Registration Dossier

Data platform availability banner - registered substances factsheets

Please be aware that this old REACH registration data factsheet is no longer maintained; it remains frozen as of 19th May 2023.

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

Diss Factsheets

Toxicological information


Currently viewing:

Administrative data

Description of key information

The key chronic study was conducted by NTP (1986). The study comprises the oral gavage administration of mixed xylenes to rats (0, 250, or 500 mg/kg/day) and mice (0, 500 or 1000 mg/kg/day) for 5 days/week for 103 weeks.

There was no evidence of carcinogenicity.
No studies are available regarding cancer in animals exposed via inhalation to mixed xylene or the individual xylene isomers.


Carcinogenicity (Oral):

No reliable oral carcinogenicity data available.

Carcinogenicity (Dermal):

No data

Carcinogenicity (Inhalation):

In a long-term cancer bioassay of the NTP increased tumour incidences were found at 750 ppm in the kidney (male and female rats), testes (male rats), lung (male mice), and liver (female mice). At 250 ppm, there was no difference between exposed and control animals. From the mode of action information available, it appears unlikely that ethylbenzene poses a carcinogenic risk for humans.

Key value for chemical safety assessment

Carcinogenicity: via oral route

Endpoint conclusion
Endpoint conclusion:
no adverse effect observed
Dose descriptor:
500 mg/kg bw/day
Study duration:
Quality of whole database:
Studies conducted in rats and mice demonstrate no evidence of carcinogenicity.

Carcinogenicity: via inhalation route

Endpoint conclusion
Endpoint conclusion:
no study available

Carcinogenicity: via dermal route

Endpoint conclusion
Endpoint conclusion:
no study available

Justification for classification or non-classification

No classification of mixed xylenes streams for carcinogenicity is warranted under GHS/CLP.



In a long-term cancer bioassay of the NTP increased tumour incidences were found at 750 ppm in the kidney (male and female rats), testes (male rats), lung (male mice), and liver (female mice). At 250 ppm there was no difference between exposed and control animals. A detailed mode of action analysis, also taking into account the high spontaneous incidence of some of the tumour types and that genotoxicity data do not indicate a direct DNA damaging effect, lead to the conclusion that the toxicological significance and relevance to human health of these findings is uncertain and it appears unlikely from the data available that ethylbenzene poses a carcinogenic risk for humans.

Therefore, the evidence is insufficient to fulfil the EU criteria for classification for carcinogenicity.

In conclusion: no classification for carcinogenicity is warranted in accordance with EU CLP (Regulation (EC) No. 1272/2008) or GHS.

Additional information

The multi-constituent substance covered by this registration comprises of individual xylene isomers (m-xylene, o-xylene, p-xylene) and ethylbenzene (>10% - <20%). The following information is available to characterise their carcinogenic mutagenic potential.

Xylenes: Non-human information

No animal studies are available on the carcinogenic effects of mixed xylene or the individual xylene isomers following dermal or inhalation exposure.

The carcinogenicity of mixed xylene (17% ethylbenzene) following oral exposure has been evaluated in chronic studies with rats and mice; however, no animal studies are available on the carcinogenic effects of the individual xylene isomers following oral exposure. Results of the chronic oral studies with mixed xylene have been negative (NTP, 1986), with no increase in tumour incidence compared with the control animals. Treatment involved administration of 0, 250, or 500 mg/kg/day doses of mixed xylene in corn oil by gavage 5 days/week for 103 weeks to groups of F344/N rats, 50 animals per group. B6C3F1 mice were treated in a similar manner but given 0, 500 or 1000 mg/kg/day of mixed xylenes in corn oil by gavage. A large number of gavage-related deaths were a confounding factor. This study did not comprehensively examine systemic effects but it did include a complete histopathological examination of all tissues as well as determination of body weight gain. Based on histopathology of all organ systems, a NOAEL of 500 mg/kg/day was observed for rats and a NOAEL of 1000 mg/kg/day was observed for mice. In conclusion there was no evidence of carcinogenicity of mixed xylenes following oral administration.

Equivocal results reported by Maltoni et al (1983, 1985) following exposure of rats to xylenes are viewed to be unreliable (IPCS, 1997) as analysis was conducted by combining all tumours; this is an unacceptable basis for analysis particularly in aged animals. In addition, no data were provided to allow an analysis on an individual tumour-type basis.

Xylenes: Human information

There is no data indicating any convincing evidence of an increased risk of cancer as a consequence of exposure to xylenes. IARC (1999) has placed xylene in Group 3: "The agent is not classifiable as to its carcinogenicity to humans". The animal data indicates that xylenes would not be carcinogenic or genotoxic in humans.

Ethylbenzene: Oral studies

Within a series of cancer bioassays on aromatics, 40 male and 40 female Sprague-Dawley rats received by gavage administration a dose of 500 mg/kg ethylbenzene in olive oil on 4-5 days per week for 104 weeks and were then kept under observation until spontaneous deaths (Maltoni et al., 1985). Olive oil alone was given to 50 male and 50 female control animals. Animals were 7 weeks old at the study begin. After 141 weeks, ethylbenzene induced an increase in total numbers of malignant tumors (35% in male and 45.9% in female rats of dosed groups versus 26.7% in male and 22.4% in female controls), the types, sites and incidences of tumors were not specified. No thymomas occurred in the 500 mg/kg groups and in control groups. The incidences of lymphoreticular neoplasms were reported to be 8.1% in male rats and 0% in female rats at 500 mg/kg in comparison to 6.7% in male control and 2% in female control animals. The study design and results were very briefly reported in this reference, contained no data on statistical evaluation, and no further details than cited here were available. Thus, this study report cannot be used for the safety assessment of ethylbenzene.


Recently severe doubts were published concerning the scientific robustness of studies from this laboratory (Schoeb et al., 2009; Ward and Alden, 2009). Although the critique related mainly to an observation of increased incidences of lymphomas after exposure to several chemicals reported in the decade between 1995-2005, it nevertheless casts serious doubts on studies in general conducted by this laboratory and on the health status of the animals used. The poor reporting of the findings after oral ethylbenzene exposure is in line with the above mentioned critique.


Ethylbenzene: Inhalation studies

In two NTP cancer bioassays on rats and mice (Chan et al., 1998; NTP, 1999) groups of 50 male and 50 female F344/N rats were exposed by inhalation to 0, 75, 250 or 750 ppm ethylbenzene (0, 0.33, 1.1, 3.3 mg/l), 6 hrs/day, 5 days/week for 104 weeks. Individual-housing animals were observed twice daily, clinical findings were recorded twice daily, body weights were recorded weekly for the first 13 weeks and monthly thereafter. Necropsy and complete histopathologic examinations were performed on all surviving and spontaneously dying animals of the chamber control and exposed groups. Sections of gross lesions, tissues masses and of more than 40 tissues/organs were examined. From the kidneys, four step sections in addition to the single standard section were prepared and examined.


In the rat study at 750 ppm the survival in males was significantly reduced (at termination 2/50 vs 15/50 in controls), while in females survival was increased (not significant). No adverse clinical findings were attributed to the ethylbenzene exposure. In males exposed to 250 or 750 ppm, body weights were reduced (up to 5 and 15%, respectively) from week 20 to the end of the study. In females all exposed groups weighed up to 5% less than the controls during the second year, but there was no dose-response.


In both males and females exposed to 750 ppm, but not to 75 or 250 ppm, the incidences of renal tubule hyperplasia were statistically significantly increased (p<0.01 for both sexes). The combined incidences from the single and 4 step sections were for males 23/50 (750 ppm) and 11/50 (controls) and for females 10/50 (750 ppm) and 1/50 (controls). The severity of nephropathy was significantly increased in males at 750 ppm and increased in females with exposure concentration. The average severity grades (1=minimal, 4=marked) of affected animals were 2.3, 2.4, 2.3, and 3.5 (p<0.01) for males and 1.3, 1.6 (p<0.05), 1.7 (p<0.01), and 2.3 (p<0.01) for females (controls, 75, 250, and 750 ppm). In males, in all experimental groups the incidence of nephropathy was high varying between 86 and 96% including the control group. In females, there was a dose related increase from 76% (controls) to 92% (750ppm). A higher incidence of transitional cell hyperplasias (renal pelvis) was present in the 750 ppm male group compared to the control group (68% vs. 24%) (p<0.01).


In the 750 ppm male group, there was a significantly increased incidence of renal tubule adenomas and combined renal tubule adenoma/carcinomas (p<0.01) based on combined original kidney sections and additional step-sectioning of the kidneys. The incidences were 0/50, 5/50, 7/50, and 20/50 for adenomas and 3/50, 5/50, 8/50, and 21/50 for adenomas and carcinomas (controls, 75, 250, and 750 ppm). The incidence of renal tubule carcinomas alone was not significantly elevated. In females, no renal tubule carcinomas were found. However, at 750 ppm, there was a significantly increased incidence (p<0.01) of renal tubule adenomas 8/50 vs 0/50 in controls. The preneoplastic finding attributable to the tumors of the renal tubule was tubular hyperplasia observed with significantly increased incidences in male and female groups exposed at 750 ppm as mentioned above. This finding was distinguished from regenerative epithelial changes commonly seen as a component of chronic progressive nephropathy. No compound induced tubule cell injury or regenerative response was observed in otherwise normal parenchyma not part of the chronic progressive nephropathy (CPN) process (Hard, 2002).


In males exposed to 750 ppm, there was a slight, but significant (p<0.01), increase in the incidence of testicular interstitial cell adenoma (44/50 vs 36/50 in controls); the historical control incidence of the test laboratory for 2-year inhalation studies was 68.7% (+/- 8.7%) with a range of 54-83%. Incidences of bilateral interstitial cell adenomas were also significantly increased (p<0.01) (40/50 vs 27/50 in controls), whereas interstitial cell hyperplasias were significantly lower (p<0.05) in 750 ppm male rats (8/50 vs 14/50 in controls).


The identified NOAEC for non-neoplastic and non-preneoplastic toxicity was 250 ppm for the rat.


The kidney slides from this study, as well as the NTP 13-week study were reexamined (Hard, 2002). He concluded that the apparent increase in renal tumors was strongly associated with chronic progressive nephropathy (CPN), a spontaneous age-related disease of rodents with no identical counterpart in humans. Apart from the three carcinomas in the high dose males, the vast majority of proliferative lesions were either small adenomas or foci of atypical tubule hyperplasia. There was also a marked exacerbation by the chemical of chronic progressive nephropathy (CPN) for males and a modest exacerbation for females: at 750 ppm end-stage CPN was found in 68% of male and 8% of female rats vs 12% (males) and 0% (females) in the control groups. Almost all of the tumors occurred in rats with advanced, usually end-stage, CPN and they were located in areas involved in the CPN process. There was a highly significant correlation between renal tumor incidence and end-stage CPN. Adjusting for end-stage CPN removed any statistically significant difference in renal tumor incidence between treated and control animals. End-stage CPN is a terminal condition where the kidneys are so morphologically altered that renal failure (as well as secondary hyperparathyroidism) occurs (Montgomery and Seely, 1990; Chandra and Frith, 1993/94) and that was the most plausible cause of deaths. There was no evidence of renal tubule injury or increased mitotic activity that might indicate to cytotoxicity/cell regeneration as a mode of action for tumor development. Although there was some evidence of a dose-related increase in hyaline droplet formation in the 13-week NTP study, it was not considered to be of the magnitude indicative of an α2u-globulin associated mechanism of renal carcinogenesis.


The association between advanced CPN and renal tubule tumor development in rats has been analyzed by Hard et al. (2012), with a conclusion (applicable to ethylbenzene) that this is not indicative of a carcinogenic effect that is relevant to human risk assessment. (Hard, C.H., Banton, M.I., Bretzlaff, R.S. et al. (2012) Consideration of rat chronic progressive nephropathy in regulatory evaluations for carcinogenicity. Tox. Sci. 132, 268-275).


B6C3F1 mice (50 mice/sex/group) were exposed by inhalation to 0, 75, 250 or 750 ppm ethylbenzene (0, 0.33, 1.1, 3.3 mg/l), 6 hrs/day, 5 days/week for 103 weeks. Exposure to ethylbenzene had no meaningful effect on survival or body weight gain.


In the lung at 750 ppm, male rats exhibited increased alveolar epithelial metaplasia (6/50 vs 0/50 in controls; p<0.05), but there was no statistically significant increase in alveolar hyperplasia. In females, no increase in the incidence of either hyperplasia or metaplasia was observed. There was increased incidence of alveolar/bronchiolar adenomas (16/50 vs 5/50 in controls; p<0.01) and of combined alveolar/bronchiolar adenoma/carcinomas (19/50 vs 7/50 in controls; p<0.01) in male mice exposed to 750 ppm but not of carcinomas (3/50 vs 2/50 in controls). The incidence of combined adenomas and carcinomas at 750 ppm (38%) was within the historical control range (10-42%). Incidences of lung tumors at 75 and 250 ppm in males or in all female exposure groups were not significantly different from the control incidences.


In female mice exposed to 750 ppm, there was a significantly increased incidence of hepatocellular adenomas (16/50 vs 6/50 in controls; p<0.05) and combined hepatocellular adenoma/carcinomas (25/50 vs 13/50 in controls; p<0.01). The incidences of adenomas (32%) and of combined adenomas and carcinomas at 750 ppm (50%) were within the historical control ranges (0-40% and 3-54%, respectively).The incidences of liver tumors in females exposed to 75 and 250 ppm were not significantly different from the control incidences. In males there was no increase in liver tumors at any exposure concentration. There were, however, increased incidences of centrilobular hypertrophy (17/50 vs 1/50 in controls; p<0.05), syncytial alteration (23/50 vs 0/50 in controls; p<0.05), and necrosis (10/50 vs 1/50 in controls; p<0.05) in the liver of males exposed to 750 ppm ethylbenzene. Syncytial alteration was defined as greatly enlarged hepatocytes containing multiple nuclei, generally five or more, either randomly scattered throughout the liver lobule or with a tendency to cluster in centrilobular areas. Hepatocellular necrosis was evident as random single cell necrosis, generally of hypertrophied cells.


There was also a statistically significant increase of follicular cell hyperplasia in the thyroid gland in both the 750 ppm males (32/50 vs 21/50 in controls) and females (35/50 vs 18/50 in controls).


A statistically significant increase of hyperplasia in the pars distalis of the pituitary gland was found in the 250 (23/50) and 750 ppm females (22/50 vs 10/50 in controls).


The identified NOAEC for non-neoplastic and non-preneoplastic toxicity was 250 ppm for the mouse.


The lung and liver sections of mice from the National Toxicology Program (NTP) two-year bioassay were reevaluated by Brown (2000). This reevaluation revealed an increased incidence of male and female mice of the 750 ppm exposure group with decreased eosinophilia of the terminal bronchiolar epithelium. Also, a dose-related increased incidence in multifocal hyperplasia of the bronchiolar epithelium with extension to the peribronchiolar alveolar epithelium was observed in all male treated groups and mid- and high-exposure females. The author noted that the necrotic hepatocytes in the high-dose males were usually that of a coagulation-type necrosis of single or small groups of cells, usually the enlarged, hypertrophied centrilobular hepatocytes. The morphology of this necrosis was histomorphologically different from apoptosis. Also the syncytial cells were not the predominant cell type with necrosis.


Summary and discussion - In carcinogenicity inhalation studies ethylbenzene was carcinogenic in F344 rats and B6C3F1 mice (NTP, 1999). Exposure to 750 ppm ethylbenzene resulted in increased tumor rates in male (kidney and testis) and in female rats (kidney), in male mice (lung) and in female mice (liver).


Since genotoxicity is not considered to be the responsible mode in the initiation of tumors (Henderson et al., 2007), other (non-genotoxic) mechanisms may be active in ethylbenzene carcinogenicity:


Ethylbenzene: Kidney tumors/male and female rats


After exposure to ethylbenzene tubular cell adenomas in male and female rats and tubular cell carcinomas in male rats were found to be associated to chronic progressive nephropathy (CPN), a spontaneous lesion of the old-aged rat. In contrast, renal tubule injury or increased mitotic activity were not observed that might indicate to cytotoxicity/cell regeneration as a mode of action for tumor development (Hard, 2002). According to Hard and Seely (2006) it is important to clearly discriminate CPN related pathological changes from genuine preneoplastic hyperplasia particularly where there has been exacerbation of CPN severity by the test chemical.


CPN is a commonly observed spontaneous disease in F344 rats, progressing with age and with preference in male rats (Montgomery and Seely, 1990; Chandra and Frith, 1993/94). CPN is not only a degenerative disease but also has regenerative aspects with a high cell proliferation rate in affected tubules (Hard and Khan, 2004). It is related to dietary protein and caloric intake as well as to the hormonal status. Chemically-exacerbated nephropathy can occur in either sex of F344 rats, but is more common in males. This lesion can result in shorter life spans in dosed animals (Montgomery and Seely, 1990) and in its advanced stages to renal tubule tumors (Hard and Khan, 2004; Hard et al., 2009).


An evaluation of 2-year carcinogenicity studies of NTP correlating the severity of CPN with the occurrence of renal tubule tumors in control male rats found a slight but statistically significant increase in CPN severity in renal tumor bearing animals compared to age matched controls without such tumaors. Therby a positive correlation between CPN and renal tubule tumor development was suggested (Seely et al., 2002). A refined grading of CPN severity suggested that this association may be higher than previously thought (Hard et al., 2009).


In a recent review Hard et al. (2009) compared rat CPN to the various types of human nephropathies. They found no entity in humans that presents with the combination or pattern of histological features found in rat CPN. There are major differences in pathology between CPN and human nephropathies. CPN in its advanced stages is associated with an increased incidence of renal tubule tumors. Chemicals that exacerbate CPN will lead to a treatment related increase of renal tumors. As there is no human counterpart of CPN, this mode of action is not relevant for species extrapolation to humans.


Renal tumor development after exposure to ethylbenzene is governed by exacerbation of CPN. There was a significant ethylbenzene-related increase in mean severity of CPN in high dose male and female rats, and a nonsignificant increase in incidence in high dose female rats. Higher grades of mean severity including secondary effects of CPN were evident in males than in females of high dose groups. This was parallel with the higher tumor incidence in males of the high dose group compared to that in the female high dose group. This supports an association of CPN and tumor growth.


The renal tubular hyperplasia was identified at increased rates in male and female rats exposed to 750 ppm and was probably the preneoplastic lesion in tumor development. The increase in kidney weight seen in subacute-subchronic repeat-dose studies possibly reflect tubular hyperplasia or hypertrophy or both, but CPN is also known to induce increase in kidney weight even in control animals (Travlos et al., 2007). The exact cause of altered weight e.g. changes in cytoplasmic and subcellular structures were not examined by appropriate methods. Although CPN is commonly viewed as an “old rat” disease, a recent evaluation of kidneys of male F344 rats from 90-day NTP studies has revealed a 100% incidence of CPN in control animals at termination of the study (4-5 months of age). Severity of CPN was associated with changes in kidney weights or urea or creatinine concentrations (Travlos et al., 2007).


In rats, ethylbenzene has no direct toxic effect on renal tubular cells up to 750 ppm after 2-year exposure and up to 1000 ppm after 92-98 days of treatment (Hard, 2002). This also suggests that the increase in mean severity grades of CPN reflects a ethylbenzene-related exacerbation of a spontaneous lesion with no toxicological significance for extrapolation to humans.


The absence of kidney tumor in the mouse gives support on the assumption of a species-specific phenomenon. Rats with moderate to severe graded CPN may differ significantly in their response to xenobiotics from rats unaffected by this lesion.


In a study comparing lung and liver microsomal metabolism of mice, rats and humans, Saghir et al. (2009) presented a metabolic scheme and indicated that side chain oxidation might be the pathway leading to rat kidney toxicity. The primary side chain metabolite is 1-phenylethanol. This substance led to kidney tumors in male F344 rats while no tumors were observed in B6C3F1 mice after gavage application over 2 years (5 d/week) at dose levels of 375 and 750 mg/kg/d (NTP, 1990). Survival of exposed male and high dosed female rats was significantly decreased and there was a high incidence of gavage related accidental deaths. Spontaneous nephropathy was observed in most male rats with greater severity in treated males. The incidence of renal tumors was statistically significantly increased in males but not in females (1/49, 13/41, 14/28 in controls, low and high dose animals; combined single and step sections of kidneys).


The contention of 1-phenylethanol being the metabolite of ethylbenzene associated with rat renal tumor formation is supported by a simplified comparison of doses given in the oral study with 1-phenylethanol (NTP, 1990) and the inhalation study with ethylbenzene (NTP, 1999). 750 ppm (corresponding to 3300 mg/m³) of ethylbenzene leading to a significant increase of renal tumors in male rats would correspond to an internal dose of approximately 430 mg/kg/d (45% inhalative absorption according to Chin et al., 1980b and 0.29 m³/kg rat respiratory volume over 6 h according to REACH R8, 2008). Taking into account that the major metabolic pathway of ethylbenzene proceeds via 1-phenylethanol and that about 25% of the absorbed dose is excreted as urinary 1-phenylethanol, a dose of 430 mg/kg ethylbenzene corresponds well with an oral dose of 375 mg/kg of 1-phenylethanol associated with renal tumors in male rats.


There are further similarities between ethylbenzene and 1-phenylethanol supporting a causal role of this metabolite for renal tumor development:


- the weight of evidenced indicates that 1-phenylethanol is not genotoxic, neither in vitro nor in vivo (Engelhardt, 2006)


- 1-phenylethanol led to an increase of male nephropathy in rats in the 2-year bioassay, although the extent of histopathology was not comparable to that carried out by Hard (2002) for ethylbenzene


- 1-phenylethanol administered over 3 days intraperitoneally (300 mg/kg/d) led to an increase of kidney weights only in male but not in female rats (Mellert, 2003)


In an oral gavage study 375 mg/kg/d over 1 and 4 weeks (750 mg/kg/d during the initial 2 days) led to a significantly increased kidney weight in male rats (females not investigated). In addition, there was a minimal to slight hyperplasia (basophilic tubules) of kidneys with an increased cell proliferation especially in the cortex and after 4 weeks of treatment (Mellert, 2004).


On the other hand, the renal carcinogenicity of 1-phenylethanol was associated by Lock and Hard (2004) (p. 242-248) with an alpha 2u globulin mode of action. In the 90 day inhalation experiment with ethylbenzene there was an indication for the accumulation of this protein; thus, some additional influence of this mechanism may be possible for ethylbenzene, too.


Finally, Lock and Hard (2004) categorized the mode of action of 69 chemicals leading to renal tubule carcinomas in 513 bioassays conducted by NCI/NTP. At that time they identified two chemicals acting via exacerbation of CPN, ethylbenzene and hydroquinone (p. 250-253). They postulated that the following criteria must be fulfilled for this mechanism to be operative:


- exacerbation of CPN to the most advanced grades, mostly end-stge CPN


- tumors must be of marginal or low incidence, predominantly adenomas of small size


- proliferative lesions must arise within CPN associated tissues


- the small lesions must not be restricted to the cortex.


All of these criteria were fulfilled for ethylbenzene. The authors concluded that this specific CPN mechanism has no relevance for extrapolation to humans.


In summary, the underlying cause for renal tubule tumor development is exacerbation of CPN by exposure to ethylbenzene, probably via side chain oxidative metabolism to 1- phenylethanol. This CPN related mechanism is not unique for ethylbenzene but has also been described with the same histopathological features for hydroquinone (Hard et al., 1997), querecitin (Hard et al., 2007), and certain pharmaceuticals that have reached the market (Hard and Khan, 2004).


Some minor uncertainty remains on the postulated causal relationship of tumor induction with the increase in mean severity of CPN: In male rats, the incidence of CPN did not increase with dose. It was already high in control male rats (94% in controls vs. 96% in male rats exposed to 750 ppm). Being already at a very high level of incidence, an increase in tumor response could not be associated to the increase in incidence of CPN. Uncertainty remains on the possible outcome in a carcinogenicity rat study on ethylbenzene using a strain with lower spontaneous rate of CPN (e.g., in Wistar rats).


These arguments do not represent significant contradictions against the hypothesis. They are judged as some remaining uncertainties on the postulated mode of action. Taking all arguments mentioned above the carcinogenic action on the rat kidney can be attributed to its mediation of the CPN. It is suggested that ethylbenzene enhances the development of CPN in F344 rats and thereby enhances a more rapid progression to renal tubular tumors, a mechanism not relevant for extrapolation to humans.


There are some kidney tumors in male rats associated with ethylbenzene exposure that were oncocytomas. Because of the low incidence (0/50, 1/50, 1/50, and 2/50 in control, low, mid, and high dose animals) a relationship to exposure is questionable and after reevaluation of the NTP study Hard (2002) only reported each one oncocytoma in the low dose male and control female group. Oncocytomas are benign tumors that occur spontaneously at extremely low incidences in rats and humans and that are characterised by abnormal mitochondria. Metastases have never been reported (Montgomery and Seely, 1990). Ethylbenzene affected metabolism in liver mitochondria (Mickiewicz and Rzeczycki, 1988). However, an association of ethylbenzene-induced increased/suppressed enzyme activity in mitochondria with the development of oncocytomas is currently, to our knowledge, unknown. If new data on ethylbenzene are generated, a concern for a presumable health risk from this rare tumor type may raise and should not be overlooked.


Ethylbenzene: Testicular tumors/male rats


The interstitial cell tumor is a common tumor type in control F344 rats if they live their natural life span. The increased rate at 750 ppm (88%) slightly exceeded that reported for historical controls of the laboratory (mean 69%, range 54%-83% (NTP, 1999)). Significantly increased rates in rats of the 750 ppm group support an association with ethylbenzene treatment. However, the relevance of this increase of tumor rates at such a high level of spontaneous incidence not accompanied by a similarly increased hyperplasia incidence remains equivocal: Ethylbenzene appeared to enhance its development in F344 rats, but not in mice.


A pharmacokinetic model was developed to further elucidate the carcinogenic effects in mice. Charest-Tardif et al. (2006) characterized the inhalation pharmacokinetics of ethylbenzene in male and female B6C3F1 mice after single and repeated exposure. They concluded that ethylbenzene pharmacokinetics is saturable at exposure concentrations >500 ppm (and therefore at 750 ppm used in the NTP cancer bioassay) but is linear at lower concentrations. The nature and magnitude of non-linearity at high concentrations should be taken into consideration when analyzing the cancer dose response data in mice.


Nong et al. (2007) developed a PBPK model for B6C3F1 mice including lungs, liver, fat, poorly and richly perfused tissues as model compartments. The model showed that hepatic metabolism alone could not adequately describe the clearance of ethylbenzene from mouse blood. Additional metabolism had to be assumed to be localized in the lung. Overall, the results indicated that the clearance rate of ethylbenzene is markedly higher in B6C3F1 mice than in rats or humans and exceeds the hepatic metabolism capacity. The evidence is consistent with a significant role for pulmonary metabolism in mice.


Ethylbenzene: Lung tumors/male mice


Alveolar/bronchiolar tumors are the most common lung tumor type occurring spontaneously or chemically induced in B6C3F1 mice.


In 750 ppm-exposed male mice, the incidences of alveolar/bronchiolar adenoma and adenoma or carcinoma (combined) were significantly greater than those in the control group but were within the historical control ranges of the laboratory (NTP, 1999). The incidence of carcinomas was not affected by treatment. In agreement to other literature data (Dixon and Maronpot, 1991) which reported mean spontaneous incidences of adenomas and carcinomas (combined) of 19 % in males and 7.3% in females of this strain, it seems questionable whether the lung tumors can be attributed to ethylbenzene. No significant increase in lung tumor rates was found in the female mice and in male and female rats.


Saghir et al. (2009) investigated species differences in metabolism of ethylbenzene in detail with lung and liver microsomes of mice, rats, and humans. The metabolic pathways of side chain oxidation (1-phenylethanol and acetophenone) and ring hydroxylation (2- and 4- ethylphenol, 2,5 - and 3,4-quinone) were analyzed. Reactive metabolites (2,5- and 3.4-dihydroxyethylbenzene-GSH)were monitored via glutathione (GSH) trapping. The results for side chain oxidation are given in the metabolism section and conversion of ethylbenzene to ring hydroxylated metabolites was much lower. Formation of 2.5-dihydroxyethylbenzene-GSH was typically 10-fold higher than that of 3.4-dihydroxyethylbenzene-GSH. Formation of 2,5-dihydroxyethylbenzene-GSH was higher by lung (highest by mouse lung) than by liver microsomes and its formation by mouse liver was higher than by rat and human liver microsomes. When 2- and 4-ethylphenol were incubated with microsomes there was a clear species difference in the rate of conversion to dihydroxylated aromatic metabolites:


2-ethylphenol to ethylhydroquinone:


– mouse>human>rat by liver microsomes


– mouse>>rat>>human by lung microsomes


4-ethylphenol to ethylcatechol:


– mouse>human=rat by liver microsomes


– mouse>>rat>>human by lung microsomes


Although ring oxidized metabolites accounted for a relatively small fraction of the overall ethylbenzene metabolism, the selective elevation in mouse lung microsomes is consistent with the hypothesis that development of lung tumors in mice is driven by the formation of ring hydroxylated metabolites in this species.


This hypothesis is supported by the study of Mellert et al. (2003a) demonstrating mouse lung cytotoxicity after ip application of 4-ethylphenol over 3 days but not after treatment with the side chain oxidized metabolites 1- and 2-phenylethanol or acetopheneone (Kaufmann et al., 2005; Mellert et al, 2002; 2002a; 2003a).


Cruzan et al. (2009) evaluated the possible mode of action for several chemicals leading to lung tumors specifically in B6C3F1 mice but not in rats. Tumors were found at the same locations for these substances, namely in the outer layer of the lung where the terminal bronchioles and alveoli intersect. The compounds considered were coumarin, naphthalene, styrene, alpha-methylstyrene, cumene, divinylbenzene, and benzofurane besides ethylbenzene. Genotoxicity was of no or only low relevance for this site of action. Although a complete picture of the mode of action had not been developed for any one of these chemicals, the data from the individual substances were synthesized and a model was developed concluding that lung tumor development in mice is driven by lung toxicity mediated by CYP 2F2 metabolism.


Lung toxicity by these chemicals occurs in the terminal bronchioles of mice but not of rats. The target cells are the Clara cells (but not the alveolar cells) and increased cell proliferation has been demonstrated by BrdU labeling after short-term exposure for several of these chemicals, including ethybenzene (Stott et al., 2003). This specific target site has been identified by the reevaluation of mouse lung tissue from the ethylbenzene bioassay by Brown (2000) showing the presence of multifocal bronchiolar/parabronchiolar hyperplasia at 750 ppm.


For coumarin, naphthalene, and styrene it was demonstrated that CYP 2F2 inhibition led to an inhibition of lung toxicity. For styrene, the ring oxidized metabolite 4-vinyphenol was found to be a potent lung toxicant (Carlson et al., 2002a) and toxicity was most probably mediated by further metabolism of 4-vinylphenol (Carlson, 2002). This finding corresponds closely to the proposal of Saghir et al. (2009) that lung tumors in mice are driven by ring hydroxylated metabolites of ethylbenzene formed in the lung. In this context it is important to note that no detectable lung toxicity was observed from exposure to side chain hydroxylated metabolites of ethylbenzene (1- and 2-phenylethanol, phenylacetaldehyde) (Carlson et al., 2002b). Similarly, lung toxicity in the terminal brochioles was not found after ip application over 3 days of 1- and 2-phenylethanol or acetophenone (Kaufmann et al., 2005; Mellert et al., 2002; 2002a; 2003a). On the other hand, ip treatment with 4-ethylphenol led to a significant increase of cell proliferation in the large and medium bronchi and terminal bronchioles (Mellert et al., 2003a).


This mechanism of lung tumor development is mouse specific and is due to preferential and lung-mediated metabolism by CYP 2F2 located in mouse Clara cells. Although CYP 2F4, the isoenzyme in rats, appears to be equally active, it occurs to a much lower extent in rat Clara cells and therefore the levels of metabolites produced are not sufficient to cause lung toxicity. Human lungs contain far fewer Clara cells and thus much lower amounts of CYP 2F1 (the isoenzyme in humans) than rats or mice. Furthermore, human lung microsomes failed to, or only marginally, metabolize these compounds. In addition, human Clara cells differ markedly from mouse Clara cells. These differences make humans much less sensitive than mice to lung toxicity due to these reactive metabolites. Thus, while lung tumors from bronchiolar cell cytotoxicity are theoretically possible in humans, it is unlikely that metabolism by CYP 2F1 would produce levels of cytotoxic metabolites in human lungs sufficient to result in lung cytotoxic responses and thus in tumors (Cruzan et al., 2009).


Ethylbenzene: Liver tumors/female mice


Similarly, there is uncertainty on the association of liver tumors in female mice with the ethylbenzene exposure. According to Haseman et al. (1998) liver adenomas/carcinomas are the most frequently occurring neoplasms in female B6C3F1 mice with a mean control incidence in chamber inhalation studies of 25.2% (range 3-54%) for adenomas and carcinomas combined. The occurrence of liver tumor in one sex of one species is a weak argument for a tumor with relatively high spontaneous rates in B6C3F1 mice. The incidences of hepatocellular adenoma and adenoma or carcinoma (combined) increased in exposed female mice only with a positive trend gaining significance at 750 ppm ethylbenzene, whereas in male mice all liver tumors were at comparable high levels without a treatment-related effect (16-34%). Again, the incidences in females did not exceed the historical control ranges of the laboratory. Early induction of liver tumors associated with higher and early mortality and the occurrence in more than one species and / or more than one sex could support a chemical-induced tumorgenicity (Maronpot and Boorman, 1982). Since increased tumor rates did not influence the survival of female mice, liver tumors in females only of this bioassay are considered a late life process.


A possible role of increased levels of reactive oxygen species (ROS) produced by ethylbenzene-mediated induction of particular cytochrome P450 enzymes has been discussed (Serron et al., 2000). For male F344 rats, increased levels of ROS, such as hydrogen peroxide, have been shown in liver microsomes after single ip injection of 10 mmol/kg ethylbenzene. No data on ROS generation are available for the mouse liver. Whether this mechanism contributes to toxicity and/or carcinogenic effect in experimental animals and its relevance for humans, is actually unknown. In this context it is important to note that an ex vivo UDS test in the liver of male and female B6C3F1 mice was negative after inhalation exposure (males up to 1000 ppm, females up to 750 ppm). Thereby a direct genotoxic effect of ethylbenzene on this target organ could be excluded (CTL, 2000),


There is good evidence that suggests that for liver tumors in female mice a phenobarbital mechanism may be operative. Whysner et al. (1996) identified the most important features associated with liver tumor induction by phenobarbital and many of these were also observed after ethylbenzene exposure. The most important key effects described for phenobarbital are:


- non genotoxicity – applies for ethylbenzene, too


- induction of CYP450 enzymes, among those especially the CYP2B subfamily – this has been demonstrated for ethylbenzene in mice after inhalation exposure to 750 ppm over 1 and 4 weeks (Stott et al., 2003)


- transient increase of cell proliferation of hepatocytes observed after a few days of exposure – ethylbenzene led to an increase of S-phase DNA synthesis at 750 ppm after 1 week of exposure, while after 4 weeks only a numerical but not statistically significant effect remained (Stott et al., 2003):


- the increase of cell proliferation is not due to a compensatory regenerative effect in the course of liver toxicity – there was no indication for liver toxicity by histopathology after ethybenzene exposure to 750 ppm over 4 weeks (Stott et al., 2003) or up to 1000 ppm over 13 weeks (NTP, 1992)


liver weight increase – after ethylbenzene exposure at 750 ppm over 1 and 4 weeks relative liver weight was significantly increased (p=0.05) (Stott et al., 2003) and after 13 weeks of exposure at 750 and 1000 ppm absolute and relative liver weights were significantly increased (p<0.01) (NTP, 1992) without adverse effects on body weights


- exposure related increase of eosinophilic (but not of basophilic) foci – in the cancer bioassay (NTP, 1999) only eosinophilic foci were reported with the following incidences: 5/50, 7/50, 6/50, and 22/50 (p<0.01) at 0, 75, 250, and 750 ppm.



In analyzing the mechanistic data for phenobarbital, Whysner et al. (1996) concluded that humans are not at risk of cancer from therapeutic doses and are resistant to the type of liver cancer found in mice exposed to phenobarbital. This should also apply to other agents with similar mechanistic characteristics. In more general terms Holsapple et al. (2006) analyzed the mode of action of rodent liver carcinogens in their relevance for human cancer risk. For a phenobarbital-like mode of action with a robust data set they concluded that the carcinogenic response in rodents is not relevant to humans.


Ethylbenzene: Pars distalis hyperplasia/female mice


Ethylbenzene affected the hypothalamus-pituitary axis of neuroendocrine regulation. Pituitary hormones were shown to interact with the cytochrome P450 expression profile in ethylbenzene exposed animals. Single and multiple ip applications of ethlybenzene led to a decrease of P450 2C11 protein and to an increase of P450 2B protein levels (Bergeron et al., 1999). These effects were differently modulated by hypophysectomy (HX) and growth hormone (GH) supplementation. HX led to a similar reduction of P450 2C11 protein as treatment with ethylbenzene, but combination of ethylbenzene and HX did not result in a further reduction. GH supplementation reversed the effect of HX and application of ethylbenzene did not decrease P450 2C11 protein in in HX/GH supplemented animals. On the other hand, ethylbenzene treatment increased CYP 2B protein in intact, HX, as well as in HX/GH supplemented rats (Serron et al., 2001). Whether the hyperplastic effect on the pituitary is mediated by any of these findings, is currently unknown.



It is concluded, that:


- Long-term inhalation exposure to ethylbenzene was carcinogenic in F344 rats and B6C3F1 mice.


- A significant increase of tumor incidences has been observed in the kidneys (renal tubule adenoma and carcinoma), testis (interstitial cell adenoma), liver (adenoma and carcinoma) and lung (alveolar/bronchiolar adenoma and carcinoma).


- There was no concordance in carcinogenic response between rats and mice. Elevated rates of kidney tumors were seen in male and female rats. Each of the other tumors occurred in one sex and in one species only.


- Genotoxicity data did not indicate a direct DNA damaging effect.


It is concluded that sufficient evidence exists that kidney tumors in the male and female rats are associated with the high strain-specific incidence of chronic progressive nephropathy (CPN) that is unknown for humans.


For tumors in the testis, liver and lung high or very high spontaneous rates occur in the mouse and rat strains used. Ethylbenzene may exert its carcinogenic action by enhancement of tumor development in genetically disposed animals or by reduction in latency periods in tumor development.


Although the detailed mechanisms underlying the increases in tumor rates are presently not clarified, it appears likely that the mode of carcinogenic action of ethylbenzene possesses species and strain specificity.


As regards the different metabolites of ethylbenzene, there is good evidence that rat kidney tumors are related to 1-phenylethanol while tumor formation in mice (especially in the lung) are driven by reactive ring hydroxylated metabolites that are formed to a much larger extent in mice than in rats or humans. Formation of these reactive metabolites by lung microsomes is much higher in mice than in rats and could not be detected in human lung microsomes.


Therefore the toxicological significance and relevance to human health of these findings is uncertain.


It appears unlikely from the data available that ethylbenzene poses a carcinogenic risk for humans exposed.


The evidence is insufficient to fulfil the EU criteria for classification for carcinogenicity.


The following information is taken into account for any hazard/risk assessment:


Ethylbenzene is carcinogenic in animals following lifetime exposure to high vapor concentrations. From the mode of action information available, it appears unlikely that ethylbenzene poses a carcinogenic risk for humans and the evidence is insufficient to fulfil the EU criteria for classification for carcinogenicity.


Justification for selection of carcinogenicity via oral route endpoint:
Results of chronic oral studies with mixed xylenes have been negative, with no increase in tumour incidence in treated rats given up to 500 mg/kg bw/d for 103 weeks or in mice following chronic oral treatments of up to 1000 mg mixed xylenes/kg bw/d (NTP, 1986).

Justification for selection of carcinogenicity via inhalation route endpoint:

In a long-term cancer bioassay with ethylbenzene, increased tumor incidences were found at 750 ppm in the kidney (male and female rats), testes (male rats), lung (male mice), and liver (female mice). At 250 ppm there was no difference between exposed and control animals. From the mode of action information available, it appears unlikely that ethylbenzene poses a carcinogenic risk for humans.