the departments of*Medicine and tPhysiology, University of Toronto, and $the Department ofMedical Physics, Toronto Hospital,Toronto, Ont.

Reprint requests to: Dr. Donald R. Crapper McLachlan, Director, Centrefor Research in Neurodegenerative Diseases, University ofToronto, 6 Queen's Park Cres. W, Toronto, ON M5S IA8

CAN MED ASSOC J 1991; 145 (7) 793- For prescribing inforr.hation see page 884

ic factors appear to be important; in one study thedisease developed in only 40% of twin pairs.4 Al-though a virus or scrapie-like infectious agent hasbeen proposed as a pathogen AD is not transmissibleto other humans or laboratory animals, and neitherviral DNA nor viral antigens have been reproduciblyfound in AD-affected tissues. Immunologic mecha-nisms have also been suggested as factors, butwhether the observed changes are cause or effect isuncertain.5

Aluminum, a widely recognized neurotoxin, hasbeen found in increased concentrations in all AD-affected tissues examined by means of methodssufficiently sensitive and appropriately applied.6

However satisfying the idea, we may not bejustified in searching for a single cause. A directexperimental approach to measuring the importanceof aluminum, or any other postulated etiologic orpathogenic factor, is not possible because AD isfound only in humans and cannot be replicatedexactly in laboratory preparations. Aluminum in-duces neither the paired helical filament configura-tion found in Alzheimer-type neurofibrillary tanglesnor the formation of senile plaques with amyloidcores. Nevertheless, it has been found in at least foursites in AD-affected brains and induces at least threecellular derangements also found in AD. However,no single observation or experimental result is, initself, conclusive.

Four independent lines of evidence implicatealuminum's role in AD: (a) toxicologic studies andlaboratory observations of the learning and memoryperformance of animals, (b) a large number ofdocumented biochemical changes at concentrationsof aluminum similar to those found in varioussubcellular compartments in AD-affected humanbrains, (c) epidemiologic evidence of the increasedincidence ofAD in relation to exposure to aluminumin drinking water and (d) the slowing of the clinicalprogression of AD by a drug that selectively removesaluminum from the body.

Effect of aluminum on learningand memory performance of animals

The direct intracranial injection of small butlethal amounts of soluble aluminum salts producesdelayed memory and learning impairment in sensi-tive species such as cats and rabbits.7-9 A dose ofintracranially injected aluminum sufficient to raisethe content in grey matter about fourfold (from 1.5,gg/g dry weight to the 50% lethal concentration ofabout 5.5 ug/g) had no immediate effect on memoryand motor tasks in trained cats; however, 7 to 10days after injection the cats began to exhibit progres-sive impairment in precise motor control duringjumping'0 and alterations in the performance of

learning and memory tasks.7 Initially the learningand memory deficits were selective: the performanceof visual discrimination tasks was not affected andthe speed of motor response not altered. The selec-tive effect of aluminum on the learning and memorysystem is element specific, because nine other triva-lent toxic metals (boron, chromium, gallium, indi-um, lanthanum, scandium, thallium, vanadium andyttrium) have been shown not to elicit the sequenceof clinical signs that follow aluminum administra-tion.'

Ten days after aluminum injection rabbits haveexhibited deficits in learning (acquisition of anactive avoidance task) and failed to retain the taskwhen tested 3 days later.'2 Impaired learning in awater maze,'3 defective classic'4 and eye-blink condi-tioning'5'6 and impaired long-term potentiation inhippocampus slices'7 (considered to be an electro-physiologic model of a learning response) have alsobeen reported in rabbits given aluminum.

Shortly after the memory defects appear, catsand rabbits exhibit progressive deterioration inmotor control, with difficulty executing body right-ing, increased muscle tone, ataxia of gait, tremors,myoclonic jerks and seizures. If the seizures are nottreated with anticonvulsant medications the animalsmay die in convulsions; treated animals may sur-vive, but with persistent severe neurologic and beha-vioural defects.

The unique, progressive clinical course after asingle intracerebral lethal threshold dose of alumi-num in susceptible mammals is marked first bydisturbance of learning and memory and then byaltered motor control, increased muscle tone, myo-clonic jerks and seizures. It resembles the clinicalcourse of AD in humans, although much shorter induration.

The intracerebral aluminum levels (5 to 6 ,ug/gdry weight'8) that produce the toxic effects in ani-mals are sometimes found in the brains of ADpatients. Approximately 23% of randomly selectedsamples of neocortical grey matter from patientswith AD have been found to contain aluminum inthe amount of 5 ,ug/g dry weight or more.'9

Aluminum-induced neurochemical changesin the brain

Aluminum affects many biochemical and neuro-chemical processes of the brain (Table 1 ). 12,17,20-105 Ofparticular importance is the observation that humanneuroblastoma cells in culture treated with low dosesof aluminum produce antigens that react with anantibody to an abnormally phosphorylated mi-crotubule-associated protein known as Tau found inAD.56 The antibody reacts specifically with ADneurofibrillary tangles. Considerable evidence now

794 CAN MED ASSOC J 1991; 145 (7) LE lerOCTOBRE 1991

I

Table 1: Actions of aluminum at the cellular level

Nuclear effectsBinds to DNA phosphate and

bases20.2'Increases histone-DNA binding22Decreases RNA inneuroblastoma23

Blocks initiation sites for RNApolymerase24

Blocks RNA polymerase activityin vitro25

Blocks ribosylation of adenosinediphosphate26

Alters sister chromatidexchange27

Alters steroid-inducedchromosome puffing28

Inhibits corticosterone binding toDNA29

Alters poly(A) RNA content in theforebrain of rabbits30

Reduces messenger RNA codingfor neurofilaments in anteriorhorn cells of rabbits31

Decreases cell division and altersDNA syntheSiS3233

Induces considerable changes inchromatin34

Alters the development of ratoffspring35

Cytoplasmic effectsInduces conformational changes

in calmodulin; blockscalmodulin-dependentcalcium-magnesium adenosinetriphosphatase (ATPase),which is important in extrusionof calcium ion from celIsW,37

Increases intracellular calciumcontent'7.38

Reduces sugar phosphorylation39Decreases respiration40Inhibits hexokinase activity41-48Stabilizes terminal phosphorylgroup on ATP4' 46

Forms long-lived complex withATP41-"

Inhibits ATP4184Inhibits brain glycolysis,depression of yeast and ratbrain cytosolic andmitochondrial hexokinaseactivity4''46

Stimulates brain pyruvatekinase46

Enhances adenylate cyclasestimulation by fluoride butinhibits activation by serotoninand guanine nucleotides inFasciola hepatica, arequirement for activation ofthe regulatory component ofadenylate cyclase byfluoride47,48

Increases the number oflysosomes; reduces thiaminepyrophosphatase andnucleotide diphosphatase inthe Golgi apparatus49

Inhibits the synthesis oftetrahydrobiopterin50

Elevates cyclic AMP and GMP*levels5'Increases ubiquitin response in

neurites of cultured nervoustissue52

Binds to ferritin and is partiallysequestered by thismechanism; may alter ironstorage53

Proliferates peroxisomes(aluminum clofibrate)54

Inactivates phosphofructokinaseand intibits hepatic glycolysis55

Induces an alteration in Tau thatis recognized by antibodies toAlz 5058

Perturbs elongation factor Tu57Inhibits protein kinase-C

activation68Inhibits proton translocatingATPases in Streptococcus andLactobacillueg

Cytoskeletal effectsInduces neurofibrillarydegeneration composed of1 0-nm fibres identical inchemical composition tonormal neurofilaments6"

Alters slow axonal transport,although disputed; has noeffect on anterogradetransport6-67

Alters phosphorylation ofcytoskeletal proteins,microtubule-associated protein2 and the 200KD component ofneurofilaments68

May have secondary effect oncyclic AMP-dependent proteinkinaseM

Promotes assembly ofmicrotubules, which are moreslowly depolymerized thanmagnesium-assembledtubules69870

Induces chronic myelopathy inrabbits7'

Induces neuronal cytoskeletallesions (through intravenousand intrathecal injections)72

Alters neurofilamentconformation in vitro73

Inhibits calpain-mediatedproteolysis; induces humanneurofilament proteins to formhigh-molecular-weightcomplexes74

Effects on membranes andmembrane-bound enzymesAlters physical properties ofmembrane lipids75

Binds to positive-charged andnegative-charged sites inmembranes in vitro76

Alters membrane structure7778Alters adenylate cyclase activity

required for activation ofregulatory component ofadenylate cyclase in vitro byfluoride47,58

Inhibits sodium-potassium-activated ATPactivity at relatively highconcentrations79

Enhances brain-specific lipidperoxidationW891

Accelerates peroxidation ofmembrane lipids stimulated byiron saltsm8'-1

Decreases activity of superoxidedismutase in rat brain'81

Alters blood-brain barrier82-5Inhibits saturatable transportsystem for N-tyrosinatedpeptides and encephalin frombrain82-85

Increases permeability ofblood-brain barrier toneuropeptideS82-85

Synaptic and neurotransmittereffectsAlters dendritic shape and

synaptic density in chronicpreparations'2,86

Blocks high-affinity uptake ofy-aminobutyric acid (GABA)and glutamate fromsynaptosomes87t8.

Blocks synaptosome uptake ofneurotransmitter aminescholine, dopamine andnoradrenaline79.89

Inhibits acetylcholinesterase23,90'9Blocks uptake of calcium and

binding of acetylcholine92Reduces glucose uptake bysynaptosomes extracted fromrat cortex93

Depresses norepinephrine anddopamine levels in cortex andactivity of enzymesdopamine-f-hydroxylase andphenylethanolamine-N-methyltransferase when fed torats with diets deficient incopper, zinc and iron9495

Reduces cholineacetyltransferase inhypoglossal nucleus and spinalcord grey matter in rabbits96-98

Inhibits fast phase ofvoltage-dependent calciuminflux into synaptosomes99

Inhibits protein phosphatase (insynaptosomal cytosolfractions)100

Is toxic to key synaptosomalenzymes (dependent onsodium-potassium,magnesium and calciumions)'01

Stimulates sodiumchloride-dependent release oftaurine and GABA in ratcortical astrocytes'02

Effects on bloodAlters activity of

cholinesterase'03,104Interacts with

transferrin'05

AMP = adenosine 3',5'-cyclic monophosphate, GMP = guanosine monophosphate.

CAN MED ASSOC J 1991; 145 (7)OCTOBER 1, 1991

indicates that these tangles are composed of apolymer of normal Tau that has undergone abnor-mal phosphorylation'06 and that aluminum inhealthy neurons induces hyperphosphorylation ofTau.

The determination of which of the other toxiceffects listed in Table 1 are important in the patho-genesis of AD will require a more advanced under-standing of the cellular and molecular characteristicsof AD and aluminum neurotoxic effects than nowpossible. The key biochemical events responsible forneuronal dysfunction and neuron death may involveenergy metabolism, calcium homeostasis, membranereceptor and channel functions or gene expression.Aluminum, or any other postulated pathogenic fac-tor, must be shown to induce changes in modelsystems similar to those observed in the humandisease and must be found at the putative site withinbrain cells in sufficient concentrations to induce thechanges observed in AD-affected tissues.'07

Abnormal accumulation of aluminum has nowbeen found in at least four sites in the AD-affectedbrain.

Neurofibrillary tangles

Different analytic methods capable of precisetissue localization'08-"'0 have shown remarkably highconcentrations of aluminum (up to 300 ,ug/g dryweight) in the bundles of paired helical filamentsthat make up AD neurofibrillary tangles. The obser-vation that aluminum induces abnormal phosphoryl-ation of Tau,56 the principal subunit of AD neuro-fibrillary tangles,'06 supports the-idea that aluminumaccumulation may occur early in the disease processand result in the formation of tangles, not late in theprocess in a terminally damaged neuron.

Amyloid cores ofsenile plaques

Focal deposits of aluminum and silicon, asaluminosilicates, are also a consistent and specificfeature of the central core of senile plaques."'"112Aluminosilicates, in vitro, can seed the formation ofpolymeric fibrillary aggregates of model peptides,including the amyloid peptide, which resemble theamyloid core of senile plaques. The amyloid precur-sor protein is considered to be the major proteinfound in senile plaques. Increased amounts of theprotein have been found in some patients withelevated brain aluminum concentrations due tochronic renal failure. In some cases of prolongeddialysis precocious development of immature senileplaques has been observed." '"2 These observations,together with the finding that aluminosilicates seedthe formation of fibrillary aggregates, support thepossibility that elevated amounts of aluminum in

serum and brain can induce some of the cellularserum and brain can induce some of the cellularresponses contributing to amyloid deposition.

Ferritin

Ferritin is an ubiquitous intracellular iron stor-age protein capable of scavenging other metals,including zinc and beryllium. The amount of alumi-num found in ferritin extracted from AD-affectedbrains was 5.6 times higher than in ferritin frommatched control preparations."3 The researchersconsidered that the increase may have been due to ageneral increase in the availability of aluminum tothe brain of patients with AD and raised the possi-bility that aluminum may release iron as Fe+++.This could facilitate the production of highly toxicfree radicals that can denature proteins and destroymembranes, thereby contributing to neuron death inAD.

Chromatin fractions

AD is associated with a change in the structureof the DNA-protein complex that constitutes thephysical matrix within which genes are expressed.''"4This change results in the reduced transcription ofcertain neuron-specific genes, including the low-molecular-weight messenger RNA for neurofilamentprotein. 15,116 Aluminum has been shown to accumu-late on DNA-containing components of the cellnucleus in the cerebral cortex."7 A ninefold increasein aluminum content is associated with the DNA-protein fraction containing repressed neuronalgenes.''I Although aluminum increases the affinity ofbinding of certain repressor proteins to DNA andcontributes to the gene repression, some other eventlikely occurs first to allow the protein to dock at aparticular DNA site.

We postulate that aluminum replaces magne-sium at a key DNA-protein binding site. Because thealuminum atotri has a smaall ionic radius and highcharge it is nearly a million times slower than themagnesium atom in dissociating from the DNA-protein complex. By replacing magnesium, alumi-num locks the repressor protein in place.

Environmental aluminum

Although aluminum is a common constituent ofthe environment it has no recognized biologic func-tion. Aluminum is absorbed primarily through thegastrointestinal tract but probably also through therespiratory epithelium and skin. Preliminary datausing accelerator mass spectrometry with alumi-num-26 ligated to citrate indicate that as much as 1%of aluminum ingested orally is absorbed into theblood stream (James Barker and J. Philip Day,

796 CAN MED ASSOC J 1991; 145 (7) LE I erOCTOBRE 1991

Department of Chemistry, University of Manches-ter, Manchester: personal communication, 1991).Aluminum and fluoride, and probably silicon, aremutually antagonistic in competing for absorption inthe gut; the more fluoride or silicon in the diet, theless aluminum absorbed.

In 1980 Shore and associates"8 did not showincreased aluminum concentrations in AD patients.However, later studies'19-121 showed significantly ele-vated aluminum levels in the serum or whole bloodof AD patients, as compared with the levels incarefully matched control patients. Aluminum ap-pears to be transported in serum and carried into thebrain by several proteins that may have different,genetically determined affinities for aluminum. Onespeculation is that individual variation in suscepti-bility to elevated aluminum concentrations may berelated to the proteins responsible for transportationin serum. A portion of aluminum in serum is carriedby high-molecular-weight proteins, including a2-mac-roglobulin, immunoglobulin, hepatoglobin, trans-ferrin and albumin. In addition, a low-molecular-weight protein (of about 18 kilodaltons) that trans-ports aluminum has recently been discovered.'22

In cerebrospinal fluid most of the aluminum isbound to these newly identified low-molecular-weight proteins. Aluminum uptake by the brain iscurrently attributed to transferrin, and the highestdensities of transferrin receptors are in regions of thebrain selectively vulnerable to AD.123 Recent evi-dence has indicated that the binding of gallium, ananalogue of aluminum, by transferrin is defective inAD.'24 Thus, a serum transport defect could contrib-ute to the accumulation of aluminum in the AD-affected brain.

Although it has been recognized for severalyears that the markedly elevated serum aluminumlevels (200 ,g/L) that occur in kidney failure mayresult in dialysis dementia, much lower concentra-tions are now recognized to be associated withimpaired cognitive function. The psychomotor per-formance of 27 patients receiving long-term hemodi-alysis who had only a mildly raised serum aluminumlevel (mean 59 [normally less than 10] Mg/L) wasimpaired, as compared with the performance ofmatched control subjects.'25 In another study'26 signsof neurologic dysfunction and impaired memoryoccurred in dialysis patients who had labile alumi-num released by desferrioxamine, an aluminumchelating agent, in challenge tests. These observa-tions further indicate that even moderate elevationsin the serum aluminum level pose a risk of cognitiveimpairment.

Sequestration of aluminum

Aluminum may be sequestered in the human

body by several mechanisms. It is injected intramus-cularly or subcutaneously into most of the world'spopulation as an adjuvant in vaccines for diphtheria,pertussis, tetanus and hepatitis and in allergenicextracts. Alum-precipitated allergenic extracts con-tain up to 850 ,ug of aluminum per dose, andinjections of ragweed pollen every 2 weeks for 2years results in a calculated dose of 44 mg ofaluminum. The injected aluminum may persist inthe tissues for weeks or years as subcutaneousnodules.'27 The granulomas comprise aluminum-lad-en histiocytes and dense lymphocytic infiltrates withgerminal centres.'27 The transportation and distribu-tion of injected aluminum in the body are unknown,but tissue depots of aluminum are likely to releasethe metal for many years. Immunologic mechanismsmay play a role in AD,5 and histiocytes that pene-trate brain tissue may serve as an additional trans-port system for aluminum, which bypasses theblood-brain barrier.

Considerable concern has been expressed aboutthe aluminum content of infant formulas and paren-teral nutrients. Bishop and collaborators'28 reportedthat a premature infant who had received an esti-mated total of 645 MAg of aluminum through parenter-al feeding had seizures and died. At autopsy thetemporal grey matter contained an average of 40.1Mg/g wet weight of aluminum, as compared with amean of 2.4 ,g/g wet weight in 12 infants dyingunexpectedly within the first year of life. This17-fold increase closely approaches that mentionedin the first reported fatal case of encephalopathyassociated with respirable aluminum.'29

X-ray energy spectroscopy has detected alumi-num in 9% of oral apical granulomas.'30 Possiblesources of aluminum in the mouth include tooth-paste, impression materials and amalgam.

Aluminum is also found in pigmented macro-phages in Peyer's patches in all patients over 6 yearsof age'3' and is presumed to be of dietary origin.Whether aluminum is sequestered in these cells frompotential neurotoxic actions is unknown. The recentdevelopment of accelerator mass spectroscopy forthe long-life isotope 26AM to be used in tracer studieswill greatly help answer the many questions concern-ing aluminum metabolism.

Epidemiologic findings

Seven studies have related elevated aluminumconcentrations in drinking water to an increasedincidence of AD.'32-'39

In a recent study'37 2792 randomly selectedsubjects aged 65 years or more were screened for ADby psychologists using dementia screening tests.'38They were then examined by senior neurologistsusing criteria established by the National Institute of

CAN MED ASSOC J 1991; 145 (7) 797OCI^OBER 1, 1991

Neurological and Communicative Disorders andStroke and the Alzheimer's Disease and RelatedDisorders Association. 140 The aluminum level indrinking water varied from 10 to 160 ,ug/L for thesample. The relative risk for probable AD at 100,ug/L was 4.53 times greater than at 10 ,ug/L. Martynand colleagues'36 reported similar results in an inci-dence study of presenile Alzheimer's disease (inpeople 65 years of age or less) in 88 counties inEngland. By examining the records of computerizedtomography (CT) units that served the districts theyidentified patients with probable AD (445 people) orpossible AD (221). The control group comprised 519people with dementia and multiple minor strokesvisible on the CT scans and 2920 with adult-onsetepilepsy without clinical dementia; the incidence ofeach of these diseases in the catchment areas wascorrelated with the aluminum content in the drink-ing water. No statistically significant relative riskbetween aluminum in the drinking water and de-mentia caused by multiple strokes or adult-onsetepilepsy was found. However, the prevalence of ADin areas with an aluminum level of 111 ,ug/L ofdrinking water was 1.7 times higher than in areaswith a level of less than 10 gg/L (confidence limits[CLs] 1.1 and 2.7).

Two epidemiologic studies in Norway used anecologic approach. Vogt'33 divided the population ofNorway into five zones according to the aluminumconcentration in the lakes, which corresponded wellwith the aluminum levels found in drinking water.The lowest concentration was 20 ,ug/L and thehighest 200 ,ug/L. The risk of death from seniledementia in the zone with the highest concentrationwas 1.48 times higher than in the zone with thelowest. Correlations between the rates of death fromAD and the aluminum concentrations were statisti-cally significant; multiple-stroke dementia and therisk of death from paralysis agitans were not relatedto the aluminum levels. Flaten'34 found a highlysignificant correlation between aluminum in pro-cessed drinking water and AD.

Table 2 lists the studies in order of increasingdiagnostic accuracy. The relative risk of AD inrelation to aluminum content of drinking water

appears to increase as the diagnostic accuracy in-creases in each of the studies.

In the United States Still and Kelly'32 examinedfirst admissions to the state mental hospitals be-tween 1971 and 1979 in three counties of SouthCarolina. Index cases were patients aged 55 years ormore classified as having primary degenerative de-mentia (considered by the authors to be predomi-nantly AD), vascular dementia, alcoholic dementiaor other forms of dementia. Fluoride levels in waterwere measured instead of aluminum levels, becausefluoride reduces aluminum uptake by the gut. Inaddition, the ingestion of aluminum compoundscounteracts dental fluorosis and results in reducedfluoride stores in teeth and bones.'4' One county had4.18 mg of fluoride per litre of drinking water; thelevels in the other two counties were 0.49 and 0.61mg/L. The incidence of primary degenerative de-mentia in the county with the highest fluoride levelwas about 20% of that in the other two; there wereno significant differences between the counties in thenumber of first admissions because of the othertypes of dementia.

A study in Newfoundland revealed clusters ofhigh rates of death from dementia, recorded ondeath certificates as an immediate, antecedent, un-derlying or contributing cause of death.'39 Indexcases were grouped according to birth place. In 1985and 1986 there was a significant excess of deaths in asmall area of Bonavista Bay that could not beexplained by differences in age, sex, ethnic origin,family origin or mobility patterns. The area wasreported to have a high aluminum concentration inthe drinking water (165 ,ug/L), the lowest pH (5.2)and the highest colour (turbidity) index in theregion. The last two conditions would increase theprobability of the formation of both polynuclearinorganic aluminum and organic aluminum ligands,which, as argued later, have greater aluminum neu-rotoxic effects than mononuclear aluminum.

A recent case-control epidemiologic survey of130 matched pairs examined the association betweenAD and the lifetime exposure to aluminum inantiperspirants and antacids.'42 For aluminum-con-taining antiperspirants the overall adjusted odds

Table 2: Relative risk of Alzheimer's disease in relation to aluminum content of drinking waterin three epidemiologic studies

Aluminum level,Diagnostic No. of uoq/L Relative

Study

Vogt1 33Martyn et al'36Michel et all3?

criteria

Death certificatesComputerized tomogramsDSM- 11138 andNINCDS-ADRDA140

subjects drinking water

18664307

2 792

20-2Q000->' 110

10-160

risk*

1.481.7

4.53*The more rigorous the diagnostic accuracy of the study, the higher the relative risk for AD.

LE Cer OCTOBRE 1991798 CAN MED ASSOC J 1991; 145 (7)

ratio for AD was 1.6 (CL 1.04 and 2.4), the riskincreasing with increased frequency of use (p =0.03), and the odds ratio for the 33% of those whoused the highest amount was 3.2. For antacids theoverall adjusted odds ratio, regardless of aluminumcontent, was 3.1, with a dose-response gradient (p =0.009), and the odds ratio for the highest tertile was11.7. When only aluminum-containing antacids wereanalysed no significant risk or dose-response trendemerged. However, in another study the brain alumi-num concentration was higher after ingestion ofantacids with a high level of aluminum. Dollingerand coworkers143 examined brain tissue specimensfrom 20 subjects scheduled for brain surgery whowere given antacids for 10 days for stress prophylax-is. Half of them received 70 mL of a high-aluminum-content antacid daily, the others an equal dose of alow-aluminum-content antacid. After the 10 days ofantacid treatment the low-dose group had a meanaluminum concentration of 0.412 ,g/g wet weight(an estimated 2.60 ug/g dry weight) and the high-dose group 1.05 ,ug/g wet weight (an estimated 5.25,ug/g dry weight). The mean aluminum level in braintissue from 20 control subjects was 0.583 Ag/g wetweight.

The new tracer techniques demonstrate thatuptake of aluminum between the bowel and theblood and between the blood and the brain occurs ata considerably higher rate than previously suspected.Accelerator mass spectrometry with the use of 26A1ligated to citrate in rats revealed that approximately1/55 000 of a single intraperitoneal injection of 26AMis incorporated into the cerebrum.'44

Epidemiologic studies demonstrate associationbut do not establish cause and effect. Nevertheless,each study reviewed reached the same conclusion:aluminum in drinking water is associated with anincreased risk of dementia of the Alzheimer type.

The most satisfactory studies would includeautopsy findings confirming the diagnosis and rulingout clusters of familial AD as confounding variables.The aluminum ligands in food, water and all sub-stances ingested or injected into people should bemeasured and correlated with the incidence of AD.Such studies will be extremely expensive to conductand will require several years to execute.

Organic and inorganic forms of aluminum

Are all forms of aluminum equally toxic? It iswell known that organic mercury, methyl mercury, ismuch more toxic than metallic mercury. The same istrue for aluminum. Some forms of aluminum, likealuminum hydroxide, are poorly absorbed from thegastrointestinal tract, whereas certain organic ligandsof aluminum, such as aluminum citrate, pass veryrapidly from the food chain into the blood.

The forms of aluminum in drinking water are ofconsiderable significance for epidemiologic studies.They are incompletely studied, and those that couldbe risk factors for AD have not been identified.Driscoll and Letterman'45 studied the chemistry andfate of aluminum in water pumped from LakeOntario and treated with alum for drinking inSyracuse, NY. They found that 52% of the alumi-num was in the form of monomeric alumino-hydrox-ide complexes, 29% was associated with organicmatter, 19% was in a fluoride complex, and a smallamount was particulate. The relative intestinal ab-sorption by mammals of these organic and inorganicforms has not been studied.

Recently, aluminum-27 magnetic resonancespectra of soil samples demonstrated the presence ofinorganic polynuclear aluminum forms, which mayaccount for up to 30% of aluminum in water.'46Since these inorganic polynuclear forms appear to be10 times more toxic than mononuclear forms incertain aquatic plants and fish their toxic effects inmammals and humans need to be investigated.Considerable seasonal variation has been observedin the distribution of aluminum forms in drinkingwater. Therefore, it is very important that thevarious organic and inorganic types of aluminum becorrelated with the incidence of AD in each of theepidemiologic studies reviewed in this article and beconsidered in future studies.

Respiratory aluminum

Rifat and associates'47 examined prolonged ex-posure to respirable aluminum to determine whetherit was associated with serious cognitive deficit.Between 1944 and 1979 finely ground alumina(particle diameter less than 2 um) was dispensedprophylactically for silicosis to groups of gold minersin northern Ontario. The alumina powder was dis-persed in the miners' change rooms for 10 to 20minutes at a concentration of 35 mg/m3 before eachunderground shift. These miners had significantlypoorer performance on cognitive tests than an age-matched group of miners who had not been exposed.The differences persisted after adjusting for potentialconfounding factors such as head injury, educationand alcohol abuse. The relative risk of severe cogni-tive deficit was 4.5 times greater among miners withmore than 20 years' exposure, 3.1 among those with10 to 20 years' exposure and 2.4 among those with 1to 10 years' exposure. Clinical and histopathologicexaminations were not done, and thus no conclusioncan be reached as to whether the cognitive deficitwas more closely related to dialysis dementia or toAD. Although long-term follow-up examinationswere not performed and brain tissue was not exam-ined, the authors have concluded that prolonged

CAN MED ASSOC J 1991; 145 (7) 799OCTOBER 1, 1991

exposure to respirable aluminum is associated withserious cognitive deficit.

Clinical progression ofAD after removalof brain aluminum

A fourth line of independent evidence has testedthe hypothesis that if aluminum is an importantpathogenic factor in AD, its removal by an ion-specific binding agent should slow the progress of thedisease. A 2-year prospective, single-blind clinicaltrial was conducted to determine whether the sus-tained use of desferrioxamine, a trivalent metal ionbinding agent, would slow the progression of thedementia.'48 A total of 48 people living at home withprobable AD were randomly assigned to three treat-ment groups: desferrioxamine, lecithin (in oral ho-meopathic doses of 1 g/d) and no treatment. Astructured performance test measuring daily livingskills was videotaped in the home and was theoutcome measure over the 2-year period. The tapeswere analysed at random by trained behaviour ratersblind to the purpose and protocol of the study. Therewas no statistical difference in the average rate ofdecline in performance between the lecithin groupand the no-treatment group. However, when datafrom these two groups were combined the average2-year decline in the desferrioxamine group was 25%of the maximum score, as compared with 57% in theno-treatment group.

A double-blind placebo-controlled multicentretrial must now be conducted to confirm these results.However, on the basis of current evidence desferri-oxamine appears to slow the progression of AD.These results support the hypothesis that aluminumis a significant toxic environmental factor in thepathogenesis of the disease.

Conclusions

Four independent lines of evidence support theconclusion that aluminum is an important risk factorin AD: (a) prolonged exposure to trace amounts ofaluminum induces cognitive deficits in experimentalanimals and humans, (b) aluminum accumulates inat least four sites in AD-affected brain tissue atconcentrations known to affect several biochemicalreactions, (c) seven epidemiologic studies have dem-onstrated.an association between AD and aluminumexposure in drinking water and antiperspirants (highlevels of aluminum in the air have been found toincrease significantly the risk of cognitive defects,but an association between this type of aluminumexposure and AD has not been investigated) and(d) treatment with a trivalent metal ion bindingagent slows, but does not arrest, the clinical progres-sion of AD.

Recommendations

* Human exposure to aluminum should belimited. After considering the available evidence aprudent person may wish to limit daily exposure toaluminum. In addition to reducing head trauma,which is considered to be a risk factor for AD,'49reduction in aluminum exposure may be the onlychange in lifestyle that offers hope of reducing theincidence of AD. Unfortunately, the sources ofaluminum are largely unknown to the public. As wellas occurring naturally in food and water, aluminumis added to drinking water, many processed foods,cosmetics, toothpaste, antiperspirants and adjuvantsin various parenteral preparations and other phar-maceutical agents. Current evidence supports thehypothesis that a major reduction in the ingestion ofaluminum would significantly reduce the incidenceof AD.

* Public policy action is required. No riskfactor other than aluminum that might be controlledby public action to reduce the incidence of AD hasyet been identified. Sources of public health infor-mation have an obligation to inform the publicabout the current state of knowledge concerning therelation between aluminum and AD so that eachindividual may make an informed decision. Addi-tional important pathogenic factors in AD will be-come recognized as knowledge grows about thisdisease.

* The aluminum content should be listed onthe packages of all substances marketed for humancontact and ingestion including processed foods,potable water, cosmetics, toothpaste and pharmaceu-tical products.

* Municipal processed water should be regulat-ed so that the aluminum concentration is less than50 ,ug/L; the long-term goal should be a concentra-tion of less than 10 ,ug/L.

* A goal for the daily intake of aluminum fromall sources by adults should be 3 mg or less.

* Further research should be conducted tounderstand fully the health risks of aluminum.

This work was supported by grants from the OntarioMental Health Foundation, the Medical Research Councilof Canada, the Department of National Health and Wel-fare, the Scottish Rite Charitable Foundation and theAlzheimer Association of Ontario.

References

1. St George-Hyslop PH, Tanzi RE, Polinsky RJ et al: Thegenetic defect causing familial Alzheimer's disease maps onchromosome 21. Science 1987; 235: 885-890

2. St George-Hyslop PH, Haines JL, Farrer LA et al: Geneticlinkage studies suggest that Alzheimer's disease is not asingle homogeneous disorder. Nature 1990; 347: 194-197

800 CAN MED ASSOCJ 1991; 145(7) LE jCr OCTOBRE 1991

3. Goate A, Chartier-Harlin MC, Mullan M et al: Segregationof a missense mutation in the amyloid precursor proteingene with familial Alzheimer's disease. Nature 1991; 349:704-706

4. Nee LE, Elvidge R, Sunduland T et al: Dementia of theAlzheimer type: clinical and family study of 22 twin pairs.Neurology 1987; 37: 359-363

5. McGeer PL, McGeer E, Rogers J et al: Anti-inflammatorydrugs and Alzheimer disease [C]. Lancet 1990; 335: 1037

6. Krishnan SS, McLachlan DR, Krishnan B et al: Aluminumtoxicity to brain. Sci Total Environ 1988; 71: 59-64

7. Crapper DR, Dalton AJ: Alterations in short term retention,conditioned avoidance response acquisition and motivationfollowing aluminum induced neurofibrillary degeneration.PhysiolBehav 1973; 10: 925-933

8. Idem: Aluminum induced neurofibrillary degeneration,brain electrical activity and alterations in acquisition andretention. Ibid: 935-945

9. Crapper DR: Experimental neurofibrillary degeneration andaltered electrical activity. Electroencephalogr Clin Neuro-physiol 1973; 35: 575-588

10. Idem: Functional consequences of neurofibrillary degenera-tion. In Gershon S, Terry RD (eds): The Neurobiology ofAging, Raven, New York, 1976: 405-432

11. Idem: Dementia: recent observations on Alzheimer's diseaseand experimental aluminum encephalopathy. In Seeman P,Brown GM (eds): Frontiers in Neurology and Neuroscience(Neuroscience Institute, University of Toronto, symposium1), U of Toronto Pr, Toronto, 1974: 97-111

12. Petit TL, Biederman GB, McMullen PA: Neurofibrillarydegeneration, dendritic dying back and learning-memorydeficits after aluminum administration: implications forbrain aging. Exp Neurol 1980; 67: 152-162

13. Rabe A, Lei M, Shek J et al: Learning deficit in immaturerabbits with aluminum-induced neurofibrillary degenera-tion. Exp Neurol 1982; 76: 441-446

14. Yokel RA: Repeated systemic aluminum exposure effects onclassical conditioning of the rabbit. Neurobehav ToxicolTeratol 1983; 5: 41-46

15. Pendlebury WW, Perl DP, Schwentker A et al: Aluminum-induced neurofibrillary degeneration disrupts acquisition ofthe rabbit's classically conditioned nictitating membraneresponse. Behav Neurosci 1988; 102: 615-620

16. Solomon PR, Pingree TM, Baldwin D et al: Disruptedretention of the classically conditioned nictitating mem-brane response in rabbits with aluminum-induced neurofi-brillary degeneration. Neurotoxicology 1988; 2: 209-222

17. Farnell BJ, Crapper McLachlan DR, Baimbridge K et al:Calcium metabolism in aluminum encephalopathy. ExpNeurol 1985; 88: 68-83

18. Crapper DR, Tomko GJ: Neuronal correlates of an encepha-lopathy associated with aluminum neurofibrillary degenera-tion. Brain Res 1975; 97: 253-264

19. Crapper DR, Krishnan SS, Quittkat S: Aluminum, neurofi-brillary degeneration and Alzheimer disease. Brain 1976; 99:67-79

20. Karlik SJ, Eichhorn GL, McLachlan DRC: Molecular inter-actions of aluminum with DNA. Neurotoxicology 1980; 1:83-88

21. Karlik SJ, Eichhorn GL, Crapper DR: Interactions ofaluminum species with deoxyribonucleic acid. Biochemistry1980; 19: 5991-5998

22. Lukiw WJ, Kruck TPA, McLachlan DR: Alterations inhuman linker histone-DNA binding in the presence ofaluminum salts in vitro and in Alzheimer's disease. Neuro-toxicology 1987; 8: 291-301

23. Miller CA, Levine EM: Affects of aluminum salts oncultured neuroblastoma cells. J Neurochem 1974; 22: 751-758

24. Sarkander HI, Balb G, Schlosser H et al: Blockade ofneuronal brain RNA initiation sites by aluminum: a primary

molecular mechanism of aluminum-induced neurofibrillarychanges. In Cervos-Navarro J, Sarkander HI (eds): BrainAging: Neuropathology and Neuropharmacology, Raven,New York, 1983: 645-647

25. Crapper McLachlan DR, Dam TV, Farnell BJ et al: Alumi-num inhibition of ADP-ribosylation in vivo and in vitro.Neurobehav Toxicol Teratol 1983; 5: 645-647

26. Crapper McLachlan DR, Farnell B, Galin H et al: Alumi-num in human brain disease. In Sarkar B (ed): BiologicalAspects of Metals and Metal Related Diseases, Raven, NewYork, 1983: 209-218

27. De Boni U, Seger M, Crapper McLachlan DR: Functionalconsequences of chromatin bound aluminum in culturedhuman cells. Neurotoxicology 1980; 1: 65-81

28. Sanderson CL, Crapper McLachlan DR, De Boni U: Alteredsteroid induced puffing by chromatin bound aluminum in apolytene chromosome of the black fly Simulium vittatum.Can J Genet Cytol 1982; 24: 27-36

29. Idem: Inhibition of corticosterone binding in vitro, in rabbithippocampus, by chromatin bound aluminum. Acta Neuro-pathol (Berl) 1982; 57: 249-254

30. Van Berkum MFA, Wong L, Lewis PN et al: Total andpoly(A) RNA yields during an aluminum encephalopathy inrabbit brains. Neurochem Res 1986; 11: 1347-1359

31. Muma NA, Troncoso JC, Hoffman P et al: Aluminumneurotoxicity - altered expression of cytoskeletal genes.Mol Brain Res 1988; 3: 115-122

32. Gelfant S: Inhibition of cell division - a critical experimen-tal analysis. Rev Cytol 1963; 14: 249-254

33. Sampson M, Clarkson D, Davies DD: DNA synthesis inaluminum treated roots of barley. Science 1965; 148: 1476-1477

34. Walker PR, LeBlanc J, Sikorska M: Effects of aluminum andother cations on the structure of brain and liver chromatin.Biochemistry 1989; 28: 3911-3915

35. Muller G, Bernuzzi V, Desor D et al: Developmentalalterations in offspring of female rats orally intoxicated byaluminum lactate at different gestation periods. Teratology1990; 42: 253-261

36. Siegel N, Haug A: Aluminum interaction with calmodulin:evidence for altered structure and function from optical andenzymatic studies. Biochim Biophys Acta 1983; 744: 36-45

37. Idem: Calmodulin dependent formation of membrane po-tential in barley root plasma membrane vescicles - abiochemical model of aluminum toxicity in plants. PhysiolPlant 1983; 59: 285-291

38. Burnatowska-Hledin MA, Mayor GH: The effect of alumi-num (Al) loading on specific tissue calcium (Ca) andmagnesium (Mg) concentrations in normal rats [abstr]. ClinRes 1982; 30: A741

39. Rorison IH: The effect of aluminum on the uptake andincorporation of phosphate by excised sanfoin roots. NewPhysiol 1965; 64: 23-27

40. Foy CD, Chaney RL, White MC: The physiology of metaltoxicity in plants. Annu Rev Plant Physiol 1978; 29: 511-566

41. Trapp GA: Studies of aluminum interaction with enzymesand proteins: the inhibition of hexokinase. Neurotoxicology1980; 1: 89-100

42. Bock JL, Ash AE: NMR and infrared spectroscopic investi-gations of the Al (III), Ga (III), and Be (II) complexes ofATP. J Inorg Biochem 1980; 13: 105-110

43. Womack FC, Colowick SP: Proton-dependent inhibition ofyeast and brain hexokinase by aluminum in ATP prepara-tions. Proc NatlAcad Sci USA 1979; 76: 5080-5084

44. Karlik SJ, Elgavish GA, Eichhorn GL: Multinuclear NMRstudies on Al (III) complexes of ATP and related com-pounds. JAm Chem Soc 1983; 105: 602-609

45. Karlik SJ, Tariene, Elgavish GA et al: Aluminum-27 NMRstudy of Al (III) interactions with carboxylate ligands. InorgChem 1983; 22: 525-529

46. Lai JCK, Blass JP: Inhibition of brain glycolysis by alumi-

OCTOBER 1, 1991 CAN MED ASSOC J 1991; 145 (7) 801

num. J Neurochem 1984; 42: 438-44547. Mansour JM, Ehrlich A, Mansour TE: The dual effects of

aluminum as activator and inhibitor of adenylate cyclase inthe liver fluke Fasciola hepatica. Biochem Biophys ResCommun 1983; 112: 911-918

48. Sternweiss PC, Gilman AG: Aluminum - a requirement foractivation of the regulatory component of adenylate cyclaseby fluoride. Proc Natl Acad Sci USA 1982; 79: 4888-4891

49. Gruca S, Wisniewski HM: Cytochemical study on the effectof aluminum on neuronal Golgi apparatus and lysosomes.Acta Neuropathol (Berl) 1984; 63: 287-295

50. Altmann P, Alsalihi F, Butter K et al: Serum alum-inum levels and erythrocyte dihydropteridine reductase ac-tivity in patients on thermodialysis. N Engl J Med 1987;317: 80-84

51. Johnson GVW, Jope RS: Aluminum alters cyclic AMP andcyclic GMP levels but not presynaptic cholinergic markersin rat brain in vivo. Brain Res 1987; 403: 1-6

52. Morandi A, Los B, Osofsky L et al: Ubiquitin and heatshock proteins in cultured nervous tissue after differentstress conditions. In Iqbal K, Wisniewski H, Winblad B(eds): Alzheimer's Disease and Related Disorders, AR Liss,New York, 1989: 819-827

53. Fleming J, Joshi JG: Ferritin: isolation of aluminum-ferritincomplex from brain. Proc Natl Acad Sci USA 1987; 84:7866-7870

54. Takagi A, Sai K, Umemura T et al: Relationship betweenhepatic peroxisome proliferation and 8-hydroxydeoxyguano-sine formation in liver DNA of rats following long-termexposure to three peroxisome proliferators: di (2-ethylhexyl)phthalate, aluminum clofibrate and simfibrate. Cancer Lett1990; 35 (1): 33-38

55. Xu ZX, Fox L, Melethil S et al: Mechanism of aluminum-induced inhibition of hepatic glycolysis: inactivation ofphosphofructokinase. J Pharmacol Exp Ther 1990; 254:301-305

56. Guy SP, Jones D, Mann DMA et al: Human neuroblastomacells treated with aluminium express an epitope associatedwith Alzheimer's disease neurofibrillary tangles. NeurosciLett 1991; 121: 166-168

57. Hazlett TL, Higashijima T, Jameson DM: Examination ofelongation factor Tu for aluminum fluoride binding sitesusing fluorescence and F-1 9-NMR methodologies. FEBSLett 1991; 278: 225

58. Cochran M, Elliott DC, Brennan P et al: Inhibition ofprotein kinase C activation by low concentrations of alumin-ium. Clin Chim Acta 1990; 194: 167-172

59. Sturr MG, Marquis RE: Inhibition of proton-translocatingATPases of Streptococcus mutans and Lactobacillus casei byfluoride and aluminum. Arch Microbiol 1990; 155: 22-27

60. Klatzo I, Wisniewski H, Streicher E: Experimental produc-tion of neurofibrillary degeneration: 1. Light microscopicobservations. J Neuropathol Exp Neurol 1965; 24: 187-199

61. Terry RD, Penia C: Experimental production of neuro-fibrillary degeneration: 2. Electron microscopy, phospha-tase histochemistry and electron probe analysis. Ibid:200-210

62. Dahl D, Bignami A: Immunochemical cross-reactivity ofnormal neuro-fibrils and aluminum-induced neurofibrillarytangles - immunofluorescence study and antineural fila-ment serum. Exp Neurol 1978; 58: 79-80

63. Selkoe DJ, Liem RKH, Yen S et al: Biochemical andimmunological characterization of neurofilaments in experi-mental neurofibrillary degeneration induced by aluminum.Brain Res 1979; 163: 235-252

64. Munoz-Garcia D, Pendlebury WW, Kessler JB et al: Animmunocytochemical comparison of cytoskeletal proteins inaluminum-induced and Alzheimer-type neurofibrillary tan-gles. Acta Neuropathol (Berl) 1986; 70: 243-248

65. Bizzi A, Crane RC, Autilio-Gambetti L et al: Aluminumeffect on slow axonal transport: a novel impairment of

neurofilament transport. J Neurosci 1984; 4: 722-73166. Kosik KS, McCluskey AH, Walsh FX et al: Axonal transport

of cytoskeletal proteins in aluminum toxicity - aluminumtoxicity in axonal transport. Neurochem Pathol 1985; 3: 99-108

67. Liwnicz BH, Kristensson K, Wisniewski HM et al: Observa-tions on axoplasmic transport in rabbits with aluminum-induced neurofibrillary tangles. Brain Res 1974; 80: 413-420

68. Johnson GVM, Jope RS: Cytoskeletal protein phosphoryl-ation is altered in the brains of aluminum-treated rats[abstr]. Presented at the 17th Annual Meeting of the Societyfor Neurosciences, New Orleans, 1987: A366. 10 (vol 13)

69. Macdonald TL, Humphries WG, Martin RB: Promotion oftubulin assembly by aluminum ion in vitro. Science 1987;236: 183-186

70. Macdonald TL, Martin RB: Aluminum ion in biologicalsystems. Trends Biochem Sci 1988; 13: 15-19

71. Strong MJ, Wolff AV, Wakayama I et al: Aluminum-induced chronic myelopathy in rabbits. Neurotoxicology1991; 12: 9-21

72. Katsetos CD, Savory J, Herman MM et al: Neuronalcytoskeletal lesions induced in the CNS by intraventricularand intravenous aluminium maltol in rabbits. NeuropatholApplNeurobiol 1990; 16: 511

73. Troncoso JC, March JL, Haner M et al: Effect of aluminumand other multivalent cations on neurofilaments in vitro: anelectron microscopic study. J Struct Biol 1990; 103: 2-12

74. Nixon RA, Clarke JF, Logvinenko KB et al: Aluminuminhibits calpain-mediated proteolysis and induces humanneurofilament proteins to form protease-resistant high mole-cular weight complexes. J Neurochem 1990; 55: 1950-1959

75. Viestra R, Haug A: The effect of aluminum 3+ on thephysical properties of membranes in Thermoplasma acid-ophilum. Biochem Biophys Res Commun 1978; 84: 134-144

76. Deleers M: Cationic atmosphere and cation competitionbinding at negatively charged membranes: pathological im-plication of aluminum. Res Commun Chem Pathol Phar-macol 1985; 49: 277-294

77. Deleers M, Servais JP, Wulfert E: Micromolar concentra-tions of A13+ induce phase separation, aggregation and dyerelease in phosphatidylserine-containing lipid vesicles. Bio-chim Biophys Acta 1985; 813: 195-200

78. Idem: Neurotoxic cations induce membrane rigidificationand membrane fusion at micromolar concentrations. Bio-chim BiophysActa 1986; 855: 271-276

79. Lai JCK, Guest JF, Leung TKC et al: The effect ofcadmium, manganese and aluminum on sodium-potassium-activated and magnesium-activated adenosine triphosphateactivity and choline uptake in rat brain synaptosomes.Biochem Pharmacol 1980; 29: 14 1-146

80. Ohtawa M, Seko M, Takayama F: Effect of aluminum onlipid peroxidation in rats. Chem Pharm Bull (Tokyo) 1983;31: 1415-1418

81. Gutteridge JMC, Quinlan GJ, Clark I et al: Aluminum saltsaccelerate peroxidation of membrane lipids stimulated byiron salts. Biochim BiophysActa 1985; 835: 441-447

82. Banks WA, Kastin AJ, Fasold MB: Differential effect ofaluminum on the blood-brain barrier, transport of peptides,technetium and albumin. J Pharmacol Exp Ther 1988; 244:579-585

83. Banks WA, Kastin AJ: Aluminium increases permeability ofthe blood-brain barrier to labelled DSIP and ,3-endorphin:possible implications for senile and dialysis dementia. Lan-cet 1983; 2: 1227-1229

84. Idem: Aluminum alters the permeability of the blood-brainbarrier to some non-peptides. Neuropharmacology 1985; 24:407-412

85. Idem: The aluminum induced increase in blood-brain barri-er permeability to delta-sleep-inducing peptide occursthroughout the brain and is independent of phosphorus and

802 CAN MEDASSOCJ 1991; 145(7) LE CerOCTOBRE 1991

acetylcholinesterase levels. Psychopharmacology (Berlin)1985; 86: 84-89

86. Uemura E, Ireland WP: Synaptic density in chronic animalswith experimental neurofibrillary changes. Exp Neurol 1984;85: 1-9

87. Wong PCL, Lai JCK, Lim L et al: Selective inhibition ofL-glutamate and gamma-aminobutyrate transporter in nerveending particles by aluminum, manganese and cadmiumchloride. J Inorg Biochem 1981; 14: 253-260

88. Sturman JA, Wisniewski HM, Shek JW: High affinityuptake of GABA and glycine by rabbits with aluminum-induced neurofibrillary degeneration. Neurochem Res 1983;8:1097-1109

89. Lai JCK, Lim L, Davison AN: Effects of Cd2+, Mn2+, andA13+ on rat brain synaptosomal uptake of noradrenaline andserotonin. J Inorg Biochem 1982; 17: 215-225

90. Marquis JK: Aluminum neurotoxicity: an experimentalperspective. Bull Environ Contam Toxicol 1982; 29: 43-49

91. Marquis JK, Lerrick AJ: Noncompetitive inhibition ofbi-aluminum, scandium and yttrium of acetylcholinesterasefrom Electrophorus electricus. Biochem Pharmacol 1983; 31:1437-1440

92. Hava M, Hurwitz A: The relaxing effect of aluminum andlanthanum on rat and human gastric smooth muscle invitro. Eur J Pharmacol 1973; 22: 156-161

93. Lipman JJ, Colowick SP, Lawrence PL et al: Aluminuminduced encephalopathy in the rat. Life Sci 1988; 42: 863-875

94. Wenk GL, Stemmer KL: The influence of ingested alumi-num upon norepinephrine and dopamine levels in the ratbrain. Neurotoxicology 1981; 2: 347-353

95. Idem: Activity of the enzymes dopamine-beta-hydroxylaseand phenylethanolamine-N-methyltransferase in discretebrain regions of the copper-zinc deficient rat followingaluminum ingestion. Neurotoxicology 1982; 3: 93-99

96. Yates CM, Simpson J, Russell D et al: Cholinergic enzymesin neurofibrillary degeneration produced by aluminum.Brain Res 1980; 197: 269-274

97. Hetnarski D, Wisniewski HM, Iqbal K et al: Centralcholinergic activity in aluminum-induced neurofibrillarydegeneration. Ann Neurol 1989; 7: 489-490

98. Simpson J, Yates CM, Whyler DK et al: Biochemical studieson rabbits with aluminum-induced neurofilament accumula-tions. Neurochem Res 1985; 10: 229-238

99. Koenig ML, Jope RS: Aluminum inhibits the fast phase ofvoltage-dependent calcium influxes into synaptosomes. JNeurochem 1987; 49: 316-320

100. Yamamoto H, Saitoh Y, Yasugawa S et al: Dephosphoryla-tion of Tau factor by protein phosphatase 2A in synaptoso-mal cytosol fractions, and inhibition by aluminum. J Neuro-chem 1990; 55: 683-690

101. Rao KSJ: Effects of aluminum salts on synaptosomal en-zymes: an in vitro kinetic study. Biochem Int 1990; 22: 725-734

102. Albrecht J, Norenberg MD: Aluminum chloride stimulatesNaCl-dependent release of taurine and y-aminobutyric acidin rat cortical astrocytes [abstr]. Neurochem Int 1991; 18:125

103. Patocka J: The influence of A13+ on cholinesterase andacetylcholinesterase activity. Acta Biol Med Ger 1971; 26:845-846

104. Marquis JK: Aluminum inhibition of human serum cholin-esterase. Bull Environ Contam Toxicol 1983; 31: 164-169

105. Trapp GA: Plasma aluminum is bound to transferrin. LifeSci 1983; 33: 311-316

106. Lee VMY, Balin BJ, Otvos L et al: A68: a major subunit ofpaired helical filaments and derivatized forms of normaltau. Science 1991; 251: 675-678

107. McLachlan DRC: Aluminum neurotoxicity: criteria for as-signing a role in Alzheimer's disease. In Lewis TE (ed):Environmental Chemistry and Toxicology of Aluminum,

Lewis Pubis, Chelsea, Mich, 1989: 299-315108. Pern DP, Brody AR: Alzheimer's disease: x-ray spectrometric

evidence of aluminum accumulation in neurofibrillary tan-gle-bearing neurons. Science 1980; 208: 297-299

109. Perl DP, Pendlebury WW: Aluminum accumulation inneurofibrillary tangle-bearing neurons of senile dementia,Alzheimer's type: detection by intraneuronal x-ray spectrom-etry studies of unstained tissue sections [abstr]. J Neuropa-thol Exp Neurol 1984; 43: 349

110. Perl DP: Aluminum and Alzheimer neurofibrillary degener-ation [abstr]. Presented at the Aluminum and Health Inter-national Symposium, Orlando, Fla, Dec 12-14, 1989

111. Candy JM, Klinowski RH, Perry EK et al: Aluminosilicatesand senile plaque formation in Alzheimer's disease. Lancet1986; 1: 354-357

112. Edwardson JA, Candy JM: Aluminum and the aetiopa-thogenesis of Alzheimer's disease [abstr]. Neurobiol Aging1990; 11:314

113. Fleming J, Joshi JG: Ferritin: isolation of aluminum-ferritincomplex from brain. Proc Natl Acad Sci USA 1987; 84:7866-7870

114. McLachlan DRC, Lukiw WJ, Kruck TPA: New evidence foran active role of aluminum in Alzheimer's disease. Can JNeurol Sci 1989; 16: 490-497

115. McLachlan DRC, Lukiw WJ, Wong L et al: Selectivemessenger RNA reduction in Alzheimer's disease. Mol BrainRes 1988; 3: 255-262

116. Clark AW, Krekoski CA, Parhad IM et al: Altered expres-sion of genes for amyloid and cytoskeletal proteins inAlzheimer cortex. Ann Neurol 1989; 25: 331-339

117. Crapper DR, Quittkat S, Krishnan SS et al: Intranuclearaluminum content in Alzheimer's disease: dialysis encepha-lopathy. Acta Neuropathol (Berl) 1980; 50: 19-24

118. Shore D, King SW, Kaye W et al: Serum and CSF aluminumand circulating parathyroid hormone in primary degenera-tive (senile) dementia. Neurotoxicology 1980; 1: 55-63

119. Naylor GJ, Smith AHW, McHarg A et al: Raised serumaluminum concentration in Alzheimer's disease. Trace ElemMed 1989; 6: 93-95

120. Van Rhijn A, Corrigan FM, Ward NI: Serum aluminum insenile dementia of Alzheimer's type and in multi-infarctdementia. Ibid: 24-26

121. Kellet JM, Taylor A, Oram JJ: Aluminosilicates and Alz-heimer's disease [C]. Lancet 1986; 1: 68,2

122. Favarato M, Mizzen C, Kruck TPA et al: Chromatographicresolution of aluminum binding components in humanserum [abstr]. Neurobiol Aging 1990; 11: 315

123. Edwardson JA, Candy JM: Aluminium and the pathogenesisof senile plaques in Alzheimer's disease, Down's syndromeand chronic renal dialysis. Ann Med 1989; 21: 95-97

124. Farrar G, Altmann P, Welch S et al: Defective gallium-transferrin binding in Alzheimer disease and Down syn-drome: possible mechanism for accumulation of aluminiumin brain. Lancet 1990; 335: 747-750

125. Altmann P, Dhanesha U, Hamon C et al: Disturbance ofcerebral function by aluminium in haemodialysis patientswithout overt aluminium toxicity. Lancet 1989; 1: 7-11

126. Sprague SM, Corwin HL, Tanner CM et al: Relationship ofaluminum to neurocognitive dysfunction in chronic dialysispatients. Arch Intern Med 1988; 148: 2169-2172

127. Slater DN, Underwood JCE, Durrant TE et al: Aluminiumhydroxide granulomas: light and electron microscopicstudies and x-ray microanalysis. Br J Dermatol 1982; 107:103-108

128. Bishop NJ, Robinson MJ, Lendon M et al: Increasedconcentration of aluminium in the brain of a parenterallyfed preterm infant. Arch Dis Child 1989; 64: 1316-1317

129. McLaughlin AIG, Kazantzis G, King E et al: Pulmonaryfibrosis and encephalopathy associated with the inhalationof aluminium dust. BrJInd Med 1962; 19: 253-263

130. Levison DA, Crocker PR, Lee G et al: Unexpected inorganic

OCTOBER 1, 1991 CAN MED ASSOC J 1991; 145 (7) 803

elements in oral lesions: results of x-ray energy spectroscopy(XES) on particulate matter in paraffin sections. J Pathol1984; 144: 119-129

131. Shepherd NA, Crocker PR, Smith AP et al: Exogenouspigment in Peyer's patches. Hum Pathol 1987; 18: 50-54

132. Still CN, Kelly P: On the incidence of primary degenerativedementia vs. water fluoride content in South Carolina.Neurotoxicology 1980; 4: 125-131

133. Vogt T: Water quality and health - study of a possiblerelationship between aluminum in drinking water and de-mentia [abstr] (Sosiale og okonomiske studier 61), CentralBureau of Statistics of Norway, Oslo, 1986

134. Flaten TP: Geographical associations between aluminum indrinking water and registered death rates with dementia(including Alzheimer's disease, Parkinson's disease andamyotrophic lateral sclerosis) in Norway. Environ GeochemHealth 1990; 12: 152-167

135. Leventhal GH: Alzheimer's Disease and Environmental Alu-minum in Maryville and Morristown, Tennessee, PhD thesis,University of Tennessee, Knoxville, Tenn, 1986

136. Martyn CN, Osmond C, Edwardson JA et al: Geographicalrelation between Alzheimer's disease and aluminium indrinking water. Lancet 1989; 1: 59-62

137. Michel P, Commenges D, Dartigues JF et al: Study of therelationship between Alzheimer's disease and aluminum indrinking water [abstr]. Neurobiol Aging 1990; 11: 264

138. Diagnostic and Statistical Manual of Mental Disorders, 3rded, rev, American Psychiatric Association, Washington,1987

139. Frecker MF: Dementia in Newfoundland: Identification of ageographical isolate? J Epidemiol Community Health (inpress)

140. McKhann G, Drachman D, Folstein M et al: Clinicaldiagnosis of Alzheimer's disease: report of the NINCDS-ADRDA Work Group under the auspices of Department ofHealth and Human Services Task Force on Alzheimer'sDisease. Neurology 1984; 34: 939-944

141. Nania JM: Effect of minerals on dental caries. In Gould RF(ed): Dietary Chemicals Versus Dental Caries (Advances inChemistry ser, 94), American Chemical Society, Washing-ton, 1970: 123-160

142. Graves AB, White E, Koepsell TD et al: The associationbetween aluminum-containing products and Alzheimer'sdisease. J Clin Epidemiol 1990; 43: 35-44

143. Dollinger HC, Zumkey H, Spieker C et al: Aluminum inantacids shown to accumulate in brain and bone tissue.Gastroenterol Obs 1986; 5: 478

144. Kobayashi K, Yumoto S, Nagai H et al: 26A1 tracer experi-ment by accelerator mass spectromety and its application tothe studies for amyotrophic lateral sclerosis and Alzheimer'sdisease. Proc Jpn Acad 1990; 66B (10): 189-192

145. Driscoll CT, Letterman P: Chemistry and fate of AI(III) intreated drinking water. JEnviron Eng 1988; 114: 21-37

146. Hunter D, Ross DS: Evidence for a phytotoxic hydroxy-aluminum polymer in organic soil horizons. Science 1991;251: 1056-1058

147. Rifat S, Eastwood MR, McLachlan DR et al: Evidenceregarding the effect of prolonged aluminium exposure oncognitive function. Lancet 1990; 336: 1162-1165

148. Crapper McLachlan DR, Dalton AJ, Kruck TPA et al:Intramuscular desferrioxamine in patients with Alzheimer'sdisease. Lancet 1991; 337: 1034-1038

149. Mortimer JA: Genetic and environmental risk factors forAlzheimer's disease: key questions and new approaches. InAltman H (ed): Alzheimer's Disease and Dementia: Prob-lems, Prospects and Perspectives, Plenum Pr, New York,1989: 85-100

[This issue's Encore selection (starting on page 823) featuresa 1936 case report on Alzheimer's disease. -Ed.]

Conferencescontinuedfrom page 769

Oct. 24-25, 1991: Gairdner Foundation LecturesUniversity of TorontoSally-Anne Hrica, Gairdner Foundation, 220-255Yorkland Blvd., Willowdale, ON M2J 1S3;(416) 493-3101, fax (416) 493-8158

Les 24 et 25 oct. 1991 : 9e Assembl&e scientifique annuelledu College des medecins de famille du Canada (sectionQuebec) - Culture et sante: defis et perspectives

MontrealMicheline Guilbault, secretaire administrative, CP 146,

Succ. Champlain, LaSalle, QC H8P 3J1; (514) 762-9889,fax (514) 762-9870

Oct. 24-27, 1991: Canadian Association of Gerontology20th Annual Scientific and Educational Meeting

Regal Constellation Hotel, TorontoCanadian Association of Gerontology, 110- 1565 Carling

Ave., Ottawa, ON K1Z 3R1; (613) 728-9347

Oct. 25, 1991: Research Day in Family MedicineRadisson Hotel, London, Ont.Margot Meijer, Thames Valley Family Practice Research

Unit, 1489 Richmond St., London, ON N6G 2M1;(519) 439-0121, fax (519) 439-0124

Oct. 25-27, 1991: Canadian Medical Society on Alcoholand Other Drugs 3rd Annual Scientific Meeting

Clarke Institute of Psychiatry and the Addiction ResearchFoundation, Toronto

Canadian Medical Society on Alcohol and Other Drugs,13-100 College St., Toronto, ON M5G 1L5;(416) 595-6000, ext. 7363, fax (416) 595-1214

Oct. 30-Nov. 1, 1991: Catholic Health Association ofCanada Administrators' Seminar - Threats to theMission of Catholic Health Care: a Strategic Response

Radisson Hotel, OttawaFreda Fraser, director of communications, Catholic Health

Association of Canada, 1247 Kilborn P1., Ottawa, ONKIH 6K9; (613) 731-7148

Nov. 1-2, 1991: Pediatric AIDS Conference (sponsored bythe Sunny Hill Hospital for Children, Vancouver, andthe Division of Continuing Education in the HealthSciences, University of British Columbia)

Coast Plaza Hotel at Stanley Park, VancouverPediatric AIDS Conference, Continuing Education in the

Health Sciences, 105-2194 Health Sciences Mall,Vancouver, BC V6T 1Z3; (604) 822-2626,fax (604) 822-4835

Nov. 1-2, 1991: Women, Food and Weight - NewPerspectives

519 Church Street Community Centre, TorontoNational Eating Disorder Information Centre, 200

Elizabeth St., Ste. CWI-328, Toronto, ON M5G 2C4;(416) 340-4156

continued on page 829

804 CAN MED ASSOC J 1991; 145 (7) For prescribing information see page 873 -