Registration Dossier

Administrative data

Link to relevant study record(s)

Description of key information

Key value for chemical safety assessment

Additional information

There are not many studies available in which the toxicokinetic properties of aluminium metal were investigated. Thus, information available on supporting substances was taken into account by read-across following a structural analogue approach as well. Aluminium compounds of similar low solubility were considered for this approach, as only those would provide a relevant comparison for bioavailability of Al ions from a metal with low solubility. It needs to be considered that the particle size and the surface area influences the bioavailability of substances of low solubility.




Considering the target substances, several studies have investigated the oral bioavailability of aluminium hydroxide, the most recent being the oral gavage study of Priest (2010) in female Sprague-Dawley rats. The whole body fractional uptake calculated for aluminium hydroxide was 0.025% (±0.041%, sd). Priest (2010) is the only study identified that investigated the bioavailability of aluminium metal or aluminium oxide, the other poorly soluble aluminium substances that are the focus of this review. The results that were reported for aluminium oxide and pot electrolyte were similar to aluminium hydroxide (0.018 ± 0.038% and 0.042 ± 0.0036%, respectively). Aluminium uptake from aluminium metal powder was below the detection limit of the method and, following a repeat of the experiment, was reported as <0.015%. Priest (2010) also studied a suite of other aluminium substances, including the more soluble salts aluminium chloride and aluminium sulphate and the organic complex aluminium citrate. These substances showed fractional uptakes of 0.054% (0.015%, sd), 0.21% (0.079%, sd) and 0.079% (0.0057%, sd), respectively. Due to the use of the same experimental methods for the different substances, the results can be quantitatively compared. The tests substances were administered without co-exposures to ligands that may influence the bioavailability. The results of Priest (2010) contribute to the evidence (as described in Krewski et al., 2007 and ATSDR, 2008) supporting a lack of dependence of aluminium whole body fractional uptake on animal species and also the low bioavailability of the target substances.

The amount and location of deposition in the respiratory tract depend on respiratory tract architecture, breathing pattern, hygroscopicity of the material and the particle size distribution. Models are available to estimate deposition in the alveolar region in rats and humans for relevant particle size distributions and ventilation patterns. 

Lung clearance and retention depend on the particle size and shape, animal species, thein vivodissolution rate and any biochemical interaction between the dissolved moiety and lung proteins (Bailey et al., 1985 a, b). The only study identified that estimated rat lung retention half-times specifically for a poorly soluble alumina (high purity, calcined boehmites; AlO(OH)-40 [MMAD 0.6 μm]; AlO(OH)-10 [MMAD 1.7 μm]) was Pauluhn (2009a). For AlO(OH)-40 [MMAD 0.6 μm], the deposited alveolar fractions were 0.105, 0.108 and 0.103 at external exposure levels of 0.4, 3 and 28 mg/m³, respectively. For AlO(OH)-10 [MMAD 1.7 μm], the corresponding deposited fractions were 0.063, 0.067 and 0.065. The lung tissue elimination half-times (t1/2) for the smaller agglomerated particulate were 56, 43 and 144 days at 0.4, 3 and 28 mg/m³. For the larger agglomerated particle size exposure, the elimination half-times were 42, 60 and 295 days. An overload effect on alveolar clearance was evident at 28 mg/m³.

Based on differences between estimated alveolar deposition and the Al excreted in urine over the 24 hour period during and following an 8 hour workshift, Riihimaki et al. (2008) estimated that 1.2% of the Al deposited in the alveolar region of the lungs of a MIG welder/grinder exposed predominantly to aluminium oxide and aluminium metal was transferred to the systemic circulation. This calculation assumes that the urinary excretion during that time period represents 100% of the systemic uptake which is not an unreasonable assumption. Although crude, this estimate is in good agreement with the radiotracer-based result of 1.9% of the initial deposited dose of 26Al2O3alumina (MMAD = 1.2 μm) obtained by Priest (2004) and results from other studies in workers (Pierre et al., 1995; Sjögren et al., 1988). The results reported by Priest (2004), Priest et al. (1998) and McAughey et al. (1998) based on measurements in two human volunteers showed that 30 to 40% of deposited particles are removed by mucociliary clearance to the gastrointestinal tract, predominantly during an initial rapid removal phase.

The available information shows low rates of transfer of aluminium to the systemic circulation following inhalation exposure to aluminium oxide. McAughey et al. (1998) calculated a half-time for pulmonary clearance by dissolution processes i.e. transfer to the systemic circulation from the lungs, on the order of 3000 days (8.25 years). Increases in the levels of Al in the urine of workers (Riihimaki et al., 2008; Schaller et al., 2007; Pierre et al., 1995; Mussi et al., 1984), however, show that exposure by inhalation does lead to transfer to the systemic circulation most likely with a significant contribution from uptake in the gastrointestinal tract following mucociliary clearance.

No data were available concerning the deposition and transfer of aluminium metal dust/powder or aluminium hydroxide powder to the systemic circulation specifically. Detailed data are lacking on their in-vivo bioavailability and biochemical reactivity in the lungs. However, as the water solubilities and mean fractional uptakes in the gastrointestinal tract of aluminium metal and aluminium hydroxide powder are similar to or less than that of aluminium oxide (Priest, 2010), the results from McAughey et al. (1998) are sufficiently representative of the rate at which these substances are transferred to the systemic circulation following inhalation exposure of similarly sized-particulates.

Considering local effects, dermal penetration and subsequent binding of aluminium with skin proteins can occur on dermal exposure to ionic aluminium complexes in solution. The exposure situations of interest involve dermal exposure to particulate forms of the target compounds. The target compounds are sparingly soluble and are unlikely to be available for adsorption and penetration when in this physical form. If dissolved in, for example sweat, and in the form of an ionic Al complex, the target compounds will exhibit only shallow penetration due to binding in the upper layers of the stratum corneum. HERAG (2007) proposed default dermal absorption factors of 0.1% for metal cations from dry dust exposures.

Flarend et al. (2001) studied the uptake of aluminium from aluminium chlorohydrate-containing antiperspirant using 26Al as a tracer. The study was carried out using two human volunteer subjects, one male and one female.0.4 mL of 21% 26Al-ACH solution was applied to an area "3x4” in the left axilla of the two volunteers. Application was done using a pre-soaked (deionized water) cotton swab. The area was allowed to air dry afterwards. After the ACH had been applied and left to dry, the area was occluded with a bandage with adhesive edges that did not contact the area of ACH application. Each morning for the next 6 days strips of tape were applied to the axilla and then stripped away, the area gently washed with towelettes–and the bandage, tape strippings and towelettes sealed in freezer bags and stored in a refrigerator until analysis. The female subject developed a mild irritation to the bandage adhesive that required cessation of their use after 4 days. Blood samples were taken by venipuncture before ACH application (0 hours) and also at 6 and 14 hours post-application; then on days 1, 2, 3, 4, 5, 6, 7, 9, 11, 14, 18, 24, 32, 42 and 53 after application. Twenty-four hour urine samples were collected daily for the first 11 days after application; then from days 13 to 14, 17 to 18, 23 to 24, 31 to 32, 41 to 42, and 52 to 53.The samples were preserved using 10 - 20% (by volume) conc. HNO3. 26Al in the samples was determined by accelerator mass spectrometry. ICP-MS was used to measure Al levels in a subset of urine samples to ensure that the amount of Al in the urine would not influence the results from the AMS analyses. Based on the amounts of 26Al in the bandages, tapes and towelettes, 48% of the Al applied to the underarm of the male subject was recovered from the skin surface in 6 days; 31% was recovered in 4 days in the female subject. Levels of 26Al in the blood showed a clear increase after the application of ACH and 26Al could still be detected 15 days after application. Although 26Al could be detected in the blood, the levels were too low for reliable estimation of the % absorbed. Results showed that 0.0082% of the estimated absorbed 26Al was eliminated in the urine of the male subject and 0.016% in the urine of the female subject. A correction factor of 0.85 was applied (assumes 80 to 90% of absorbed aluminium is eliminated in urine over the period of 40 days (Priest et al. 1995; Hum Exp Toxicol 14: 287 - 293 cited) and a factor of 2 to account for two underarms. In conclusion, aluminium is absorbed into the systemic circulation on single occluded application of aluminium chlorohydrate to underarms. Based on urine measurements, 0.01% of the applied aluminium was absorbed showing that aluminium does not cross the dermal barrier effectively.

Speciation in biological fluids

Chen et al. (2010) studied the retention behavior of Al present as Al-citrate (Al-cit) and Al-transferrin (Al-Tf) complexes at pH 7.4 in serum from one healthy volunteer compared to serum from chronic hemodialysis patients (n=2) and synthetic human serum (n=2). Serum from a healthy participant was received from the Hospital of Wuhan University (Wuhan, China), and chronic hemodialysis patient serum samples were received from Wuhan Union Hospital (Wuhan, China). A new method using high performance liquid phase chromatography (HPLC)–ultraviolet visible (UV)/inductively coupled plasma mass spectrometry (ICP-MS) with C18 column dynamically coated with zwitterionic bile acid derivative, 3-[(3-cholamidopropyl)- dimethylammonio]-1-propanesulfonate (CHAPS) was described and it was used to study simultaneous quantification of both high molecular weight Al and low molecular weight Al species. The limits of detection (LODs, 3δ) for Al-Tf and Al-Cit were 0.74 and 0.83 ng/mL, respectively, and the relative standard deviations (RSDs) were 2.8% and 3.0% (c= 50 ng/mL, n = 7), respectively. All standard solutions and reagents were of analytical reagent grade. To avoid environmental Al contamination, all laboratory ware was made of polyethylene or Teflon and rinsed with 10% nitric acid. Chromatographic separation conditions (mobile phase pH, buffer concentration, column temperature and mobile phase flow rate, retention behavior) for the various Al species were optimized and controlled. In serum from the healthy volunteer, the concentrations of Al- Cit and Al-Tf were 2.2 and 27.9 ng/mL, respectively, with sum of Al-Cit and Al-Tf of 30.1ng/mL. In serum from hemodialysis patients, the concentrations of Al-Cit and Al-Tf ranged from 43.7–50.8 ng/mL and 288.7–291.6 ng/ml, respectively, and the sum of Al-Cit and Al-Tf was 317.0-333.5 ng/mL. The results show that in human serum, 87.4–91.1% Al was present as large-molecule Al-transferrin, while small molecule Al-complex compounds represented 7.1–15.2% of the total Al. Despite the observation that total serum Al levels in healthy person were approximately 10 times less than those in hemodialysis patients, the distribution of Al (Al-Tf and Al-Cit) in serum from healthy volunteer was identical to that in serum from the hemodialysis patients. The Chen et al. (2010) findings are supported by data from Harris et al. (1992; 2003) in that at physiological concentrations of Tf and Cit between 40 mol/L (Zhu et al., 1996) and 100 mol/L [Harris, 1992), about 93% of the total serum Al was bound to Tf (Harris, 2003). When serum Al-Tf levels were some 10 fold greater in hemodialysis patients, Al-Cit levels were almost 20 times higher in the same samples compared to serum from healthy volunteer. Limitations: very few subjects and samples were included that reduced the statistical power of the provided results and conclusions; the report lacked detail on participants characteristics (gender, age, severity and stage of disease), insufficient detail was given on sample collection, processing, holding times and description of the HPLC method and detection conditions. A Reliability Score 2 was assigned.

Wu et al. (2012) evaluated the effect of glutamate (Glu) and citrate (Cit) on the absorption and distribution of Al as AlCl3·6H2O, monosodium Cit (C3H4OH(COOH)2COONa), and monosodium Glu (C5H8NNa4·H2O) (purity > 99.99%) in rats under both in–vitro and in-vivo conditions. For the in-vitro experiment, 18 adult male Sprague–Dawley rats (average weight of 250±15 g) were randomly divided into three groups. The entire small intestine was dissected to obtain five to eight centimeter sections of the duodenum, jejunum, and ileum. These sections were cultured with 20 mmol AlCl3, 20 mmol AlCl3+20 mmol Cit, or 20 mmol AlCl3+20 mmol Glu and the luminal solution was obtained for Al determination. In a sub-chronic study, 24 adult male Sprague–Dawley rats (127 ± 10 g) were divided into four groups and fed the following diets: no supplemental Al or Glu added (control), AlCl3(1.2 mmol), AlCl3(1.2 mmol) + Cit (1.2 mmol), or AlCl3(1.2 mmol) + Glu (1.2 mmol) daily for 50 days. The laboratory stock diet contained Al at 987.9 mg/kg (no data provided on Al concentrations in drinking water). Feed intake and body weight were measured at the beginning and termination of the study. At the end of study, blood samples were obtained for biochemical analyses and Al was measured in the brain, kidney, liver, and bone using inductively coupled plasma mass spectrometry (ICP-MS) (no limit of detection was provided).

A significant increase of Al levels in the duodenum, jejunum, and ileum compared to the AlCl3group was observed following the administration of AlCl3+Cit and AlCl3+Glu based on wet weight (p<0.05). Significantly (p < 0.05) greater Al levels in the duodenum (based on dry weight) in AlCl3+Cit and AlCl3+Glu groups were noted. The Al absorption rate in isolated small intestine was ranked as: duodenum > jejunum > ileum. 

There were no changes in body weight gain in the AlCl3-fed animals compared to the control; however,food consumption was reduced significantly in all groups given AlCl3 compared to the control rats.The authors stated that the “growth rate” was decreased in the AlCl3 and AlCl3+Glu –fed rats, but data provided in Table 2 of the publication do not support this conclusion. There was statistically significant increase in Al levels in erythrocytes, kidney, and liver among the rats fed AlCl3but there was no such increase in brain or bone. Higher Al levels compared with the control rats were measured in the brain, kidney, bone, and liver in the rats fed AlCl3+Cit, but rats fed AlCl3+Glu accumulated the highest Al levels in erythrocytes, brain, and kidney. Compared to the rats fed AlCl3, administration of AlCl3+Glu led to a significant increase in Al content in red blood cells, brain, and kidney (p<0.01) while AlCl3+Cit –treated rats had a significant increase of Al levels in the kidney and bone (p<0.01). Compared with the AlCl3+Cit group, administration of AlCl3+Glu increased significantly (p<0.01) Al retention in red blood cells yet this was associated with reduced Al in the kidney and bone. Aluminum levels in the AlCl3+Glu group were the highest in the brain and the lowest in the liver; by way of comparison, rats fed the AlCl3+Cit accumulated the highest Al levels in kidney and bone (p<0.01). The authors concluded that the results indicated that Cit and Glu enhanced Al absorption in the intestine in-vitro, and feeding Al along with monosodium Glu increased Al concentrations in red blood cells, brain, and kidney. The study lacked concurrent control groups fed citrate alone and glutamate alone and the experimental design limits conclusions that can be drawn regarding the effect of combined sub-chronic citrate/glutamate exposure on Al bioavailability and distribution. Measures of Al in only erythrocytes may not account fully for gastrointestinal uptake of Al consumed as its chloride salt. Other limitations include: limited reporting of study design and results (acclimatization, adaptation and housing of animals, the actual and total Al doses consumed were not reported (high Al content was reported in the diet – about 1000 mg/kg diet yet there were no data on Al concentrations in drinking water); there was neither rationale for dose selection nor details available regarding diet preparation, added concentrations or stability of the test substances once mixed into the food); no explanation of measures to prevent laboratory Al contamination from external sources was provided. All of these factors limit the usefulness of the findings for hazard identification. Based on the overall study design, a Klimisch Score of 3 was assigned.

Rawy et al. (2012) reported the acute oral toxicity and associated kinetic parameters [including the maximum concentration (Cmax), maximum time to peak concentration (Tmax, days), the elimination rate constant (Lz, day 1), the elimination half-life time (t1/2,days), mean residence time (MRT, days) and clearance rate (Cl, L/day)] of Al in the liver, kidney, brain, intestine (µg/g wet wt.) and serum Al (µg/mL)] after a single oral (gavage) dose of Al chloride hexahydrate (AlCl3x 6H2O) in adult male albino rats (120 ±20g) throughout 1, 3, 7, 14 and 28 days (n=5 animals per group). The administered dose was 0.07 g Al/kg bw/day (1/50 of oral LD50 established by the authors at 3.5 g AlCl3/kg bw).The control rats received deionized water. Al content in the liver, kidney, brain, intestine and serum was determined at different study periods by using inductively coupling plasma mass spectrometry (limit of detection was not reported). The maximum Al concentrations (µg/g wet weight) following 28 days of post-exposure period were detected in intestine (2.59 ± 0.09), kidney (2.48 ± 0.06), brain (1.88 ± 0.06), liver (1.37 ± 0.09), and serum 2.18 ± 0.09 (µg/mL), respectively. The longest half-life time (T1/2, day) were observed in the brain (85.71 ± 17.40), liver (30.71 ± 7.19), kidney (28.59 ± 4.08), serum (19.08 ± 1.38) and intestine (16.54 ± 0.47), respectively. The slowest clearance rates (L/day) of Al were recorded in the brain (0.334 ± 0.01), kidney (0.723 ± 0.07), serum (1.121 ± 0.05), intestine (1.148 ± 0.03) and the liver (1.389 ± 0.18). The elimination rate constant was the lowest in the brain (0.01), then in the liver (0.03) and kidney (0.03) and the highest averages were detected in intestine (0.04) and serum (0.04). The maximum mean residence time (day) were recorded in the brain (124.22 ± 25.14) followed by the liver (45.47 ± 9.92), kidney (43.33 ± 5.93), serum (28.52 ± 2.05), intestine (25.14 ± 0.74). The authors reported that administration of a single oral (gavage) dose 70 mg Al/kg (administered as AlCl3) increased Al levels in blood and different organs of rats which persisted 28 days of post-exposure period. The highest levels of Al accumulated in the brain, followed by the liver, kidney and intestine. It should be noted that the overall results must be interpreted with caution due to the study limitations: no details were available regarding Al content in the diet and drinking water, no details were provided regarding the administered dosing solution (the vehicle used to prepare the AlCl3solutions, volume given, pH of the AlCl3solutions). No descriptions were given of measures taken to preclude Al contamination during the collection and handling of the tissue and blood samples. There are few miscalculations in the Al levels in the internal organs and blood (Table 2) which also limited a confidence in the reported results and provided conclusions. A Klimisch Score of 3 was assigned. The Rawy et al. (2012) results are of only limited utility for human health risk assessment.


Systemic Distribution & Metabolism

Aluminium Speciation in Blood & Aluminium Kinetics in Blood after Exposure

After a bolus oral gavage dose of aluminium citrate or aluminium hydroxide together with citrate (Priest et a., 1994), maximum serum concentrations occur during the first hour post-dosing.After intake, aluminium blood concentrations fall rapidly but renal excretion is maintained during this period. The evidence suggests that a large fraction of available aluminium is initially bound to low molecular weight species that can move out into tissue fluids. Aluminium in this form can move easily between tissue fluids and blood and remains available for excretion allowing maintenance of the early excretion levels. In contrast, in blood itself most aluminium is bound to transferrin which is retained in blood vessels. With time the tissue fluid pool is depleted and renal clearance rates decrease as it becomes more difficult first to excrete aluminium that is bound to higher molecular weight proteins and then aluminium that is bound to red blood cells.


Aluminium Levels and Accumulation in the Brain

In summary, a small fraction (on the order of 10-7– 10-8) of orally administered Al was found in the rat brain by Fink et al.(1994)and Jouhanneau et al. (1997). After intravenous infusion, Yokel et al. (2001b) found that the Al concentration in the brain of rats reached its peak (~0.005% of the administered dose per gram of brain tissue) on day 1, regardless of the form of the Al compound (Al transferrin or Al citrate). Yokel et al. (2001b) also showed that Al was retained in the brain for longer than 256 days after administration and that the Al chelator, desferrioxamine, modestly enhanced Al elimination from the brain. The terminal brain half-times were 150 days and 55 days, respectively, for rats that did not receive and received DFO.

Potential mechanisms of transport of Al [administered as the citrate] across the BBB have been examined in several in-vitro and in-vivo studies; these have shown that Al transport cannot be explained solely by diffusion and that the process is carrier-mediated (Allen et al., 1995; Ackley and Yokel, 1997; Yokel et al., 2002; Nagasawa et al., 2005). A monocarboxylate transporter (Ackley and Yokel, 1997; Yokel et al., 2002) and system Xc- (Nagasawa et al., 2005) has been suggested as a potential carrier of Al across the BBB. Al uptake by brain cells is likely to occur through a transferrin - transferrin receptor system (Roskams and Connor, 1990).

Based on the results of their experiments with intranasal administration of Al salts to New Zealand rabbits, Perl and Good (1987) suggested that Al may enter the central nervous system directly via nasal-olfactory pathways. The available evidence does not suggest that this is an important route of exposure for aluminium, however (Krewski et al., 2007; Perl and Good, 1987).


Aluminium Levels in Different Regions of the Brain

Healthy humans

Markesbery et al. (1984) analyzed trace element concentrations in various brain regions in 28 neurologically normal adults aged up to 85 years and also in seven infants. The measurements were conducted by instrumental neutron activation analysis. Brain Al concentrations increased with increasing age. The mean Al concentration was 0.467±0.033 µg/g wet weight in adults and 0.298±0.05 in infants. Based on information presented on the graph, mean Al concentrations in the brain were ~0.35 µg/g in the age group 20-39 years, around 0.43-0.47 µg/g in the age groups 40-59 and 60-79 years, and ~0.70 µg/g in those older than 80 years. Mean concentrations were highest in the globus pallidus (0.893 µg/g), putamen (0.663 µg/g) and middle temporal lobe (0.654 µg/g), and lowest in the superior parietal lobule (0.282 µg/g). Based on the review of available literature (Priest, 2004), normal human brain Al levels are within the range <1 to ~5 µg/g, with most results being closer to the lower end of this range.


Animal studies

Yuan et al. (2012) evaluated the distribution and concentrations of Al in neonatal rat brain following Al treatment. Groups of 14-16 postnatal day 3 (PND 3) rat pups (total n =46) received intraperitoneal injections of aluminum chloride (AlCl3) at 0 (control group), 7 (low dose group) or 35 mg Al/kg bw/day (high dose group) during 14 days. The control pups received injections of normal saline. The Al content was studied in the cerebral cortex, hippocampus, diencephalon, cerebellum, brain stem, pituitary, and olfactory bulb in the Al treated (n=8 and n=7 in the low and high dose groups, respectively) and control rats (n=8) using flameless atomic absorption spectrophotometer with a graphite furnace (the analytical limit of detection was not reported). After injections of AlCl3for 2 weeks, no significant differences were observed in body weight gain in both groups of the Al treated pups; however, cerebellar weight was significantly lower in the AlCl3-treated groups compared to the control (p<0.05) and the brain stem weights were significantly lower in the high dose group than in the control group (p<0.05). The weights of cerebral cortex, hippocampus, diencephalon, brain stem, pituitary gland and olfactory bulbs were not different from the control. The relative cerebellar weight (average cerebellar weight to body weight ratio) was significantly lower in the high dose group than in the control group (p < 0.05). The Al levels in rats given 7 mg Al/kg bw/day were significantly higher than the control - in the hippocampus (751.0 ± 225.8 versus 294.9 ± 180.8 ng/g; p < 0.05), diencephalon (79.6 ± 20.7 versus 20.4 ± 9.6 ng/g; p < 0.05), and cerebellum (144.8 ± 36.2 versus 83.1 ± 15.2 ng/g; p < 0.05). In summary, Yuan et al. (2012) found increased Al levels in the hippocampus, diencephalon, cerebellum, and brain stem of preweaning animals following repeated intraperitoneal injections of 7 and 35 mg Al/kg bw/day. The levels of Al in the diencephalon, hippocampus, cerebellum, and brain stem of the high dose group were increased 4.0, 2.5, 1.7, and 1.3-fold, respectively, compared to the control. The study has a number of limitations including failure to justify the administered doses, frequency of injections, no information was provided on the Al content in the laboratory stock diet and drinking water, preparation of dosing solutions, the volume and pH of the administered AlCl3solutions, feeding and housing conditions, availability of maternal care for neonates, brain samples collection, no measures to prevent environmental Al contamination of the samples were reported. No measurements of Al levels in blood and urine and/or neurobehavioral testing were conducted. A Klimisch Score of 3 was assigned. 


Dialysis encephalopathy

Elevated concentrations of Al were observed in the frontal cortex of 21 uraemic patients (McDermott et al., 1978). The concentrations were significantly higher in the 7 patients with dialysis encephalopathy (mean 15.9 ±10.5 µg/g dry weight) than in the 12 patients on dialysis but without encephalopathy (4.4±2.7 µg/g), and then in the 2 uraemic patients who were not dialyzed (2.7±1.4 µg/g). In patients with dialysis encephalopathy, the mean Al concentration in the grey matter was about 3 times higher than that in the white matter (20.6±16.4 µg/g and 6.9±5.3 µg/g, respectively). This difference was small in the other two groups of patients.

Alfrey et al. (1980) determined tissue Al concentrations (by atomic absorption spectrometry) in 38 dialyzed uraemic patients dying of dialysis encephalopathy, 57 dialyzed uraemic patients dying of other causes, 30 non-dialyzed uraemic patients and 36 control subjects who had had no known illness and died of external causes. Tissue Al levels were increased in all uraemic patients. Al levels in brain grey matter were 24.5±9.9 mg/kg dry tissue in patients who died of dialysis encephalopathy (significantly higher than in all the other groups), 8.5±3.5 mg/kg in dialyzed uraemic patients who died of other causes, 4.1±1.7 mg/kg in non-dialyzed uraemic patients and 2.4±1.3 mg/kg in the control subjects. There was no correlation between brain Al and duration of dialysis in patients with encephalopathy. For any duration of dialysis, patients dying of encephalopathy had higher brain Al levels than patients dying of other causes.


Aluminium Accumulation in Organs and Tissues from the Systemic Circulation

Priest (2004) reviewed data collected by the International Commission on Radiological Protection on aluminium content in organs and tissues (ICRP publication 23). Excluding the lung and lymph nodes (which can contain un-dissolved Al that is not part of the systemic Al body burden); only the liver, connective tissues, skin and skeleton concentrate more Al than the body average. Data for the skin might not be valid because of the likelihood of external contamination. Other data on Al tissue distribution in workers occupationally exposed to Al and members of the public (also summarized by Priest, 2004) are generally consistent with those reported by the ICRP. The skeletal system is the primary site of accumulation in humans. As discussed in the section on inhalation, aluminium also accumulates in the lungs due to inefficient transfer to the systemic circulation.


The injection studies by Guo et al. (2005 a, b) with aluminium chloride (intraperitoneal and subcutaneous, respectively) and Llobet et al. (1995) using intraperitoneal injection of aluminium nitrate show that systemic aluminium can be transferred to the testes when administered at high doses. The study by Ondreicka et al. (1966) shows that Al levels may increase in the testes on chronic oral exposure. Older rats may show increased accumulation relative to adult and young rats (Gomez et al., 1997). Further information on the toxicokinetics of aluminium in the testes is not currently available.

Urinary excretion is the main route by which systemic aluminium is eliminated from the body (Priest, 2004; Talbot et al., 1993, 1995). The filtration of aluminium from the blood into urine occurs in the kidney glomerulus and its efficiency shows a dependency on the chemical species to which the Al is complexed in the blood (Shirley and Cote, 2005).


Alemmari et al. (2012) studied Al in the liver, bile, serum and urine of newborn piglets following parenteral Al administration. Pigs were selected because of the similarity of their hepatic physiology to that of humans (Pond et al., 2001). The 3- to 6-day-old domestic pigs (20 piglets per group) received daily intravenous injections of 1.5 mg/kg per day aluminum chloride hexahydrate for 1-, 2-, 3-, or 4 weeks. The Al dose and route of exposure were chosen as being sufficient to induce hepatotoxicity (Klein et al., 1987). The control group (20 piglets per group) received daily intravenous injections of physiological saline. All piglets had free access to food and water. Aluminum detection in the serum, urine, liver and bile at each study period was conducted using inductively coupled plasma mass spectrometry. During the study period, serum Al levels increased 3.92-4.66 times that in the control, but the difference in serum Al between the Al- treated groups over time was not significantly different. Aluminum concentrations in urine of Al-treated piglets also increased over time (from 8.3 times after 1 week to 38.7 times at week 4 compared to the control group). Most notable was the observation that there was no correlation between Al levels in serum and Al levels in urine. A statistically significant increase in the mean Al concentration in the liver increased proportionally with increased duration of exposure; the highest Al content was observed in the 4-week group where the mean hepatic Al concentration was more than 3-fold that in the 1-week group. There was a direct correlation between the hepatic Al concentration and total bile acid in the serum (r = 0.67; p = 0.001). In addition, the urinary Al concentration was proportional to the hepatic Al content (r = 0.73; p =0.001), but there were no correlations between liver injury and serum Al concentrations. Compared to the saline controls, numerous electron-dense lysosomes accumulated Al were observed in the hepatocytes of the Al-injected pigs. No Al deposits were seen in the livers of control pigs. The authors concluded that repeated intravenous injection of high Al dose (1.5 mg/kg per day as aluminium chloride hexahydrate) caused hepatotoxicity in newborn piglets. No correlations were established between Al concentrations in serum and urine or between serum Al and hepatotoxicity. The relevance of the data obtained in pigs to human health risk assessment is not clear. A Klimisch Score of 3 was assigned.


Inhalation exposure

Buclez and Lafitte-Rigaud (2007) summarized and reported results from two large occupational investigations of atmospheric concentrations and kinetics of Al in different Al occupational settings that involved various Al forms and exposure conditions. The first study included bauxite mines, alumina refineries, aluminium fluoride plants, Al smelters (old unhooded and new technology smelters), Al power plants and Al welding.  A description of the results of this study, originally reported by Pierre et al. (1995), was provided in the previous RSI (2010) report on toxicokinetics.

A second study was conducted in a new technology smelter and a rolling mill. Exposure of workers at the smelter/potlines (196 participants and 175 from 640 total employees, respectively) and casthouse/rolling mill (375 participants and 341 participants from 1420 employees, respectively) to alumina, Al fluoride, cryolithe, electrolytic bath compounds, Al dust in casthouse and Al metal particulates was assessed. Atmospheric total and soluble Al concentrations and Al levels in blood, urine, and saliva were measured. Analyses of total Al in air were conducted by absorption atomic absorption spectrophotometry (AAAS). Collection of all biological samples was performed in a clean room. Bottles for blood, urine and saliva collection were pre-rinsed and carefully cleaned in nitric acid and distilled water. Analysis of biological samples was conducted by electrothermal atomic absorption spectrometry under strictly controlled conditions. The differences in Al levels for exposed and non-exposed workers and the relation between Al levels in blood, urine and saliva with atmospheric measurements were investigated. The main criteria for identification of biological indices of Al exposure included sufficient sensitivity of the endpoint and acceptable variability (reproducibility of the measurements in urine and saliva).

Atmospheric Al levels (mg Al/m3, total=115 measurements) in different departments at the smelter were (mean±SD): casthouse 0.12±0.122 (n=49 measurements); potroom0.35±0.29 (n=44); rodding shop 0.11±0.01 (n=2); crushed bath 0.41±0.37 (n=2); research laboratory 0.15±0.14 (n=8); central maintenance workshop 0.54±0.81 (n=8).Atmospheric Al levels (mg Al/m3, total=194 measurements) in different departments at the casthouse/rolling mill were (mean±SD): cold rolling 0.035±0.016 (n=13); hot rolling 0.087±0.081 (n=10); maintenance cold rolling 0.036±0.010 (n=6); maintenance hot rolling 0.089±0.063 (n=13); central maintenance workshop 0.074±0.071 (n=11); recycling 0.181±0.119 (n=25); finishing 0.035±0.014 (n=13); varnishing 0.044±0.024 (n=16); slab casthouse 0.335±0.735 (n=87); general services 0.021±0.015 (n=5). Unfortunately, it is not clear from the description whether the Al concentrations in workplace air were measured in the worker breathing zone (personal samples) or at fixed stations in the plants and refineries (area samples).

Urinary Al concentrations varied during different phases of the study and depended upon the activities in which employees were engaged; at the smelter/potlines the Al values (expressed as µg Al/g creatinine) were: potroom 21.2 (phase D1), 21.1 (phase D2), 17.9 (phase D3) and 16.2 (phase D4) (n=42 measurements ); casthouse 10.4 (phase D1), 8.9 (phase D2), 7.4 (phase D3) and 10.8 (phase D4) (n=59); central maintenance workshop 10.4 (phase D1), 9.6 (phase D2), 6.7 (phase D3) and 7.8 (phase D4) (n=18). Urinary Al levels during different phases of the study in different departments at the casthouse/rolling mill were (µg Al/g creatinine): casthouse 6.0 (phase D1), 5.7 (phase D2), 9.1 (phase D3) and 6.7 (phase D4) (n=69); rolling 4.2 (phase D1), 4.7 (phase D2), 7.8 (phase D3) and 7.5 (phase D4) (n=30);shearing6.9 (phase D1), 4.5 (phase D2), 4.6 (phase D3) and 5.5 (phase D4) (n=38). There was a significant increase in urinary Al (p=9x1011), blood Al (p=0.00003) and salivary Al levels (p=2.8x1010) among Al-exposed compared to non-exposed potroom workers. Statistically significant relationships between atmospheric Al measurements and urinary Al and salivary Al levels (p=0.0089 and p=0.0013, respectively) were observed; in contrast, there was no significant relationship between blood Al and atmospheric Al concentrations. The text was not clear whether blood samples refers to serum or plasma and if the numbers of urine samples refers to employees who gave urine samples for each area or phase or whether those values refers to total numbers of samples measured.

Sensitivity analysis revealed that 22 and 95 µg/g of urinary Al was associated with exposure to 1 mg/m³ total Al and 1 mg/m³ soluble Al, respectively, and that salivary Al at 87 µg/L was associated with exposure to 1 mg/m³ total Al in workplace air. However, reproducibility was sufficient only for urinary Al as those results were stable during all phases of the study (D1, D2 and D3, D4), the stability of salivary Al during the second phase of the study (D3, D4) was not consistent. There was a statistically significant difference in urinary Al levels (p=0.004) and salivary Al levels (p=2.8x0.00001) between casthouse workers exposed to Al versus workers who did not encounter airborne Al; however, no relationship was observed between Al atmospheric measurements and urinary and salivary Al. No significant difference in blood Al between exposed versus non-exposed casthouse workers was observed. No statistically significant differences in urinary Al, blood Al and/or salivary Al were found between Al exposed and non-exposed rolling mill workers, shearing workers, or maintenance workers, but a significant relation was observed between Al atmospheric measurements and urinary Al in shearing workers. The authors concluded that their findings suggest that only urinary Al, alone or associated with airborne Al measures, is of value in monitoring Al exposure to soluble Al (as occurs in smelting operations). In other activities such as casting and rolling, only atmospheric Al measures were of value and that Al concentrations in saliva were not conclusive due to large and rapid variations. The authorsconcluded thatthe results using saliva were not reliable as a biological indicator of exposure (BEI) due to the marked variations. As a point of information, all Al levels in workplace air were less than the current Threshold Limit Values. It should be noted that only a presentation is available for review. A preliminary Reliability Score of 4 (not assignable). 

Weinbruch et al. (2010) investigated the size, morphology, elemental composition and hydroscopic properties of potroom aerosol particles using high-resolution environmental scanning microscopy. The study was designed to investigate and to define a mechanism for transport of highly water soluble gases (HF, SO2) into the alveolar portion of the lower respiratory tract. Aerosol samples in the breathing zone of workers were obtained in the Søderberg and Prebake potrooms of a primary Al smelter in Norway. Particulates found in Søderberg process air were soot (60.4%), calcium sulfate (14.4%), silicates (12.0%), aluminium oxides/cryolite mixtures (7.2%); in Prebake potroom air the main particulates were comprised of aluminium oxides/cryolite mixtures (56.3%), soot (36.4%) and aluminum oxides (2.8%). At high relative humidity (comparable with that in the human respiratory tract), cryolite and related particles either underwent partial transformation leading to a water droplet with an insoluble core (cryolite and aluminum oxide/cryolite mixture) or formed thin water films on the surface (aluminum oxide). Pure Al oxide particles developed water films at lower relative humidity compared to the mixed aluminium oxide/cryolite particles. The authors suggested that the presence of small droplets or thin films of water on particles offered a mechanism for transport of highly water soluble gases (HF, SO2) into the alveolar region of the lung. The fraction of HF transferred to the liquid aerosol phase was estimated by a one-dimensional mass balance model and the lung was described by a Weibel tree model using morphological data for adult males assuming a respiratory rate of 30 L/min. It was estimated that during peak potroom air exposures, a significant fraction of gaseous HF (up to 10% during peak exposure) was transported into the lower respiratory tract via the aerosol. For SO2, this fraction appeared to be approximately 1%. Overall, these data indicated that HF may penetrate deeper into the lung in the presence of soluble particles or transferred on particles that form surface water films compared to vapor phase HF alone. The authors acknowledged that the significance of such a transport mechanism currently is not known. A Reliability Score is 2 was assigned for this mechanism of action study.


Abramson et al. (2010) conducted a cohort study at two Australian Al smelters to investigate relationships between occupational exposures to the airborne contaminants (fluorides, sulphur dioxide (SO2), coal tar pitch volatiles (as benzene soluble fraction, BSF), oil mist and inhalable dust) and changes in respiratory function associated with occupational asthma over time. A total of 446 new employees (77% of the 583 eligible workers) participated in the study from 1995 to 2003. The results indicated that long-term occupational SO2exposure was significantly associated with wheeze and chest tightness, bronchial hyper-responsiveness (BHR) to methacholine challenge, increased airflow limitations (reduced forced expiratory volume in 1 second/forced vital capacity ratio) and a longitudinal decline in pulmonary function. A dose-response relationship was also found between occupational exposure to airborne fluorides and the same outcomes, but the statistical association was reduced compared with that for SO2. These results might be taken to suggest that despite of the deeper penetration, long-term pulmonary irritation associated with HF tended to be more modest than SO2.


Dermal exposure

Human studies

There is uncertainty regarding long-term health consequences in relation to Al translocation and its distribution in the body (Guillard et al.,2012). Mitkus et al. (2011) estimated and compared the Al body burden in infants contributed by Al-containing vaccine adjuvants (assuming complete and instantaneous absorption) with the body burden of Al contributed by the diet (formula or breast milk) throughout the first 400 days of life. In summary, the authors reported that there is little risk for aluminum toxicity even with maximal exposures to Al adjuvant. The conclusions provided by the authors should be considered with caution because the differential kinetic behavior of Al following subcutaneous or intramuscular injections and dietary Al exposure can be anticipated. The United States Food and Drug Administration (2003) stated: “Generally, when medication and nutrition are administered orally, the gastrointestinal tract acts as an efficient barrier to the absorption of aluminum, and relatively little ingested aluminum actually reaches body tissues. However, parenterally administered drug products containing aluminum bypass the protective mechanism of the gastrointestinal tract and aluminum circulates and is deposited in human tissues.” Mitkus et al. (2011) suggested that while Al accumulates in brain over one’s lifetime, “the concentration of aluminum in brain is lower than that in many other tissues of the body (e.g., liver, spleen), and only 1% of whole-body aluminum is present in the brain or central nervous system at any given time.” However, “if the same is true for the infant brain, and whether such an accumulation is benign, remains to be determined” (Center for Trace Element Studies, 2011).


Pregnant and lactating women, children

Human studies

Mitkus et al (2011) updated estimates of the Al body burden contributed by dietary Al (formula or breast milk) throughout the first year of life compared with baseline Al levels at birth. The authors estimated a total Al level in blood at birth of 384 µg, or about 0.4 mg Al. According to the authors, the “background” Al body burden in infants should be considered ‘low’ because the “placenta partially protects the developing fetus from exposures from the mother during pregnancy” (Yokel, 1985; Cranmer et al., 1986). No empirical Al transplacental transfer data in either humans or laboratory animals were included to support that conclusion.

Kruger et al. (2011) measured Al in approximately 160 samples of human placenta (placenta bodies, placenta membranes, and umbilical cords) using electrothermal atomic absorption spectrometry (ETAAS). The placentas were collected between 1993 and 1998, and analyzed from 2007 to 2009. The method detection limit (MDL) (3s) was 0.25 µg/g. These samples were analyzed in duplicate, with two replicates per each injection. Placentas were collected shortly after delivery. A total of 167 women (age ranged from 15 to 38; mean = 24 years) participated in the study; 48% were African-American, 35% were Caucasian, and 13% were Hispanic. At least 43% of these women reported tobacco smoking at some point during pregnancy consumed at an average rate of 3 (± 5) cigarettes smoked per day. This population and qualifications for participation in the study were as described previously by Schell et al. (2003) as reported in Kruger et al. (2011). Placental analyses revealed geometric mean concentrations of approximately 0.5 µg Al/g in placenta bodies (0.56 µg/g, range 0.25–4.3 µg/g dry weight; n = 165) and membranes (0.53 µg/g, range <0.25–9.2 µg/g dry weight; n = 155). The mean Al concentration in the umbilical cord was 0.27 µg/g (range <0.25–1.8 µg/g dry weight, n = 154) and Al was detected in 95% of placenta bodies and 81% of placenta membranes, but in only 46% of the umbilical cords. Of those detected values, umbilical cords contained only about half the Al concentrations found in placental bodies and membranes. The Al concentrations detected in the cord and placenta body were significantly different (p < 0.001) and the Al concentrations found in the cord and membranes were also significantly different (p<0.001). However, no significant difference was observed between Al concentrations in the placenta body and membrane (p>0.05). These results indicate that placenta body and membrane tissue accumulate Al to a greater extent compared to the umbilical cord tissue. According to the authors, these differences might be explained by poor transport across the placenta, or, perhaps, by the absence of suitable binding sites in cord tissue. The Al levels found by Kruger et al. (2011) in the placenta were similar to the concentrations previously reported (mean 1.5 µg/g dry weight, range 0.68–5.44 µg/g) by Ward et al. (1987), who employed neutron activation analysis to measure placental Al in samples from the United Kingdom (cited by Kruger et al., 2011). Different analytical methods, different human populations (i.e., of different geographical, lifestyle and dietary characteristics), and different decades could explain the relatively small differences in results between these two studies. As a point of reference, the Al concentrations in placenta were comparable with Al levels reported in other soft tissues (Bush et al., 1995). The results suggest that, while the placenta may serve as a partial barrier to Al exposure duringin uterodevelopment, in approximately 50% of cases the fetus may still be exposed to Al from maternal blood. Unfortunately, no data were provided on Al levels in maternal blood and no descriptions were given of measures taken to preclude Al contamination during the collection and laboratory handling of the tissue. No information provided regarding the health of the participants, their dietary or occupational exposure to Al which limited the utility of reported findings. A Reliability Score of 3 was assigned to this study.

Mannelo et al. (2011) measured Al in milk from 45 healthy lactating mothers. Blood samples were collected from 15 of those subjects. Milk was obtained at each of three stages of lactation: colostrum (3–5 days after birth, n=14), intermediate milk (8–13 days after birth, n=19) and mature milk (3–6 weeks after birth, n=12). Milk samples were frozen, then centrifuged and after the top lipid layer was removed, the clear supernatants were analyzed. In healthy women, the mean Al content in milk was 24.8 ± 0.8 μg/L (range 11–36 μg/L). The Al level (mean± SD) in milk was higher than in serum (24.8 ± 0.8 vs 5.6 ± 0.5 μg/L, p < 0.001). Mean Al concentrations in milk at different stages of lactation were 23.4 ± 2.0 μg/L in colostrum (n = 14), 25.5 ± 1.2 μg/L in intermediate milk (n = 19), and 25.0 ± 1.0 μg/L in mature milk (n = 12); no significant differences in Al concentrations were found between the different stages of lactation. Limited social and demographic information was provided (no information was provided regarding health status of the mothers, possible exposure to Al through medications, diet and drinking water and no details of the mother’s occupational history was given). The fact, no information was provided regarding the selection process for breast milk donors; the limited number of participants reduced the statistical power of the results and conclusions. A Reliability Score of 3 was assigned for this study.



Animal studies

Bone is the main tissue where Al accumulates which determines in large the total Al body burden (Krewski et al., 2007). Three studies with the similar design were conducted in China to determine the effects of the prolonged oral Al exposure on Al levels in bone and its influence on bone formation.

Li et al. (2010) studied the effects of chronic aluminum (Al) exposure on Al levels in bone (femur) and (costal) cartilage in rats. Healthy Wistar male and female rats (4-week-old, 10 animals per group) were given drinking water (free access) containing Al chloride (400 mg/L, Al3+/L) (actual dose is not clear) during 150 days. The control group received distilled water. Ten rats from each group were sacrificed every 30 days (days 30, 60, 90, 120 and 150). All animals were kept under controlled environmental conditions and had free access to water and standard pellet feed (no data available on Al content in the diet and drinking water). The Al levels of serum, bone and cartilage were measured using a PEAA800 flame atomic absorption spectrophotometer (no level of detection provided). The body weights of Al-treated male and female rats were significantly lower than the control rats (p<0.01) from day 60. The serum levels of Al were significantly (p < 0.01) higher in the Al-treated male rats from days 30 to 150 compared to the control group with dose-effect relationship. The Al levels in the femur were significantly (p < 0.01) higher in the Al-treated male and female rats from days 30 to 150 compared to the control groups. The concentrations of Al in costal cartilage were significantly higher in the Al-treated male and female rats from days 30 to 150 compared to the control. The results indicate that prolonged oral exposure to Al (administered as Al chloride) at 400 mg/L led to Al accumulation in serum, bone and cartilage. There were no significant differences between Al levels in bone and cartilage in male or female rats given the same Al concentration and there were no statistically significant differences in serum Al which indicates that both sexes responded in a similar manner. No Al accumulations in serum, bone and cartilage of the control rats were observed. The major limitations of this study include: failure to report Al levels in the diet, the pH of the drinking solution was not reported, no data were available regarding drinking water consumption, no description of precaution measures to avoid sample contamination by Al from environmental sources were provided, high serum Al levels in the control group indicate that external contamination may have occurred or the background serum Al levels were high. It is not clear if the bone and cartilage Al concentrations were expressed on a wet or a dry sample basis which limits comparisons with data provided by other researchers. A Klimisch Score of 3 was assigned to this study.

Li et al. (2011a) examined the effects of oral exposure to the acidic Al-containing drinking solution on bone Al, Ca and P content during bone formation in young rats. Healthy male Wistar rats (4-week-old, 74–96 g, 10 animals per group) were given drinking water containing 100 mg/L AlCl3·6H2O (99%purity) for 150 days. The pH of the AlCl3solution was adjusted to 5.6. The control group was given distilled water (pH 7.0) under the same experimental conditions. All rats had free access to water and standard pellet diet (no details regarding Al content in the diet and drinking water were reported). The Al content in femur was measured using a PEAA800 flame atomic absorption spectrophotometer (limit of detection was not reported). The body weights of the Al-treated rats were significantly lower than the control from day 60 (p<0.05) and on days 90–150 (p < 0.01). The serum pH of Al-treated rats was significantly lower at the end of treatment, on day 150 (p < 0.05). The mean Al content in femurs increased in the Al-treated rats and was significantly higher (p < 0.01) than in the control group from day 30. The concentrations of bone Ca and P were significantly lower in the Al-treated group than in the control group on day 150 (both p < 0.05). It should be noted that the bone Al levels in the control rats were almost identical to the Al levels reported for the control rats by Li et al. (2010). The authors concluded that the results suggest that long-term oral exposure to Al (administered as Al chloride, 100 mg/L) with drinking water (pH 5.6) led to Al accumulation in bones, inhibition of bone formation and bone loss in young rats. The major limitations are similar to the limitations previously described for Li et al. (2011a) study; however, measures to prevent environmental Al contamination of the samples were reported.


Li et al. (2011b) investigated the effects of the prolonged consumption of Al in drinking water on the content of Ca, P, and Mg during bone formation in young rats. Healthy male Wistar rats (4-week-old, 74–96 g., 10 animals per group) were given drinking water containing AlCl3(430 mg/L, Al3+/L) for 150 days (the actual dose was not provided). The control group was given distilled water under the same experimental conditions. All rats had free access to water (0 mg Al/L) and standard pellet feed (2.87±0.39 mg Al/kg). Experiments were carried out in accordance with the guiding principles in the use of animals in toxicology adopted by the Chinese Society of Toxicology. The Al contents in right femurs were measured using a PEAA800 flame atomic absorption spectrophotometer (limit of detection was not reported). The body weights of Al-treated male rats were significantly lower than the control rats (p<0.01) from day 60. The mean Al content in bone increased in the Al-treated rats and it was significantly (p < 0.01) higher than in the control group on day 30, 60, 90, 120 and 150. Bone Ca, Mg, and P levels increased in both control and Al-treated groups in parallel with their growth, but the Ca, Mg, and P levels were significantly lower in the Al-treated group compared to the control group on days 120–150 (p<0.05). The similar results were reported Li et al. (2011b). The concentrations of Zn, Fe, Cu, Mn, Se, B, and Sr in bone in the Al-treated group were also significantly lower compared to the control group from day 60. The authors suggested that prolonged administration of a high concentration of AlCl3in drinking water can increase Al levels in bone and inhibit bone mineralization through disruption of trace elements metabolic pathways. The authors reported measures implemented to prevent Al laboratory contamination of the biological samples. The major limitations are similar to the limitations previously described for Li et al. (2011) study. A Klimisch Score of 3 was assigned to this study.

It should be noted that in spite of the fact that rats in the Li et al. (2010; 2011a, b) studies received different concentrations of AlCl3(100, 400 or 430 mg/L) over 5 months, Al accumulation in bone was very similar at all corresponding time points; for example, on day 150 the Al levels in femurs were 603, 508 and 491% compared to the controls, respectively, with the highest accumulation observed at the lowest administered dose, this sheds some doubt on a dose related response.

Hirayama et al. (2011) studied the influence of age on the background concentrations of Al and 28 other elements in the rat femur. Female Wistar rats (4 weeks old, n=5 per group) were reared from 4 weeks to 113 weeks on the CE-2 (CLEA JAPAN) laboratory stock diet and tap water. Elemental Al concentrations in the diet and in tap water were 60 and 0.038 µg/L, respectively. Five rats were sacrificed at 5, 9, 13, 17, 21, 25, 33, 42, 50, 59, 68, 77, 85, 95, 105, and 113 weeks of age (4 rats at 113 weeks). Bone Al was determined by inductively coupled plasma mass spectrometry (ICP-MS) with an analytical limit of detection at 5 µg/L. The concentrations of Al were expressed on a wet sample basis; the mean Al concentration in bone (n= 5) at 17 weeks was 0.31±0.14 µg/L and for all ages (n=79) the mean was 0.53 µg/L. Al, together with Fe, Cd, Ba, W, and Tl showed no significant time variation (p > 0.1, ANOVA), but Cu, Zn and Mo showed increasing trends with age. Uptake rates of Al into bone varied between individual rats and the values of Al, V, Ni, Ag, Cd, Sb, W, Tl, and U in age-matched rats were distributed across a broad range. The concentrations of Na, Mg, P, K, Ca, Mn, Cu, Zn, Se, Rb and Sr in the rats with the same age were confined to narrow ranges. A Klimisch Score of 2 was assigned to the Hirayama et al. (2011) report.

Zhu et al. (2012) investigated the influence of chronic Al consumption on Al accumulation in rat spleen. Healthy male Wistar rats (5 weeks old, 110-120 g, 10 animals per group) were given AlCl3 at doses of the 0, 64.18 (Low Dose, LD), 128.36 (Mid Dose, MD) and 256.72 (High Dose) mg/kg bw in drinking water for 120 days. Rats had unlimited access to water and standard feed. The Al concentration in spleen was measured using graphite furnace atomic absorption spectrophotometry, and concentrations of Fe, Cu and Zn were determined using flame atomic absorption spectrophotometry. No deaths were observed in the Al treated rats during the experimental period. The body weights of all AlCl3-treated rats were significantly lower than in the control rats. Concentrations of Al and Cu in the spleen of rats increased with the increased doses of AlCl3. The concentrations of Al in the low, mid, and high dose groups were significantly higher than in the control rats (p < 0.05, p<0.01 and p<0.01, respectively). The concentrations of Cu in the mid dose (P < 0.05) and high dose groups (p < 0.01) were significantly higher than those in controls. Fe and Zn concentrations in the spleen decreased with the increased doses of AlCl3. Fe concentrations in the mid dose and high dose groups were significantly lower compare to the control groups (p < 0.01). The concentrations of Zn in the mid dose (p < 0.05) and high dose (p < 0.01) groups were significant lower compared to the control groups. The authors suggested that prolonged administration of a high dose of AlCl3in drinking water can increase Al concentrations in spleen and disturb the balance of trace elements in the spleen. The major limitations of this study include: no data were available on Al levels in the diet and drinking water, the pH of the drinking solution and drinking water consumption were not reported, there were no measures of Al in serum to determine absorbed dose, no description of precaution measures to avoid sample contamination by Al from environmental sources was provided. A Klimisch Score of 3 was assigned to this study.


The objective of the study from Atomic Energy Canada (2010) was to measure the fraction of aluminium that enters the bloodstream of the rat following the ingestion of aluminium citrate, aluminium chloride, aluminium nitrate; aluminium sulphate, aluminium hydroxide, finely divided aluminium metal, powdered pot electrolyte, FD&C Red 40 aluminium lake, SALP, Kasal, sodium aluminium silicate.The test materials were prepared using 26Al as a radioactive tracer.Aluminium citrate, aluminium chloride, aluminium nitrate; aluminium sulphate were used as aqueous solutions. Aluminium hydroxide, aluminium oxide, SALP, Kasal, and sodium aluminium silicate were suspended in water with added 1% carboxymethylcellulose (to maintain a suspension). The solutions and suspensions were administered through feeding tubes.The particle sizes of FD&C red 40 aluminium lake, powdered pot electrolyte and aluminium metalwere too large to pass through feeding tubes; they were mixed with honey and added to the back of the rat tongue. An initial experiment was conducted to measure the fraction of bloodstream aluminium that is retained by the rats by day 7 post-injection.Twelve rats were injected intravenously with 0.5 ml of aluminium citrate solution containing 0.19 ng of 26Al.Six control animals received citrate injections containing no 26Al. The animals were sacrificed on day 7 post-injection. To address issues related to possible contamination of samples by external radionuclide from urine and faeces, in six rats the retained aluminium fraction was determined in short carcasses excluding tissues potentially contaminated by urine and faeces (the pelt, gastrointestinal tract, paws, feet and heads). In the other six rats, the retained aluminium fraction was determined in full carcasses (except pelts).The fraction of 26Al uptake excluded by the analysis of the reduced samples was determined by comparing the results for short carcasses with the results for full carcasses.The resulting correction factor was then used in the main study (ingestion) to determine Al content in the full carcass from the Al content in the short carcass.

In the main (ingestion) study each compound was administered to 6 rats. Six control animals received water. Seven days after the administration, the rats were sacrificed, their short carcasses were ashed in a muffle furnace, and a white ash was sent for analysis to.At the university, a known amount of stable isotope 27Al was added to each sample, the samples were dissolved in acid, and aluminium was extracted by precipitation. The 26Al:27Al ratio was determined by accelerator mass spectrometry (AMS).The amount of 26Al in each sample was calculated and corrected to account for the amount discarded with the unanalyzed tissues.The fraction of 26Al absorbed was calculated by reference to the 26Al administered and the 26Al fraction retained at 7 days post-injection (determined in the initial experiment).

The highest fractional uptake of 26Al (~0.21%) was seen for aluminium sulphate and the lowest (~0.02%) for aluminium oxide with 10-fold difference between the two values.The insoluble compounds (hydroxide, oxide and powdered pot electrolyte) administered as suspensions were less bioavailable than soluble compounds.The results for D&C Red 40 aluminium lake and for sodium aluminium silicate were closer to the results for soluble salts, which the authors explain by possible release of 26Al from particulates by partial dissolution in the gastrointestinal tract. The bioavailability of Al metal, SALP and Kasal could not be determined because the amount of 26Al present in the samples was not sufficient to determine the 26Al/27Al ratio.A reanalysis is being conducted. The authors suggest that the bioavailability of aluminium metal particles may be considerably lower than that of soluble aluminium compounds. There is a figure given for aluminium metal bioavailability as a "lower than" number.

The authors compared the results of these analyses with the results of human volunteer studies using 26Al-labelled compounds and found that the results were consistent. It was concluded that the compounds tested “present no unique biological hazard as a consequence of their bioavailability” and that the rat is a suitable experimental model for studying metal bioavailability relevant to humans. The study was conducted according to the principles of Good Laboratory Practice.The protocol was subjected to internal and external peer review. A Klimisch score 1 was assigned to this study.


Talbot et al. (1993, 1995) observed an average of 59 (sd, 10) % of 26Al excreted in urine during the first 24 hours after intravenous injection and 72 (sd, 7) % - by day 5 in a study with six healthy, male volunteers. A marked inter-individual variability in the pattern of early urinary excretion of 26Al was observed. The cumulative fraction of 26Al excreted by day 5 was in the range from 62% to 84%. The mean renal clearance rate was 16 litres of whole blood per day (sd, 10) and the range 5 to 33 suggesting a change in the binding of Al with increasing time in the blood. Based on the series of 26Al radiotracer experiments conducted in human volunteers at the Harwell Laboratory, UK, Priest (1997) reported that 85 to 90% of aluminium is excreted in urine during the first day after IV exposure.


Results from several studies support an enhancement of excretion when aluminium is introduced into the blood already complexed with citrate (Lote et al., 1992, 1995; Maitani et al., 1994). Studies in humans that have investigated whether co-administration of silicate with aluminium enhances the efficiency of excretion have produced conflicting results, however (Jugdaosingh et al. 2000, King et al., 1997; Bellia et al., 1994, 1996; Birchall et al., 1996).


A dose-dependence of efficiency of initial urinary excretion has been observed that may result from the formation of high molecular weight aggregates at high blood aluminium concentrations (Xu et al., 1991; Yokel and McNamara, 1988). Studies that have been conducted on the rate of elimination of aluminium in animals and humans have shown half-times that increase with the duration of follow-up indicating the presence of several compartments.The longer elimination half-times, reflecting slower elimination from compartments other than blood, do not show evidence of dose-dependence.


Greger and Radzanowski (1995) observed differences in the accumulation and kinetics of aluminium administered by oral gavage to Sprague-Dawley rats of different ages in a 44 day experiment. The growing rats accumulated more aluminium in their tibias on day 1 of the experiment and also retained more aluminium over the 44-day experimental period. Half-times in the bone of the ageing rats were several times greater than in the growing and mature rats (173 days versus 38 and 58 days in the growing and mature rats, respectively). Iron status may also affect the distribution and retention of aluminium. Greger et al. (1994) observed more aluminium in the livers but less aluminium in the tibias and spleens of male anaemic rats compared with male normal rats but these differences were not significant at all time-points.


As filtration in the kidneys is the main route by which systemic aluminium is excreted, individuals with renal insufficiency will retain more aluminium. The immature renal systems of neonates, particularly those who are pre-term, put them at increased risk of accumulating higher levels of aluminium (Krewksi et al., 2007). 

Genius et al. (2011) measured the concentrations of the 18 trace elements including Al in serum, urine and sweat to assess the rate and extent of inorganic elements elimination through excessive sweating. Nine men and 11 women (mean 44.5 ± 14.4 and 45.6 ± 10.3 years of age, respectively) participated in the trial. Blood, a first-morning urine sample and sweat were collected from 10 healthy participants (6 men from 21 to 57 years of age and 4 women from 43 to 50 years of age) and 10 participants (3 men from 38 to 61 years of age and 7 women from 25 to 68 years of age) with various diseases (diabetes, obesity, hypertension; rheumatoid arthritis; addiction disorder; bipolar disorder; lymphoma; fibromyalgia; depression; chronic fatigue; diabetes, fatigue, obesity; chronic pain, cognitive decline). Venous blood samples were obtained using vacutainer blood-collection sets and participants collected first-morning urine samples on the same day that blood samples were collected. Sweat was collected from pre-washed skin at any site on the body within 1 week before or after blood collection. No specific guidance was given regarding the duration of sweating or regarding the type or location of exercise. Ten participants collected sweat during an infrared sauna session; seven collected sweat during a regular steam sauna and three collected sweat during and immediately after exercises. Samples were analyzed by inductively coupled plasma sector field mass spectrometry using synthetic blanks, method blanks, control samples, and standards matching the sample solutions. The analytical limit of Al detection in serum, urine and sweat was 2 µg/L. Performance of the analytic procedure was controlled by analysis of reference materials (International Atomic Energy Agency, bovine blood) and control samples (trace elements in serum and urine from Sero AS, Norway). The median Al concentration in serum (n=20 samples) was 185.0 µg/L (0.19 mg/L), the median Al concentration in urine (n=20 samples) was 899.50 µg/L (0.9 mg/L), and the median Al concentration in sweat (n=20 samples) was 4650.0 µg/L (4.6 mg/L). The median serum Al in participants who suffered from various medical conditions was 161 µg/L and that in healthy controls 226 µg/L. The median Al concentration ratio in urine/serum was 3.9 (n=20) and that in sweat/serum was 21.6 (n=20). Comparison of median excretion efficiency of Al through sweat and urine found that sweat was a more efficient Al elimination pathway (excretion efficiency = 5.3).

The Genius et al. (2011) study is the first investigation that measured Al and trace elements simultaneously in three body fluids in the same group of participants, thus enabling comparisons between urine and sweat excretion. It was evident that for some inorganic elements including Al, excretion via sweat (induced under aggravated conditions) was more efficient than via urine. The study limitations include: the high serum and urine Al concentrations reported for healthy individuals, no information was provided if vacutainer blood-collection sets were validated for trace elements testing as glass tubes are coated with silicone and micronized silica particles to accelerate clotting, the study design precluded assessment of Al elimination via feces, milk, saliva or other bodily fluids. Questions remain whether the fluid released on sweating may represent a combination of perspiration, sebum and matter released directly from skin and not exclusively from sweat and the possibility that sweat originating from different parts of the body may be comprised of different substances at different concentrations. Moreover, Al excretion rates may also vary with sweating duration and frequency which were not considered in the study. Despite the authors’ stated that during all stages of sample preparation and analyses precautions were taken to avoid sample contamination by Al from environmental sources, the possibility of inadvertent Al laboratory contamination of blood, sweat and urine samples remains. Serum creatinine was not measured at the time of sweat collection. It was not clearly reported how the first-morning urine was collected and the samples handled. In addition, limited details were available regarding socio-demographic characteristics, selection of the participants of the study including inclusion/exclusion criteria, lack of information regarding dietary, pharmaceutical and occupational or environmental exposures to Al. The relatively small (20) number of participants decreased statistical power and the utility of the reported findings and conclusions. A Reliability Score of 3 was assigned.


Newton and Talbot (2012) conducted a reanalysis of data reported in their earlier publication (Priest et al., 1995/Priest, N.D., Newton, D., Talbot, R.J. and Warner, A.J. (1995). The original report described the disposition of Al as radioactive aluminium-26 (26Al) in volunteer who had received intravenous injection of 26Al-citrate. A detailed account of the administration procedures was given previously (Priest et al., 1995). Briefly, 0.7 mg of 26Al, carrier free as Al citrate, was injected into an antecubital vein of a healthy Caucasian male (height 1.83 m, weight 77 kg and age 41 years) who had given informed consent. The study was approved by AEA Technology’s independently constituted ethics committee. The elimination rates of aluminum were reported on the basis of the time-profiles of aluminum in blood, feces and urine up to 13 days and follow up to 2.5 years on the whole-body retention of the tracer and on its concentration in blood. In the later stages of the study, the residuals approached the limit of detection which complicated reliable quantification of the remaining body deposits; as a result, details of the extended investigation were not published. According to the re-analysis by Newton et al. (2012), only 6% of the daily loss at 3350 days could be explained by urinary and faecal excretion. Recently, Genius et al. (2011) found median Al concentrations of 4.6 mg/L in sweat and 0.9 mg/L in urine for 20 volunteers. Based on the revision of the available data and supplemented by new data on Al kinetics (Genius et al., 2011), Newton et al. (2012) suggested that additional mechanisms (transdermal losses) may contribute to the clearance of systemic Al at longer time intervals and mitigate its long-term accumulation from chronic exposure. According to Newton et al. (2012), losses via sweat might provide a credible explanation of data reported by the ICRP (1975) that the estimated balance was 10 times greater for daily perspired Al (1 mg) than that amount found in urine (0.1 mg), although the estimate of Al elimination in body sweat was extrapolated from observations made under unusual conditions (Consolazio et al., 1964 cited by Newton and Talbot, 2012). However, it should be noted that because the results of the original study (Priest et al., 1995) and Newton and Talbot (2012) study, respectively, are based on only one participant, the overall statistical significance of the exposure-effect associations is rather weak, and the overall results are only suggestive. For purposes of the present assessment, the results have limited utility for human health risk assessment within the scope of REACH. 



At present, there is no PBPK model that incorporates aluminium inputs from both the inhalation and oral routes. Data available at the time of development of these models did not sufficiently characterise rates of transfer to, from and within the brain, transfer to the foetus in utero, or transfer to and distribution in the suckling infant. Uncertainties in the speciation of Al in the blood and its effect on transfer to different tissues and renal elimination currently limit the models.