McCollom Walton #WALTON


That aluminum is invariably present in the animal body has long been known. Its percentage in different tissues varies considerably and it appears that the total amount in the body increases with age.

Research concerned with aluminum has been directed mainly towards the answer of one or the other of two problems: First, whether or not aluminum, introduced into the body in amounts larger than would be furnished by natural foods and drinks, is toxic; and second, whether or not aluminum is of value in hemoglobin synthesis.

Concerning the toxicity of aluminum there is a wide variety of opinions. At one extreme is the view of Gies and his coworkers, who believe that aluminum compounds when present in the diet are absorbed into the body with subsequent harmful effects. The other extreme view is represented by those who believe that aluminum has a definite physiological function. That Osborne and Mendel thought they obtained better growth of rats when traces of iodine, manganese, fluorine and aluminum were added to their "artificial protein-free milk" than they did without these elements, has already been mentioned. Daniels and Hutton have suggested also that aluminum may be essential to reproduction in rats. An intermediate view is represented by the report of the Referee Board of Consulting

Scientific Experts headed by Ira Remsen and by that of McCollum, Rask, and Becker who hold that aluminum is not toxic, nor does it interfere with growth or reproduction.

That aluminum is toxic when injected directly into the blood stream was demonstrated, at least for the rabbit, by Siebert and Wells (77). The chief question, therefore, seems to be whether or not aluminum compounds are absorbed from the alimentary tract. Smith (78) fed bread made with and without alum baking powder, to groups of pigs and obtained no differences in rate of growth; neither did he find aluminum in the organs of the pigs. Mackenzie (54) also fed aluminum compounds to pigs with no harmful effects. Taylor (84) claims that aluminum is not absorbed into the blood stream. This view is also held by McCollum, Rask, and Becker (51) who state further that, by the use of the spectroscope, aluminum is a constituent of neither animal nor plant matter.

There is considerable evidence, however, indicating that aluminum is absorbed, but not to any great extent. The literature to 1928 has been reviewed by Smith (79) and by Myers and his associates (64). Myers and his associates conclude that traces of aluminum are present in the tissues normaIly and that the amount is slightly increased on a diet containing considerable aluminum. Growth is apparently unaffected by aluminum feeding. It has been shown that ingested aluminum compounds are excreted almost entirely in the feces. It is then, perhaps, the almost non-absorbability of these compounds that renders them harmless.

Concerning the value of aluminum in hemoglobin synthesis, uniformly negative results have been obtained. Thus we can conclude that there is no experimental evidence indicating that the body has a requirement for aluminum and, although this element accumulates slightly in the tissues with age, there is little likelihood that animals are harmed in any way by its consumption in natural foods, drinks, or the usual mineral supplements.


Dr Judie Walton
Australian Institute for Biomedical Research

1. Introduction   
Soluble aluminium is classified as a definite neurotoxin to humans (Simonsen et al., 1994). Some of the first evidence indicating the neurotoxicity of aluminium was reported in 1921 by Spofforth in The Lancet:

The patient was in a state of great exhaustion and suffering from very severe and persistent vomiting.... I suspected metallic poisoning and later sent a specimen of his urine to Messrs. Thomas, Bourlet, and Newman, analytical chemists who reported that it contained a large amount of aluminium, also of phosphates. The patient said he had been dipping red-hot metal articles, contained in an aluminium holder into concentrated nitric acid.... It caused loss of memory, tremor,
jerking movements and impaired coordination....J Spofforth, LRCP, MRCS

Kopeloff et al. (1942) experimentally demonstrated the neurotoxicity of aluminium. Equal
concentrations of either aluminium, silver, or copper were applied directly to the surface of monkeys' brains. The silver and copper applications produced little effect but the aluminium caused severe convulsions, coma and death within a few days.

Toxicity is a product of three factors: 
1) degree of individual susceptibility; 
2) duration of exposure; 
3) concentration of the bioavailable toxin; i.e. that fraction absorbed across the gastrointestinal or lung linings, or skin into the bloodstream where it is then available for uptake by the brain and other tissues. 

Individual susceptibility is the most variable of these factors. Identical exposure of two individuals may yield total resistance in one and severe symptoms or death in the other. Aluminium bioavailability depends on its ability to be absorbed which, in turn, is strongly influenced by its solubility. When either aluminium chloride or aluminium sulphate (alum) is added to pure water their solubilities are, to a large extent, governed by pH.

Aluminium speciation and solubility in surface waters and biological solutions are much more complicated than in pure water. Alum treatment, in a rapid filtration process, removes particulate forms of aluminium from water and this generally decreases the total aluminium content of the water. However, several drinking water studies have shown that, in many cases, alum-treatment increases the level of soluble monomeric inorganic aluminium in the finished water supply (Kopp, 1970; Tran, et al., 1993; Zhang, et al., 1994; Shovlin, et al., 1993; Miller, et al., 1984). This soluble fraction is potentially more bioavailable, and thus more toxic than the particulate aluminosilicates removed by filtration.

2. Aluminium absorption and bioavailability

2.1 Study 

We carried out bioavailability experiments in rats to investigate how various foods, beverages, and other factors influence aluminium absorption from ingested drinking water (Walton et al., 1994). Some animals were given a pharmacological dose of alum (8 mg) diluted in 2 ml purified water. Others were given the same dose of alum diluted in 1 ml water together with 1 ml of either a beverage or a puréed food. Hourly plasma and urine samples were taken for aluminium absorption and excretion measurements using graphite furnace atomic absorption spectroscopy (GFAAS).

We found that alum co-exposure with beverages was more likely to increase aluminium absorption from water than its co-exposure with foods. Lemon juice produced the largest increase (1700%) followed by orange juice (1260%). Coffee, tomato juice and wine also significantly enhanced aluminium absorption from water.

Meat (beef) and wheat products produced a modest inhibition of aluminium absorption. We note that wheat contains high levels of phytic acid and silica which can complex with the trivalent metal ions.

A l u m i n i u m


Other gastrointestinal factors reported to affect aluminium absorption include molecules formed in, and secreted into, the digestive tract such as mucus and bicarbonate. Aluminium absorption increases when iron, magnesium and/or calcium are deficient.

Three main findings came out of our study.
* Aluminium absorption levels are disproportionate to ingested levels. This is because various forms of aluminium differ greatly in their solubilities and extent of absorption.
When we gave rats equal concentrations of aluminium either in the form of an alum solution, or as an aluminium hydroxide antacid tablet pulverized in water, aluminium absorption from the alum
source was 2500% higher than from the antacid source. 

* Synthetic aluminium food or beverage additives are less well incorporated, more easily extracted, and more likely to be absorbed than aluminium sources naturally contained in foods and beverages. Alum-treated water is a synthetic beverage. Tea contains high levels of aluminium (e.g., 5 mg/L) bound to polyphenolic ligands which inhibit its absorption (Powell et al., 1993). Fresh orange juice by itself naturally contains about 26 mg/L aluminium but the amount absorbed was too low to be detectable with GFAAS. However, when orange juice concentrate was reconstituted with water containing alum, its citric and ascorbic acids increased aluminium absorption from the water by 1260%. Likewise, aluminium contained in biscuits made with aluminium-based baking powder was readily solubilised and absorbed.

* More aluminium is absorbed from alum-treated water drunk on an empty stomach than on a full stomach. If orange juice reconstituted with alum-treated water was drunk on an empty stomach, aluminium absorption was greatly enhanced This probably involves a pH effect on aluminium solubility since food neutralizes gastric acidity whereas beverages have little buffering capacity. 

Previous estimates of aluminium bioavailability have been based on the total amount of aluminium humans consume each day. These amounts are about 10-15mg from food, according to Greger (1985), and about 0.15mg from drinking water.

This oversimplified basis disregards the proportion of soluble aluminium in the ingested substance. Moreover, it fails to take into account the complex effects that other nutrients exert on aluminium absorption.
Daily oral intake levels of aluminium are poor indicators of aluminium absorption and bioavailability. The only way to estimate aluminium bioavailability is to measure it over time, starting within 30 minutes of ingestion and continuing at 30-60 minute intervals for several hours thereafter, during which time the plasma and urinary aluminium levels peak and then fall.

2.2. Study 2

Metal metabolism is similar in rats and humans and aged rats serve as a model for aged humans. Rats that are 24-27 months of age are approximately equivalent to 72-81 years old humans whereas those that are 7- 10 months are similar to 21-30 years old humans. We compared aluminium absorption and excretion in old and young rats of these ages.

* We found that the older individuals had less citric-acid stimulated uptake of aluminium into the blood from water than the younger ones but they also excreted the absorbed aluminium more slowly. Humans are reported to lose 50% of their kidney glomeruli between the ages of 30 and 85 years (Hamburger and Crozier, 1979) so their ability to filter the blood should be similarly decreased in old

A l u m i n i u m


2.3. Study 3

We also carried out experiments to determine where in the gastrointestinal tract, aluminium was absorbed. The rats were anaesthetised and a segment of the digestive tract was surgically tied off to form a pocket in either the duodenum, jejunum, ileum, colon, or stomach. The pocket was filled with alum-treated water prior to collecting hourly blood and urine specimens. 

* Our results showed that aluminium can be absorbed anywhere along the digestive tract, even from the colon. The jejunum had the highest rate of aluminium absorption. The stomach had a different pattern of aluminium absorption than the intestinal segments. Several laboratories have shown that most aluminium, upon absorption into the bloodstream, rapidly
becomes protein-bound for transport through the vasculature (e.g. Day et al., 1991).

Fraction % Al Bound
High Molecular weight fraction
* Transferrin-association 80
* Albumin-association 10
* Other 5
Low molecular weight fraction 
* (citrate/free/etc) 5

As bioavailable aluminium circulates, it is removed from the bloodstream by the kidneys and other organs. Most aluminium excreted from the body is in the low molecular weight fraction which is filterable by the kidneys. Under some conditions, a fraction of the urinary aluminium can be returned to the circulation during its passage through the proximal tubules (Bumatowska-Hledin et al., 1985).
Aluminium is also excreted from the blood into the bile and returned to the intestine (Klein et al., 1983). Our results suggest that a fraction of this biliary aluminium may return to the bloodstream by reabsorption from the intestine below the bile duct entrance. Transferrin serves as the main carrier for aluminium in the bloodstream. It also serves as an iron-transporting protein but is typically only 30% saturated by iron.

3. Aluminium and the brain
Transferrin receptors are found in capillaries of the brain but not of other tissues (Jefferies et al., 1984). These receptors facilitate the uptake of transferrin-bound metals across the blood-brain barrier into the brain tissue. Tissue culture experiments show that brain cells are able to concentrate high levels of aluminium available to them. In a 25 mmol aluminium solution, 60-70% of the aluminium was taken up by the cells (Shi and Haug, 1990). In another experiment, in one week of incubation, intracellular levels reached 10-20 mmol Al while extracellular levels were held at 100 mmol with an Al-EDTA complex (Guy et al., 1991).

3.1. Study 4

We used an ultra-sensitive technique, accelerator mass spectrometry plus 26aluminium (26Al), to track the movement of physiological amounts of aluminium from drinking water into the brains of rats (Walton et al., 1995). Two weeks after the rats were given 25mg/L 26Al in water, they were overdosed with anaesthetic and a saline solution was used to flush the blood from their brains. 

* The results showed that trace levels of aluminium entered the brain from the equivalent of a single glass of water.

A l u m i n i u m


Figure 1: Some potential sources of bioavailable aluminium

* Alum-treated drinking water
* Food additives:
- free-flowing agent in salt
- rising agent in some baking powders
- hardening agent for pickles, candied fruits
* Many antacids.
* Buffered aspirins
* Abrasive in some toothpastes
* Beer & soft drinks in aluminium drink cans*
* Food cooked in aluminium trays and foils
* Use of aluminium cookware and kitchen utensils
* Use of carbonated drink makers
* Aluminium components in coffee percolators
* Aluminium anode rods in hot water heaters
* Aluminium implants (hip replacement, facial implants, dentistry)
* Some medical treatments
- alum irrigation for bladder haemorrhage, rectal prolapse
* Cosmetics
* Vaccines (aluminium added to increase response)
* Aluminium fumes (e.g. from welding)
* Styptic pencils
* Most deodorants
* Dusting powder in rubber gloves, condoms, other sanitary goods
* Pesticide containing aluminium phosphide
* Aluminium cans are resin-lined but carbonated drinks attack the lining and corrode the aluminium during storage.

According to our calculations, the fraction of the ingested aluminium dose absorbed into the brain from drinking water ranged from 10-7 to 10-8. Other investigators have calculated that about 10-3 to 10-4 of an ingested dose of water-borne aluminium is absorbed across the gastrointestinal barrier into the blood. Kobayashi has calculated that about 10-4 of an injected aluminium dose is taken up from the blood into the rat brain (Kobayashi et al.. 1990). Our measurements of aluminium absorption from the gastrointestinal tract into the brain are therefore consistent with other available data.

* If a comparable amount of aluminium is taken up into human brain as into rat brain, the amount of neurotoxic aluminium entering the human brain from drinking water over a lifetime could accumulate to significant levels and might lead to brain cell damage. Drinking water is only one source of bioavailable aluminium and is unlikely to be the sole contributor to brain aluminium (see Fig. 1).

* In our 26Al experiment, we noted that some brains took up more than ten times as much aluminium than others. We have hypothesized that these uptake differences may be due to genetic variation arising at the gastrointestinal level. Some types of highly inbred mice are known to absorb more aluminium than others (Fosmire et al., 1993). Our Wistar rats were outbred, and were therefore genetically non- identical. In our experiments of gastrointestinal absorption of aluminium, and in others where data are listed, large ranges of absorption values are characteristic, both for rodents and for humans (e.g. Van der Voet et al., 1989; Taylor et al., 1992; Weberg et al., 1986). Such absorption differences give

A l u m i n i u m


rise to higher or lower blood aluminium levels which in turn affect the amount of aluminium available for uptake by the brain. Aluminium mechanisms of cellular toxicity primarily arise from its physical and binding properties, due to its ionic radius and charge influence. The ionic radii of aluminium, iron, magnesium, and calcium, are: Al, 0.054nm; Fe, 0.065nm; Mg, 0.072nm; Ca, 0.100nm.

The aluminium ion is approximately the same size as the ferric ion and smaller than magnesium and calcium ions. The transferrin example illustrates how aluminium utilises molecules which normally serve iron to access brain cells. Being somewhat smaller, aluminium can replace magnesium in many biological systems and it competes with calcium for phosphate and small ligands (Meiri et al, 1993).
However, it binds many anions much more strongly than the essential metals. For example aluminium binds ATP 107 times more strongly than magnesium (Martin, 1986), thus interfering with reactions which require readily reversible dissociation. This is basically the same mechanism whereby lead is likely to kill cells (Walton, 1973).

Aluminium produces toxic effects at the cell membrane in several ways: 
1) by altering the physical properties of the membrane; 
2) by interfering with the function of voltage-activated ionic channels;
3) by altering the secretion of transmitters. Within the cell, aluminium can affect many key processes in the nucleus, cytoplasm, and mitochondria, including: 1) glucose metabolism, 2) signal transduction, 3) neurotransmitter synthesis, 4) phosphorylation and dephosphorylation of cytoskeletal proteins, 5) slow axonal transport of neurofilament proteins, and 6) inhibition of nucleotide activity (Meiri et al., 1993). Al(OH)3 and other insoluble species are reported to be the active aluminium forms in some pathological conditions (Zhang and Colombini, 1989).

Conditions of calcium or magnesium deficiency, which promote high aluminium uptake, are associated with neurological deterioration. ALS/Parkinsonism-dementia (primarily found on Guam) is an example which affects humans (Garruto, 1991), and grass tetany similarly affects sheep (Allen et al., 1984; Dennis, 1971). Without intervention, both conditions are usually progressive, resulting in death.

4. Alzheimer's Disease
Alzheimer's disease (AD) is distinguished from ordinary senile dementia by the following characteristics. Initially, the affected person has difficulties with word-finding, recent memory, and learning. As the disease progresses, it involves a relentless progression towards profound and utter stupefaction (Alzheimer, 1907). Larger than usual numbers of plaques and tangles form in the brain, particularly the cerebral cortex and the hippocampus.

Neurofibrillary tangles develop within pathologically-altered neurons. The basic subunit of these tangles consists of tau, a microtubule-associated protein. In AD, normal tau becomes replaced by abnormal tau which is hyperphosphorylated. Accumulation of abnormally phosphorylated tau precedes the formation of neurofibrillary tangles in AD (Bancher et al., 1989). Due to their hyperphosphorylation, the tangles are resistant to breakdown by proteases and are able to outlast the cells. The plaques are intercellular accumulations of a biological rubbish, largely consisting of an aberrant peptide called b-amyloid. The amyloid precursor protein (APP) is a stress protein which accumulates in nerve axons when the intra-axonal transport is impaired (Shigamatsu and McGeer, 1992). It may also accumulate under other conditions of damage or altered expression. The APP is cleaved and intercellular b-amyloid peptides are aggregated to form solid plaques. Once compact, the b-amyloid matrix becomes a sink for associated molecules such as apolipoprotein E, a1-chymotrypsin, heparin sulphate, proteoglycans, and ubiquitin (Shin et al., 1994).

AD also involves a profound loss of cells in some brain regions, notably in the hippocampus and cerebral cortex, and decreased activity of many neurotransmitters: acetylcholine, norepinephrine, serotonin, somatostatin, g-aminobutyric acid, and glutamate (Beal et al., 1989).

A l u m i n i u m


Several causative agents for AD have been proposed. At least one of these is genetic in nature. The amyloid hypothesis proposes that mutation leads to faulty genes and gene products which set up a cascade that results in AD. The DNA molecule which codes for b-amyloid was cloned in 1987 and its products have been intensively studied. The features of the cascade postulated to link amyloid with AD are largely unknown.

Identical twin studies in which one twin develops AD and the other does not, establish that its development is strongly influenced by an environmental factor which is not easily explained by genetics alone. Therefore, other researchers have looked to the environment for possible causative agents. Viruses have been proposed as a cause of AD. In the 1980s, many attempts were made to experimentally transmit AD from an affected brain to an unaffected brain. These attempts were all unsuccessful. Consequently, the virus hypothesis has few remaining proponents. Zinc is another environmental agent suggested as a cause of AD because it is important to memory formation (Constantinidis, 1991). Memory is impaired when zinc is deficient. Compared to Alzheimer incidence, zinc deficiency is rare. The numbers of people who have zinc deficiency and AD do not match.

5. Discussion of evidence against aluminium as a cause of Alzheimer's Disease
Aluminium has long been considered a candidate cause of AD and this association has been historically controversial. Before considering the evidence that supports this association, we will consider the evidence opposing aluminium as causal to AD.
* Eight population studies have found positive correlation between higher levels of aluminium in the drinking water and increased incidence of AD; two have not. Some of the eight positive studies have been criticised for methodological flaws. The two negative studies are also flawed. One of them (Wettstein et al, 1991) compared two concentrations of total aluminium that were both below 0.1 mg/L which, according to McLachlan (presentation at the current meeting) is the "no-effect level". In the second, alum treatment was introduced into the drinking water supply only three years preceding the study (Wood et al., 1988). All ten studies are flawed in that they are based on total aluminium levels rather than on waters having high versus low soluble (inorganic monomeric) aluminium levels.
* Wisniewski et al. (1984) have noted that neurofibrillary tangles produced by aluminium injection into rabbit brain differ from those that occur in humans with AD. This may in part be due to a species difference. Cultured human brain cells exposed to aluminium for several weeks, stain with an antibody which specifically recognises the phosphorylated tau protein of AD neurofibrillary tangles (Guy et al., 199 1).
* Wisniewski (1995) has also noted that the tau protein change which occurs in dialysis patients exposed to high levels of aluminium is somewhat different than the tau protein found in patients  with AD. On the other hand, another study has shown that the abnormal tau proteins in AD and ALS/Parkinsonism-dementia of Guam (which is also associated with high aluminium uptake) have the same structure even though they are distributed differently within the brain (Buee-Scherrer et
al., 1995).
* In a similar vein, some symptoms of the aluminium-associated conditions of dialysis dementia and ALS/Parkinson-dementia are known to differ from each other and also from AD. It is common for metal toxins to produce different signs and symptoms when the rate, amount, route, or other metal intake condition varies. For example, lead poisoning can produce an encephalopathy in children who eat lead paint chips and lowering of the intelligence quotient in others who are exposed to high levels of leaded petrol fumes.
* The report by Landsberg et al. (1992) has been widely quoted by scientists as having been unable to detect any aluminium in the senile plaques characteristic of AD (e.g. Doll,1993). Newspaper accounts and other publications have gone even further to claim that this work has eliminated

A l u m i n i u m


aluminium as a primary cause of AD (e.g. Alcoa Insight Newsletter, 1995). In actuality, Landsberg et al. gave evidence of having found aluminium in some unstained plaques although not in their cores. Other investigators have found aluminium in unstained plaque cores (Candy et al., 1989). 

The significance of the Landsberg et al. (1992) finding should be reconsidered in view of the large number of errors the paper contains and the fact that it shows inconsistencies with a companion paper produced by the same authors on the same subject at around the same time (Landsberg et al., 1993). This second paper states that the authors were unable to unambiguously identify and analyse plaques in their unstained tissue. It also reports having found aluminium in more than twice as many plaques and plaque cores of stained AD brain as in comparable samples described
in the other report. This is despite the fact that they rate the sensitivity of their nuclear microscope at 50 ppm in the latter publication and at 15 ppm, three times more sensitive, in the former one. No explanation was offered for these discrepancies.
* Population surveys have found no difference between numbers of AD patients and controls who have used antacids for several months. This is not surprising given that digestive disorders are very common in elderly people. These results have little if any scientific value because they are based on third party (hearsay) impressions of use instead of actual measurements over time. They even fail to distinguish between heavy usage and occasional usage.

6. Evidence for aluminium as a cause of Alzheimers's Disease
What is the evidence supporting aluminium as a cause of AD?
* Aluminium has been used to clarify water for centuries on a small scale, originally in a slow filtration method. In 1880, Frankfurt am Main, Germany, was one of the first cities to use a rapid water filtration machine. About 20 years later, a 51-year-old woman from Frankfurt developed the first known case of AD. When he reported it in 1907, Dr Alois Alzheimer wrote: "The case presented even in the clinic such a different picture, that it could not be categorised under known disease headings, and also anatomically it provided a result which departed from all previously known disease pathology." 

Describing the neurofibrillary tangles, he said,
"As these fibrils stain with dyes different from normal neurofibrils, a chemical change to the fibril substances must have taken place." Two others of the first 10 AD patients confirmed by 1911 also came from Frankfurt (Alzheimer, 1911). By 1925, a total of 33 confirmed cases of AD had been described in the medical literature and the following year a report in The Lancet described AD with the words, "This condition is rare" (James, 1926).
* Bauxite mining and large-scale aluminium production began around 1900 and the consumption increased greatly thereafter, especially during the world wars. Following the end of the Korean war, in the late 1950s, aluminium producers entered the domestic market producing drink cans and food packaging (Plunkert, 1993). These products, if heated, or in prolonged contact with foods and liquids, tend to corrode and are potential sources of bioavailable aluminium (Kandiah and Kies, 1994; Seruga et al., 1994; Liukkonen and Piepponen, 1992). Since 1980, there has been a sharp upturn and exponential increase in the incidence of AD (NCHS, 1986). In the USA alone, over the past 10 years, the AD incidence has increased from 1.2 million cases in 1984 to 4 million cases in 1994. According to Alzheimer's Disease International, there are now 15 million total AD cases worldwide. This upturn in AD incidence follows the expansion of aluminium into the food and drink market by approximately 15-20 years. The association may be coincidental; or perhaps not.
* Some evidence suggests that AD patients absorb more aluminium. Thus, according to Taylor et al (1992), AD patients under 77 years have higher aluminium blood levels, following an aluminium citrate drink than their age-matched controls.

A l u m i n i u m


* The significance of our finding that trace amounts of radioactive aluminium can enter the brain from the equivalent of a single glass of water proves that living brain tissue actively removes aluminium from the bloodstream. This 26Al data contradicts the suggestion that brain aluminium may merely be "a footprint" resulting from passive aluminium absorption into cells already compromised by disease or death.
* Aluminium concentrates in brain regions known to be affected in AD: the cerebral cortex, hippocampus, and the amygdala (Edwardson, 1992b). It specifically affects pyramidal cells and spares interneurons as in AD (Kowall et al, 1989).
* Aluminium has been located in the brain lesions of AD. It has consistently been found to associate with neurofibrillary tangles in the brain tissue of Alzheimer patients (e.g. Perl and Pendlebury, 1992). It has also been found in stained and unstained plaques. As mentioned above, some investigators have reported finding aluminium in the unstained cores of the plaques (Candy et al., 1989) while another group has not (Landsberg et al., 1992).
What is known about the participation of aluminium in plaque and tangle formation?
* Aluminium non-enzymatically phosphorylates human tau in vitro (Abdel-Ghany et al., 1993). It induces the aggregation of paired helical filament tau. Aluminium stabilises paired helical filament tau, increasing its resistance to proteolytic degradation. Injection of paired helical filament tau plus aluminium into rat brain induces stable co-deposits of b-amyloid, ubiquitin, apolipoprotein E, and a1-chymotrypsin (Shin et al, 1994).
* Aluminium interferes with neurofilament transport. Normally, neurofilaments are continually synthesised and transported along axons. Aluminium phosphorylates the neurofilament proteins and alters their ability to transport along the axon. This disruption causes amyloid precursor protein to accumulate within the axon and to distend it (Shigamatsu and McGeer, 1992).
Aluminium is also known to aggregate b-amyloid (Scott et al., 1993).
* Many dialysis patients are subjected to prolonged aluminium exposure. AD-like changes have been found in their processing of tau protein. In their grey matter, normal tau protein becomes depleted and abnormal tau increases in association with aluminium concentration (Harrington et al., 1994). Some of the younger dialysis patients have neurofibrillary tangles and about 30% of the patients have plaques (Edwardson et al, 1992a).
* Animal studies with aluminium have repeatedly shown AD-type behavioural changes. Prolonged aluminium exposure results in learning and memory deficits (Bilkei-Gorzo, 1993). These deficits have correlated with NFTs, large decreases in synaptic density, or impaired acetylcholine metabolism (Mieri, 1993).
* Finally, aluminium injection into the rabbit brain reduced the activity of five of the six
neurotransmitters known to decline in AD. These include acetylcholine, serotonin,
norepinephrine, g-aminobutyric acid and glutamate activities but somatostatin activity was unchanged. (Beal et al., 1989).

7. Conclusions
If the evidence for the proposed candidates is weighed, there is much more evidence for the aluminium hypothesis than for any of the other candidate causes (Table 1). This together with other known actions of aluminium, some which are discussed in the present paper, lead us to the following conclusions:
* Aluminium is a neurotoxin which, upon ingestion, can directly enter the brain. The long-term health consequences of this are currently unknown but must be regarded as a potential risk.* Historically, the availability of aluminium in drinking water geographically correlates with but precedes the first identified cases of AD by 20 years or more. * The recent rise in AD incidence lags an increased human exposure to ingestible forms of aluminium by a period of approximately 20 years.

A l u m i n i u m


* Bioavailable aluminium is the only agent which has been demonstrated to experimentally induce alterations in the molecular biology, pathology, and behavioural function which are consistent with the pathological features of Alzheimer's disease.
In view of the great human and social impact of Alzheimer's disease, and various other disease states with which aluminium is associated, bioavailable aluminium must be urgently considered as a risk factor to health.

Table 1. Alzheimer's Disease features and evidence that proposed causes produce those features

Evidence from proposed causes 
AD Features Gene/Protein Virus Zinc Al
Parallel increases in incidence/exposure no n/e n/e yes
Behavioural decrements:
* memory n/e n/e yes yes
* cognition n/e n/e yes yes
Principal affected areas:
* cerebral cortex yes n/e no yes
* hippocampus yes n/e yes yes
* amygdala yes n/e yes yes
Presence of agent in human AD NFTs no n/e no yes
* NFT formation:
* straight filaments no no no yes
* paired helical filaments no no no no*
* Phosphorylated tau epitopes no no no yes
Presence in:
* AD plaque generally yes n/e yes yes
* in plaque cores yes n/e n/e e
* Enhanced expression of amylod precursor protein yes n/e yes yes
* Aggregated amyloid into plaques n/e n/e yes yes
Reduced neurotransmitter activities:
* acetylcholine n/e n/e n/e yes
* serotonin n/e n/e n/e yes
* somatostatin n/e n/e n/e no
* norepinephrine n/e n/e n/e yes
* glutamate n/e n/e n/e yes
* g-aminobutyric acid n/e n/e n/e yes

n/e - no evidence; e- equivocal * early evidence of PHF formation in dialysis patients exposed to high aluminium levels (Harrington et al.,

8. Recommendations
On the basis of these conclusions, we make the following recommendations:
* An immediate review should be undertaken to identify all significant sources of bioavailable aluminium and to institute measures which will lower their levels in the diet.

A l u m i n i u m


* The 0.2 mg/L WHO guideline value for aluminium in drinking water, based on aesthetics, should be addressed for consumer health protection. The aesthetic level is too high and fails to distinguish between soluble, potentially toxic forms, and insoluble forms. A health-based guideline, utilising current available data, should be established for the level of soluble aluminium in drinking water. Current data indicate that the health guideline level for soluble aluminium in water supplies should preferably be 0. 002 mg/L but no more than 0.010 mg/L.
* Improve labelling requirements:
- Use the word "aluminium" rather than a code for aluminium food additives in view of its neurotoxin status.
- Identify aluminium additives at all stages of food production: e.g. baking powder, self-raising
flour, bread baked with aluminium-containing baking powder or flour, etc.
- Label aluminium cans and food containers with a non-encoded "use by" date which can be understood by consumers.

Alcoa (1995). Aluminium and drinking water. Alcoa Insight Newsletter, April, 1995, vol. 3 (Special Edition).
Allen, VG, Robinson, DL and Hembry, FG (1984). Effects of ingested aluminum sulfate on serum magnesium and the possible relationship to hypomagnesemic tetany. Nutr Rep Int 29:107 -113.
Abdel-Ghany, M, El-Sebae, AK, and Halloway, D (1993). Aluminum-induced nonenzymatic phospho-incorporation into human tau and other proteins. J Biol Chem 268:11976-11981.
Alzheimer, A (1907). Ueber eine eigenartige Erkrankung der Hirnrinde. Zentralblatt fur Nervenheilkunde und Psychiatrie 30:177-179.
Alzheimer, A (1911). Uber eigenartige Krankheitsfalle des spateren Alters. Zeitschrift fur die gesamte Neurologie und Psychiatrie 4:356-385.
Bancher, C, Brunner, C, Lassmann, H, Budka, H, Jellinger, K, Wiche, G, Seitelberger, F, Grundke-Iqbal, K and Wisniewski, HM (1989). Accumulation of abnormally phosphorylated tau precedes the formation of neurofibrillary tangles in Alzheimer's disease. Brain Res 477:90-99.
Beal, NE, Mazurek, NE, Ellison, DW, Kowall, NW, Solomon, PR, and Pendlebury, WW (1989). Neurochemical characteristics o aluminum-induced neurofibrillary degeneration in rats. Neuroscience 29:339-346.
Bilkei-Gorzo, A (1993). Neurotoxic effect of enteral aluminium, Food Chem Toxicol 31: 357-361.
Buee-Scherrer, V, Buee, L, Hof, PR, Leveugle, B, Gilles, C, Loerzel, AJ, Perl, DP and Delacourte, A
(1995). Neurofibrillary degeneration in amyotrophic lateral sclerosis/Parkinsonism-Dementia
Complex of Guam Immunochemical characterization of tau proteins. Am J Pathol 68: 924-932.
Burnatowska-Redin, MA, Mayor, GH and Lau, K (1985). Renal handling of aluminum in the rat: clearance and micropuncture studies. Amer J Physiol 249: F192-197.
Candy, JM, Oakley, AE, Mountfort, S et al. (1989). The distribution of aluminium in the forebrain of chronic renal dialysis patients in relation to Alzheimer's disease using SIMS. Secondary Ion Mass Spectrometry (SIMS) 4:323-326.
Constantinidis, J (1991). The hypothesis of zinc deficiency in the pathogenesis of neurofibrillary tangles. Med Hypothesis 35:319-323.
Day, JP, Barker, J, Evans, LJA, Perks, J, Seabright, PJ, Ackrill, P, Lilley, JS, Drumm, PV, Newton, GWA (1991). Aluminium absorption studied by 26Al tracer, The Lancet, 337:1345.
Dennis, EJ (1971). Magnesium deficiency and grass tetany. Is aluminium a key? Fret Sol 15: 44.
Doll, R (1993). Review: Alzheimer's disease and environmental aluminium. Age and Ageing 22: 138-153.

A l u m i n i u m


Edwardson, JA, Candy, JM, Ince, PG, McArthur, FK, Morris, CM, Oakley, AE, Taylor, GA and Bjertness (1992a). Aluminium accumulation, b-amyloid deposition and neurofibrillary changes in the central nervous system. Aluminium in Biology and Medicine. Wiley, Chichester (Ciba Foundation Symposium 169), pp. 165-185.
Edwardson, JA, Ferrier, IN, McArthur, FK, McKeith, IG, McLaughlin, I, Morris, CM, Mountfort, SA,
Oakley, AE, Taylor, GA, Ward, MK and Candy, JM (1992b), Alzheimer's disease and the aluminium hypothesis. Aluminum in Chemistry Biology and Medicine (Nicolini, M, Zatta, PF and Corain, B, eds.), N.Y.: Raven Press, pp. 85-96.
Fosmire, GJ, Focht SJ, and McCleam, GE (1993). Genetic influences on tissue deposition of aluminum in mice. Biol Trace Elem Res 37:115-121.
Gamto, RM (1991). Pacific paradigms of environmentally-induced neurological disorders: clinical, epidemiological and molecular perspectives. NeuroToxicology 12:347-377.
Greger, JL (1985). Aluminum content of the American diet. Food Technol 39:73-80.
Guy, SP, Jones, D, Mann DMA and Itzhaki, RF (1991). Human neuroblastoma cells treated with aluminium express an epitope associated with Alzheimer's disease neurofibrillary tangles. Neurosci Lett 121:166-168.
Harrington, CR, Wischik CM McArthur, FK, Taylor, GA, Edwardson, JA and Candy, JM (1994). Alzheimer's-diseaselike changes in tau protein processing: association with aluminium accumulation in brains of renal dialysis patients. The Lancet, 343: 993-997.
James, GWB (1926). The treatment of senile insanity, The Lancet, 820-821.
Jefferies, WA, Brandon, MR, Hunt, SV, Williams, AF, Gatter, KC and Mason, DY (1984).
Transferrin receptor on endothelium of brain capillaries. Science 312: 162-163.
Kandiah, J and Kies, C. (1994). Aluminum concentrations in tissues of rats: effect of soft drink packaging Biometals 7:57-60.
Klein, AM, Bumatowska-Hledin, MA, Kovan, J and Mayor, GH (1983). Reduced fecal aluminum excretion in rats following parathyroid hormone or bile duct ligation. Kidney Int. 23:215.
Kobayashi, K Yumoto, S, H Nagai, Hosoyama, Y, Imamura, M, Masuzawa, S, Koizumi, Y and Yamashita, H (1990). 26Al tracer experiment by accelerator mass spectrometry and its application to the studies for amyotrophic lateral sclerosis and Alzheimer's disease. Proc Japan Acad 66 Ser B: 189-192.
Kopeloff, LM, Barrera, SE and Kopeloff, N (1942). Recurrent convulsive seizures in animals produced by immunologic and chemical means. Am J Psychiatr 98:881-886.
Kopp, JF (1970). Occurrence of trace elements in water. Proc Third Annual Conf Trace Substances in Environmental Health, Columbia: Univ Mo, pp 59-73.
Kowall, NW, Pendlebury, WW, Kessler, JB, Perl, DP and Beal, MF (1989). Aluminum-induced neurofibrillary degeneration affects a subset of neurons in rabbit cerebral cortex, basal forebrain and upper brainstem. Neurosci. 29:329-337.
Landsberg, JP, McDonald, B, Grime, G, and Watt, F (1993). Microanalysis of senile plaques using nuclear microscopy. J Geriatr Psych Neurol 6:97-104.
Landsberg, JP, McDonald, B and Watt, F (1992). Absence of aluminium in neuritic plaque cores in Alzheimer's disease, Nature 360:65-68.
Liukkonen, LH and Piepponen, S (1992). Leaching of aluminium from aluminium dishes and packages. Food Addit Contam. 9:213-23.
Martin, BR (1986). The chemistry of aluminum as related to biology and medicine. Clin Chem. 32:1797-1806.

A l u m i n i u m


Meiri, H, Banin, E, Roll, M and Rousseau, A (1993). Toxic effects of aluminium on nerve cells and synaptic transmission. Neurobiol. 40:89-121.
Miller, RG, Kopfler, FC, Kelty, KC, Stober, JA and Ulmer, NS (1984). The occurrence of aluminum in drinking water, J Amer Water Works Assoc. 76:84-91.
National Center for Health Statistics(1986). TRANSAX: The NCHS system for producing multiple- cause-of-death statistics. Hyattsville, MD: National Center for Health Statistics.
Perl, DP and Pendlebury, WW (1992). Aluminum (Al) accumulation in neurofibrillary tangle (NFT) bearing neurons of senile dementia Alzheimer's type (SDAT)--Detection by intraneuronal X-ray spectrometry studies of unstained tissue sections. Proc Am Assoc Neuropathol, p. 349.
Plunkert, PA (1993). Aluminum. Mineral Prices in the US through to 1991. USBM, pp. 1-4.
Powell, JJ, Greenfield, SM, Parkes, HG, Nicholson, JK and Thompson, RPH (1993). Gastrointestinal availability of aluminium from tea. Fd Chem Toxic 31:449-454.
Scott, CW, Fieles, A, Sygowski, LA and Caputo, CB (1993). Aggregation of tau protein by aluminum. Brain Res. 628:77-84.
Seruga, M, Grgic, J, Mandic, M (1994). Aluminium content of soft drinks from aluminium cans. Z
Lebensm Unters Forsch 198:313-316.
Shi, B and Haug, A (1990). Aluminum uptake by neuroblastoma cells. J Neurochem 55:551-558.
Shigematsu, K and McGeer, PL (1992). Accumulation of amyloid precursor protein in damaged neuronal processes and microglia following intracerebral administration of aluminum salts. Brain Res. 593:117-123.
Shin, R-W, Lee, VM-Y, and Trojanowski, JQ (1994). Aluminum modifies the properties of Alzheimer's disease PHFt proteins in vivo and in vitro. J Neurosci 14 :7221-7233.
Shovlin MG, Yoo, RS, Crapper-McLachlan, DR, Cummings, E, Donohue, JM, Hallman, WK,
Khachaturian, Z, OrmeZavaleta, J, Teefy, S (1993). Aluminium in drinking water and Alzheimer's disease; a resource guide. AWWA Research Foundation and the American Water Works Association, Denver, CO, pp. 7-8.
Simonsen, L, Johnsen, H, Lund, SP, Matikainen, E, Midtgard, U, Wennberg, A (1994).
Methodological approach to the evaluation of neurotoxicity data and the classification of neurotoxic chemicals. Scand J Work Environ Health 20:1-12.
Spofforth, J (1921). Case of aluminium poisoning. The Lancet, 18 June 1921.
Taylor, GA, Ferrier, IN, McLoughlin, M, Fairbairn, AF, McKeith, AG, Lett, D and Edwardson, JA (1992). Gastrointestinal absorption of aluminium in Alzheimer's disease: response to aluminium citrate. Age and Ageing 21:81-90.
Tran, T, MacLennan, I and Persi, S (1993). Aluminium speciation analyses in water from Prospect Prototype Water Treatment Works. AWT, Science & Environment, Sydney Water Board report.
Van der Voet GB, Van Ginkel, MF and De Wolff, FA (1989). Intestinal absorption of aluminium in rats: stimulation by citric acid and inhibition by dinitrophenol. Tox Appl Pharmacol 99:90-97.
Walton, J (1973). Granules containing lead in isolated mitochondria. Nature 243: 100-101. Walton J, Hams, G and Wilcox, D (1994). Bioavailabilily of aluminium from drinking water: co- exposure with foods and beverages. Research Report No. 83, Melbourne: Urban Water Research Association of Australia.
Walton, J, Tuniz, C, Fink, D, Jacobsen, G, and Wilcox, D (1995). Uptake of trace amounts of aluminum into the brain from drinking water. Neurotoxicology 16: 187-190.
Weberg, R and Berstad, A (1986). Gastrointestinal absorption of aluminium from single doses of  aluminium containing antacids in man. Eur J Clin Invest 16:428-432.

A l u m i n i u m


Wettstein, A, Aeppli, J, Gautschi, K and Peters, M (1991). Failure to find a relationship between mnestic skills of octogenarians and aluminum in drinking water. Int Arch Occup Environ Health 63:97-103.
Wisniewski, NM (1994). Aluminium, tau protein, and Alzheimer's disease, The Lancet 344:204-205.
Wisniewski, HM, Merz, PA and lqball, K (1984). Ultrastructure of paired helical filaments of Alzheimer's neurofibrillary tangle. J Neuropathol Exp Neurol 43:643.
Wood, DJ, Cooper, C, Stevens, J and Edwardson, J (1988). Bone mass and dementia in hip fracture patientsfrom areas with different aluminium concentrations in water supplies. Age and Ageing 17:415-419.
Zhang, P, McCormick, M and Hughes, J (1994). Behaviour of aluminium during water treatment. Research report No. 85, Melbourne: Urban Water Research Association of Australia.
Zhang, DW and Colombini, M (1989). Inhibition of aluminium hydroxide of the voltage dependant closure of the mitochondrial channel, DAC. Biochim. biophys Acta 991:68-78.