Experimental Studies

This chapter provides a review of the in vitro and in vivo animal studies completed since the publication of the NIOSH criteria document on asphalt fumes in 1977. Ideally, these studies should provide definitive data regarding the genotoxicity, carcinogenicity, and other toxic effects of asphalt-based paints and asphalt fumes generated during paving and roofing operations. Because of the difficulty in obtaining a sufficient quantity of paving and roofing asphalt fumes in the field, however, many of the studies reviewed used laboratory-generated asphalt fumes.

6.1 Genotoxicity

Since publication of the NIOSH criteria document [1977a], genotoxic effects were described in the following studies: National Toxicology Program (NTP) [NTP 1990], Blackburn and Kriech [in AI 1990a], Machado et al. [1993], Reinke and Swanson (laboratorygenerated asphalt fumes) [1993], Qian et al. [1996], Schoket et al. [1988a,b], Toraason et al. [1991], and Wey et al. [1992]. However, genotoxic effects were not observed by Reinke and Swanson in a study of fumes collected from a hot-mix asphalt storage tank [1993].

6.1.1 Mutagenic Effects

The NTP evaluated the mutagenic potential of roofing asphalt fume condensate fractions and neat (unfractionated) asphalt fumes from the study by Sivak et al. [1989]. Sivak et al. heated Type III roofing asphalt to 316 /C (601 °F) to generate fume condensates and then separated the condensates using HPLC. Five fractions, designated A through E, of the condensates and unfractionated asphalt fumes were examined for mutagenic potential using the Ames Salmonella mutagenicity assay. The chemical composition of each fraction is provided in section 6.2.1. Fractions B and C and recombined fractions A through E were reported as positive, fractions A and D and the unfractionated fumes were weakly positive, and fraction E was negative. Positive responses were observed only with metabolic activation (S9) [NTP 1990; Zeiger 1990].

The same fractionated asphalt fume condensates from the study by Sivak et al. [1989] were tested by Blackburn and Kriech [in AI 1990a] using the modified Ames Salmonella mutagenicity assay. The results were consistent with those of the NTP study. Mutagenicity indices of 21 asphalt fume samples collected under a variety of conditions ranged from 0 to 8.8, with an average of 4.7. These indices were approximately 150-fold less than the indices for coal-tar-pitch fumes.

Machado et al. [1993] evaluated the mutagenic activity and PAH content of laboratory-generated fumes from a variety of asphalts. Materials examined included two Type III roofing asphalts representing different crude petroleum sources. One of the roofing asphalts was similar to the asphalt (air-blown using a ferric chloride catalyst) examined by Niemeier et al. [1988] and Sivak et al. [1989]. Machado et al. also evaluated 18 paving asphalts representing 14 crude petroleum sources and various processing conditions and a Type I coal-tarpitch fume. Fume condensates were examined for mutagenic activity with a modified Ames Salmonella mutagenicity assay (Table 6-1) and for PAH content. The fume generation temperature of all roofing materials was either 232 /C or 316 /C (450 °F or 601 °F) and that of all paving materials was 163 /C (325 °F). One sample of paving material was heated to 221 /C (430 °F). Machado et al. reported that all samples tested showed weak-to-moderate mutagenic activity (Table 6–1). Moreover, the mutagenic responses to the asphalt fume condensates were approximately 100-fold less than mutagenic responses to the coal-tar-pitch samples.

Results of the analyses for PAH content, as measured by HPLC fluorescence, of the roofing and paving asphalts, coal-tar pitch, and their fume condensates were as follows. Concentrations of individual PAHs in samples of roofing and paving asphalt and asphalt fume condensates were less than 50 ppm by weight. Most concentrations of individual PAHs in roofing asphalt or fumes were less than10 ppm, and all concentrations in paving asphalts or fumes were less than 2 ppm. Concentrations of individual PAHs in the coal-tarpitch samples were 100- to 1,000-fold higher than in the roofing and paving samples. For example, B(a)P was detected in all samples examined; maximum concentrations in asphalt, coal-tar pitch, and asphalt and coal-tar-pitch fume condensates were approx imat e ly6 ppm, 18,000 ppm, 2.8 ppm, and 480 ppm, respectively.

Machado et al. attempted to correlate mutagenicity indices with PAH content. The correlation coefficient of the pooled data varied from 0.17 to 0.86, depending upon which samples were included in the analysis. The investigators also suggested that the crude petroleum source, along with processing conditions, had some influence on the PAH content of the various materials tested.

Reinke and Swanson [1993] compared the chemistry of PAHs and S-PACs and mutagenic potential of field- and laboratory-generated asphalt fumes from an asphalt cement. Temperatures ranged from 146 to 157 /C (295 to 315 °F) for the field samples and from 149 to 316 /C (300 to 601 °F) for the laboratory samples. The field asphalt fume condensates were collected from the headspace of an asphalt storage tank at a hot-mix asphalt production plant. Fumes were collected into a cold trap for approximately 36 continuous hours.

A summary of the chemical analyses (GC/MS) for PAHs and S-PACs and the modified Ames Salmonella mutagenicity assay is provided in Table 6–2. The mutagenicity index of the storage tank headspace asphalt fumes was between 0 and 1, while the mutagenicity indices of the asphalt fumes generated in the laboratory at 149 and 316 °C (300 and 601°F) were 5.3 and 8.3, respectively. The authors noted positive trends between mutagenicity indices and the percentage of $three-ring PAHs and S-PACs. They suggested that the increased mutagenicity of the fumes generated at 316 /C (601 °F) could be attributed to its increased content of four-ring S-PACs (Table 3–7).

A study was undertaken in Europe by De Méo et al. [1996a] to compare the mutagenic potential of fume condensates generated at 160 and 200 °C (320 and 392 °F) from coal tar and two paving asphalts (45/60 pen and 160/210 pen) in a modified Ames Salmonella mutagenicity assay [De Méo et al. 1996b]. Modifying the procedure of Brandt et al. [1985], the authors generated coal tar and asphalt fume condensates (a mix of the vapor phase and the particulate phase) that they considered to be representative of fumes produced in the field [1996a]. The condensates were tested for mutagenic activity in the presence and absence of metabolic activation (S9) using the Salmonellatester strains TA98, TA100, YG1041, and YG1042. All fume condensates were mutagenic to all bacterial tester strains only in the presence of metabolic activation. The mutagenic potencies of the coal-tar fume condensates were 15- to 600-fold higher than those of the asphalt fume condensates. The authors further investigated the effect of these fume condensates on in vitro DNA-adduct formation; these results are presented in section 6.1.3.

Robinson et al. [1984] examined the mutagenic potential of several asphalt-based paints using the Ames Salmonella mutagenicity assay. None of the asphalt-based paints demonstrated mutagenic activity in either the presence or absence of metabolic activation (S9).

6.1.2 Chromosomal Aberrations

Condensates of Type I and Type III roofing asphalt fumes generated in the laboratory at 316±10 °C (601 °F) using the same methodology as in Sivak et al. [1989] caused a dose-related increase in micronucleus formation in exponentially growing Chinese hamster lung fibroblasts (V79 cells) [Qian et al. 1996]. The results of immunofluorescent antibody staining showed that both roofing asphalt fume condensates induced mainly kinetochore-positive micronuclei (68 to 70%). The authors suggested that Type I and Type III roofing asphalt fume condensates are aneuploidogens and possess some clastogenic activities. The asphalt fume condensates cause mainly cytogenetic damage by spindle apparatus alterations in cultured mammalian cells.

Reinke and Swanson [1993] tested three paving asphalt fume condensates generated in the field and in the laboratory in an unspecified chromosomal aberration assay. Results were negative. The authors suggested that "the absence of positive findings" may be explained by the fact that this assay was not "optimized for petroleum asphalt fumes."

6.1.3 DNA-Adduct Formation

De Méo et al. [1996a] also tested coal tar and paving (45/60 pen and 160/210 pen) asphalt fume condensates generated at 160 and200 °C (320 and 392 °F) for their ability to produce DNA adducts in vitro when added to calf thymus DNA. DNA-adduct formation was assessed by 32P-postlabeling; the fume con-densates were diluted in acetone to a final con-centration of 20 µg in 1 mL. B(a)P was used as the positive control. The authors reported that all of the fume condensates induced DNA-adduct formation. No specific DNA adducts were identified. The authors further noted that the patterns of the autoradiograms of the DNA adducts demonstrated qualitative differences, indicating qualitative differences in the nature of the compounds in the coal tar and asphalt fume condensates responsible for the for-mation of these adducts.

As a continuation of the De Méo et al. study [1996a], Genevois et al. [1996] tested the same coal tar and paving asphalt fume condensates for their ability to induce DNA-adduct for-mation in vivo. They applied 100 µL (about 100 mg) of the undiluted fume condensates to the shaved dorsal skin of BD4 rats (three rats per group) twice 2 days apart; three untreated rats served as the control group. Twenty-four hours after the last treatment, all of the rats were sacrificed, and skin, lungs, and lymphocytes were collected. DNA-adduct formation was assessed by 32P-postlabeling. DNA adducts were found in skin, lungs, and lymphocytes of all the treated rats, but no specific DNA adducts were identified. These in vivo data are in agreement with De Méo et al.’s in vitro data and indicate qualitative differences in the nature of the compounds in the coal tar and asphalt fume condensates responsible for the induction of the DNA adducts.

De Méo et al. and Genevois et al. also analyzed the fume condensates for PAH content using HPLC-fluorescence. The data indicated that while large amounts of unsubstituted PAHs are present in the coal-tar fumes, these compounds are only minor constituents of the asphalt fume condensates.

In two studies, Schoket et al. reported upon the formation of DNA adducts in (1) mice [1988a] and (2) adult and fetal human skin samples maintained in short-term tissue culture [1988b]. In both studies, asphalt- or creosote-based paints or pharmaceutical-grade coal tar were applied topically, and DNA-adduct formation was assessed by 32P-postlabeling. The results in both studies suggested that a variety of adducts were formed from the three materials; however, no specific DNA adducts were identified.

The DNA-adduct concentrations found in mouse skin 24 hours after a single application of the test materials are listed in Table 6–3. Mice that received multiple applications of the three agents showed accumulations of DNA adducts in both skin and lung tissue. The DNA-adduct concentrations observed in the multiple- treatment studies were reported in graphic form only; the adduct concentrations found in the lungs were consistently lower than those in the skin. The authors concluded that the detection of DNA adducts in the lungs demonstrated that PAHs in the three agents were absorbed from the skin and metabolically activated in organs distant from the site of application.

DNA-adduct concentrations found in adult human skin 24 hours after a single application of the test materials are listed in Table 6–4. The authors concluded that similar amounts of DNA adducts are formed from these materials in both mouse and human skin.

6.1.4 Intercellular Communication

The five asphalt roofing fume fractions used by Sivak et al. [1989] were tested for inhibition of intercellular communication. The inhibition of intercellular communication by a tumor promoter is believed to isolate an initiated or preneoplastic cell from the growth regulatory signals of surrounding cells, leading to the development of neoplasia. All fractions inhibited intercellular communication inChinese hamster lung fibroblasts (V79 cells) [Toraason et al. 1991]. The greatest activity occurred in fractions D and E, and the least activity in fraction A.

Similarly, Wey et al. [1992] examined the effect of these fractions on intercellular communication in human epidermal keratinocytes. All asphalt roofing fume fractions inhibited intercellular communication in a concentration-dependent fashion.

6.2 Carcinogenicity

Since publication of the NIOSH criteria document [1977a], there have been several reports of carcinogenicity in mice following applications of laboratory-generated asphalt roofing fume condensates [Thayer et al. 1981; Niemeier et al. 1988; Sivak et al. 1989, 1997], raw roofing asphalt [Sivak et al. 1989, 1997], and asphalt-based paints [Robinson et al. 1984; Bull et al. 1985] to the skin of mice. However, in another study [Emmett et al. 1981], raw roofing asphalt applied dermally to mice was not carcinogenic.

6.2.1 Roofing Asphalt Fume Conden-sates and Raw Asphalt

Niemeier et al. [1988] Niemeier et al. [1988] investigated the tumorigenicity of fume condensates generated in the laboratory at 232 and 316 °C (450 and 601 °F) from Type I and III roofing asphalt and Type I and III coal-tar pitch. All fume samples were cryogenically collected. Fumecondensates were applied biweekly to the skin of male CD-1 (nonpigmented) and C3H/HeJ (pigmented) mice for 78 weeks. Eighteen groups of 50 mice per strain received these ap-plications singly or in combination. Half of each group was exposed to simulated sunlight.

Tumors were produced by fume condensates of both types of asphalt (Tables 6–5 and 6–6) and both types of coal-tar pitch. The majority of benign tumors were papillomas; the majority of malignant tumors were squamous cell carcinomas. Both strains of mice exposed to asphalt fumes had significantly (P=0.01) more tumors than the control groups, although the C3H/HeJ mice demonstrated a greater tumorigenic and carcinogenic response to both asphalt and coal-tar-pitch fume solutions than did the CD-1 mice. The C3H/HeJ mice showed a significant increase (P=0.01; Fisher-Irwin exact test) in tumorigenic response for both types of condensed asphalt fumes generated at 316 °C (601 °F) compared with tumors gene-r a t e d a t 232 °C (450 °F). The mean time to tumor appearance was longer for all groups of CD-1 mice compared with the correspondingC3H/HeJ groups. The mean latency period ranged from 39.5 to 56.1 weeks among the C3H/HeJ groups and from 47 to 76.5 weeks among the CD-1 groups treated with roofing asphalt fume condensates. Mean latency time increased with simulated sunlight, which generally inhibited tumorigenic response, possibly because of photo oxidation or photodestruction of the carcinogenic components of the test materials. Niemeier et al. [1988] concluded the following:

• Unlike the carcinogenic activity of coal-tar pitch, that of asphalt fume condensates could not be explained by B(a)P (or PAH) content.
• The carcinogenic activity of the asphalt fume condensates may have been due to the high concentrations of aliphatic hydrocarbons, which have co-carcinogeniceffects.
• Higher generation temperatures may have increased carcinogenic effects.

Sivak et al. [1989, 1997] Sivak et al. [1989, 1997] heated Type III roofing asphalt from the same lot used by Niemeier et al. [1988] at 316 °C (601 °F), generated fume condensates, and separated them by HPLC (see Belinky et al. 1988 for a description of this procedure). The chemical composition of fractions A through E, as analyzed by GC/MS, is provided in Table 3–8. Raw roofing asphalt, neat asphalt fumes, asphalt heated to 316°C (601°F) with fumes allowed to escape, reconstituted asphalt fumes, and the asphalt fume fractions—individually and in various combinations, were then tested for their carcinogenic and tumor-promoting activity. Fractions A through E were dissolved in a 1:1 solution of cyclohexane and acetone to yield concentrations proportional to their presence in the unfractionated (neat) asphalt fume condensate, i.e., 64.1%, 8.3%, 10.5%, 11.5% and 5.6%, respectively, and were applied biweekly to 40 groups of male C3H/HeJ mice and two groups of male Sencar mice (30 mice per group) for 104 weeks (2 years). The Sencar mice were included to allow for possible genetic variation and susceptibility to tumor development.

Raw roofing asphalt, neat asphalt fumes, asphalt heated to 316°C (601°F) with fumes allowed to escape, reconstituted asphalt fumes, and the asphalt fume fractions—individually and in various combinations, were then tested for their carcinogenic and tumor-promoting activity. Fractions A through E were dissolved in a 1:1 solution of cyclohexane and acetone to yield concentrations proportional to their presence in the unfractionated (neat) asphalt fume condensate, i.e., 64.1%, 8.3%, 10.5%, 11.5% and 5.6%, respectively, and were applied biweekly to 40 groups of male C3H/HeJ mice and two groups of male Sencar mice (30 mice per group) for 104 weeks (2 years). The Sencar mice were included to allow for possible genetic variation and susceptibility to tumor development.

Table 6–7 shows all the treatment groups, the number of papillomas and carcinomas per group, the number of tumor-bearing mice, and the average time (in weeks) to carcinoma de-velopment. The raw roofing asphalt and neat asphalt fumes induced carcinomas (local skin cancers) in three of 30 and 20 of 30 C3H/HeJ mice, respectively. However, the heated asphalt with fumes allowed to escape did not induce any tumors. Fractions B and C induced carcinomas in 10 of 30 and 17 of 30 C3H/HeJ mice, respectively, while fractions A, D, and E failed to induce any carcinomas when applied alone. All the combinations of the fractions induced tumors only if they included B or C. The A and D combination, the A and E combination, and the A, D, and E combination failed to induce any tumors. Furthermore, fractions A, D, and E failed to act as either tumor promoters or co-carcinogens. Fourteen of the 30 Sencar mice treated with the asphalt fume condensate developed carcinomas.

As noted in the preceding paragraph, only fractions B and C, whether applied alone or in combination, elicited tumor responses. Frac-tions B and C contained PACs that included PAHs, S-PACs, and O-PACs, such as alky-lated aryl thiophenes, alkylated phenanthrenes, alkylated acetophenones, and alkylated dihy-drofuranones. Fraction B contained most of the S-PACs, and only a few were carried over to fraction C. Fraction C contained a small amount of four-ring PACs (refer to Table 6–7). Sivak et al. [1989, 1997] stated the need for additional co-carcinogenesis and tumor-promotion experiments using a wider range of experimental variables, further chemical sepa-ration of fractions B and C, more short-term genotoxicity assays, and additional carcino-genicity assays to identify biologically active materials in the roofing asphalt fume condensates.

Emmett et al. [1981] In an earlier study, Emmett et al. [1981] examined the carcinogenic potential of a standard roofing asphalt (asphalt type not provided) dissolved in redistilled toluene at a 1:1 ratio by weight. Fifty milligrams of this solution was applied twice a week to the shaved intrascapular region of the back of 50 male C3H/HeJ mice. The vehicle control group of 50 mice received 50 mg of toluene twice a week, and the positive control group received 50 mg of 0.1% B(a)P in toluene twice a week. The dosing regimen continued for 80 weeks or until a skin lesion was diagnosed as a papilloma; when a papilloma progressed and was diagnosed grossly as a carcinoma, the tumor-bearing mouse was sacrificed and autopsied. Selected histopathological examination of the tumors confirmed the gross diagnosis.

No tumors were observed in the mice treated either with roofing asphalt or toluene. Twentysix of the first group and 37 of the second group survived 60 or more weeks. Seventynine percent of the mice treated with B(a)P developed tumors. Chemical analysis by gas chromatography (with an electron capture detector) of the raw roofing asphalt indicated that B(a)P concentrations were below the level of detection, i.e., <0.0004%.

6.2.2 Asphalt-Based Paints

Robinson et al. [1984] examined the effects of four formulations of asphalt-based paints (lab-eled A through D) and three formulations of coal-tar-based paints (labeled E, F, and G) using female Sencar mice in mouse skin bioassays. All formulations except G had been used to prevent corrosion in drinking water distribution systems [Alben 1980; Miller et al. 1982]. The asphalt-based paints were formulations containingxylene, or xylene and mineral spirits with be-tween 89% and 98% cutback asphalt.

The asphalt- and coal-tar-based paints were evaluated not only for their potential tumor-initiating ability, but also for their ability to function as complete carcinogens. Both the coal-tar and asphalt paint formulation groups initiated tumor development in mouse skin. The activity exerted by the coal-tar paints (data not presented) was approximately 100-fold greater than the activity exerted by the asphalt-based paints. Table 6–8 presents data demon-strating the tumor-initiating activity of the asphalt-based paints, provides gross tumor observations, and classifies tumors examined histologically [Robinson et al. 1984; Bull et al. 1985]. Animals receiving the initial 200-µL dose of the asphalt solutions showed a statistically significant increase (P<0.05; multiple-cell chi square analysis) in both the number of tumor-bearing animals and number of tumors per animal compared with animals treated with mineral spirits [Bull et al. 1985]. However, the tumor response induced by the coal-tar paints was greater than that induced by the asphalt-based paints, even though the vol-ume of the coal-tar paints (0.2-20 µL) was less than the volume of the asphalt-based paints (200 µL).

Only coal-tar paint formulation E and asphalt paint D were analyzed for their ability to act as complete carcinogens. Two microliters of coal tar and 200 µL of asphalt D were applied to 40 female Sencar mice once a week for 30 weeks; the mice were sacrificed after 52 weeks. Under the experimental conditions provided, only coal-tar formulation E acted as a complete car-cinogen. It induced the development of 171 tu-mors (papillomas) in 83%, or 33 of 40, of the mice; 10% of the mice had carcinomas (four animals had five carcinomas). Of the mice that had been treated with asphalt D, only one in 40 (3%) developed a tumor (papilloma), while three of 40 mice in the group treated with mineral spirits developed papillomas.

Chemical analyses of coal-tar formulations E, F, G, and H and asphalt-based paints A and D, which are also used to prevent corrosion in drinking water systems, were conducted using GC/MS. Results of the analyses indicated that PAH concentrations were high in coal-tar formulations E, F, G, and H, and very low in asphalt-based paints A and D. This observation is based on the five biologicallyactive PAHs (chrysene, benz[a]anthracene, B(a)P, benzo[e]pyrene, and phenanthrene)found in coal-tar-based paints. The concentrations of these five compounds as a percentage of total PAHs were 44% in coal tar E, 42% in coal tar F, and 33% in coal tar G. Of these five PAHs, only trace amounts ofphenanthrene (<0.01%) were found in both asphalt-based paints.

Robinson et al. [1984] concluded that the four asphalt-based paints tested contained chemicals capable of initiating tumors in mice, and that a number of these tumors were carcinomas. However, they were not found to be complete carcinogens.

6.3 Conclusions

The following conclusions concerning the adverse effects of asphalt fumes, raw asphalt, and asphalt-based paints are based on the results of the preceding experimental studies.

6.3.1 Asphalt Fumes

Asphalt fumes are comprised of complex chemical mixtures generated by the volatilization of asphalt. In attempts to simulate occupational exposure to asphalt fumes in experimental animals, investigators developed several methods for generating asphalt fumes in the laboratory. It has yet to be determined whether these fumes are representative of the fumes to which workers are exposed during the manufacture and application of asphalt products (see section 3.4.3). Currently, available data generated using asphalt fume condensates (fumes were collected above the asphalt surface inside a hot-mix asphalt storage tank) are limited. A comparison of the biologic activity of these storage tank fumes and asphalt fume condensates generated in the laboratory at typical paving temperatures indicated that (1) fumes collected from a hot-mix asphalt storage tank were not mutagenic (Table 6–2) and (2) the laboratory-generated fumes were mutagenic [Reinke and Swanson 1993]. In other studies, paving and roofing asphalt fumes generated in the laboratory under a variety of conditions were also mutagenic [Machado et al. 1993; NTP 1990; AI 1990a; De Méo et al. 1996a]. These results indicate that asphalt fumes collected above the asphalt surface inside a hotmix asphalt storage tank and laboratorygenerated asphalt fume condensates may not induce similar biologic activity. In addition, fumes generated in the laboratory from two paving asphalts at 160 and 200 /C (320 and 392 °F) induced DNA-adduct formation in vitro and in vivo [De Méo et al. 1996a; Genevois et al. 1996].

Two studies examined the carcinogenic potential of roofing asphalt fume condensates generated in the laboratory at temperatures approximating those observed in typical and worst-case roofing operations [Niemeier et al. 1988; Sivak et al. 1989, 1997]. The data indicated that roofing asphalt fume condensates generated in the laboratory and applied dermally cause benign and malignant skin tumors in several strains of mice. Furthermore, additional data supportive of carcinogenicity demonstrated that these and similarly derived laboratory roofing asphalt fume condensates are mutagenic in the Ames Salmonella mu-tagenicity assay [NTP 1990; AI 1990a; Machado et al. 1993], induce micronuclei formation [Qian et al. 1996], and inhibit intracellular communication in mammalian cells [Toraason et al. 1991; Wey et al. 1992]. Differences in chemical composition and physical characteristics have been noted between roofing asphalt fumes collected in the field and those generated in the laboratory (see chapter 3, Kriech and Kurek [1993]). How-ever, it is not known if these differences are responsible for the genotoxic and carcinogenic effects reported in the preceding experimental studies. Although no animal studies have ex-amined the carcinogenic potential of asphalt fumes collected during roofing operations, the carcinogenic response using laboratory-gen-erated asphalt fumes suggests a potential hazard to roofers.

Since no animal studies have examined the carcinogenic potential of either field or laboratory-generated samples of paving asphalt fume condensates, no definitive determination can be made about the carcinogenic potential of paving asphalt fume condensates in experi-mental animals. However, the positive muta-genic responses obtained using laboratory-generated paving asphalt fumes are a cause for concern and support the need for further research.

6.3.2 Raw Asphalt and Asphalt Paints

Conflicting results from two separate studies [Sivak et al. 1989,1997; Emmett et al. 1981] were obtained when raw roofing asphalts were applied to the skin of mice. The raw roofing asphalt used by Sivak et al. [1989, 1997] was weakly carcinogenic and caused malignant skin tumors, while the raw roofing asphalt used by Emmett et al. [1981] did not. Available data also indicate that several for-mulations of asphalt-based paints caused benign and malignant skin tumors in mice [Robinson et al. 1984; Bull et al. 1985]. However, these paints were not mutagenic in the Ames Salmonella mutagenicity assay either with or without metabolic activation (S-9). Several other asphalt-based paints were positive in another type of genotoxicity assay, i.e., DNA-adduct formation, which is postulated to be one of the steps responsible for mutagenesis and carcinogenesis [Schoket et al. 1988a]. These asphalt-based paints also caused the formation of DNA adducts in the skin and lungs of treated mice and in fetal and adult human skin cultures [Schoket et al. 1988a,b].

The results are conflicting as to the carcinogenicity of raw roofing asphalt. One study reported a weak carcinogenic response [Sivak et al. 1989, 1997], while another study reported no carcinogenic response [Emmett et al. 1981]. However, the data indicate that the asphalt paint formulations used in the preceding studies [Robinson et al. 1984; Bull et al. 1985; Schoket et al. 1988a,b] are carcinogenic and exert some genotoxicity. Although no published data exist that examine the carcinogenic potential of asphalt-based paints in humans, NIOSH concludes that asphalt-based paints are potential occupationalcarcinogens.