Gas oils (petroleum), thermal-cracked, hydrodesulfurized

Carcinogenicity

Description of key information

Catalytically cracked light cycle oil with 8.7% 3 to 7 ring polycyclic aromatic compounds was carcinogenic when applied repeatedly to mouse skin. The evidence from 3 studies with cracked gas oils suggests that they are carcinogenic.

Key value for chemical safety assessment

Carcinogenicity: via oral route

Endpoint conclusion

Endpoint conclusion:

no study available

Carcinogenicity: via inhalation route

Endpoint conclusion

Endpoint conclusion:

no study available

Carcinogenicity: via dermal route

Endpoint conclusion

Endpoint conclusion:

adverse effect observed

Study duration:

chronic

Species:

mouse

Justification for classification or non-classification

Following dermal application tumours were observed. The percentages of tumour bearing animals indicate that the development of tumours is likely not dependent on dermal irritation. Data from the experiments conducted with cracked gas oils indicate that they are carcinogens, therefore they are classified as Carcinogenic Cat 1B, H350, according to EU CLP Regulation (EC No. 1272/2008).

Additional information

Mice were exposed to a catalytically cracked light cycle oil (CAS No. 64741-59-9), as part of a study to investigate the influence of skin irritation on tumour development (Klimisch score = 2, EMBSI, 1996). The dosing regimen used in this study was designed to allow investigation of the carcinogenic potential of gas oil streams under conditions that produced varying levels of skin irritation through manipulation of the applied concentration and frequency of treatment. The treatment regime was designed to deliver a constant total weekly dose of each distillate while varying the extent of any local skin irritation by manipulating test material concentration and frequency of application. For example the 100 % (neat) material would be applied once a day whilst the 50% concentration would be applied twice a day to achieve an equivalent dose of test material.

Dermal irritation was observed in animals treated with test material and the positive and negative control substances, and was generally greater after treatment with neat distillate twice per week compared with more frequent treatment with lower concentrations. Concentration-dependent increases in dermal irritation were observed in all of the treatment groups (0.28, 1.59, 2.40). The pattern of mortality in animals from the 28.5% treatment groups was generally similar to that of the mineral oil (negative) control group, with slightly higher mortality in the 50% or 100% groups. This was statistically significant in the case of the cracked gas oil, where estimated median survival was around 60-80 days less than that of the controls. Survival in the positive control group (median 300 days) was also significantly lower than the controls (median 582 days).

There were no significant treatment-related gross findings at necropsy other than skin irritation and masses (tumours) at the treatment site. Tumours (predominately papillomas or squamous cell carcinomas) occurred in all groups treated with test material, affecting 1, 17 and 7 mice from the high, intermediate or low dose groups respectively with a time to first tumour of approximately 300 to 650 days.

The catalytically cracked light cycle oil, containing moderate amounts of 3 to 7 ring polycyclic aromatic compounds, was a dermal carcinogen with a short time to tumour development, following dermal treatments that did not cause appreciable skin irritation. These findings are consistent with a genotoxic mechanism of action. The reason for the failure of undiluted test material to produce clear evidence of carcinogenicity, despite the much greater incidence of tumours in the 50% group, is unclear but it may be due to widespread epidermal necrosis destroying initiated cells. Hence the evidence suggests that cracked oils are dermal carcinogens acting via a genotoxic mechanism.

Additional data support that cracked gas oils are carcinogens (Biles et al., 1988; API, 1989; Broddle et al., 1996;).  This information is presented in the dossier.  In Biles et al. (1988), the dermal carcinogenicity of 10 middle distillate samples, including a light catalytic cycle oil with a boiling point <338 °C, was tested in this study.  Some individual PAH analysis was provided and very high phenanthrene levels were found in the sample. In view of some large sensitivity values (e. g. Chrysene <108 ppm), however, it would seem that the analysis method was insufficiently sensitive to measure the critical levels of 4-6 ring PAHs that may have been present in the samples. The results obtained with the test material are: median survival of 78 weeks, 250 animals bearing tumours (4%), 90 days to first tumour, and median time to first tumour of 140 days. Clear evidence of skin irritation/skin injury was seen with all the test materials including hyperplasia, hyperkeratosis, dermatitis, epidermal degeneration and epidermal necrosis. Tumour incidences in internal organs were reported to be sporadic and, except for hepatocellular carcinoma, to be of low frequency. No treatment related increase in tumours in internal organs was reported. It was noted by the authors that there was a significant increase in skin tumour yield with the majority of samples tested, but that non-neoplastic dermal changes including hyperplasia may have contributed to the tumourigenic response. In view of the questionable adequacy of the PAH analysis and the high levels of phenanthrene and pyrene found in some samples it is uncertain however whether a genotoxic mechanism can be ruled out.  

In API (1989), the dermal carcinogenicity of eleven refinery streams, including three gas oils (one straight run and two cracked), was investigated in mice.  One of the cracked gas oils showed 78% of mice developing tumours (72% malignant, 6% benign), while the other cracked gas oil showed 52% incidence of tumours (46% malignant, 6% benign).  The only conclusions drawn by the authors was that the test materials were dermatotoxic and carcinogenic and that the degree of these effects varied greatly. However examination of the tumour data suggests that the first cracked gas oil might be active as a genotoxic carcinogen as it produced a high incidence of tumours with a short latent period. The results with the second gas oil are suggestive of a possible combination of genotoxic and non-genotoxic mechanisms.

Justification for selection of carcinogenicity via dermal route endpoint:

One of 3 well conducted skin cancer studies

Carcinogenicity: via dermal route (target organ): other: skin