Ammonium hydrogendifluoride

Toxicological Summary

Workers - Hazard via inhalation route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

2.3 mg/m³

Most sensitive endpoint:

repeated dose toxicity

DNEL related information

Modified dose descriptor starting point:

NOAEC

Acute/short term exposure

Most sensitive endpoint:

irritation (respiratory tract)

DNEL related information

Modified dose descriptor starting point:

NOAEC

Local effects

Long term exposure

Most sensitive endpoint:

irritation (respiratory tract)

Acute/short term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

3.8 mg/m³

Most sensitive endpoint:

irritation (respiratory tract)

DNEL related information

Dose descriptor starting point:

NOAEC

Workers - Hazard via dermal route

Systemic effects

Long term exposure

Hazard assessment conclusion:

exposure based waiving

Most sensitive endpoint:

skin irritation/corrosion

Acute/short term exposure

Hazard assessment conclusion:

exposure based waiving

Most sensitive endpoint:

skin irritation/corrosion

DNEL related information

Local effects

Long term exposure

Hazard assessment conclusion:

no-threshold effect and/or no dose-response information available

Most sensitive endpoint:

skin irritation/corrosion

Acute/short term exposure

Hazard assessment conclusion:

no-threshold effect and/or no dose-response information available

Most sensitive endpoint:

skin irritation/corrosion

Additional information - workers

Background: read-across justification

Limited toxicological data are available for the substance AMBI (ammonium hydrogendifluoride). AMBI will dissociate under physiological conditions with the generation of hydrogen, fluoride and ammonium ions; read-across is therefore proposed to studies with soluble fluoride compounds (hydrogen fluoride, sodium fluoride); ammonia and ammonium salts. The toxicity of AMBI can therefore be characterised by read-across to these supporting substances and without the need for additional testing.

Toxicokinetics

The basic toxicokinetics of the ionic species fluoride and ammonium are well characterised and are summarised separately.

Ammonia (ammonium): basic toxicokinetics

Ammonia is generated by the bacterial flora of the gastrointestinal tract (~4 g/day) and as a very small water-soluble molecule, is likely to be rapidly and extensively absorbed. The results of a study in the rat (Schaerdel et al, 1983) indicate that gaseous ammonia is absorbed into the bloodstream following inhalation exposure; this is consistent with its water solubility and small molecular size. Ammonium is distributed to all tissues in the body and is capable of crossing the blood-brain barrier. No studies of metabolism are available, however the physiological role of ammonia as a product of normal metabolism (protein catabolism) is very well characterised. Ammonia is rapidly detoxified in the liver by the urea cycle. The urea cycle consists of five enzymes: carbamoylphosphate synthetase I (CPS I), ornithine transcarbamylase (OTC), argininosuccinate synthetase (ASS), argininosuccinate lyase (ASL) and arginase. The initial reaction of the urea cycle is the formation of carbamoyl phosphate from ammonia and bicarbonate, a reaction catalysed by CPS I, which requires N-acetylglutamate as an allosteric cofactor. Condensation of carbamoyl phosphate with ornithine yields citrulline (by OTC); this in turn condenses with aspartate to give argininosuccinate (by ASS), a reaction that requires the cleavage of two further high-energy phosphate bonds. Argininosuccinate is hydrolysed to fumarate and arginine (by argininosuccinase). Arginine is cleaved by arginase to give urea and ornithine. OTC, like CPS I, is also a major mitochondrial protein; the remaining enzymes are in the cytoplasm of hepatocytes. Urea synthesis cannot therefore be saturated at realistic substrate concentrations. The urea cycle is high capacity and the capacity also increases short-term and long-term in reponse to increased demand (e.g. in response to dietary protein intake). It is estimated that some 7 to 25% of the ammonia delivered via the portal vein escapes periportal urea synthesis and is used for glutamine synthesis. Ammonia is rapidly detoxified in mammals by conversion to urea by the urea cycle in liver cells and is subsequently excreted (as urea) in urine following glomerular filtration. Ammonium ions (NH4 +) are also excreted by the kidney. Hepatic excretion of urea (15 -30% of that generated) results in the generation of ammonia by the gastrointestinal flora, and subsequent re-absoprtion.

The body excretes approximately 30 g urea/day; urea is synthesised by the hepatic urea cycle from ammonia generated by protein catabolism. Therefore it can be calculated that the body typically produces 16 g/day ammonia, although this figure is influenced by the amount of dietary protein. Therefore it can be concluded that exposure to ammonia at quantities in this range will be without toxicological effect. Ammonia is toxic and therefore exposures greatly exceeding the normal production may overwhelm the normal detoxification mechanisms. The normal serum level of ammonia is 0.15 -100 ug/dL. Assuming a blood volume of 5L (55% serum), this is equivalent to approximately 0.4 -2.75 mg ammonia present in the blood of a normal adult at any point in time. However it is notable that, due to the generation of ammonia by gastrointestinal tract bacteria, the concentration of ammonia in the hepatic portal circulation is much higher than that in the rest of the systemic circulation.

Fluoride: basic toxicokinetics

The absorption of inorganic fluoride across mucous membranes is passive and is independent of the fluoride source. Inhalation absorption of fluoride is likely to be extensive. Following dermal exposure to AMBI, absorption is of fluoride likely to be minimal expect in cases where the normal skin structure is compromised as a consequence of the corrosive effects of the substance. The absorption of fluoride following oral administration of AMBI has not been investigated, but is likely to be rapidly absorbed.  Absorbed fluoride is distributed primarily in the blood, 75% in the plasma and 25% associated with erythrocytes. Half of the plasma fluoride may be bound to organic molecules. Fluoride is rapidly distributed and is sequestered in the bones and teeth, where exchange with hydroxyl groups results in incorporation into the bone and tooth structure. Levels of fluoride in bones and teeth are directly correlated with exposure levels. Fluoride is excreted rapidly as a consequence of glomerular filtration, with a plasma half-life of 2-9 hours. The half-life for skeletal fluoride in humans is reported to be 8-20 years.

Dermal absorption

Following dermal exposure to AMBI, absorption of fluoride is likely to be minimal expect in cases where the normal skin structure is compromised as a consequence of the corrosive effects of the substance. A number of human case studies with related compounds such as HF have noted elevated plasma fluoride concentrations and/or systemic fluoride poisoning in burns cases, indicating the ability of fluoride to be absorbed through damaged skin. Ammonia is a small water soluble molecule; high water solubility indicates that significant dermal absorption is unlikely under normal conditions. However the substance is corrosive, and it is possible that significant absorption may occur in situations where the integrity of the skin barrier is compromised by severe local reactions, such as in burns cases.

Acute toxicity

The acute oral LD50 of the substance in the rat is reported to be 130 mg/kg bw (Musch & Hoffer, 1990). Ammonium hydrogendifluoride is listed on Annex I of Directive 67/548/EEC with classification as Corrosive (R34) 'Causes burns'. Testing for acute dermal and inhalation toxicity is not required and cannot be justified for reasons of animal welfare.

Irritation / corrosion

No data are available: waivers are proposed for skin irritation and eye irritation endpoints based on the existing R34 classification of the substance. The substance dissolves readily in aqueous (i.e. physiological) environments to produce hydrofluoric acid which is likely to be responsible for the corrosive effects.

Sensitisation

 

No data are available: a waiver is proposed. The substance is listed on Annex I of Directive 67/548/EEC with classification as (R34) 'Causes burns'. Testing for skin sensitisation is not justified on scientific grounds or for reasons of animal welfare. The local dermal effects of the substance will be dominated by irritation/corrosion and sensitisation is considered to be unlikely. Additionally it is noted that there is no evidence that ammonia/ammonium or fluoride are skin sensitisers.

No data are available for AMBI, however comprehensive read-across data are avaialable for other soluble inorganic fluoride salts, ammonia and ammonium salts.

Repeated dose toxicity

Read-across data for fluoride compounds

Repeated dose oral toxicity

No studies have been performed with the substance, however comprehensive data are available for the read-across substance sodium fluoride. The repeated dose oral toxicity of ammonium hydrogendifluoride and NaF are considered to be essentially identical, with the exception of likely irritant/corrosive effects of ammonium hydrogendifluoride at high dose levels. With the exception of corrosivity, the repeated dose oral toxicity of ammonium hydrogendifluoride will be due to fluoride, therefore read-across from the comprehensive NTP dataset with the soluble salt NaF is appropriate. In a 14 -day range-finding study with NaF in the rat, mortality was seen at drinking water concentrations of 400 and 800 ppm. Signs of toxicity (reduced weight gain, reduced water consumption, lethargy and dehydration) were noted in surviving animals in these groups. The NOAEL for this study was 200 ppm. In a 14-day range-finding study in the mouse, mortaility was seen at the highest dose level of 800 ppm; signs of toxicity (reduced weight gain, abnormal gait and posture, reduced water consumption) were also apparent at this dose level. A NOAEL of 400 ppm is determined for this study. In a 6 -month rat study, the effects of exposure to NaF were limited to reduced weight gain, dental fluorosis, thickening and ulceration of the gastric mucosa at the highest dose level of 300 ppm; gastric effects were also seen at 100 ppm. The fluoride content of plasma, bone and teeth increased with dose levels. The NOEL for this study was 30 ppm, however these local effects are not considered to be relevant for the risk assessment therefore a NOAEL of 100 ppm can be determined. In a 6 -month nouse study, mortality attributable to acute nephrosis was seen at the highest dose level of 600 ppm. Skeletal effects were seen in males at the lowest dose level of 50 ppm.

Repeated dose dermal toxicity

No studies are available. The effects of dermal exposure will be dominated by local irritation / corrosion. There is no evidence of significant dermal absorption of AMBI under exposure conditions where the integrity of the skin barrier is maintained. Testing for repeated dose dermal toxicity can therefore be waived on scientific grounds and for reasons of animal welfare. The effects of repeated inhalation exposure to the read-across substance hydrogen fluoride (HF) have been adequately characterised; the effects of repeated exposure to fluoride are also well characterised.

Repeated exposure inhalation toxicity

In a published study (Sadilova et al, 1974), female rats were exposed to 1 mg/m3 HF 6 hours/day for 1 month. Effects were noted on the teeth, bones and respiratory tract. Two proprietary studies (Plackeet al;1990, 1991) are summarised in the EU RAR, however the data owner is unknown and therefore there is no access to these studies. The EU RAR summary states an overall NOAEL for repeated inhalatory exposure in male and female rats was identified as 0.72 mg/m³(actual HF concentration) for a 6 hours per 5 days per week for 91 days exposure regimen. No adverse effects were noted at this concentration. At higher concentrations death, tissue irritation, dental malformations, haematological and biological changes and changes in several organ weights were observed.

Summary

Effects of repeated fluoride exposure in experimental animals were seen on the teeth, bones, respiratory tract and kidney. Evidence from epidemiological studies in humans also indicate that prolonged exposure to fluoride causes dental and skeletal effects.

Read-across data for ammonia and ammonium compounds

Repeated dose oral toxicity

A 90 -day study in the rat with ammonium sulphate showed only minor effects at high dose levels (diarrhoea, renal pathology); a NOAEL of 886 mg/kg bw/d was determined, equivalent to 225 mg/kg bw/d ammonia (Tagaki et al, 1999). A NOAEL of 256 mg/kg bw/d (equivalent to 67 mg/kg bw/d) was determined for 1 -year and 2 -year studies by the same group (Ota et al, 2006).

Repeated dose dermal toxicity

No data are available.

Repeated exposure inhalation toxicity

A number of non-standard studies of various duration and in different species are available. The data indicate that the primary effect of exposure to inhaled anhydrous ammonia is local irritation of the respiratory tract. In a 5-week study in pigs, ammonia concentrations had a highly significant adverse effect upon feed consumption and average daily weight gain. However, there was no significant effect upon efficiency of food conversion. During both trials the high ammonia levels appeared to cause excessive nasal, lacrimal and mouth secretions. This was more pronounced at 100 and 150 ppm than at 50 ppm. Autopsies carried out on three animals showed no significant gross or microscopic differences related to ammonia level. Cultures of Corynebacterium and Pasteurella were obtained from swabs of the ethmoid turbinates from two animals removed from the compartment maintained at 150 ppm and one animal maintained at 100 ppm. There was no evidence of these bacteria in turbinate swabs from other animals (Stombaugh et al, 1969). Sherman and Fischer rats were exposed to environmental ammonia, derived from natural sources for 75 days, or to purified ammonia for 35 days. Rats were either inoculated intranasally withM. pulmonisprior to exposure, or left untreated. The average ammonia concentrations were 105 mg/m³for 75 days and 175 mg/m³for 35 days exposure. Ammonia exposure (from either source) significantly increased the severity of the rhinitis, otitis media, tracheitis and pneumonia (including bronchiectasis) characteristic of murine respiratory mycoplasmosis (rats infected withM. pulmonis). The prevalence of pneumonia showed a strong tendency to increase directly with environmental ammonia concentration (Brodersonet al,1976). Twenty seven male rats, along with 27 age and weight matched controls, were exposed to atmospheric ammonia gas at a concentration of 350 mg/m³for up to 8 weeks. The rats were sacrificed after different exposure times. Nasal irritation began on the fourth day. After 3 weeks continuous exposure exposed rats showed nasal irritation and inflammation of the upper respiratory tract. The number of pulmonary alveolar macrophages was similar to that in the controls. After 8 weeks none of the inflammatory lesions were present (Richardet al,1978). Weatherby (1952) exposed twelve male guinea pigs (plus 6 controls) were exposed to anhydrous ammonia gas for up to 18 weeks (6 hours per day, 5 days per week). The average concentration in air was 119 mg/m³. Four experimental and 2 control animals were sacrificed at 6 week intervals throughout the study. There were no significant findings at necropsy after 6 and 12 weeks exposure. In animals sacrificed after 18 weeks, there was mild congestion of the liver spleen and kidneys, with degenerative changes in the adrenal glands, and hemosiderosis in the spleen indicating hematotoxicity. There was cloudy swelling in the epithelium of the proximal tubules of the kidney as well as albumin precipitation in the lumen with some casts.

In a 50-day study (Stolpe & Sedlag, 1976), male Wistar rats were exposed to two concentrations of ammonia gas, continuously for 50 days. Concurrent controls remained untreated. There was no mortality at either concentration (35 or 63 mg/m³), and no treatment-related clinical effects were observed. Body weight gain and food intake, as compared to control values, was not significantly affected by ammonia exposure. At 63 mg/m³rats showed increased haemoglobin and haematocrit levels compared to controls. The NOAEC was 35 mg/m³

 

Conclusion of read-across studies

AMBI will dissociate under physiological conditions to form fluoride and ammonium ions. The effects of repeated exposure to AMBI at high concentrations will be dominated by local corrosivity and irritation. Ammonium is of relatively low systemic toxicity, therefore the critical systemic toxic effects of AMBI will be due to fluoride.

Genetic toxicity

Genetic toxicity in vitro

No evidence of mutagenicity was seen in a guideline-compliant GLP Ames tests performed with AMBI (Herbold, 1988). No evidence of mutagenicity was seen with sodium fluoride in an Ames test (NTP, 1990). No evidence of mutagenicity was seen in a mammalian cell mutation assay (V79/HPRT) with sodium fluoride. This study was performed only in the absence of metabolic activation, however this deviation is not considered to be critical as the test substance is not metabolised. A positive result with sodium fluoride is reported in a mouse lymphoma assay (NTP, 1990). Sister chromatid exchange and chromosomal aberrations are reported in an additional NTP study performed with sodium fluoride.

No evidence of mutagenicity was seen in a guideline-comparable Ames test performed with anhydrous ammonia (Shimizu et al,1985). Similarly, there was no evidence of mutagenicity in a non-standard study using E. coli (Szybalski, 1958).

Genetic toxicity in vivo

Gerdes (1971) reports a marginally (but not statistically significant) postive response to HF in a study in Drosophila melanogaster; positive effects in Drosophila are also reported by Mohamed et al(1971). The significance of these results is unclear; the EU RAR for HF considers the findings of these studies to be inconclusive. Zeiger et al(1994) report no evidence of clastogenicity, even at dose levels causing severe toxicity, in a well-conducted mouse study performed with sodium fluoride in which chromosomal aberrations and micronucleus formation was assessed. In contrast, a poorly reported inhalation exposure study performed with HF (Voroshilinet al, 1975) reports clastogenicity in the bone marrow of exposed rats but no dominant lethal effect in exposed mice. No evidence of an increase in the incidence of micronucleated polychromatic erythrocytes was seen in a mouse micronucleus assay performed with the read-across compound ammonium chloride (Hayashiet al, 1988).

Genetic toxicity: conclusion

The EU RAR concludes that, while the dataset on the genotoxicity of the read-across substance HF is limited, studies with sodium fluoride are also informative as for all substances target tissues will exposed to fluoride (either free or bound to organic molecules). The EU RAR therefore reviews the available data for NaF and HF and concludes that fluoride does not interact directly with DNA and is not genotoxic when administered via an appropriate route (i.e. by oral or inhalation exposure).

Ammonia is a simple molecule and does not possess any structural alerts for genotoxicity. Ammonia is present at relatively low levels in the systemic circulation as a consequence of protein catabolism (largely in the liver) and is also present at higher levels in the hepatic portal circulation due to the breakdown of urea by gastrointestinal bacteria. The ubiquitous presence of ammonia in the leads to the conclusion that it is unlikely to be genotoxic. The WHO evaluation (EHC 54, 1986) concludes that there is no evidence that ammonia is mutagenic in mammals. A UK Health Protection Agency (HPA) evaluation similarly concludes that ammonia does not have significant mutagenic potential.

 

 

Carcinogenicity

No studies are available for AMBI. High quality NTP studies in the rat and mouse are available for the read-across substance, sodium fluoride. Data are also available for ammonia and ammonium salts. The effects of chronic AMBI exposure will be dominated by local effects at the site of contact (irritation/corrosion), therefore performing studies cannot be supported for scientific reasons and also on animal welfare grounds. Once absorbed into the body, AMBI will dissociate into its constituent ions and systemic toxicity will be due to fluoride. The analogous behaviour of sodium fluoride (or any other water-soluble fluoride salts) means that read-across from NaF to AMBI is scientifically justified.

Studies in the rat: sodium fluoride

The NTP rat study showed evidence of an effect of sodium fluoride administration on the bones and teeth, consistent with the findings of other studies. There was no effect on survival in this study; bodyweights, food and water consumption, haematological and clinical chemistry paramaters and organ weights were unaffected by treatment. Serum, urine and bone fluoride concentrations were increased in all treated groups; the urine calcium concentration was also marginally higher in females at the highest dose level. Osteosclerosis was seen in females at the highest dose level. The incidence of osteosarcoma was increased in males at the intermediate dose level (2%) and the high dose level (4%) but was within the historical range (0 -6%; mean 0.5%). The NTP concluded that the study provides 'equivocal evidence' for carcinogenicity in male rats. An additional carcinogenicity study with NaF in the rat is available (Maureret al,1990). No evidence of carcinogenicity was seen in this study, at dose levels sufficient to cause toxicity.

Studies in the mouse: sodium fluoride

The NTP mouse study showed evidence of an effect of sodium fluoride administration on the teeth, consistent with the findings of other studies. There was no effect on survival in this study; bodyweights, food and water consumption, haematological parameters and organ weights were unaffected by treatment. Clinical chemistry revealed elevated ALP activity in females at the highest dose level. Microscopic findings were limited to dentine dysplasia in male mice at 175 ppm. There was no evidence of carcinogenicity in either sex. An additional carcinogenicity study with NaF is available (Maureret al,1993). A high level of osteosarcomas was seen in all (control and treated) groups in this study, a finding which was attributed to infection with a retrovirus. No conclusion on the carcinogenicity of sodium fluoride can be drawn from this study.

Ammonia studies

Ammonia/ammonium is present in the body as a normal product of metabolism (protein catabolism). It is therefore considered unlikely that this simple molecule is carcinogenic. Nevertheless some data are available and are evaluated.

No evidence of carcinogenicity was seen in a rat dietary study with ammonium sulphate (Ota et al, 2006). The NOAEL for this study was 0.6% (dietary level) equivalent to 256 and 284 mg/kg bw/day in males and females respectively [67 and 74 mg/kg bw/d ammonia equivalents]. In a non-standard mechanistic assay, Tsujiet al (1992) exposed MNNG-initiated rats to 0.01% ammonia solution via drinking water. Gastritis was seen in all animals, indicating a local irritant effect. The incidence of gastric tumours was increased in treated animals, suggesting that ammonia may be acting as a promoter of carcinogenesis. Solutions of hydrazine as 0.001%, methylhydrazine as 0.01%, methylhydrazine sulfate as 0.001%, and ammonium hydroxide as 0.3, 0.2 and 0.1% were administered continuously in the drinking water of 5- and 6-week-old randomly bred Swiss mice for their entire lifetime. Similarly ammonium hydroxide as a 0.1% solution was given to 7-week-old inbred C3H mice. Hydrazine and methylhydrazine sulfate significantly increased the incidence of lung tumors in Swiss mice, while methylhydrazine enhanced the development of this neoplasm by shortening its latent period. The ammonium hydroxide treatments in Swiss and C3H mice were, however, without carcinogenic effect, and did not inhibit the development of breast adenocarcinomas in C3H females, which are characteristic of these animals. The present study thus proves for the first time the carcinogenicity of methylhydrazine, provides further evidence of the tumor-inducing capability of hydrazine by itself and negates the possibility that the metabolite of hydrazine, ammonium hydroxide, could interfere in the development of neoplasia (Toth, 1972).

Reproductive toxicity

Under physiological conditions, AMBI will dissociate with the production of fluoride and ammonium ions. No studies of reproductive toxicity are available for AMBI, however read-across to other soluble fluorides, ammonia and soluble ammonium salts is appropriate

Fluorides: published studies

Araibi et al(1989) report adverse effects on the fertility of male rats administered sodium fluoride in the diet at concentrations of 100 and 200 ppm. Exposure resulted in a reduction in succesful matings and reduced litter size; findings were associated with a reduction in seminiferous tubule diameter and a thickened peritubular membrane. The numbers of tubules containing spermatozoa were decreased and serum testosterone levels were also reduced. Chinoy & Sequeira (1989) report alterations in the histoarchitecture of the testes in mice gavaged with sodium fluoride at dose levels of 10 and 20 mg/kg bw/d for 30 days. Findings were characterised by severe disorganisation and denudation of germinal epithelial cells of the seminiferous tubules, absence of sperm from the tubular lumen, reduced in epithelial cell height, nuclear pkynosis, denudation of cells and absence of sperm occurred in the cauda epididymis. The effects seen after 30 days administration were reversible. Chinoyet al(1992) report reduced fertility in male rats administered sodium fluoride by gavage at dose levels of 5 and 10 mg/kg bw for 30 days. Findings were accompanied by reduced sperm count and motility and various biochemical changes in the testes. The results of these studies are consistent, however their value and reliability is significantly compromised by the absence of any information on the fluoride levels in diet and/or drinking water. The actual levels of fluoride exposure cannot be accurately assessed. It is also notable that the findings of these published investigative studies of non-standard design contrast with the total absence of reproductive toxicity at comparable dose levels in the FDA studies reported below. Messer et al(1973) investigated the reproductive toxicity of sodium fluoride in a two-generation study in which female mice were administered the test material in the drinking water at dose levels of 0, 50, 100 or 200 ppm. A progressive decline in litter production was seen in the control group. All females administered 200 ppm fluoride died over the study period; only a small number of litters were produced at the 100 ppm. It is suggested that a level of 50 ppm sodium fluoride (equivalent to approximately 7.5 mg/kg bw.d fluoride) is required to maintain reproductive capacity in female mice. In a 3 -generation mouse study (Tao & Suttie, 1976), no effects of fluoride on reproduction were seen. The study is of limited value, however the authors suggest that the effects of fluoride seen in the study of Messeret al(1973) was due to the influence of fluoride on teh absorption of iron from a low iron diet.

Fluorides: FDA studies

The effects of sodium fluoride administration on spermatogenesis in rats were investigated in a two-generation study (Sprando et al, 1997). In contrast to the previous studies, no effects were observed on reproductive organ weights, sperm parameters or biochemical parameters at dose levels of up to 250 ppm (drinking water). Additional deatiled investigations by the same authors did not reveal any effects on spermatogenesis in F1 males (Sprandoet al, 1998). No effects on reproduction were seen at the highest dose level of 250 ppm in a guideline-comparable two-generation rat study (Collinset al, 2001). In a further FDA study designed primarily to assess the potential effects of fluoride on spermatogenesis (as indiceted in various published studies), Sprando et al (1996) demonstrated that injection of sodium fluoride into the rat testis was without effect on spermatogenesis.

In contrast to the other studies which report effects of fluoride on male fertility and spermatogenesis, no effects were observed in the FDA studies following extensive investigation. The two-generation FDA study is of standard design and is comprehensively reported, and it is notable in these studies that thr contribution of diet and drinking water to the total fluoride intake was assessed. The EU RAR for HF also considers the data available for the reproductive toxicity of NaF and concludes that the FDA studies are key, for reasons of design, reporting and control of fluoride levels. The EU RAR concludes that the NOAEL for reproductive toxicity is 250 ppm NaF, which corresponds to approximately 10 mg/kg bw/d fluoride. The absence of any apparent effects on the reproductive organs in chronic toxicity and carcinogenicity studies is also notable.

Ammonia/ammonium studies

A guideline-comparable two-generation study with ammonium perchlorate did not identify any effects on reproductive parameters at dose levels of up to and including 100 mg/kg bw/day. The study did identify effects on the parental thyroid associated with perchlorate exposure, however findings are not attributable to ammonium. The results of the study therefore confirm that exposure to ammonium is not associated with reproductive toxicity (York et al, 2001). There is no evidence that exposure to ammonium ions causes reproductive toxicity. Inhalation exposure to ammonia will result in an equilibrium in the blood (at physiologically relevant pH) between non-ionised ammonia (NH3) and ionised ammonium (NH4+) in a ratio of approximately 1:100. The same equilibrium will exist in animals orally exposed to ammonium salts, therefore read-across is appropriate. Human maternal blood contains measurable levels of ammonia as a consequence of protein catabolism; the blood in the hepatic portal circulation contains much higher levels of ammonia due to its generation from urea by the gastrointestinal flora. Ammonia is rapidly and effectively detoxified in the liver by the urea cycle and also via additional pathways, therefore will not accumulate and is unlikely to cause any reproductive effects.

Developmental toxicity

Fluoride studies

In a rat developmental toxicity study (NTP, 1994; Heindelet al, 1996), maternal toxicity (transiently reduced bodyweight gain) was apparent at the highest dose level of 300 ppm sodium fluoride (in drinking water), equivalent to 13 mg/kg bw/d fluoride. No evidence of developmental toxicity was seen at this dose level. No clear evidence of developmental toxicity was seen in an FDA rat study (Collinset al,1995) at dose levels of up to 250 ppm sodium fluoride in drinking water (equivalent to 12.3 mg/kg bw/d fluoride). Maternal toxicity in this study was limited to reduced food intake at the highest dose level. No evidence of developmental toxicity was seen in a rabbit study (NTP, 1993; Heindelet al, 1996) at dose levels of up to 400 ppm sodium fluoride (equivalent to 14 mg/kg bw/d fluoride from all sources).

Ammonia/ammonium studies

No evidence of developmental toxicity was seen in a guideline-compliant rabbit study with ammonium perchlorate (York et al, 2001) at the highest dose level of 100 mg/kg bw/day. The influence of ammonium ions on foetal development was investigated in mice in a non-standard study involving in vitro exposure prior to transplantation into dams (Lane & Gardner, 1994). Examination on gestational day 15 showed an apparent relationship between the duration of exposure and the incidence of exencephaly. Embryos that were cultured with various concentrations of ammonium ion before being transferred to recipient dams showed increased incidence of exencephaly and a decreased percentage of implantation sites with increased ammonium concentration. It is unclear how embryos might be exposed to ammonia or ammonium in vivo or if in vivo exposure would affect foetal development and implantation in a way similar to that described in this study. No evidence of foetal toxicity was seen in a study in pigs (Diekman et al, 1993) exposed to maternally toxic concentrations of ammonia by inhalation. Although the design of the study is somewhat limited, it can be concluded that the relatively low concentrations of ammonia required to induce local irritant effects are very unlikely to cause systemic toxicity or any developmental toxicity.

Conclusion

There is no evidence that exposure to ammonium ions causes specific developmental toxicity in vivo. Inhalation exposure to ammonia will result in an equilibrium in the blood (at physiologically relevant pH) between non-ionised ammonia (NH3) and ionised ammonium (NH4+) in a ratio of approximately 1:100. The same equilibrium will exist in animals orally exposed to ammonium salts, therefore read-across is appropriate. Human maternal blood contains measurable levels of ammonia as a consequence of protein catabolism; levels of ammonia in foetal blood are slightly higher. The blood in the hepatic portal circulation contains much higher levels of ammonia due to its generation from urea by the gastrointestinal flora. Ammonia is rapidly and effectively detoxified in the liver by the urea cycle and also via additional pathways, therefore will not accumulate and is unlikely to cause any developmental toxic effects at relevant exposure levels.

Observations in humans

The effects of chronic exposure to fluorides is well characterised. The predominant findings of skeletal and dental fluorosis reflect the incorporation of fluoride into the bones and teeth. The findings of case studies involving AMBI indicate that toxicity is due to a combination of local corrosivity at the site of contact and systemic fluoride toxicity.

 

DNEL derivation

Data on the systemic effects of the substance (AMBI) are limited to a study of acute oral toxicity. Based on a review of the available toxicity data for hydrogen fluoride, other soluble inorganic fluoride salts, ammonia and ammonium salts; it is concluded that the most appropriate read-across compound for AMBI is hydrogen fluoride (HF). These substances are both corrosive and water soluble; under physiological conditions, AMBI will generate ammonium and HF, with subsequent dissociation of HF to form hydrogen and fluoride ions. The toxicity of fluoride is well characterised. By contrast, ammonium salts are of low toxicity. It is therefore clear that the systemic toxicity of AMBI will be due to the contribution of fluoride; the effects of fluoride are critical to DNEL derivation.

The critical effects of AMBI on workers are local irritation and corrosion (relevant for both dermal and inhalation exposure) and systemic fluoride toxicity (relevant for inhalation exposure). The most sensitive endpoint for systemic fluoride toxicity in experimental animals and in epidemiology studies is skeletal fluorosis. Although systemic fluoride toxicity has been reported in animals and man following dermal exposure to HF, this is considered to be unlikely unless the integrity of the skin is compromised (i.e. in accidental exposure resulting in skin burns). The potential for systemic toxicity resulting from dermal exposure to AMBI can therefore be eliminated by the use of appropriate engineering controls and PPE.

 

The SCOEL have recommended (1998) IOEL values of 1.5 mg/m3 (8-hour TWA) for inorganic fluorides (as F-) and 2.5 mg/m3 (15-minute STEL) specifically for HF.  They concluded that the 8-hour TWA was sufficient to protect against systemic effects (fluorosis) and that the STEL value for HF was adequate to limit peaks of exposure which could result in irritation. Based upon the study of Largent and Columbus (1960), conducted in volunteers exposed for 6 h/d for 10-50d, a STEL (15 mins) of 3 ppm (2.5 mg/m3) was proposed for hydrogen fluoride to limit peaks in exposure which could result in irritation. As AMBI will generate HF following inhalation exposure, this value is also considered to be appropriate for AMBI. Correcting for the fluoride content of AMBI results in a STEL of 3.8 mg/m3. Correcting the TWA value of 1.5 mg/m3 for fluorine content results in a value of 2.3 mg/m3.

 

 

 

 

 

 

 

 

 

 

 

 

 

General Population - Hazard via inhalation route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.045 mg/m³

Most sensitive endpoint:

repeated dose toxicity

Acute/short term exposure

DNEL related information

Local effects

Long term exposure

Most sensitive endpoint:

repeated dose toxicity

Acute/short term exposure

Hazard assessment conclusion:

no-threshold effect and/or no dose-response information available

DNEL related information

General Population - Hazard via dermal route

Systemic effects

Long term exposure

Most sensitive endpoint:

repeated dose toxicity

Acute/short term exposure

Hazard assessment conclusion:

exposure based waiving

Most sensitive endpoint:

skin irritation/corrosion

DNEL related information

Local effects

Long term exposure

Hazard assessment conclusion:

no-threshold effect and/or no dose-response information available

Most sensitive endpoint:

skin irritation/corrosion

Acute/short term exposure

Hazard assessment conclusion:

no-threshold effect and/or no dose-response information available

General Population - Hazard via oral route

Systemic effects

Long term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.015 mg/kg bw/day

Most sensitive endpoint:

repeated dose toxicity

DNEL related information

Modified dose descriptor starting point:

NOAEL

Acute/short term exposure

Hazard assessment conclusion:

DNEL (Derived No Effect Level)

Value:

0.015 ng/kg bw/day

DNEL related information

Additional information - General Population

DNEL values for the general public are of limited relevance, as direct exposure is not predicted. The substance is not used by the general population in the scenarios envisaged and exposure will only therefore occur indirectly via the environment. DNEL values for acute exposure of the general popualtion are therefore not derived. AMBI will react rapidly in the environment to form fluoride and ammonium ions and will further interact with other ionic species naturally present in the environment. Indirect exposure to the substance per se will therefore not occur; exposure will be to its ultimate dissociation products, fluoride and ammonium. Due to the low inhereant toxicity and rapid biodegradation of ammonium, exposure is not considered further. It is considered that the relevant indirect exposure will be limited to fluoride.

DNEL for systemic effects

The critical systemic effect for fluoride exposure is skeletal fluorosis. The SCOEL have recommended an IOEL value of 1.5 ppm F- (1.8 mg/m3) to protect against the systemic effects of fluoride exposure in workers.

Dermal DNEL (systemic)

Dermal exposure of the general public is not predicted under the use scenarios envisaged and, in any case, must be minimised by the use of protective equipment. Dermal absoprtion is not likely except in cases of exposure where the integrity of the skin is compromised (i.e. burns); this will not occur as the general population is not directly exposed to the substance.

Inhalation DNEL (systemic)

An inhalation DNEL for the general public can be derived by the application of an additional assessment factor of 2 to take into account potential additional intra-species variation and an additional factor of 2 to take into account relative breathing rates and the duration of exposure. This approach results in a DNEL of 0.45 mg/m3 F-. However the potential fluoride exposure resulting from this DNEL is equivalent to 9 mg/day, which exceeds the upper tolerable daily intake of 7 mg fluoride (EFSA, 2008). An alternative approach to deriving a systemic DNEL would be to allow inhalation exposure to account for 10% of the estimated total daily fluoride intake of 6 mg/day. On this basis, an inhalation DNEL of 0.03 mg/m3 f- (0.045 mg/m3 AMBI) can be derived and is considered to be adequately protective.

Oral DNEL (systemic)

On the same basis (i.e. allowing for fluoride intake to acccount for 10% of background), an oral DNEL of 0.015 mg/kg bw/d can be derived (assuming bodyweight of 60 kg).

DNEL for local effects

Dermal DNEL (local)

The critical local effect for short-term and long-term dermal exposure is irritation / corrosion, however this cannot be quantified and therefore a DNEL is not derived. As dicussed above, exposure of the general population will be limited to very low levels of indirect exposure to fluoride via the environment; the corrosive effects associated with higher levels of the substance is not likely to occur.

Inhalation DNEL (local)

The critical local effect of inhalation exposure to AMBI is respiratory tract irritation. However the general population will not be directly exposed to the substance; exposure will be limited to indirect exposure (via the environment) to low levels fluoride and therefore respiratory tract irritation is not predicted. An inhalation DNEL for local effects is not derived: the inhalation DNEL derived for systemic effects is considered to be protective.