A GHS-consistent approach to health hazard classification of petroleum substances, a class of UVCB substances.

The process streams refined from petroleum crude oil for use in petroleum products are among those designated by USEPA as UVCB substances (unknown or variable composition, complex reaction products and biological materials). They are identified on global chemical inventories with unique Chemical Abstract Services (CAS) numbers and names. The chemical complexity of most petroleum substances presents challenges when evaluating their hazards and can result in differing evaluations due to the varying level of hazardous constituents and differences in national chemical control regulations. Global efforts to harmonize the identification of chemical hazards are aimed at promoting the use of consistent hazard evaluation criteria. This paper discusses a systematic approach for the health hazard evaluation of petroleum substances using chemical categories and the United Nations (UN) Globally Harmonized System (GHS) of classification and labeling. Also described are historical efforts to characterize the hazard of these substances and how they led to the development of categories, the identification of potentially hazardous constituents which should be considered, and a summary of the toxicology of the major petroleum product groups. The use of these categories can increase the utility of existing data, provide better informed hazard evaluations, and reduce the amount of animal testing required.

Light white oils exhibit low tissue accumulation potential and minimal toxicity in F344 rats.

Female F344 rats were fed diets containing 0.02%, 0.2%, or 2.0% white mineral oil for 90 days. There were no gross or microscopic differences in target organs at the 0.02% level. In the higher-dose groups, relative liver and mesenteric lymph node (MLN) weights were increased, and MLN inflammation was observed. At the 2% level, there was very limited evidence of microgranuloma formation in the liver but at a lower incidence and at lesser severity than has been reported in studies of C22-C25 oils. Analysis of liver extracts from treated rats revealed that C15-C20 constituents were underrepresented by comparison to their corresponding concentrations in the test oil. These results provide evidence that although hydrocarbons with carbon numbers < C20 are absorbed, they are not preferentially retained and do not contribute to inflammatory processes in liver.

Asphalt fume dermal carcinogenicity potential: II. Initiation-promotion assay of Type III built-up roofing asphalt.

Clark et al. (accepted for publication) reported that a sample of field-matched fume condensate from a Type III built-up roofing asphalt (BURA) resulted in a carcinogenic response in a mouse skin bioassay, with relatively few tumor-bearing animals, long tumor latency and chronic skin irritation. This mouse skin initiation/promotion study was conducted to assess possible mechanisms, i.e., genotoxic initiation vs. tumor promotion subsequent to repeated skin injury and repair. The same Type III BURA fume condensate sample was evaluated in groups of 30 male Crl:CD1® mice by skin application twice per week (total dose of 50 mg/week) for 2 weeks during the initiation phase and for 26 weeks during the promotion phase. Positive control substances were 7,12-dimethylbenz(a)anthracene (DMBA, 50 µg applied once) as an initiator and 12-O-tetradecanoyl-13-acetate (TPA, 5 µg, applied twice weekly) during the promotion phase. During the 6 months of study with the asphalt fume condensate, eight skin masses were observed when tested for initiation, five of which were confirmed microscopically to be benign squamous cell papillomas. Only two papillomas were observed when tested for promotion. There was no apparent relationship between skin irritation and tumor development in this study. These results are more indicative of genotoxicity rather than a non-genotoxic mode of action.

Asphalt fume dermal carcinogenicity potential: I. dermal carcinogenicity evaluation of asphalt (bitumen) fume condensates.

Asphalt (bitumen) fume condensates collected from the headspace above paving and Type III built up roofing asphalt (BURA) tanks were evaluated in two-year dermal carcinogenicity assays in male C3H/HeNCrl mice. A third sample was generated from the BURA using a NIOSH laboratory generation method. Similar to earlier NIOSH studies, the BURA fume condensates were applied dermally in mineral oil twice per week; the paving sample was applied 7 days/week for a total weekly dose of 50 mg/wk in both studies. A single benign papilloma was observed in a group of 80 mice exposed to paving fume condensate at the end of the two-year study and only mild skin irritation was observed. The lab generated BURA fume condensate resulted in statistically significant (P<0.0001) increases in squamous cell carcinomas (35 animals or 55% of animals at risk). The field-matched BURA condensate showed a weaker but significant (P=0.0063) increase (8 carcinomas or 13% of animals) and a longer average latency (90 weeks vs. 76 for the lab fume). Significant irritation was observed in both BURA condensates. It is concluded that the paving fume condensate was not carcinogenic under the test conditions and that the field-matched BURA fume condensate produced a weak tumor response compared to the lab generated sample.

Paving asphalt products exhibit a lack of carcinogenic and mutagenic activity.

A paving asphalt and a vacuum residuum (derived from crude oil by atmospheric and subsequent vacuum distillation and used as a blend stock for asphalt) were tested in skin carcinogenesis assays in mice and in optimized Ames assays for mutagenic activity. In the skin cancer tests, each substance was applied twice weekly for 104 weeks to the clipped backs of groups of 50 male C3H mice. Neither the paving asphalt nor the vacuum residuum (30% weight/volume and 75% weight/weight in US Pharmacopeia mineral oil, respectively) produced any tumors. The positive control benzo[a]pyrene (0.05% w/v in toluene) induced tumors in 46 of 50 mice, demonstrating the effectiveness of the test method. Salmonella typhimurium tester strain TA98 was used in the optimized Ames assay to evaluate mutagenic potential. Dimethylsulfoxide (DMSO) extractions of the substances were not mutagenic when tested up to toxic limits. Thus, under the conditions of these studies, neither the paving asphalt nor the vacuum residuum was carcinogenic or mutagenic.

Development of a screening tool to prioritize testing for the carcinogenic hazard of residual aromatic extracts and related petroleum streams.

Residual aromatic extracts (RAE) are petroleum substances with variable composition predominantly containing aromatic hydrocarbons with carbon numbers greater than C25. Because of the high boiling nature of RAEs, the aromatics present are high molecular weight, with most above the range of carcinogenic polycyclic aromatic hydrocarbons (PAHs). However, refinery distillations are imperfect; some PAHs and their heteroatom-containing analogs (collectively referred to as polycyclic aromatic content or PAC) may remain in the parent stream and be extracted into the RAE, and overall PAC content is related to the carcinogenic potential of an RAE. We describe here a real-time analytical chemistry-based tool to assess the carcinogenic hazard of RAE via the development of a functional relationship between carcinogenicity and boiling point. Samples representative of steps along the RAE manufacturing process were obtained from five refineries to evaluate relationships between mutagenicity index (MI), PAC ring content and gas chromatographic distillation (GCD) curves. As expected, a positive linear relationship between MI and PAC ring content occurred, most specifically for 3-6 ring PAC (R2=0.68). A negative correlation was found between MI and temperature at 5% vaporization by GCD (R2=0.72), indicating that samples with greater amounts of lower boiling constituents were more likely to be carcinogenic. The inverse relationship between boiling range and carcinogenicity was further demonstrated by fractionation of select RAE samples (MI=0.50+0.07; PAC=1.70+0.51wt%; n=5) into low and high boiling fractions, where lower boiling fractions were both more carcinogenic than the higher boiling fractions (MI=2.36±0.55 and 0.17±0.11, respectively) and enriched in 3-6 ring PACs (5.20+0.70wt% and 0.97+0.35wt%, respectively). The criteria defining carcinogenicity was established as 479°C for the 5% vaporization points by GCD, with an approximate 95% probability of a future sample having an MI below the recommended limit of 0.4 for RAEs. Overall, these results provide a cost-efficient and real-time tool by which the carcinogenic potential of RAEs can be assessed at the refinery level, ultimately providing a means to readily monitor and minimize the carcinogenic potential of RAEs.

Results of chronic dietary toxicity studies of high viscosity (P70H and P100H) white mineral oils in Fischer 344 rats.

Two-year dietary studies were conducted to determine the chronic toxicity and its reversibility, and the carcinogenicity of P70(H) and P100(H) white mineral oils in Fischer-344 rats (F-344). The studies were identical in design and followed the Organization for Economic Cooperation and Development, Guidelines for Testing Chemicals, Guideline 453, 1981. Additional endpoints evaluated were: (1) extent of mineral hydrocarbon deposition in liver, kidneys, mesenteric lymph nodes, and spleen of female rats at 3, 6, 12, 18 and 24 months, and (2) reversibility of effects following cessation of exposure. Dietary concentration were 60, 120, 240, and 1,200 mg/kg/day, adjusted periodically to account for bodyweight changes. Study results were consistent with preceding subchronic studies. No treatment-related mortality, neoplastic lesions, or changes in clinical health, hematology, serum chemistry, or urine chemistry were evident in any group administered either white oil. Statistically significant higher food consumption was noted in the 1,200 mg/kg group males and females exposed to either white oil and statistically significant higher body weights were noted in the 1,200-mg/kg males during the latter portion of the P100(H) study. Higher mesenteric lymph node weights were accompanied by increased severity of infiltrating histiocytes. This occurred to a greater extent with the P70(H) than the P100(H) oil. No other histopathology of significance was observed. Mineral hydrocarbons were detected in the liver following exposure to either oil. Maximal concentrations of mineral hydrocarbons in the liver were similar with both oils but occurred more rapidly with the P70(H) oil. Liver mineral hydrocarbon content returned to near-background levels during the reversibility phase. In conclusion, lifetime exposer of F344 rats to P70(H) and P100(H) white oils resulted in only minimal findings and with no consequence to clinical health. Thus, under the conditions of these studies, the No Observable Adverse Effect Level (NOAEL) for these studies was considered to be 1,200 mg/kg/day.