Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

The new ECHA CHEM database has been released by ECHA, and it now contains all REACH registration data. There are more details on the transition of ECHA's published data to ECHA CHEM here.

Diss Factsheets

Administrative data

Description of key information

Based on read-across

Oral: NOAEL (chronic, rat) 30 mg Al/kg bw/day as aluminium citrate.

Inhalation: LOAEC (subchronic, rat) 70 mg Al/m³ as aluminium oxide.

Key value for chemical safety assessment

Repeated dose toxicity: via oral route - systemic effects

Endpoint conclusion
Dose descriptor:
30 mg/kg bw/day
Study duration:

Repeated dose toxicity: inhalation - systemic effects

Endpoint conclusion
Dose descriptor:
70 mg/m³
Study duration:

Additional information

There are no studies available on the repeated dose toxicity of aluminium hydroxide by the oral, inhalation and dermal route.

In terms of hazard assessment of toxic effects, available data on the repeated dose toxicity of other aluminium compounds was taken into account by read-across following a structural analogue approach, since the pathways leading to toxic outcomes are likely to be dominated by the chemistry and biochemistry of the aluminium ion (Al3+) (Krewski et al., 2007).



A GLP study was performed using aluminium chloride basic in accordance with OECD Test Guideline (TG) 422 "Combined Repeated Dose and Reproductive/Developmental Screening Test" (Beekhuijzen 2007). No mortality or clinical signs of intoxication were observed in male and female Wistar rats due to treatment with Al chloride basic at dose levels of 40, 200, and 1000 mg/kg body weight which contribute 0, 7.2, 36 and 180 mg Al/kg bw/day, respectively.

Treatment with Al chloride basic by oral gavage revealed paternal toxicity (irritation effect on glandular stomach mucosa, local effect) at 1000 mg/kg in both the male and female Wistar rats. Based on findings observed macroscopically (red foci or thickening of the grandular mucosa of the stomach) and supported by microscopic examination, the maternal/parental No Observed Adverse Effect Level (NOAEL) for local toxic effects on stomach was established at 200 mg/kg and LOAEL – at level 1000 mg/kg, for both males and females.

Several statistically significant changes in clinical biochemistry parameters were observed at 1000 mg/kg: decreased Hb level in males, MCHC in both Al treated males and females, decreased alkaline phosphatase activity, decreased total protein and albumin levels in blood serum and increased potassium level. Decreased Hb levels were observed in two other doses in males but no dose response relationship was observed. Lack of relevant baseline values for the observed clinical data limit the interpretation of the results.The authors consider clinical biochemistry and haematology changes observed at 1000 mg/kg to be of slight nature and generally within the range expected for rats of this age and strain.Because any morphological correlates were absent, these changes were considered not indicative of organ dysfunction and not of toxicological significance.

No reproduction, breeding and early post-natal developmental toxicity was observed in rats at 1000 mg/kg body weight for males and females. Based on the reported results, a NOAEL for reproduction, breeding and early post-natal developmental toxicity was suggested at a level of 1000 mg/kg bw, the highest dose tested in this study.


In another study (Hicks et al., 1987) Male Sprague-Dawley rats were exposed to aluminium-compounds with diet. The animals were randomly assigned to five groups, 25 animals in each. The groups received 1) basal diet (control), 2) aluminium hydroxide (302 mg Al/kg body weight), 3) KASAL -the basic form of sodium aluminium phosphate containing ≈6% of Al (141 mg Al/kg body weight), 4) KASAL II - the basic form of sodium aluminium phosphate containing ≈13% of Al (67 mg Al/kg body weight) and 5) KASAL II (288 mg Al/kg body weight). Treatment continued for 28 days, during which the animals were observed twice daily for their behaviour, signs of toxicity, and mortality. General physical examinations, body weight and food consumption measurements were performed weekly. After 28 days of treatment, 15 animals from each group were killed. Blood was collected from 5 rats of each group for blood cell counts, haemoglobin concentration, haematocrit and serum chemistry measurements. These rats were subjected to gross necropsy and histopathological examination. Femurs from 10 rats were taken for possible aluminium analysis; femurs from 5 rats were analyzed for Al concentrations. Five rats were allowed to recover for 2 months and five rats for 5 months after termination of the treatment. During these recovery periods, the rats received the basal diet and were observed daily; body weight and food consumption were measured monthly. Femurs were collected at autopsy from these rats for aluminium analysis. During the entire experimental period, no mortality was reported and no treatment-related clinical signs were observed. All clinical observations were characteristic of male Sprague-Dawley rats of relevant age. There were no significant group differences in body weight, food and water consumption and haematological parameters.A mild (2 - 4%) but significant increase in serum sodium level was observed in all treated animals.However, all increased sodium levels were within the range of historical control for rats of the same age in the laboratory. A significant 16% increase in absolute kidney weight was reported in the group of rats receiving KASAL II at 67 mg Al/kg bw. This increase appeared not to be treatment-related because no such increase was seen in the group of animals treated with this substance at 288 mg Al/kg bw. There were no other significant group differences in organ weights. All lesions seen at microscopic examination were “those normally expected for young adult male Sprague-Dawley rats” (Hicks et al., 1987). No lesions suggestive of a treatment-related effect were seen. Aluminium concentrations in all femur samples from all groups were < 1 ppm and most were below the limit of detection or quantification.The distribution of samples in which Al was not detectable, was detectable but not quantifiable or was quantifiable, was similar in all the groups. It should be noted that this comparison was based on small numbers of samples from each group (5). Al was quantifiable in all 5 samples from animals treated with Al(OH)3, in 2 samples from the control animals and in none of the samples from animals treated with KASAL or KASAL II. The results of this study provide no evidence for significant deposition of Al in the bone and no evidence for adverse effects induced by Al hydroxide or basic food grade sodium aluminium phosphate (KASAL and KASAL II) during 28-day dietary administration at daily doses up to ≈300 mg Al/kg body weight. 


Sodium aluminium phosphate was administered to beagle dogs with diet at concentrations 0% (control), 0.3%, 1.0% and 3.0% for 6 months (Katz et al., 1984). There were no significant group differences in body weight throughout the experiment. Reductions in mean body weight occurred in all groups during week 27, which the authors attributed to “pretermination tests and increased handling by technicians.” No treatment-related clinical signs and no ocular changes in any of the animals were observed. In most weeks, treated male and female dogs consumed less food than control dogs. In male animals, none of the differences in mean food consumption values was statistically significant. In females, significant reductions occurred “sporadically”. The authors did not consider these differences in food consumption as “toxicologically significant”, the conclusion that was supported by the absence of corresponding reduction in body weights. The treatment did not have any effect on haematological and blood biochemistry parameters, urinalysis results and results of analysis for occult blood in faeces. There were no significant differences in mean organ weights between the treated groups and the control group. Gross pathology and histopathology findings were in the “normal range of variations for dogs of this strain and age”; no treatment-related lesions were observed. The results of this study provide no evidence for toxicity of acidic form of sodium aluminium phosphate during 6-month administration at concentrations up to 3% in the diet.


Aluminium citrate was administered to ten female Sprague Dawley rats with drinking water at a concentration of 80 mmol/L for 8 months (equivalent of 6 - 11 mmol Al/kg bw/day) (Vittori et al. 1999). Plasma iron concentration and total iron-binding capacity were not different in the control and the Al treated rats, indicating that the Al treated animals were not depleted of iron. There were no significant group differences in blood urea concentration, which suggests that kidney function was not altered by Al administration. Significantly lower haematocrit and blood Hb concentration were observed in the Al treated rats than in the control rats. Significantly higher reticulocyte count, abnormal erythrocyte morphology, a significant inhibition of (late colony-forming unit-rethroid, CFU-E) growth and a significant reduction of 59Fe uptake in the bone marrow were reported in the Al treated rats. Plasma haptoglobin concentration was significantly lower in the Al treated animals than in the control animals. This and the presence of abnormal erythrocytes in the Al treated rats are indicative of intravascular haemolysis. Scanning electron microscopy combined with EDAX detected Al inside circulating erythrocytes with abnormal shape from animals in the Al treated group. Al concentrations in the bone, spleen, liver, kidney and plasma were significantly higher in the Al treated group than in the control group. No significant group difference in brain Al concentrations was seen. There was no correlation between plasma Al concentrations and Al levels in the organs or any other biochemical data. The results of this study suggest that Al may affect erythropoiesis in rats with normal renal function.


A recent combined one-year developmental and chronic neurotoxicity study with Al-citrate (Alberta Research Council Inc, 2010) may be of interest for the evaluation of the neurotoxicity of aluminium hydroxide, aluminium oxide, and Al-metal taking into consideration the tenfold lower bioavailability of Al-hydroxide compared to Al-citrate and excluding effects that can likely be related to the salt rather than the cation. The study was conducted according to OECD TG 426 and GLP, and the exposure covered the period from gestation day 6, lactation and up to 1 year of age of the offspring. Pregnant Sprague-Dawley dams (n=20 per group) were administered aqueous solutions  via drinking water of  3225 mg/Al citrate/ kg bw/day (300 mg Al/kg bw/day); 1075 mg/Al citrate/kg bw/day (100 mg Al/kg bw/day); 322.5 mg/Al citrate/kg bw/day (30 mg Al/kg bw/day). The highest dose was a saturated solution of Al-citrate. Two control groups received either a sodium citrate solution (citrate control with 27.2 g/L, equimolar in citrate to the high dose Al-citrate group) or plain water (control group). The Al citrate and Na-citrate were administered to dams ad libitum via drinking water from gestation day 6 until weaning of offspring. Litter sizes were normalized (4 males and 4 females) at postnatal day (PND) 4. Weaned offspring were dosed at the same levels as their dams. Dams were sacrificed at PND 23. At PND 4  1 male and 1 female pup of each litter  were allocated to 4 testing groups: D23-sacrifice group for pre-weaning observations and D23 neuropathology, D64, D120 and D365 postweaning groups for post weaning observations and neuropathology at the respective days of sacrifice. Endpoints and observations in the dams included water consumption, body weight, morbidity and mortality and a Functional Observational Battery (FOB) (GD 3 and 10, PND 3 and 10). Pups were examined daily for morbidity and mortality. Additional neurobehavioral tests were performed at specified intervals and included, T-maze, Morris water maze, auditory startle, and motor activity. Female pups were monitored from PND26 for vaginal opening, male pups from day 35 for preputial separation. Clinical chemical and haematological analysis was performed for each group on the day of scheduled sacrifice. Al-concentrations were determined in blood, brain, liver, kidney, bone and spinal cord tissues by inductively coupled plasma mass spectrometric analysis. Further metals such as iron, manganese, copper and zinc were also determined. The pathological investigation includes rain weight and neuropathology. Statistical analyses were performed using the SAS software release 9.1. Data collected on dams and pups were analysed separately. All analysis on pups was performed separately for each sex. Statistical significance was declared from P ≤ 0.05.

Results: Dams: Eight high dose dams developed diarrhoea. In the Na-citrate group one dam stopped nursing and the pups were euthanized. No significant differences between mean body weights of dosed animals compared to controls were observed during gestation and lactation. During gestation and lactation low and mid dose group animals consumed considerably more fluid than controls and high dose group animals. This is not considered treatment related as there was no dose response. In all animals the target dose was exceeded during lactation due to the physiologically increased fluid consumption.

Pups: During the pre-weaning phase weights of mean body weights of male and females in the sodium citrate and high dose group were significantly lower than the untreated controls. This suggests a citrate rather than Al-related effect. No differences between treated and control animals were observed in the FOB. No other clearly treatment related effects were observed pre-weaning.

F1-postweaning: General toxicity

No significant differences in body weights throughout the study were observed between low and mid-dose animals sodium-citrate and untreated controls. High dose males had significant lower body weights than controls by PND 84. These animals also had clinical signs. At necropsy urinary tract lesions were observed in the animals of the high dose group, most pronounced in the males, hydronephrosis, uretal dilatation, obstruction and/or presence of calculi. All high dose males were sacrificed on study day 98. The effect is probably due to Al-citrate calculi precipitating in the urinary tract at this high dose level. This effect is related to the citrate salt and cannot be attributed to the Al-ion. Female high dose animals showed similar urinary tract lesions, but with a lower incidence and severity. Urinary tract lesions were also observed in single mid dose males, but also in a few sodium citrate and control animals. Fluid consumption during the study was increased in the sodium citrate and Al-citrate groups (in particular high and mid dose) compared to controls. This is probably due to the high osmolarity of the dosing solutions. However, the consumed dose levels decreased in all dose groups during the study. In the beginning the target dose was considerably exceeded, while versus the end of the study it was considerably below the target dose.  According to the authors the assigned dose levels still remain valid.

Developmental landmarks:

In sodium citrate controls and high dose males and females the number of days to reach preputial separation or vaginal opening was longer than in untreated control animals. This may be related to the lower body weights in these animals at the respective time-point. As the sodium citrate group showed similar retardation this effect cannot be allocated to the aluminium cation.

Neurobehavioral testing

No consistent treatment related effects that could be related to Al-ion exposure were observed in the FOB. No treatment related effects on autonomic or sensimotoric function were observed in the study. A weak association between Al exposure and reduced home cage activity, a very weak association with excitability, some association with neuromuscular performance were reported but according to the authors this may also be related to group differences in body weight, and an association with physiological function and is thus not considered clearly treatment related. No treatment related effect on general motor behavior was observed. No clearly treatment related effect on auditory startle response was observed. There was no evidence of any treatment related effect on learning and memory in the Morris Water Maze test and no clearly treatment related effects in the T-maze test. Hind limb grip strength and to a lesser extend foot splay were reported to be reduced compared to controls in high and mid dose male and female animals, more pronounced in younger than in older  rats. However, the observed effects can be related to the lower body weights of the individual animals undergoing this test. No details on the individual findings and historical control data are available. It can therefore not be concluded with certainty that the observed neuromuscular effects are primary effects of the treatment and attributable to Al3+. The NOAEL was reported based on this effect as 30 mgAl/kg bw in a conservative approach.

Haematology: No clinically significant differences in hematology were observed at the investigation on day 23. In day 64 and 120 females and day 64 males the high dose group showed slight reduction in hematocrit (males only), mean hemoglobin and mean corpuscular cell volume. No such changes were observed in the 364 day group.

Clinical chemistry: while a number of borderline statistically significant changes were observed, such as globuline levels, alkaline phosphatase and glucose in the high dose group little or no biological significance is associated with them. Elevated creatinine and urea levels in Day 64 males are consistent with the renal toxicity observed in these animals.

Organ weights: Brain weights did not differ among the groups, with two exceptions in the day 64 group males brain weights were significantly lower than controls. In the 120 day female high dose group brain weights were also significantly lower than controls. These findings were not reproduced at the other sacrifice times. Brains to body weight ratios were not significantly different and the lower brain weights can be attributed to the body weight.

Pathology: The main pathology findings were the renal lesions with precipitates in the urinary tract and secondary lesions such as hydronephrosis and uretal dilatation   in particular in the high dose group males and to a lesser extend females. Fluid colonic content was also observed in some high dose animals, in particular males. According to the authors the test item clearly precipitated in the urinary tract causing stone formation and blockage and resulted in fluid colonic content. No other macroscopic effects were observed in other organs.

Histopathology: No treatment related histopahological effects were observed in the nervous system at any time point.

Aluminium concentrations in different organs were dose related. Tissue concentrations were highest in blood, and then in decreasing order brainstem, femur, spinal cord, cerebellum, liver cerebral cortex.

A conservative NOAEL of  322 mg Al-citrate/kg bw  corresponding to 30 mg Al/kg bw was derived from this study (with a bioavailability correction this would correspond to ca. 300 mg Al from Al(OH)3).

The most important effects were however related to a precipitation of the citrate in the kidneys and urinary tract and this effect is not related to the Al3+ ion.  The effects on grip strength and foor splay observed can also not be attributed unequivocally to Al-exposure as they may have been secondary to the general toxicity and body weight differences between treated and control animals undergoing this test.





Human Studies

Aluminium powder

The majority of published human studies of lung effects on exposure to aluminium powder were conducted prior to 1970 (Krewski et al., 2007 (review); Doese, 1938; Goralewski, 1939 to 1948 in Perry, 1947; Koelsch, 1942; Meyer and Kasper, 1942a,b; Crombie et al., 1944; Mitchell et al., 1959, Mitchell et al., 1961; McLaughlin et al., 1962). Effects were associated with intermittent use of a more permeable and possibly biologically active petroleum-based mineral oil coating in place of stearine (Dinman et al., 1987). The small, cross-sectional study (n = 62) by Kraus et al. (2006) provides some evidence for the development of lung pathology (small, round opacities in the upper lung; a thickening of the interlobular septae) consistent with alveolitis without fibrotic activity on exposure to aluminium powder (respirable size range; with diameters smaller than 5μm). Exposure duration in this study ranged from 78 to 360 months. Multivariate logistic regression showed a significant independent association between Al levels in urine and the occurrence of abnormal high resolution computed tomography (HRCT) findings (OR = 1.008, 1.002 - 1.013; 95% CI; p < 0.006; with adjustment for age, time of exposure, smoking habits, vital capacity, FEV1/VC, and resistance). 


“Aluminium” dust

Miller et al.(1984) observed pulmonary alveolar proteinosis in a 44-year old male who had been exposed to high levels of aluminium-containing dust during 6 years as a rail grinder. A recent report (Cai et al., 2007) reported granulomatosis lung disease in a 50-year old woman who had worked in a metal reclamation factory and been exposed to high levels of aluminium dust. Energy dispersive X-ray analysis of the granuloma tissue showed high concentrations of aluminium. Separation of an effect specific to aluminium from an effect due to high doses of dust, or in fact, aluminium oxide, is not possible based on these studies.



Aluminium smelters - occupational asthma

Donoghue et al. (2010) studied occupational asthma among employees in Al pre-bake smelters of Australia and New Zealand from 1991 to 2006 and examined relations between asthma in highly exposed workers and potroom air contaminants. The authors collected asthma incidence each year by a survey of seven Al smelters using diagnostic criteria developed in 1990 by the Australian Aluminium Council Health Panel. Regular medical surveillance, including respiratory questionnaires and spirometry, was conducted at all smelters with intervals from 3 months to 2 years between examinations depending upon job type and duration of employment. No information was available on ages of the workers, gender or range of length of employment. Asthma cases were identified by surveillance following development of symptoms or a few of the cases were diagnosed by a family physician. Pre-placement criteria and assessment of individual suitability for jobs with exposure to potroom dust, fumes, and gases were introduced before the study period; these criteria evolved over the course of the study and these criteria were not uniform at all smelters. These parameters included a history of asthma beyond childhood, reduced forced expiratory ratio (FER) and evidence of reversible airway obstruction. In some smelters, assessment of non-specific bronchial hyper-responsiveness using methacholine challenge was performed. Incidence rates for occupational asthma were calculated for each smelter and for all smelters combined and the data were presented for each year of the study. All cases of occupational asthma identified among smelter employees (regardless of job category) were divided by the total number of smelter employees (regardless of job category) and the incidence rates were expressed as the number of cases per 1,000 employees per year. These annual surveys also obtained data on the work areas in which asthma cases were reported, but due to limited data on employees in each work area, incidence rates by work area were not calculated. Employees who worked ‘‘in close proximity to pot fume or bath material for several hours a week as part of their normal job” (e.g., potrooms, potroom services, rodding, potlining, cryolite recovery, scrubbing, and alumina) were defined as the highly “bath exposed” workers. Exposure data were based on personal sampling of inhalable particulate, respirable particulate, particulate fluoride (F), gaseous hydrogen fluoride (HF) and total F for potroom employees charged with anode changing as it was the most consistent job across all of the smelters. Exposure data were collected from the breathing zone of potroom employees (the numbers of employees were not provided) under the supervision of qualified occupational hygienists for each year (1996 – 2006), but the study design was such that use of personal respiratory protection was not taken into account.

The statistical significance of changes in exposure concentrations (mg/m³) of inhalable particulate, respirable particulate, particulate fluoride, gaseous hydrogen fluoride, total fluoride across all Al smelters during the study period was assessed by regression P-values and Spearman’s correlation coefficients were calculated for correlations between the incidence rate and each exposure variable. A total of 329 cases of occupational asthma were identified and the highest rate occurred in 1992 (9.46/1,000 per year), but this declined to 0.36/1,000 per year in 2006. This amounted to a 96.2% reduction in asthma incidence. Of the 329 cases, 180 (55%) occurred in potroom production employees and of the total at least 243 of those cases (74%) occurred in employees who were assigned duties in the ‘‘bath exposed’’ areas.The mean proportion of all employees who were ‘‘bath exposed’’ over the period 1991 – 2006 was 50% (2,916/5,827) (no further details provided). The median values of the geometric mean concentrations of inhalable particulate, respirable particulate, particulate F, gaseous HF and total F across all seven of the Australian and New Zealand smelters in the worker’s breathing zone declined over the study period. Statistically significant correlations were observed between the reductions in the incidence rate of asthma and reductions in total respirable particulate, total F, particulate F and gaseous HF. The correlation coefficient was greatest for total F (rs= 0.497).

The Donoghue et al. (2010) results demonstrate reductions in occupational asthma among employees of seven New Zealand and Australian Al smelters from 9.46 per 1000 employees per year in 1991 to 0.36 per 1000 employees per year in 2006. Moreover, this reduction was correlated with reductions in the geometric mean of total F in the breathing zone among employees undertaking anode changing (rs = 0.497, p < 0.001). 

A number of the potroom exposure control measures were implemented during the study period (1991 - 2006) and these included an increased focus on standardized work practices, exposure monitoring programs were improved, hooding of pots was performed, ore and fluoride delivery systems were enclosed and enclosed overhead crane cabins with air purification were added and quality control on anode manufacture was improved to reduce replacement of failures. In addition, crucible cleaning and maintenance operations were isolated from potrooms, anode butt cooling was conducted in areas remote from the main potroom aisleway (a practice employed in only some smelters), exhaust ventilation of pots and forced draft ventilation were added in some smelters, natural dilution ventilation of potrooms was augmented by automated fume control in some smelters, an increased focus on consistent alumina/bath anode covers with no holes in the crust, improved process control to minimize anode effects, reduced process upsets that required intervention, improved fume system maintenance to minimize fume system outages, real-time gaseous HF fluoride monitoring in potroom roves with warning signals, controlled sweeping for fugitive dust on floors (e.g., prohibitions on compressed air for sweeping), automated anode butt cleaning and an increased focus on housekeeping. Among the improvements were: additional education on potential health impacts and work practices to minimize exposures, mandatory respiratory protection with clear rules, introduction of enhanced respirator selection and use of powered air purifying respirators (PAPR), quantitative respirator fit testing, education on respirator use, dedicated respirator maintenance and cleaning centers in some smelters and biological monitoring of urinary F for assessment of respiratory protection in some smelters. The Al smelting process was improved over time by blending low sulfur coke with normal coke to reduce SO2 emissions.

There is no question that misclassification of workers can influence the outcome of these types of studies be they prospective or retrospective in nature. Concern can also arise regarding comparisons based on incomplete or highly variable exposures within jobs or between all workers with asthma. If in these circumstances only “bath exposed” workers are considered, this selection can introduce non-differential misclassification of equal or perhaps even greater gravity to the use of area sampling data or failure to measure workplace levels at all.

Fluoride exposure has previously been associated with development of asthma symptoms and non-specific bronchial hyper- responsiveness in cohort studies of Norwegian potroom workers (Kongerud et al., 1991; 1994; Søyseth et al., 1994). However, occupational exposure data for these older studies were incomplete. The strongest epidemiological evidence is likely to come from an inception cohort study of new workers with detailed longitudinal characterization of personal exposures to the host of airborne materials found in Al smelters (Abramson et al., 1989).

Strengths of the Donoghue et al. (2010) study design include: consistent diagnostic criteria that were applied throughout the study period, prospective collection of asthma incidence and collection of personal samples for the highly exposed workers. While the results are most encouraging, the study does suffer from a number of limitations including: data on cases of occupational asthma come from different sources (onsite medical centers and family physicians and it is not clear whether family physicians applied the same criteria to diagnose occupational asthma); the proportion of cases diagnosed by family physicians is not reported; missing data on numbers of asthma cases and/or numbers of employees for some years; the pre-placement criteria were different at different smelters and “evolved during the study period”; the pre-placement criteria applied only to jobs with highest potential for potroom exposures but the incidence rates of occupational asthma were calculated for all employees regardless of job category; there was no description of date(s) when pre-placement examinations were introduced in smelters and the possible impact of worker reassignment, migration out or replacement of the “bath-exposed” workers on the incidence rate was not discussed; lack of data on potential confounding factors (employee turnover rates, age, tobacco consumption). Although it is not clearly stated whether the mean Al exposures for anode changers represent full-time shift measurements or whether it represents the highest short-term or transient peak exposures to dust and gas, the values are likely to be 8-hour time-weighted-averages (the usual way in which occupational exposures of this type are reported). There may be concern regarding correlations between exposure among the most highly exposed employees and the overall rate of occupational asthma given that no data on the asthma rates among the highly exposed workers were presented. The limitations in exposure measures and failure to account for worker migration out of the industry provide the basis for a Reliability Score of 2. In any case no conclusion on a possible effect of Al-exposure can be drawn from this study.


Aluminum smelters exposure to airborne contaminants and respiratory outcomes  

Abramson et al. (2010) investigated the relationships between occupational exposures to airborne contaminants (total F, gaseous fluoride, sulphur dioxide (SO2), coal tar pitch volatiles (as benzene soluble fraction (BSF), oil mist and total inhalable dust) and changes in respiratory function over time. Following a cohort employed at two Australian Al smelters where 446 new employees (77% of the 583 eligible workers) were examined at regular intervals over a period of 9 years (from 1995 to 2003), all participants completed an interviewer-administered questionnaire, pulmonary function tests and skin prick testing for common aeroallergens. Most of the workers were under 35 years old, with a median age for men of 30 (IQR 23-36) years and women of 31 (IQR 25-37) years. At baseline interview, wheeze and chest tightness were reported by 22.6% and 10.6% participants, respectively, and 9.4% were diagnosed with asthma. Baseline pulmonary function findings were within normal values, but 57% were atopic on skin prick testing. At smelter A and smelter B, the proportion of current male tobacco smokers was 32% and 23%, respectively, and the proportions of current female smokers were 39% and 28%, respectively. The proportion of former (19%) and never smokers (51%) was similar between sites.

A task exposure database (TED) was used by site industrial hygienists to record routine air monitoring for the airborne fluorides, SO2, coal tar pitch volatiles (as BSF), oil mist and inhalable dust. Based on these data, a Task Exposure Matrix (TEM) was constructed for each contaminant for each full shift task at each site for each year of follow-up of the study (described in Benke et al., 2000). All individual exposures were assigned using the arithmetic mean for each job/task combination and those values were expressed in mg/m³. By combining the TEM with each worker’s job history, the cumulative exposure between interviews (years 3 mg/m3) was calculated for each constituent.

Data were analyzed with the Generalized Estimating Equations (GEE) method to account for the correlation among repeated measurements from each participant during follow up. Logistic models were used in analyses of health symptoms and BHR data and these models were adjusted for age, gender, age and gender interaction, tobacco smoking, smelter site, atopy and interaction between gender and smoking. Linear models were used in analyses of lung function and these models were also adjusted for age, gender, age and gender interaction, smoking, smelter site, height at entry interview and interaction between gender and smoking. Nearly all (98%) of the production, office (96%) and maintenance (95%) workers remained in the same group over the follow-up period. Cumulative exposures were described by tertile (but those data presented in the publisher’s appendix Table E2 were not available for review).

The highest prevalence and widest range of exposures were found for inhalable dust, but the empirical data were not presented. The 95thpercentiles for all airborne constituents considered in the Abramson et al. (2010) study were less than the current Australian Exposure Standards for an 8 h shift; only 1.1% of the cohort was exposed to total inhalable dust above the standard of 10 mg/m3. Asthma symptoms (wheeze and chest tightness) were associated with cumulative exposures to SO2 (p < 0.001 and p < 0.01, respectively), inhalable dust (p < 0.002 and p < 0.02) and coal tar pitch volatiles as the benzene soluble fraction (BSF) (p < 0.005 and p < 0.04). Fluoride (no details on the type of fluoride, e.g., gaseous or total) exposure was associated with wheeze (P < 0.04), but not with chest tightness.The authors reported that the association between wheeze and gaseous fluoride (OR 1.19, 95% CI 1.01 to 1.41 per tertile) was very similar to that for total fluoride (but those data were presented in publisher’s appendix Table E7 not available for review). When fluoride and inhalable dust were analyzed simultaneously for wheeze, only the coefficient for inhalable dust remained statistically significant (data from the publication referenced at appendix Table E4 were not available).Airflow limitation [the lower values of the forced expiratory ratio (FEV1/FVC)] was associated with higher cumulative exposures to coal tar pitch BSF (p < 0.03), fluoride (p < 0.04) and SO2 (p < 0.001). Longitudinal changes in pulmonary function (decline in FEV1 and FVC) were significantly associated with cumulative exposure to fluoride (p < 0.001 and p = 0.002, respectively), inhalable dust (p = 0.02 and p < 0.001, respectively) and SO2 (p = 0.001 and p = 0002, respectively).

Statistically significant associations were largely confined to male employees. The authors reported that the findings for gaseous F were similar to those for total F; however, data referenced at appendix Table E9 of the publication were not presented.The results suggested that the likelihood that bronchial hyper-responsiveness (BHR) increased significantly with cumulative exposure to coal tar pitch BSF (p = 0.03), fluoride (p = 0.03), inhalable dust (p = 0.001), SO2 (p = 0.009) and oil mist (p < 0.001).Bronchial hyper-responsiveness was associated with current tobacco smoking in females and atopy in both genders, but those data (referenced at publication appendix Table E11) were not presented. 

The strengths of the Abramson et al. (2010) study include the inception cohort design and robust assessment of workplace area exposures. The study is limited by possible misclassification of Al exposure including lack of personal sampling data (the results given in Benke et al., 2000; 2001). The data were expressed only as inhalable (total) dust and there was no differentiation between metallic Al and Al oxides. In addition, the fact data referenced in Tables E2-E11 were not available for review may limit interpretation of the results. The analyses were not adjusted for the use of personal protective equipment (PPE) and thus the reliability of the exposure data reported may be called into question. 

The Abramson et al. (2010) data establish relationships between cumulative occupational exposures to a number of airborne contaminants and occupational asthma among Al smelter workers. There was a significant association between chronic occupational SO2 exposure and wheeze and chest tightness, BHR reactions to methacholine challenge, reduced FEV1 and a longitudinal decline in lung function.A concentration-response relationship could be seen between fluoride exposure and those same outcomes, but the association was less evident. There was also a significant relationship between long-term cumulative exposure to inhalable dust and asthma-associated symptoms (wheeze and chest tightness), the longitudinal decline in pulmonary function (decline in FEV1 and FVC values) and increased BHR with the greater associations for wheeze and consecutive changes in FVC and BHR. Exposure to the BSF of coal tar pitch volatiles was also associated with increased asthma, airflow limitations and BHR, but coal tar pitch volatiles are not recognized as respiratory sensitizers. Oil mist exposure was also associated with increased BHR.Although many of the exposures were highly correlated, further statistical analyses suggested that of the known respiratory irritants, SO2 was more likely than fluoride to be responsible for many of the symptoms observed.In summary, chronic exposure to elevated levels of inhalable dust, SO2 and fluoride were the most important determinants associated with decrements in pulmonary function among these Al smelter employees. Again, no correlation to Al-levels was made and the study does not establish any relationship with Al-exposure.

Aluminum smelters - cardiopulmonary toxicity 

Friesen et al. (2010) studied acute and chronic polyaromatic hydrocarbon (PAH) exposure in relation to cardiopulmonary mortality in a cohort of 6,423 men and 603 women who worked for 3 or more years at an Al smelter in British Columbia, Canada. The authors linked data for the cohort to national mortality rates for the years 1957 to 1999 and examined exposure-response regarding the incidence of chronic respiratory diseases (COPD) and cerebrovascular disease. Work histories were abstracted from company records. Smoking status (ever smoking, never smoking, and unknown) was obtained through self-administered questionnaires sent to current workers, pensioners or to their survivors if the worker was deceased. Mortality of the cohort for select causes of death was compared with that of the whole of British Columbia using standardized mortality ratios (SMRs) adjusted for age, gender and time period. The development of a benzo(a)pyrene-based [B(a)P] quantitative job exposure matrix was described by Friesen et al. (2006). To create the [B(a)P] job-exposure matrix as a surrogate for total PAH exposures, statistical models were developed to derive annual arithmetic mean B(a)P levels for each operation and maintenance job in smelter potrooms that were based on personal exposure measurements collected from 1977 – 2000. This matrix accounted for different rates of exposure and declines in concentrations over time and potline. Exposure estimates for jobs without measurements were extrapolated from exposure estimates from the statistical models by adjusting for the amount of time worked in areas where ambient B(a)P levels had been determined.  Job- and time-period-specific B(a)P exposures levels were linked to each employee’s work history for calculation of cumulative and current B(a)P exposures. The models included smoking status and time-dependent covariates for years (5-year categories), time since first employed (years; continuous) and work status (employed at smelter: yes/no). For cerebrovascular disease and COPD, exposures were examined using cumulative B(a)P metrics (0-, 2-, 5-, and 10-year lags). The participants were workers who had a mean age of 32.4 years (range: 18–65), who were employed an average of 14.5 years (range: 3 – 45 years) and who contributed an average of 23.5 years (maximum 47) to study follow-up. Workers who died of ischemic heart disease (IHD) were more likely to have ever smoked than the average worker in the cohort (65% – 70% vs. 57%). The all-cause mortality SMR was less than that for the province’s population for both males (SMR = 0.87, 95% confidence interval (CI): 0.82 - 0.92) and females (SMR = 0.85, 95% CI: 0.63 - 1.11). Ischemic heart disease mortality (n = 281) was associated with cumulative historic B(a)P exposure (hazard ratio = 1.62, 95% confidence interval: 1.06, 2.46) in the highest category (> 66.7μg/m3-year). However, the higher hazard ratio for IHD found was for chronic B(a)P exposure and it was restricted to those who were in active employment (adjusted for smoking status and calendar year) - 2.39 (95% confidence interval: 0.95, 6.05) and who also had the highest cumulative B(a)P exposures (> 66.7μg/m3-year). The higher B(a)P exposures also had the widest confidence intervals. The stronger associations observed during employment suggest that cardiovascular effects may be reversible after termination of employment at the smelter even after adjusting for tobacco consumption. 

The Friesen et al. (2010) results suggest that cumulative workplace air PAH exposures in this Al smelter declined over time. One strength of the Friesen et al. (2010) study was its longitudinal design; however, the B(a)P exposures were not well characterized and exposure to Al and other constituents in smelter air was not taken into account. The authors elected to rely on mortality rates rather than morbidity for cardiopulmonary outcome and used semi-quantitative exposure estimates which may not reflect the actual PAH exposures. Failure to account for the host of factors known to influence cardiovascular health including diet, physical activity and the medical history of the participants are serious deficiencies. This study suggests there may be an association between heart disease and chronic PAH exposure in Al smelter workers. The study does not include an analysis of Al-exposure and related possible effects and is therefore not useful for the risk assessment of Aluminium.


Animal Studies

The high doses of particulates applied in animal studies tend to lead to overload-related effects (ATSDR, 2008; ILSI, 2000). Animal studies administering dust by intratracheal instillation (ITI) are not useful as sources of dose-descriptors for the inhalation route of exposure as some responses may result from the un-physiologic mode of administration. ITI studies can be useful, however, for screening and comparative ranking of particles for toxic effects. Several studies have shown interspecies differences in pulmonary reaction on exposure to aluminium metal and alumina (Engelbrecht et al., 1959; Gross et al., 1973; Christie et al., 1963).

Gross et al. (1973) did not observe development of alveolar proteinosis or thickening of alveolar walls in rats, hamsters or guinea pigs exposed to Al2O3dust (66% < 1 μm). Pigott et al. (1981) reported no evidence of fibrosis in a repeated dose inhalation study that administered alumina fibres (Saffil) at levels between 2 and 3 mg/m³. The only pulmonary response observed was the occurrence of pigmented alveolar macrophages. 

Ess et al. (1993) investigated the subacute and chronic effects of short-term ITI administration (50 mg total dose, five 0.1 mL injections of suspension in sterile saline over a period of 2 weeks) of five smelter-grade and two laboratory-grade aluminas to Sprague-Dawley rats. These doses were sufficient to overload clearance mechanisms. All the dusts led to an inflammatory reaction in the alveoli evidenced through significantly elevated BALF total protein, LDH and sustained increases in PMN compared with the saline control. Only the laboratory grade aluminas showed signs of fibrosis (collagen) one year post-instillation, however.     

Adamcakova-Dodd et al. (2010) studied the pulmonary response following sub-acute inhalation exposure to aluminium nanowhiskers in mice. Aluminium nanowhiskers have been used in manufacturing processes as catalyst supports, flame retardants, adsorbents, or in ceramic, metal and plastic composite materials. Male mice (C57Bl/6J) were exposed to aluminium nanowhiskers for 4 h/day, 5 days/week for 2 or 4 weeks in the dynamic whole body inhalation exposure chamber. Control animals were exposed to a comparable sound level (80 dB) and laboratory air. The primary dimensions of these nanowhiskers were 2-4 nm x 2800 nm (Sigma-Aldrich) and the test nanomaterial [Al(OH)3:AlOOH] contained 35% Al.  The average concentration of aluminium nanowhiskers in the chamber was 3.3 ± 0.6 mg/m³ and median particle size diameter was 154.1 ± 1.6 nm. Both groups of mice were killed within 2 or 4 weeks of exposure at which time bronchoalveolar lavage (BAL) fluid was analyzed for differential and total cells, total protein, activity of lactate dehydrogenase (LDH) and cytokines. Total and differential white blood cells in the BAL fluid were counted. Lungs were processed for histopathology and pulmonary mechanics measurements were taken (flexiVent). The Al content in the lungs, heart, liver, spleen, kidney and brain was determined by ICP-OES. It was reported that the total number of cells as well as number of macrophages in BAL fluid was double that in mice exposed for 2 weeks and 6 times higher in mice exposed for 4 weeks, compared to air sham controls (p<0.01 and p<0.001, respectively). However, no neutrophilic inflammation in BAL fluid was found and the percentage of neutrophils was below 1% in all groups. No significant differences were found in total protein, activity of LDH, or cytokine levels (IL-6, IFN-γ, MIP-1α, TNF-α, and MIP-2) in BAL fluid between shams and Al-treated mice. In summary, sub-acute (2 or 4 weeks) inhalation exposures to Al whiskers increased the numbers of macrophages in BAL fluid. No other inflammatory or adverse responses were observed. Currently available information does not permit characterization of the nanowhiskers used in this study as particles or fibers.

This study provides new data on solubility and dissolution process of Al in biological fluids under different physiological conditions.However, the applicability of reported findings for manufactured Al2O3 nanomaterials to the bulk Al powders and dust is not clear.Only an abstract is available which limited interpretations of the study results. A Klimisch Score 4 (not assignable) was considered an appropriate for this study.


In vitro cytotoxicity studies

Aluminium metal reacts rapidly with air to form an aluminium oxide coat. Thus, exposure of tissues or cells to zero valence aluminium metal is unlikely by inhalation unless the aluminium powder is coated with a substance that acts as an effective barrier to oxidation in air but does not act as an effective barrier in the lung. The in-vitro studies by Wagner et al. (2007) and Braydich-Stolle et al. (2010) provide some evidence for a difference in cytotoxic effect between Al-NPs (2-3 nm oxide coat) compared with Al2O3-NPs. Al-NPs also showed an effect on macrophage phagocytosis of bacteria and particulates under the experimental conditions. The importance of this effect in-vivo in humans for larger particle sizes with thicker oxide coats is unclear. The utility of in-vitro studies for predicting the pulmonary toxicity profile in-vivo remains limited due to the dependence of biological effects, deposition, retention and inflammatory response on particle surface and physico-chemical characteristics (Sayes et al., 2007). 


Mechanism of Action

Overall, the results of the available in-vitro studies described earlier support the low cytotoxicity of poorly soluble aluminium oxide. Aluminium hydroxide and the closely related oxyhydroxide are similarly poorly soluble. These substances can be considered PSPs i.e. poorly soluble particulates of low cytotoxicity.



The current weight of evidence does not support a chemical-specific hazard on inhalation exposure to alumina (aluminium oxide, aluminium hydroxide) as experienced by the worker population. Gross et al. (1973) and Pauluhn (2009a) are considered the most adequate studies from which to obtain a dose descriptor to form the basis for a DNEL for repeated dose toxicity (inhalation, local effect) for these substances. The NOAEC from Gross et al. (1973), a subchronic study, for aluminium oxide (mean diameter 0.8 µm) is 70 mg/m³. The NOAEC from Pauluhn (2009a; sub-acute study; MMAD=1.7μm; agglomerated nanomaterials) for aluminium oxyhydroxide is 3 mg/m³ for a range of sensitive endpoints.



No animal studies are available in which the repeated exposure toxicity of aluminium has been investigated.

Repeated dose toxicity: inhalation - systemic effects (target organ) respiratory: lung

Justification for classification or non-classification

According to DSD (67/548/EEC) or CLP (1272/2008/EC) classification criteria for repeated dose toxicity, no classification is required.