Aluminium sodium dioxide

Endpoint summary

Key value for chemical safety assessment

Effects on fertility

Additional information

There is no information available on the toxicity to reproduction or development of sodium aluminate.

Sodium aluminate is a strong base (pH >11.5) and thus corrosive to the skin and mucous membranes. The introductory sections to Annexes VII-X of the REACH regulation (EC No. 1907/2006) point out that in vivo testing with corrosive substances at concentration/dose levels causing corrosivity shall be avoided. Furthermore, due to the corrosive properties of sodium aluminate, repeated human exposure by any route is considered not significant as the corresponding operational conditions and necessary risk management measures (use of PPE) to avoid contact are in place under normal handling and use conditions. Additionally, before new tests are carried out data from structurally related substances (read-across approach) shall be assessed first.

In terms of hazard assessment of toxic effects at potentially non-irritant concentrations, available data on the toxicity to reproduction/development 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; ATSDR, 2008).

A combined 28-day repeated dose toxicity study with the reproduction/developmental toxicity screening test (OECD guideline 422, GLP compliant) was conducted in rats exposed to aluminium chloride basic via oral gavage (Beekhuijzen, 2007). This study is considered the most relevant for the assessment of potential fertility effects, and is supported by the available information on several other published studies on reproduction and/or developmental toxicity of aluminium compounds, which have been extensively reviewed and are summarised further below (Krewski et al., 2007; ATSDR, 2008).

In the Beekhuijzen (2007) study, four groups of 20 Wistar rats (10 per sex) were exposed to aluminium chloride basic at 0, 40, 200 and 1000 mg/kg bw/day, corresponding to 0, 3.6, 18 and 90 mg Al/kg bw/day. Males were exposed for 28 days (starting 2 weeks prior to mating, during mating and up to termination). Females were exposed for 37-53 days (starting 2 weeks prior to mating, during mating, during post-coitum and during at least 3 days of lactation).

Histopathological assessment revealed a mild to moderate subacute inflammation of the glandular stomach mucosa and minimal to moderate superficial mucosal eosinophilic spheroids in both sexes at 90 mg Al/kg bw/day. In males, these findings supported the macroscopic findings of the glandular mucosa/limiting ridge at this dose level. The mucosal eosinophilic spheroids are apparently intracellular degenerative products of the superficial mucosa and possibly associated with inflammation below the base of the mucosa. These findings are indicative of local irritating properties of aluminium chloride basic.

The slightly lower body weight and food intake of females at 90 mg Al/kg bw/day recovered to control levels as treatment progressed. No toxicological significance was therefore ascribed to these changes.

Changes in clinical pathology parameters at 90 mg Al/kg bw/day were of a slight nature and generally within the range expected for rats of this age and strain. Also, any morphological correlates were absent. Therefore, these changes were considered not indicative of organ dysfunction and to be of no toxicological significance.

There were no (further) changes for mortality, clinical signs, functional observations, organ weights, reproduction breeding data and pup development that were considered to be an effect of treatment.

In conclusion, treatment with aluminium chloride basic by oral gavage in male and female Wistar rats at dose levels of 40, 200 and 1000 mg/kg bw/day, corresponding to 3.6, 18 and 90 mg Al/kg bw/day, revealed parental toxicity at the highest dose level comprising local stomach effects. No reproduction, breeding and developmental toxicity was observed for treatment up to the highest dose level.

Based on the findings on the stomach observed macroscopically, the parental NOAEL for local effects was established at 18 mg Al/kg bw/day. The parental NOAEL for systemic toxicity was 90 mg Al/kg bw/day. The reproduction, breeding and developmental NOAEL was established at 90 mg Al/kg bw/day, the highest dose tested in this study.

No effects on the length of the gestation period in female rats exposed to up to 3225 mg/kg bw/day of aluminium citrate (equivalent to 300 mg Al/kg bw/day) were reported in the combined developmental neurotoxicity/chronic study (ToxTest Alberta Research Council, 2009) described in the "Developmental toxicity" section below.

The results of a two generation reproduction toxicity study in Crl:CD(SD) rats with aluminium sulfate (Al2(SO4)3) administered with drinking water were described by Hirata-Koizumi et al. (2011a).

The study was conducted according to the OECD Guideline 416 and GLP. Twenty-four Crl:CD(SD) rats per sex and group (F0 and F1 generation) were given Al2(SO4)3 (AS) in pH 3.57-4.20 drinking water beginning at 5 weeks of age for 10 weeks until mating, during mating, throughout gestation and lactation. Litters were normalized on PND 4. In the F1 generation, 24 male and 24 female weanlings were identified as parents on PNDs 21 to 25, ensuring an equal distribution of body weights across groups. Drinking water provided to the F1 offspring contained the identical AS concentrations as those of their parents. These animals were then mated, and followed through gestation and lactation until sacrifice on PND 26. Each female was mated with a single male receiving the same AS drinking water concentration; if successful mating did not occur (as evidenced by sperm in a vaginal smear or presence of a vaginal plug) within the two week mating period, then the female was put in with another male from the same group who had mated successfully.

Aluminium sulfate was dissolved in deionized water at 0, 120, 600 and 3000 ppm. Deionized water used to make up the drinking water contained < 5 mg/L total Al. Calculated mean Al2(SO4)3 intakes during the treatment period were 8.6, 10.7, 14.4 and 15.3 mg/kg bw/day in the 120 ppm group, 41.0, 50.2 71.5 and 74.2 mg/kg bw/day in the 600 ppm group and 188, 232, 316 and 338 mg/kg bw/day in the 3000 ppm group in F0 males, F1 males, F0 females and F1 females, respectively. Corresponding mean intakes of Al were 1.36, 1.69, 2.27 and 2.41 mg/kg bw/day in the 120 ppm group, 6.47, 7.92, 11.28 and 11.7 mg/kg bw/day in the 600 ppm group and 29.7, 36.6, 49.8 and 53.3 mg/kg bw/day in the 3000 ppm group in F0 males, F1 males, F0 females and F1 females, respectively. The Al content of the diet was 25-29 mg/kg and the mean Al intakes from the laboratory chow alone were 1.6, 1.9, 2.2 and 2.3 mg/kg bw/day in the F0 males, F1 males, F0 females and F1 females, respectively.

Drinking water consumption was reduced significantly compared to the controls in males and females of all treatment groups. This reduction was clearly concentration-dependent and the authors attributed this important observation to avoidance of the water because of its acidic character (pH 3.57-4.20). Among rats given 3000 ppm, body weights, body weight gains and food consumption were reduced significantly compared to the concurrent F0 male and female controls for up to 3 weeks after the start of exposure. Food consumption decreased significantly in the F1 generation. In F0 and F1 females, there was a dose-related decrease in food consumption during the third week of lactation that was statistically significant at 600 and 3000 ppm. There were no adverse effects at any concentration on estrous cycle parameters, copulation, fertility, gestation or delivery indices in either the F0 or F1 generations. There were no significant changes in sex ratios. In the AS-treated males there were no significant treatment-related effects compared to the control animals regarding numbers of testis sperm, the percentage of mobile sperm, swimming speed or sperm morphology after ingestion of up to 38.5 mg Al/kg bw/day.

The body weights of F1 males and females given 3000 ppm AS were significantly less on PND 21 compared with that of rats given deionized drinking water. Similar reductions were seen in the F2 male and female pups although the difference did not achieve statistical significance in the F2 males. No significant differences were observed for birth weight. No significant treatment-related effects were reported for age at completion of pinna unfolding, age at incisor eruption, eye opening or anogenital distance in the F1 and F2 male pups and the F1 female pups. In the female F2 pups, the completion time to pinna unfolding on PND 2 was delayed significantly in the 600 ppm group, but there were no significant treatment-related differences in righting reflex (PND 5), negative geotaxis reflex (PND 8) or mid-air righting reflex (PND 18). There were no significant treatment-related differences in the time of preputial separation in the F1 generation. In the F1 females vaginal opening was significantly (p < 0.05) delayed in the 3000 ppm group (31.4 ± 1.7 versus 29.5 ± 2.1 days in control). At the time of vaginal opening, there were no significant differences in body weight between the control group and those given 3000 ppm AS (119.0 ± 13.3 versus 109.6 ± 11.6 g).

No significant differences were observed in male and female F1 rats regarding spontaneous locomotor activity at 4 weeks of age or in performance on the water-filled T-maze test at 6 weeks of age compared to their respective controls. In the males, there were no significant changes in elapsed time or number of errors in trials on days 2 to 4; in the females, the elapsed time and number of errors were significantly less than controls among animals given 600 ppm on day 2, but no differences were observed on days 3 and 4 and there was evidence for a concentration-response relationship.

In the F0 and F1 adults, there were no treatment-related changes in the internal organs noted at necropsy and no histopathological alterations of the reproductive organs were observed. In F0 males given 3000 ppm, the absolute and relative liver weights were reduced significantly relative to the control; the absolute spleen weight was also significantly lower at 3000 ppm. There were no significant changes in organ weights in the F0 females. In the F1 males, the absolute weights of the kidneys at 3000 ppm and testes at 600 ppm were reduced, but there were no significant differences in the relative weights. The F1 females showed no significant treatment-related differences in relative or absolute organ weights.

The F1 male pups had significantly lower relative liver weights and significantly higher relative brain weights. In the F2 generation, the males had significantly lower relative thymus and spleen weights and a significantly higher relative brain weight; the F2 females had significantly lower relative liver weights and significantly higher relative brain weights. Necropsy of the F1 and F2 weanlings found no dose-related lesions and there were no treatment-related alterations in the histology of the liver or spleen.

Hirata-Koizumi et al. (2011a) identified a LOAEL for parental systemic toxicity and offspring developmental toxicity of 3000 ppm of AS in drinking water for Crl:CD(SD) rats (31.2-55.6 mg Al/kg bw/day or 188-338 mg Al2(SO4)3/kg bw/day) based on reduced body weight gains in the parents and decreased body weight gain, delays in sexual maturation and reduced liver and spleen weights in the F1 and F2 offspring at 3000 ppm. The authors proposed the NOAEL for parental systemic toxicity and offspring developmental toxicity of 600 ppm (8.06-14.0 mg Al/kg bw/day or 41.0-74.2 mg Al2(SO4)3/kg bw/day. It must be noted here that there were no indications of an adverse effect of ingested AS on reproductive parameters (estrous cycles, numbers of primary follicles, copulation index, cauda epididymal sperm/g, numbers of testis sperm, progressively motile sperm, sperm swimming speed and pattern, percentages of abnormal sperm and male and female fertility) in either generation or at any dose. Therefore, the NOAEL for fertility is considered to be 3000 ppm of AS in drinking water for Crl:CD(SD) rats (31.2-55.6 mg Al/kg bw/day or 188-338 mg Al2(SO4)3/kg bw/day).

The major findings of this include decreased drinking water consumption for both sexes in all AS groups, variable reductions in food consumption, reduced body weight in preweaning animals at 3000 ppm, delayed sexual maturation of the female F1 offspring at 3000 ppm, and decreased absolute liver, epididymides, thymus and spleen weight in the offspring at 3000 ppm. As stated above, the authors proposed a LOAEL for aluminium sulfate for parental systemic toxicity and reproductive developmental toxicity of 3000 ppm and NOAEL 600 ppm. However, the authors state, correctly, that because “paired-comparison data are not available to assess the effects of decreased water intake in the absence of AS exposure” there is a possibility that observed developmental effects are the results of reduced water consumption and the authors suggested their NOAEL was conservative.

The statistically significant delay in F1 females’ vaginal opening (29.5 ± 2.1 in controls and 31.4 ± 1.7 days in the highest dose group) was not accompanied by adverse changes in oestrous cyclicity, anogenital distance or further reproductive performance. The authors concluded it is unlikely that “Al to have a clear impact on the hormonal event”. The AS levels added to drinking water by Hirata-Koizumi et al. (2011a) were 190, 946 and 4700 times greater than Al levels found naturally in drinking water (ca. 0.1 mg/L; WHO, 2003).

Interpretation of the results is difficult due to the clear effect of AS treatment on fluid consumption. Addition of the test substance to drinking water at high concentrations led to reduced pH (3.57 to 4.2) and this appears to have reduced the palatability of the drinking water. At these AS levels, the F0 and F1 females also decreased their food consumption relative to the controls during week 3 of lactation. As a result, due to decreased drinking water consumption and decreased food consumption of F0 and F1 dams during the later stages of lactation, it is not possible to conclude with certainty whether the observations reported were associated with Al or represent secondary effects due to maternal dehydration and reduced nursing that may have influenced pup weight on PND 21. Because the effects reported could be related to decreased maternal fluid consumption, the utility of this study for risk assessment is limited.

In another study, Hirata-Koizumi et al. (2011b) investigated the potential reproductive toxicity of aluminium ammonium sulfate (NH4)[Al(SO4)2] in a GLP and OECD Guideline 416-compliant 2 generation reproductive toxicity study.

Aluminium ammonium sulfate (AAS) was dissolved in deionized water at 0, 50, 500 or 5000 ppm. The Al concentration in the deionized water was < 5 µg/mL and the Al content of the diet was 22-29 mg/kg. Calculated mean (NH4)[Al(SO4)2] intakes during the treatment period were 3.78, 4.59, 6.52 and 6.65 mg/kg bw/day in the 50 ppm group, 33.5, 41.8, 58.6 and 61.9 mg/kg bw/day in the 500 ppm group and 305, 372, 500 and 517 mg/kg bw/day in the 5000 ppm group in F0 males, F1 males, F0 females and F1 females, respectively. Corresponding mean intakes of Al were 1.98, 2.35, 2.89 and 3.10 mg/kg bw/day in the 50 ppm group, 5.35, 6.57, 8.81 and 9.36 mg/kg bw/day in the 500 ppm group and 36.3, 44.2, 59.0 and 61.1 mg/kg bw/day in the 5000 ppm group in F0 males, F1 males, F0 females and F1 females, respectively. The Al content of the diet was 22-29 mg/kg and the mean Al intakes from the laboratory chow alone were 1.56, 1.83, 2.20 and 2.39 mg/kg bw/day in the F0 males, F1 males, F0 females and F1 females, respectively.

Groups of 24 male and 24 female Crl:CD (SD) rats (F0 generation) were given AAS in drinking water from 5 weeks of age for 10 weeks prior to mating, during mating and gestation, when the parental males were culled, and for the females through weaning. Litters were normalized to 8 pups on PND 4. At weaning, 24 males and 24 females were selected to serve as the F1 generation and they were given AAS in drinking water for 10 weeks prior to mating, during mating and gestation, and for the females through weaning, as for the F0 generation. Exposure of the F1 weanlings occurred at the same concentrations as those of their parents.

Drinking water consumption was reduced at all concentrations compared to that of the concurrent controls. These reductions were clearly concentration-dependent and the reduction was significant at 500 and 5000 ppm in males and females of the F0 and F1 generations. These reductions were significant at 50 ppm in the F0 males and at some intervals during AAS exposure of the F0 and F1 females.

A transient decrease in food consumption was observed in the 500 and 5000 ppm groups and in body weight in the 5000 ppm group. One F1 male in the 500 ppm group died, but that death was not considered treatment-related.

There were no significant effects of AAS consumption on the estrus cycle. The authors reported no differences for copulation, fertility index, gestation index, precoital interval, gestation length, number of implantations, live pups delivered or delivery index, sex ratios of pups or viability during the preweaning period in females compared to the control. There were no significant differences between control and AAS-treated groups regarding the numbers of testis and cauda epididymal sperm, percentage of motile and progressively motile sperm, sperm swimming patterns and speed or the numbers of morphologically-abnormal sperm. Moreover, there were no significant differences in the numbers of primordial follicles in the F1 ovaries between animals given 5000 ppm AAS and those consuming deionized water.

In the F1 and F2 pups, there were no treatment-related differences in numbers of offspring with congenital malformations, sex-ratio or viability on PND 0, 4 or 21. Reduced body weights were reported in the F1 male and female pups at 5000 ppm, but not in lower dose groups. The F1 male pups had a significantly lower body weight on PND 21, F1 female pups on PND 14 and 21, and F2 male and female pups on the PND 26. In female F1 pups, vaginal opening was delayed significantly among those whose mothers consumed 5000 ppm AAS (mean ± S.D: 32.3 ± 1.8 days vs. 30.2 ± 2.1 days in control), but their body weights were not significantly different from those of the concurrent control at the time of vaginal opening.

Absolute weights of testes and epididymis of the F1 and F2 male pups at 5000 ppm were lower than control. Absolute weights of the uterus were significantly lower in the F1 female pups, and absolute weights of the ovary and uterus were significantly lower in the F2 females. Histopathological examination revealed no treatment-related changes in the reproductive organs. Hirata-Koizumi et al. (2011b) considered these findings secondary to the decreased body weights and attributed the reductions in growth and development of the offspring “to the astringent taste of AAS which would decrease the palatability of drinking water in the AAS-treated groups”.

Spontaneous locomotor activity was no different among F1 males from dams given AAS in drinking water and those whose mothers consumed deionized water alone. There was some variation in activity among the F1 females.

The results presented by Hirata-Koizumi et al. (2011b) provide no evidence that prolonged consumption of AAS has an adverse impact on copulation, fertility and reproductive success in male and female Crl:CD(SD) rats consuming up to 517 mg AAS/kg bw/day. In discussing their data, Hirata-Koizumi et al. (2011b) concluded that “copulation, fertility or gestation indices were not affected up to the highest dose tested at which average Al intake from food and drinking water was estimated to be 36.3-61.1 mg Al/kg bw/day.”

The authors identified a LOAEL of 5000 ppm for both parental toxicity and offspring developmental toxicity (based on decreased body weight gain, decreased food consumption and increased relative kidney weight in F0 females and F1 males and females; reduced preweaning body weight gain in F1 male (at PND 21) and female (PND 14, 21) pups, delay in the vaginal opening in F1 female pups, potentially attributed to inhibition of growth and decreased organ weights in F1 and F2 male and female offspring). The suggested LOAEL level corresponds to 36.3-61.1 mg Al/kg bw/day or 305-517 mg AAS/kg bw/day. The NOAEL for both parental toxicity and offspring developmental toxicity from the Hirata-Koizumi et al. (2011b) study is 500 ppm which corresponds to 5.35-9.36 mg Al/kg bw/day or 33.5-61.9 mg AAS/kg bw/day. Based on the lack of effects on the reproduction parameters examined, the NOAEL for fertility is considered to be 5000 ppm, equivalent to 36.3-61.1 mg Al/kg bw/day or 305-517 mg AAS/kg bw/day.

The major findings reported by the authors were decreased body weight in preweaning animals, delayed maturation of the female offspring, and decreased organ weight in the offspring. However, because the effects could be related to decrease fluid consumption and limited nursing ability of dams, the utility of this study for risk assessment is limited.

No adverse reproductive effects were seen in male Sprague-Dawley rats, as assessed by plasma gonadotropin levels, histopathological evaluation, and serial matings, following exposure to 70 mg Al/kg/day as aluminium chloride in drinking water for up to 90 days (Dixon et al. 1979); this dose does not include base dietary aluminium.

Several other studies have evaluated reproductive effects after short-term oral exposure to aluminium compounds in animals (ATSDR, 2008). An increased incidence of resorptions occurred in female BALB/c mice treated with 41 mg Al/kg/day as aluminium chloride by gavage (aluminium in base diet not reported) on gestation days 7-16 (Cranmer et al. 1986). No reproductive effects were observed in female Sprague-Dawley rats exposed to 158 mg Al/kg/day as aluminium hydroxide or aluminium citrate by gavage and base diet from gestation day 6 to 15 (Gomez et al. 1991), or in THA rats treated with 73.1 mg Al/kg/day as aluminium chloride by gavage (aluminium in base diet not reported) from gestation day 7 to 16 (Misawa and Shigeta 1992). In a study of female reproductive system development (Agarwal et al. 1996), offspring of rats that were gavaged with aluminium lactate on gestation days 5-15 showed a transient irregularity of the oestrus cycle (increased number of abnormal cycle lengths) at 250 mg Al/kg/day; doses as high as 1000 mg Al/kg/day did not affect other end points (gonad weights, anogenital distance, time to puberty, duration of induced pseudopregnancy, or numbers of superovulated oocytes). The inconsistent findings summarised above may reflect differences in susceptibility among different strains/species of animals or compound differences in toxicity or bioavailability as well as deficiencies in study design. Additionally, because levels of aluminium in the base diet were not reported by Agarwal et al. (1996), Misawa and Shigeta (1992), or Cranmer et al. (1986), the doses in these studies are likely to underestimate actual aluminium intake.

In a combination short-term- and intermediate-duration study, no adverse effects on fertility or other general reproductive indices were found in female rats that were exposed to 38-77 mg Al/kg/day as aluminium nitrate by gavage and base diet for 14 days prior to mating with males that were similarly treated for 60 days pre-mating (Domingo et al. 1987). These exposures were continued throughout mating, gestation, parturition, and weaning and caused a reduction in the growth of the offspring in all treated groups, but the effects were negligible and transient (slight decreases in body weight, body length, and tail length observed on postpartum days 1 and 4 were no longer evident at time of weaning).

Mating success (numbers of litters and offspring) was not affected in a three-generation study with Dobra Voda mice that were exposed to 49 mg Al/kg/day as aluminium chloride in drinking water and base diet over a period of 180-390 days (Ondreicka et al. 1966). No reproductive effects were observed in pregnant Swiss Webster mice that consumed 250 mg Al/kg/day as aluminium lactate throughout gestation and lactation (Golub et al. 1992a). However, an alteration in gestation length was observed in pregnant Swiss Webster mice that consumed 155 mg Al/kg/day as aluminium lactate in the diet during gestation and lactation (Donald et al. 1989). The effect on gestation length was small but statistically significant; all litters in the control group (7.5 mg Al/kg/day) were born on gestation day 18, whereas 4 of 17 litters exposed to ≥ 155 mg Al/kg/day were born earlier or later (gestation days 17, 19, or 20).

Short description of key information:
Based on the available information on soluble and insoluble aluminium compounds (read-across), there are no indications for adverse effects on reproduction following repeated oral exposure to aluminium-containing substances. Furthermore, due to its corrosive properties, human exposure to sodium aluminate by any route is unlikely to occur under normal working conditions as it should be avoided by implementation of the corresponding operational conditions and necessary risk management measures (use of PPE).

Effects on developmental toxicity

Description of key information

Based on the available information on soluble and insoluble aluminium compounds (read-across), there are no consistent indications for adverse developmental effects following repeated oral exposure to aluminium-containing substances. Effects reported include decreased pup survival/increased pup mortality, decreased growth, delayed maturation, and impaired neurodevelopment (see also Neurotoxicity for further details on the latter). Adverse developmental effects have been reported after subchronic and chronic oral exposure to aluminium dose levels usually higher than 100 mg Al/kg bw/day.
The most comprehensive available study however did not indicate a clear developmental toxic effect at dose levels up to 3225 mg/kg bw/day of aluminium citrate (equivalent to 300 mg Al/kg bw/day) in rats exposed via drinking water.
Due the corrosive properties of sodium aluminate, human exposure by any route is, however, unlikely to occur under normal working conditions as it should be avoided by implementation of the corresponding operational conditions and necessary risk management measures (use of PPE). Therefore, human exposure to aluminium dose levels associated with possible developmental effects is unlikely to occur by repeated exposure to sodium aluminate.

Additional information

A developmental neurotoxicity and chronic toxicity study similar to OECD guidelines 426 and 452 and in compliance with GLP was conducted with aluminium citrate (ToxTest.Research Council Inc., 2009). Pregnant Sprague-Dawley dams (n = 20 per group) were administered aqueous solutions of aluminium citrate at 3 dosage levels (nominal - 322.5, 1075 and 3225 mg/kg bw/day, equivalent to 30, 100 and 300 mg Al/kg bw/day, respectively). Two control groups received either a sodium citrate solution (citrate control with 27.2 g/L) 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. Pups were assigned to one of four cohorts (80 males, 80 females): a pre-weaning cohort that was sacrificed at PND 23, and cohorts that were sacrificed at PND 64, PND120 and PND 364.

Endpoints and observations in the dams included water consumption, body weight, a Functional Observational Battery (FOB), morbidity and mortality. Endpoints were assessed in both female and male pups that targeted behavioural ontogeny (motor activity, T-maze, auditory startle, the Functional Observational Battery (FOB) with domains targeting autonomic function, activity, neuromuscular function, sensimotor function, and physiological function), cognitive function (Morris swim maze), brain weight, clinical chemistry, haematology, tissue/blood levels of aluminium and neuropathology at the different dose levels and time points PND 23, 64, 120 and 364.

Statistical analyses were undertaken according to intention-to-treat, with appropriate consideration of multiple testing issues and, through the study design, also the unit of analysis. Censored analyses using survival analysis (Fixed Effects Partial Likelihood) were required for the grip strength measurements due to an equipment-defined maximum value. Females and males were analysed separately.

There were no significant Al-citrate treatment-related effects on mean body weights observed in the dams during the gestation and postnatal periods. The Na-citrate group, however, was significantly lighter than the control group on PND 15 (7.3%; p = 0.0316). Eight dams in the high dose aluminium group were found to have diarrhoea compared with none in the other treatment groups. The low and mid-dose Al-citrate groups consumed more water than the control group but the high dose group did not, suggesting that the effect was not simply due to treatment. There were no significant treatment-related differences in gestational length. There were no consistent treatment-related effects observed for the FOB tests in the dams. Due to the differences in water consumption, the % of target dose differed between groups and with time through the study. In the high dose group of dams, the actual dose (expressed as equivalent aluminium dose) during the first week of gestation was 200 mg Al/kg bw/day, 67% of the target dose (300 mg Al/kg bw/day). In the last week before weaning (and sacrifice), the actual dose received by the dams was close to 175% of the target dose. Statistical analyses comparing the actual doses received by the low, mid- and high- Al-citrate treatment groups showed that the order of the dose groups was maintained, however.

The most notable treatment-related effect observed in the offspring was renal pathology - hydronephrosis, ureteral dilation, obstruction and presence of calculi - most prominently in the F1 adult males. Higher mortality and significant morbidity were observed in the F1 adult males in the high dose group; leading to euthanization of this group at ca. study day 89. Clinical observations that showed a relationship with treatment, either directly or secondary to renal failure, were poor coat, weight loss, and haematuria. Diarrhoea was also observed. These signs were found only in the high dose Al-citrate treatment group. Haematuria was also observed in some animals in the Na-citrate group in the Day 364 cohort. Dosing with Al-citrate was associated with a reduction in body weight compared to controls. The results in the Day 364 cohort show a clear, consistent effect on post-weaning body weight in the high dose Al-citrate group in both male and females of the F1 generation. In the Day 120 cohort males, the mid-dose animals were significantly lighter than the controls. An effect of Na-citrate was observed in the female of the F1 generation in the Day 364 cohort. Overall, dosing of animals with aluminium citrate led to higher fluid consumption than in the control animals. Dosing with Na-citrate was associated with a significant increase in fluid consumption relative to that of the controls in most cohorts, with the exception of the Day 64 cohort females (fluid consumption was significantly lower in the Na-citrate group) and the Day 364 males (no significant difference between the two groups). The animals' fluid consumption varied with time and, in mature animals, was less than expected (120 mL/kg bw/day) with implications for the actual dosage of test item received. Despite the deviations from the target dose, the low-, mid- and high-dose groups showed the required trend of lowest to highest maintaining statistically significant group differences in dose levels. For most of the study period, the actual dose received was less than the target dose in all treatment groups.

In the female pups, the mean number of days to reach vaginal opening was 31.3 (± 2.1, sd) in the control group and 39.7 (± 5.6, sd) in the high dose Al-citrate group, a significant difference (p < 0.0001), but a similar difference was seen for the Na-citrate control group. In males, the mean number of days to reach preputial separation was 39.6 (±2.1, sd) in the control group and 42.5 (±3.2, sd) in the high dose group, also a significant difference in the pair-wise comparisons (p<0.0001);a similar delay was also observed in the Na-citrate control group. Delayed development of both male and female pups was observed in the high dose Al-citrate group and also in the Na-citrate group.Additionally, a litter effect was noted by the authors.The effect is considered treatment-related but whether the effect is secondary to decreases in body weight is not clear, however.In addition, as an effect was observed in the Na-citrate group, the role of aluminium in causing this developmental effect can neither be concluded nor excluded.

FOB observations showed no clear treatment-related effect among the neonatal Day 364 cohort pups that were assessed at PND 5 and 11 or in the juvenile pups assessedca.PND 22. In the adult F1 animals, the data provide little evidence for an Al effect on the autonomic function domain, the sensimotor function domain, or excitability. Significant wasting (physiological domain), was observed in the high dose females and appears related to treatment. However, a similar effect was observed in the Na-citrate controls as well.Characteristics of defecation (number of boluses) also showed differences with treatment. In addition, there was limited evidence of effects on activity/well-being of the pups at the high dose as reflected in fur appearance, deposits and rearing. There was some evidence for dose-response relationships between neuromuscular measurements – hind-limb and fore-limb grip strength and Al-treatment in both males and females, although some of the effects may be secondary to body weight changes. Furthermore, the FOB endpoint most consistently associated with Al-citrate treatment, grip strength, measurements showed considerably variability and a consistent ordering of the Al-treatment group responses (dose-response) was not observed at all time points. No consistent treatment-related effects were observed in ambulatory counts (motor activity) in the different cohorts. No significant effects were observed for the auditory startle response, T-maze tests (pre-weaning Day 23 cohort) or the Morris Water Maze test (Day 120 cohort).

Haematology parameters showed no significant treatment-related effects in the Day 23 cohort. In the Day 64 cohort, however, both males and females showed low grade microcytic anaemia (significantly lower mean cell volume, mean cell haemoglobin, and haematocrit). The anaemia had resolved by the end of the study in the Day 364 cohort females. Clinical chemistry results showed serum chemistry changes associated with aluminium toxicity such as elevated alkaline phosphatase and serum calcium. The authors state the levels still remained within the normal range. Effects were most pronounced in the Day 64 cohort animals. By Day 364 in the females, alkaline phosphatase levels did not differ significantly between the treatment groups.

Whole body Al levels in neonatal pups from high dose females and males were greater than those in the control groups. There were no significant sex differences. These results suggest transfer of Al from dams to pups in utero, although a contribution from breast milk PND 0 to 4 is also possible. Concentrations of Al in bone showed the strongest association with Al dose and some evidence of accumulation over time in all of the Al-treated groups. Of the central nervous system tissues, Al levels were highest in the brainstem. Although levels of Al were relatively low in the cortex (< 1µg/g), they were positively associated with Al levels in the liver and femur. In females, Al levels in the high dose group remained elevated relative to the other groups at all time points suggesting that accumulation might have occurred.

Pathological examinations showed clearly that urinary tract pathology was a treatment-related effect. The only other treatment-related effect reported on necropsy was watery, tan-coloured fluid in the digestive tract in some high dose animals, more frequently in the Day 64 group.None of the lesions seen on histopathological examination of brain tissues of the Day 364 group was treatment-related and, as these were also seen in the control group, were likely due to ageing.

This study has many strengths. It was conducted according to GLP with a design based on OECD TG #426. The study used adequate numbers of animals and randomization to reduce bias, assessed endpoints in both female and male offspring, and studied a wide range of neurotoxicity endpoints. Haematology, clinical chemistry, pathology and general toxicity endpoints were also assessed. Three dose levels were used although the highest was close to the MTD.Although representative of actual human exposures, extending the period of exposure beyond weaning until day 364 leads to ambiguity in interpretation of the results as effects observed later in the study may have resulted from either later exposures or exposures during periods critical for development. There were a number of deviations from protocol that are clearly described in the study report. Overall, these deviations were unlikely to have impacted the results of the study.

The results from this study are informative for neurotoxic effects due to combined prenatal and chronic postnatal exposure of rats to high doses of aluminium (30 mg Al/kg bw/day, 100 mg Al/kg bw/day and 300 mg Al/kg bw/day). As the offspring were dosed during the whole post-weaning period, it is difficult to differentiate between developmental or direct toxicity after weaning, however. Urinary tract pathology was observed in rats in the high dose group, more frequently and more severe in the males. The study showed no evidence of an effect of Al-citrate on memory or learning but a more consistent effect was observed in endpoints in the neuromuscular domain.

The ambiguity as to the critical period of exposure and the time-varying water consumption complicate the derivation of a point-of-departure from this study. A LOAEL for kidney toxicity and possible neurotoxicity (based on the effects on grip strength) of 1075 mg/kg bw/day of aluminium citrate (equivalent to 100 mg Al/kg bw/day) for aluminium toxicity was assigned. The critical neurological effect was a deficit in fore- and hind-limb grip strength in the mid-dose group, supported by evidence of dose response – although the results were variable and may be related to body weight effects as well –and observed effects in the mid-dose animals: urinary tract lesions at necropsy (4 males, 1 female); body weight (mid-dose males weighed less than controls in the Day 120 cohort); defecation (more boluses produced by females in the mid-dose group compared with the controls); urination (mid-dose males produced more urine pools than controls); tail pinch (mid-dose females displayed more exaggerated responses); foot-splay (mid-dose females had significantly narrower foot-splay than the controls); and the albumin/globulin ratio (Day 64 mid-dose males had a greater mean ratio than the controls).

Consequently, the NOAEL for neurotoxicity was 322.5 mg/kg bw/day of aluminium citrate (equivalent to 30 mg Al/kg bw/day). Furthermore, based on the lack systemic effects in general and of effects on gestational length in particular and the absence of clinical signs in maternal animals, 3225 mg/kg bw/day of aluminium citrate (equivalent to 300 mg Al/kg bw/day) was considered the maternal NOAEL for reproduction toxicity.

As no effects were observed in the F1 generation up to the juvenile age, the effects observed in this study appear to be related to repeated direct substance intake of the F1 generation rather than to intra-uterine or lactational exposure.

Sodium aluminate is stable only under alkaline conditions. At physiological pH, sodium aluminate rapidly decomposes to sodium and hydroxyl ions and various aluminium species, in particular insoluble compounds such as aluminium oxide (Al₂O₃) and hydroxide (Al(OH)₃). Therefore, studies investigating potential reproductive/developmental effects of aluminium hydroxide (Al(OH)₃) were taken into account, since this is a relevant aluminium species humans would likely be exposed to following handling of sodium aluminate in solutions at non-irritating concentrations.

Domingo et al. (1989)investigated the embryotoxic and teratogenic potential of Al(OH)₃administered orally to pregnant Swiss mice. Mated female mice (20 animals per group) were administered (oral, gavage) 0, 66.5, 133 or 266 mg Al(OH)₃/kg bw/day (equivalent to 23, 46, and 92 mg Al/kg bw/day) from gestation day 6 through 15. Dams were sacrificed on gestation day 18. No sign of maternal toxicity was observed in any group based on changes in maternal weight gain, food consumption and gross signs of abnormalities at post-mortem examination. The number of total implantations, the foetal sex ratio, body weights and lengths of foetuses were not significantly affected at any of the administered doses of aluminium hydroxide. The number of early resorptions/litter was increased in all Al (OH)₃treated groups (3.0 - in the 23 mg Al/kg group, 2.4 - in the 46 mg Al/kg group, and 1.3 – in the 133 mg Al/kg group versus 0.4 in the control group) and the number of live foetuses decreased in all groups (11.1 in the control group, 9.4 in the 23 mg Al/kh group, 9.2 in the 46 mg Al/kg group and 9.8 in the 92 mg Al/kg group) (n = 18-20 litters per group). Observed effects were not considered as treatment related effects as there was no dose-response relationship observed. The Al-treated foetuses did not exhibit any marked differences in external malformations, internal soft-tissue abnormalities or skeletal abnormalities compared to the controls. The NOAEL was considered to be 92 mg Al/kg based on lack of embryo/foetal toxicity or teratogenicity. The authors suggested that the lack of developmental toxicity of Al(OH)₃was likely due to lower gastrointestinal absorption of this compound compared with other forms of aluminium.

A similar study was conducted byGomez et al. (1990)in rats. Aluminium hydroxide was administered by gavage (2 times, daily) to pregnant Sprague-Dawley rats at dose levels of 192 (n = 18 animals per group), 384 (n = 18 animals per group) and 768 (n = 10 animals per group) mg/kg (equivalent to 66.5, 133 and 266 mg Al/kg bw/day, respectively) from day 6 through 15 of gestation. The animals were killed on day 20 of gestation. No adverse effects were reported on animal appearance, behaviour, maternal body weight, or absolute and relative organ weight (uterine, kidney and liver). No differences were observed for haematological and biochemical parameters but detailed results for these outcomes were not provided in the publication. Although not statistically significant, the incidence of early resorptions was higher in all Al (OH)₃-treated groups than in the control group(0.4 - in the 46 mg Al/kg group, 1.3 - in the 92 mg Al/kg group, and 0.6 – in the 266 mg Al/kg group versus 0.0 in the control group). Increased post-implantation loss (%) was observed compared to the control group(3.6 - in the 46 mg Al/kg group, 12.5 - in the 92 mg Al/kg group, and 5.0 – in the 266 mg Al/kg group versus 0.6 in the control group). Observed changes were not considered as treatment related effects because no relationship to dose was observed. Increased post-implantation loss (2.2 times compared to the control group) was observed only in the dose 92 mg Al/kg group. Statistically significant decrease in maternal food consumption was not associated with decreased maternal body weight and no dose-response relationship was found. No Al-treatment related effects were observed on critical gestational parameters such as number of litters, corpora lutea, number of total implantations, number of live foetuses, sex ratio, or foetal body weight at any dose administered. During foetal examination, no external and visceral abnormalities or skeletal malformation was detected. No significant differences in placental concentrations of aluminium were observed between the different groups. The NOAEL was considered to be 266 mg Al/kg bw/day based on lack of embryo/foetal toxicity or teratogenicity.

The influence of citric acid on the embryonic and/or teratogenic effects of high doses of Al(OH)₃in rats was investigated byGómez et al. (1991). Three groups of pregnant rats were administered daily doses (gavage) of Al(OH)₃(384 mg/kg bw/day, equal to 133 mg Al/kg bw/day , n = 18), aluminium citrate (1064 mg/kg bw/day, n = 15), or Al(OH)₃(384 mg/kg bw/day, equal to 133 mg Al/kg bw/day) concurrently with citric acid (62 mg/kg bw, n = 18) on gestational days 6 to 15. A control group received distilled water during the same period (n = 17). There were no treatment-related differences on critical gestational parameters such as numbers of litters, corpora lutea, number of total implantations, number of live foetuses, sex ratio, or foetal body weight in the group treated with Al(OH)₃. No external and visceral abnormalities or skeletal malformation were detected on foetal examination. Maternal and foetal body weights were significantly reduced, the number of foetuses with delayed sternabrae and occipital ossification was significantly increased (p<0.05), the number of foetuses with absence of xiphoides was increased in the group treated with Al(OH)₃and citric acid as compared to the control group. No significant differences in the number of malformations were detected between any of the groups (authors did not provide the quantitative data).

Colomina et al. (1992)evaluated the influence of lactate on developmental toxicity attributed to high doses of Al(OH)₃in mice. Oral (gavage) daily doses of Al(OH)₃(166 mg/kg bw, n = 11), aluminium lactate (627 mg/kg b, n = 10), or Al(OH)₃(166 mg/kg bw) with lactic acid (570 mg/ kg bw, n = 13) were administered to pregnant mice from gestational day 6 to 15. An additional group of mice received lactic acid alone (570 mg/kg bw). A control group (n = 13) received distilled water during the same period. No signs of maternal toxicity (no statistically significant changes in food consumption, maternal body and organ weight) were observed in the dams treated with Al(OH)₃. No statistically significant treatment-related differences oncritical gestational parameters such as number of litters, corpora lutea, number of total implantations, number of live foetuses, sex ratio, or foetal body weight were observed in the Al (OH)₃-treated group and no external abnormalities or skeletal malformation were detected on foetal examination. However, aluminium concentrations were significantly higher in the bones of dams, and aluminium was detected in the whole foetus of the Al(OH)₃-treated animals. Concurrent administration of Al(OH)₃and lactic acid resulted in significant reductions in maternal weight compared to the control group. In the group given lactate only, aluminium was detected in whole foetuses; however, this was not statistically different from the mean level found in the control group. Aluminium lactate administration resulted in significant decreases in maternal body weight and food consumption, foetal body weight accompanied by increases in the incidence of cleft palate. Delayed ossification was also observed in the aluminium lactate-treated animals. Although not statistically significant, the incidence of skeletal variations was higher in the group concurrently administered Al(OH)₃and lactic acid than in the control group. No other signs of developmental toxicity were detected in the Al (OH)₃and lactic acid group.

In a similar experiment,Colomina et al. (1994)assessed the effect of concurrent ingestion of high doses of Al(OH)₃and ascorbic acid on maternal and developmental toxicity in mice. Three groups of pregnant mice were given daily doses (gavage, 2 times daily) of Al(OH)₃(300 mg/kg bw or 103.8 mg Al/kg), ascorbic acid (85 mg/kg bw), or Al(OH)₃concurrent with ascorbic acid (85 mg/kg bw) from gestational day 6 to day 15. A fourth group of animals received distilled water and served as the control group. The animals were killed on gestation day 18. The number of litters, corpora lutea, number of total implantations, number of live foetuses, sex ratio, and foetal body weight did not differ between the control and Al(OH)₃-treated groups. No external and visceral abnormalities or skeletal malformations were detected on foetal examination. Placenta and kidney concentrations of aluminium were significantly higher in mice receiving Al(OH)₃and Al(OH)₃plus ascorbic acid than in controls. No information was provided on the number of dams and litters in the Al-treated and control groups.

In addition to the comprehensive developmental neurotoxicity and chronic toxicity study (ToxTest. Alberta Research Council Inc., 2009) and the available studies on animals exposed to aluminium hydroxide described above, a large number of studies have examined the developmental toxicity of aluminium in rats and mice. A variety of effects have been reported, but were not consistent, including decreased pup survival/increased pup mortality, decreased growth, delayed maturation, and impaired neurodevelopment (see Discussion on Neurotoxicity for further details on the latter) (ATSDR, 2008).

Increases in pup mortality, typically occurring within the first 4 postnatal days, were observed in rats exposed to 155 mg Al/kg/day as aluminium chloride in the diet on gestational days 8-20 (Bernuzzi et al., 1986). Another study found a decrease in the number of live pups per litter and an increase in the number of dead young per litter on PND 21 in the offspring of rats administered via gavage 51 mg Al/kg/day as aluminium nitrate for 14 days prior to mating, on gestation days 1-20, and lactation days 1-21 (Domingo et al., 1987). The gavage administration route may have influenced the results of this study; other studies involving exposure to aluminium nitrate, aluminium citrate, or aluminium lactate via drinking water or diet have not reported increases in mortality at doses as high as 330 mg Al/kg/day as aluminium lactate in the diet on gestation days 1 through PND 35 (Colomina et al., 2005; Golub et al., 1992, 1995, 2001; McCormack et al., 1979).

Numerous studies have reported decreases in pup body weight gain (Bernuzzi et al., 1986,; Colomina et al., 2005; Domingo et al., 1987a, 1987b; Golub and Germann 2001; Golub et al., 1992; Misawa and Shigeta, 1992; Paternain et al., 1988; Sharma and Mishra 2006). Since some of these studies did not report the aluminium content of the basal diet, their usefulness in establishing dose-response relationships is limited. With few exceptions, most studies have shown that aluminium does not adversely affect birth weight in the absence of effects on maternal body weight (Colomina et al., 2005; Donald et al., 1989; Golub and Germann 2001; Golub et al., 1992, 1995; McCormack et al., 1979). The possible exception to this finding was decreases in birth weight observed in the offspring of rats administered aluminium nitrate via gavage at doses of ≥ 38 mg Al/kg/day on gestation day 1 through lactation day 21 (Domingo et al., 1987b) or 77 mg Al/kg/day on gestation day 14 through lactation day 21 (Domingo et al., 1987a); neither study reported whether there were significant effects on maternal body weight gain. Paternain et al., (1988) also reported a decrease in pup body weight in rats receiving gavage doses of 38 mg Al/kg/day as aluminium nitrate on gestation days 6-14; a decrease in maternal weight gain was also reported at this dose level.

Although most studies did not find effects on birth weights, several studies did find decreases in post-birth pup body weights; however, this finding was not consistent across studies. Lower pup body weights starting on PND 10 were observed in mouse pups exposed to aluminium during gestation only, during lactation only, or during gestation and lactation (Golub et al., 1992); a decrease in maternal body weight gain was observed in the dams exposed during lactation. This study suggests that aluminium may influence growth directly and may not be only related to changes in maternal body weight during lactation. Similarly, decreases in body weights were observed on PND 12, 16, and 21 in the pups exposed to 100 mg Al/kg/day as aluminium nitrate in the drinking water (with added citric acid) on gestation day 1 through lactation day 21; a decrease in maternal food and water intake was also observed at this dose level (Colomina et al., 2005). A third study found decreases in pup body weight at PND 21 in mice exposed to 130 mg Al/kg/day as aluminium lactate in the diet on gestation day 1 through PND 35 (Golub and Germann 2001). The lower body weights were still present at 5 months of age even though aluminium exposure was stopped on PND 35; an increase in food intake was also observed in these animals. In contrast to these studies, no adverse effects on body weight were observed in mouse pups exposed to 330 mg Al/kg/day as aluminium lactate in the diet on gestation day 1 through PND 21 or 35 (Donald et al., 1989; Golub et al., 1995).

Gestational exposure to aluminium does not appear to result in an increase in the occurrence of malformation and anomalies, although reductions in ossification were observed (Sharma and Mishra 2006). Delays in ossification were observed at doses that also resulted in decreases in pup body weight. Some alterations in physical maturation were observed in rats exposed to aluminium nitrate in drinking water (with added citric acid) on gestation day 1 through lactation day 21 (Colomina et al., 2005). The observed effects included significant delay in vagina opening at 50 or 100 mg Al/kg/day, testes descent at 100 mg Al/kg/day, and incisor eruption in males at 50 mg Al/kg/day.

No effects on days to pinna detachment or eye opening were observed. No delays on pinna detachment, eye opening, or incisor eruption were observed in rats administered via gavage 73 mg Al/kg/day as aluminium chloride (aluminium content of the diet not reported) on gestation days 8-20 (Misawa and Shigeta, 1992).

Justification for classification or non-classification

Corrosivity is the dominating effect of sodium aluminate. Therefore, it is unlikely that exposure to aluminium dose levels related to developmental effects (mostly > 100 mg Al/kg bw/day via the oral route) would be achieved by exposure to sodium aluminate via any route. Classification of sodium aluminate for reproductive toxicity including effects on or via lactation is not necessary.