2-butoxyethanol

Endpoint summary

Description of key information

ORAL:

Guinea pig: 1200, 1414mg/kg.  Average approx 1300mg/kg

Rat LD50: 615 (female), 880 (male), 1480, 2420 (male), 2600 (male).  500-1500 (female), 1000-2000 (female)  Average: ~1400mg/kg.

Mouse: 1230, 1519-2005.  Average approx 1500mg/kg

Rabbit:  LD100<650mg/kg

Dog: LD0>650mg/kg

Human: LOAEL=400mg/kg

INHALATION (Saturated vapour pressure estimated at 3.9mg/l, excluding unreliable data):

Guinea pig: LC0 (4hr) >2.25mg/l, LC0 (1hr) >3.2-3.4mg/l, LC0(7hr)>400ppm (2mg/l), LC0(7hr)>400ppm (2mg/l), LC0(7hr)>400ppm (2mg/l)

Rat: LC50 >4.9mg/l (3hr), >3.9mg/l (4hr), 2.2-2.4mg/l (4hr) , >4.26mg/l (7hr). female (4hr) ~900ppm, male (7hr) >900ppm, >1.44mg/l(3hr)

Dog: LC0(7hr)>400ppm (2mg/l), Dog: LC0(7hr)>400ppm (2mg/l), Dog: LC0(7hr)>400ppm (2mg/l)

Rabbit LC50(7hr) ~ 400ppm (2mg/l) (based on an average of 3 replicates)

DERMAL (excluding unreliable data)

Guinea pig: 230-300, >1200 (possibly >2000), >2000mg/kg

Rat: LD50>2000mg/kg (occlusive and semiocclusive conditions.)

Rabbit: LD50 (mg/kg) = 435, 612, 405, 567, 841 (occlusive), >2000 (semiocclusive).

Key value for chemical safety assessment

Acute toxicity: via oral route

Endpoint conclusion

Endpoint conclusion:

adverse effect observed

Dose descriptor:

LD50

Value:

1 300 mg/kg bw

Quality of whole database:

Assessment based on choice of most appropriate species to represent toxicity to humans (guinea pig)

Acute toxicity: via inhalation route

Endpoint conclusion

Endpoint conclusion:

no adverse effect observed

Quality of whole database:

Assessment based on choice of most appropriate species to represent toxicity to humans (guinea pig)

Acute toxicity: via dermal route

Endpoint conclusion

Endpoint conclusion:

no adverse effect observed

Quality of whole database:

Assessment based on choice of most appropriate species to represent toxicity to humans (guinea pig)

Additional information

2 -butoxyethanol is data rich for the end point of acute toxicity. There is data on multiple species (including human data in chapter 7.10.3), with multiple studies on many of the species and all routes are well covered. The data, when assessed in conjunction with other data shown in chapter 7.9.3, provides a clear picture of the relative hazard posed to each species from acute exposure and the likely hazard to humans. As demonstrated elsewhere in this dossier, of the animals models, the guinea pig is the species that most closely represents the toxicity of 2 -butoxyethanol to humans. This is because the normally preferred species, mouse, rat and rabbit, are susceptible to haemolytic effects triggered by exposure to 2 -butoxyethanol, whilst the guinea pig and human are resistant to this effect. Thre is reliable data on the guinea pig for all exposure routes, therefore this is the data that is used to assess acute toxicity to humans.

ORAL ROUTE

Two studies are available in Guinea-pigs, one of which is a relatively recent GLP guideline study. The LD50 calculated were 1414 and 1200 mg/kg. The same clinical signs and pathology than other species tested were seen in these studies. Necrosis and hemorrhage of the gastric mucosa was also seen. A value of 1300mg/kg is used the average result from the two studies available.

INHALATION ROUTE

A reliable GLP and guideline guinea pig study is available that showed no substance related adverse effects following four hours exposure to the maximum attainable vapour nose only exposure of 2.25mg/L. Another GLP study is available in guinea pigs and this showed no adverse effects for exposures at the maximum practical whole body vapour concentration achievable (around 3.2mg/l) but the exposure time was only one hour. Another study in guinea pigs showed no effects from two 7 hour exposures to 400ppm (2mg/l) repeated one week apart. This same result was observed in two further repeats of the study using 2 -butoxyethanol from different commercial suppliers. One of these studies was continued for 5 days of continuous daily 7 hour exposure to the same concentration and no adverse effects were seen. Note that extrapolation of these results to 4 hours using the Haber equation would be equivalent to an LC0 of ~2.4mg/L. There is strong evidence from this species is that the acute toxicity hazard to humans is minimal by the inhalation route.

DERMAL ROUTE

In guinea-pigs, variations were seen depending on the studies. LD50 ranged from 230 mg/kg to >2000 mg/kg. The most recent study was performed according to guideline and GLP. This study showed a LD50 of greater than 2000 mg/kg. For this study, no adverse effects were described (local or systemic). Another study produced a result consistent with this. The exceptionally low figure (for the guinea pig) was from a much older, pre-guideline study. The relative ages of these studies should be taken into account when weighting their relative importance and the fact that this LD50 is out of line with other, more recent, data available for guinea pigs, including from the oral route. The older study is considered unreliable and the LD50 likely to be can be considered to be greater than 2000mg/kgbw.

HUMAN DATA

Acute human toxicity data comes mainly from children accidental ingestion or adult suicide attempts made with mixtures containing 2 -butoxyethanol and from one study on human volunteers by inhalation. For oral route case reports, ingested doses are difficult to evaluate because of the lack of data concerning the body weight of all patients and the exact ingested dose, but a semi-quantitative

estimation of the ingested doses was made for each case. The range of doses which lead to clinical symptoms varies between 0.5 and 4.5 g/kg bw. In all cases, patients exhibited CNS depression and metabolic acidosis. Signs of haemolysis were seen in some but not all cases and this finding was not systematic (showing that humans are much more resistant to haemolysis than rodents). After a first acute ingestion, a second administrator some days later did not exhibit the same symptoms. In all reported cases the patients totally recovered after treatment. According to this data, a conservative LOAEL of 400mg/kg bw can be taken into account for acute toxicity effects by oral route in humans. It would be appropriate to convert this by route to route extrapolation to predict toxicity by other routes.

Haematotoxicity was not observed in a volunteer study, in which humans were voluntarily exposed to up to 200 ppm (~1.0mg/L) for 4 h

Justification for selection of acute toxicity – oral endpoint
The guinea pig is considered the most appropriate model for assessing acute toxicity to humans.

Justification for selection of acute toxicity – inhalation endpoint
Based on human data and extrapolation of oral route information, it is concluded that adverse acute toxicity effects in humans cannot be seen at exposures up to and including the saturated vapour concentration.

Justification for selection of acute toxicity – dermal endpoint
The guinea pig is considered the most appropriate model for assessing acute toxicity to humans.

Justification for classification or non-classification

Review of all the relevant data

When considering the acute toxicity data for 2–butoxyethanol (EGBE), it must be borne in mind that the purpose of classification is to indicate the hazard to humans. Therefore in interpreting the data on animals, the relative sensitivity of the test animals and humans should be considered as part of the scientific judgement to determine which information is most appropriate to use to determine the classification. There is data presented elsewhere in this Chemical Safety Assessment that clearly indicates that the most important short term acute systemic toxicity effect from exposure to EGBE is hemolysis and that humans along with guinea pigs and dogs are particularly resistant to the effect compared to rats and rabbits which are the preferred test species for acute toxicity testing. The data supporting these conclusions regarding the relative sensitivity is summarised point by point here:

• Hemolysis of red blood cells is caused by butoxy acetic acid (BAA), the primary metabolite of EGBE. (Carpenter et al, 1956; Ghanayem, 1989; Ghanayem et al, 1987; Ghanayem et al, 1993). This metabolite is produced by the alcohol and aldehyde dehydrogenase enzyme system (Ghanayem et al, 1987) in the liver. Rodents, guinea pigs and humans are therefore all capable of converting EGBE to BAA.
• BAA appears to increase the fragility of red blood cells in some species, leading them to rupture when passing through the vascular system (Ghanayem, 1989).
• BAA produces more hemolysis in older animals compared with younger animals. This was demonstrated to be due to the increased fragility of older red blood cells (present in higher numbers in old animals). Animals that had been bled, triggering the increased production of new red blood cells, were less susceptible to the haemolytic activity of BAA; in these animals the LD50 of EGBE was higher than in animals of the same age that had not been bled; i.e. the haemolytic activity is a critical component of the acute toxicity of EGBE (Ghanayem et al, 1990; Ghanayem et al, 1992; Sivarao et al, 1995).
• There is a clear species difference in susceptibility to the hemolysis caused by BAA. Acute toxicity studies in rats and rabbits show clear evidence of hemolysis. Studies in guinea pigs do not. In vitro studies of red blood cell hemolysis using rodents, rabbits, guinea pigs, cats, dogs, pigs and primates demonstrated that the red blood cells of rodents and rabbits appear to be more sensitive to this effect whereas guinea pigs and humans are significantly less sensitive (Ghanayem et al, 1993; Udden et al, 1994; Udden, 2000).
• Physiologically based pharmacokinetic models of EGBE demonstrate that it would be almost impossible to achieve a high enough plasma concentration of EGBE in humans to trigger hemolysis of red blood cells (Lee, 1998). This is consistent with the few reported cases of acute human intoxication with EGBE where there was no reported evidence of haemolytic activity (Bauer et al., 1992; Butera et al., 1996; Burkhart et al., 1998; Dean et al, 1992; Gijsenbergh et al., 1989; Gualtieri et al, 1995; Gualtiere et al, 2003; Hung et al., 2010; McKinney et al., 2000; Rambourg-Schepens et al., 1988).

From these findings it can be concluded that in rats and rabbits, the critical effect leading to mortality following an acute exposure is hemolysis of red blood cells. Human red blood cells are significantly less susceptible to BAA mediated hemolysis, therefore the LD50 values derived from rat and rabbit studies are not reliable predictors of acute toxicity potential in humans; they will significantly over-predict the likely toxicity at a given dose.

Guinea pig red blood cells appear to be similar to human red blood cells regarding their sensitivity to BAA mediated hemolysis. This similarity in sensitivity to hemolysis makes the guinea pig a more appropriate model for assessing the acute toxicity potential of EGBE in humans (Gingell et al., 1998).

Other toxic modes of action should be considered. In cases of human poisoning incidents involving exposure to EGBE, the primary toxic effect was metabolic acidosis, likely resulting from high concentrations of BAA in the blood (Bauer et al., 1992; Burkhart et al., 1998; Butera et al., 1996; Dean et al, 1992; Gijsenbergh et al., 1989; Gualtieri et al 1995, 2003; Hung et al., 2010; McKinney et al., 2000; Rambourg-Schepens et al., 1988). There were no reports of hemolysis in these cases. Metabolic acidosis, if severe enough, can produce mortality and this mode of action is likely to be responsible for acute toxicity observed in guinea pigs following oral exposure to EGBE, but this effect only occurs at much higher doses than those which cause haemolysis in sensitive species. Accordingly, death was not reported in any of the case reports, in which humans accidentally or intentionally ingested doses up to 4200 mg/kg EGBE. In all cases, recovery was complete without any irreversible effects.

Interpretation of data

The acute toxicity of EGBE is mediated through the primary metabolite, BAA. There are two critical toxic effects, hemolysis of red blood cells and metabolic acidosis. Human red blood cells are not sensitive to the haemolytic activity of BAA, therefore this toxic effect is of negligible relevance to humans. Humans are however sensitive to metabolic acidosis and this effect has been observed in poisoning incidents involving EGBE. The choice of model to use as a surrogate to estimate toxicity to humans is important because haemolysis in sensitive species occurs at much lower doses than in humans. Use of a haemolysis sensitive species would over-estimate human toxicity. According to paragraph 3.1.2.2.1 of Annex I of regulation 1272/2008, whilst the rat and rabbit are the preferred species to evaluate acute toxicity, `when experimental data for acute toxicity are available in several animal species, scientific judgement shall be used in selecting the most appropriate LD50 value from among valid, well-performed studies‘. Of all the animal models, the guinea pig is most similar to humans regarding its resistance to hemolysis and susceptibility to metabolic acidosis. Therefore, based on clear and compelling scientific evidence, the guinea pig is the most appropriate animal model for assessing the acute toxicity potential of humans (Gingell et al., 1998, Boatman 2013, ECHA Risk Assessment Committee, 2018).

Assessment for classification - Approach using animal data.

The above arguments support the case that the guinea pig is the most appropriate animal model to use to classify for acute toxicity effects in humans. Reliable and robust studies exist for guinea pig for all three routes of exposure. These lead to the following conclusion as the appropriate classification to reflect the hazard to humans:

• ORAL ROUTE: LD50 ~1300mg/kg. Under the EU CLP regulation 1272/2008 (CLP), the LD50 clearly falls into the range for a category 4 classification. (All data available in REACH registration dossier and listed in the appendix to this document). An ATE of 1300mg/kg is appropriate
• DERMAL ROUTE: LD50>2000mg/kg, implying no classification under the CLP regulation. (All data available in REACH registration dossier and listed in the appendix to this document).
• INHALATION ROUTE: LC0>=maximum attainable vapour concentration (1h, 4h and 7h exposures across multiple studies, LC0>2mg/l. Since the LC50 cannot be reached at the saturated vapour concentration, no classification is warranted.

Taking into account the toxic effects seen from acute exposure and the relative sensitivities of humans versus other species, the evidence suggests that 2-butoxyethanol does not warrant classification for the inhalation and dermal route and that only classification as acute toxicity category 4 is appropriate for the oral route. This is the conclusion drawn in a review of the data by Boatman et al (2013).

Assessment for classification - Approach using human data

According to the guidance on the application of the CLP criteria, human data can be considered for use in classification decisions. Such data (e.g. case reports) should report severe effects after a single exposure. A weakness of such data is often the lack of exposure or dose information. Useful data from multiple sources has been identified that reports both adverse effects and also includes exposure information. Such information can therefore be considered robust enough to use as a starting point for an alternative approach to derive an appropriate classification and is discussed in more detail below.

There is a significant amount of acute oral toxicity data available in humans primarily through case reports of suicide attempts made using cleaners and other products containing 2-butoxyethanol. It is notable that none of these attempts resulted in death and all patients made full recoveries. The available published data is shown in the table below. The table only includes data on adults. Data is also available on children (who also all survived) but the dose information is too imprecise for use here:

Sex	Age (years)	Estimated dose (g/kg)	Reference
Male	53	0.65	Bauer (1992)
Male	47	0.57	Butera (1996)
Male	53	1.8 - 3.0	Hung (2010)
Female	51	0.4 – 1.2	McKinney (2000)
Male	19	2.0 - 4.2	Burkhart (1998)
Male	18	1.1 – 1.5	Gualtieri(1995)
Female	23	1.0	Gijsenber(1989)
Female	50	0.5 – 1.0	Rambourg-Schepens(1988)

All of these studies are reported in detail in the IUCLID dossier.

With the exception of the Hung data, which post dates it, all of this information was reviewed in the EU risk assessment of 2-butoxyethanol published in 2006. The following is an extract from the risk assessment conclusions of a review of this data:

Acute human toxicity data comes from children accidental ingestion or adult suicide attempts made with mixtures containing EGBE and from one study on human volunteers by inhalation. For oral route case reports, ingested doses are difficult to evaluate because of the lack of data concerning the body weight of all patients and the exact ingested dose, but a semi-quantitative estimation of the ingested doses was made for each case (see table 4.7). The range of doses which lead to clinical symptoms varies between 0.5 and 4.5 g/kg bw. In all cases, patients exhibited CNS depression and metabolic acidosis. Signs of haemolysis were seen in some cases but this finding was not systematic (this showed that human is much more resistant to haemolysis than rodents). After a first acute ingestion, a second administrator some days later did not exhibit the same symptoms, this finding was also seen with animals in some studies. In these cases, EGBE was ingested together with other substances (ethanol and/or unknown substances) that could have some influence on the symptoms seen. Between 0.5 and 1.5 g/kg bw the patients totally recovered after treatment. According to this data, a LOAEL of 400 mg/kg bw can be taken into account for acute toxicity by oral route in humans in the risk characterisation section. If a risk characterisation by dermal and inhalation route is needed for acute effects, kinetic data is sufficient to extrapolate this oral LOAEL to inhalation and dermal LOAEL. Human data is preferred for risk characterisation, especially for EGBE because of its haematotoxicity more marked in animals than in humans. It should be noted that this is a worst case estimation derived from the McKinney paper in which the possible range of exposure was between 0.4 and 1.2 g/kg bw.

The conclusion of a LOAEL of 400mg/kg seems reasonable and conservative based on the available data.

The guidance on the application of the CLP criteria states that human data should be used to derive that ATE (single or dose range expected to cause mortality) without any adjustment for comparison with the classification criteria. For the oral route, the ATE based on the LOAEL above would therefore be 400mg/kg. This falls in the range for classification as category 4 under the EU CLP regulation 1272/2008 (CLP). This is clearly conservative as it is based on a LOAEL and not mortality or an LC50, which would be significantly greater than this.

This data can be converted by route to route extrapolation to assess appropriate classifications for other routes of exposure.

For the inhalation route, it is informative to compare the dose that would derive from exposure to the saturated vapour concentration of 4.8mg/l. Based on table R8.2 from the guidance on information requirements and chemical safety assessment, human respiratory volume can be considered to be 0.2L/min/kg. For a 4 hr exposure (the normal duration for an acute inhalation study), this becomes 48L/4hr/kg. At a maximum achievable concentration of 4.8mg/l, this would result in a maximum theoretical dose of 230mg/kg (or 138mg/kg assuming 60% absorption) which is clearly well below the LOAEL derived for human. On this basis, it becomes apparent that it is not possible to reach harmful levels of exposure to 2-butoxyethanol by the inhalation route, so classification is not required.

A similar approach is also possible for the dermal route. Assuming 100% absorption by the oral route and 30% by the dermal route would mean that the equivalent LOAEL by the dermal route would be 1333mg/kg. A comparison with the classification criteria would lead to the conclusion that as 'harmful' under directive 67/548 (DSD) and category 4 under the EU CLP regulation 1272/2008 (CLP) is appropriate based on the face value of this data. However, this is clearly conservative as it is based on a LOAEL and not mortality or an LC50, which would be significantly greater than this. It would be reasonable to assume that the lethal dose would be at least a factor of 2x higher than the LOAEL, which would imply an extrapolated ATE by the dermal route of no classification, in line with the data from animals which show similar toxicity sensitivities to humans.

Conclusions

The use of acute toxicity data from humans and from animal species appropriate to predict toxicity in humans (i.e. eliminating from consideration species which show toxic effects not relevant to humans) leads to a consistent conclusion that 2-butoxyethanol should be classified for acute toxicity category 4 under the CLP regulation (H302) but does not require classification by the inhalation or dermal route.

Concentration limits

Based on the data from the guinea pig, an ATE of 1200mg/kg would be appropriate for the classification of mixtures for acute toxicity by the oral route based on the lowest available LC50.