National Toxicology Program

LITERATURE REVIEW

OF HEALTH EFFECTS CAUSED BY

OCCUPATIONAL EXPOSURE TO ASPHALT FUMES

This Interim Review Produced by NIOSH in Support of
Nomination to the National Toxicology Program

6/23/97

CHEMICAL AND PHYSICAL PROPERTIES*

*Information obtained from Sax and Lewis [1987].

________________________

Physical state at room temperature Black or dark-brown solid or viscous liquid

Solubility in water at 20°C Insoluble

Solubility in organ solvents Carbon disulfide

Definition of asphalt -

PRODUCTION, USE, AND POTENTIAL FOR OCCUPATIONAL EXPOSURE

Paving asphalts are manufactured principally by simple atmospheric distillation or by atmospheric distillation followed by fractionation under vacuum. They may also be manufactured by solvent precipitation and mild partial air-blowing. Roofing asphalts are generally produced by atmospheric or vacuum distillation followed by air-blowing [NAPA 1994].

Most of the asphalt produced in the United States is used in paving and roofing. Only about 1% is used for waterproofing, dampproofing, insulation, paints, and other activities [AI 1990a]. The National Occupational Exposure Survey (NOES) [NIOSH 1983] estimates that during the period 1981-83, more than 473,000 U.S. employees were potentially exposed to asphalt. Table 3 presents the 10 industries and the 10 occupations (excluding janitors) with the most employees potentially exposed to asphalt.

Paving Asphalt

Of the three types of asphalt products used in the construction of paved surfaces in the United States: asphalt paving cements (hot-mix asphalt or HMA), cutback asphalts, and asphalt emulsions, HMA (asphalt mixed with mineral aggregate) accounts for 85% of the total used. Cutback asphalts and asphalt emulsions are used for road sealing and maintenance, and account for 4% and 11% respectively, of the total used. Currently, about 4,000 HMA facilities and 7,000 paving contractors employ nearly 300,000 employees in the United States [AI 1990a].

Roofing Asphalt

Four types of asphalt (I through IV) are used in roofing products in the United States. The type of asphalt used is determined by the grade or slope of the roof. For example, Type I roofing asphalt, often referred to as "dead level," has a low softening point and is used on surfaces with a grade of 0.5 inch per foot or less. Types II and III roofing asphalt are typically used on roofs with slopes of 0.5 to 1.5 and 1 to 3 inches per foot, respectively. Type IV roofing asphalt (a hard asphalt with a high softening point) is used on roofs with a grade of 2 to 6 inches per foot [ASTM 1992]. In 1990, an estimated 46,000 on-roof employees were exposed to asphalt fumes in the United States, and about 6,000 to 12,000 employees were exposed in approximately 120 plants manufacturing asphalt roofing shingles and rolls and modified bitumen² roofing products [AI 1990a].

General Exposure

The major route of occupational exposure to asphalt fumes (e.g., paving, roofing, and asphalt-based paints) is by inhalation; they may also be absorbed through the skin. A summary of representative information on the occurrence of asphalt fumes in the workplace is presented in Table 3 and Table 4.

Dermal exposure to asphalt fumes has been examined using skin wipes (see Table 5). Skin wipe samples were collected at various worksites (e.g., refineries, HMA facilities, paving and roofing sites, and roofing manufacturers) and analyzed for PAHs [AI 1991]. The PAH concentrations determined from postshift samples ranged from 2.2 to 520 ng/cm² (see Appendix A).

Exposure Limits

The Occupational Safety and Health Administration (OSHA) currently has no permissible exposure limit (PEL) for asphalt fumes. In 1989, OSHA announced that it would delay a final decision to establish a PEL for asphalt fumes because of complex and conflicting issues submitted to the record [54 Fed. Reg. *2641]. The PEL originally proposed to reduce the potential carcinogenic risk of occupational exposure to asphalt fumes was 5 mg/m³ as an 8-hr TWA. In 1992, OSHA published another proposed rule for asphalt fumes that included a PEL of 5 mg/m³ (total particulates) for general industry and for the maritime, construction, and agricultural industry [57 Fed. Reg. 26182]. Comments are still being received by OSHA and a final decision is pending.

In a 1977 criteria document, NIOSH established a recommended exposure limit (REL) of 5 mg/m³ as a 15 min ceiling for up to a 10-hr work shift, during a 40-hr workweek, to protect against irritation of the serous membranes of the conjunctivae and the mucous membranes of the respiratory tract [NIOSH 1977a]. In 1988, NIOSH testimony to the Department of Labor and OSHA recommended that asphalt fumes be considered a potential occupational carcinogen [NIOSH 1988]. This recommendation was based on information presented in the 1977 criteria document [NIOSH 1977a] and a study by Niemeier et al. [1988] showing that exposure to condensates of asphalt fumes caused skin tumors in two strains of mice.

The American Conference of Governmental Industrial Hygienists (ACGIH) threshold limit value (TLV^®) is 5 mg/m³ as an 8-hr TWA and was recommended to reduce the risk of possible carcinogenicity [ACGIH 1991]. Australia, Belgium, Denmark, and the United Kingdom have also limited occupational exposures to asphalt fumes to 5 mg/m³ as an 8-hr TWA. Additionally, the United Kingdom has established a short-term exposure limit (STEL) of 10 mg/m³. Germany currently rates asphalt fumes as "suspected of having a carcinogenic potential [ILO 1991].

STUDIES OF GENOTOXICITY AND CARCINOGENICITY (ANIMALS)

Mutagenic Effects

The five fractions of laboratory-generated roofing asphalt fume condensates and unfractionated asphalt fumes used by Sivak et al [1989] (see description of Sivak study under Carcinogenic Effects) were examined for their mutagenic potential in Salmonella. Fractions A through E combined, and fractions B and C were positive; fractions, A, D, and neat asphalt fumes were weakly positive; and fraction E was negative [NTP 1990]. Positive responses required exogenous metabolic activation. The same fractionated asphalt fume condensates from the Sivak et al.study [1989] were also tested using a modified Ames assay [ Blackburn and Kriech 1990] and the results were comparable to those of the NTP [1990] study.

Eight asphalt fume samples collected on teflon filters at HMA plants as part of an Interagency Agreement with the Federal Highway Administration (FHA) were tested for mutagenic activity in a Salmonella mutagenicity assay. Preliminary results indicate that there was no mutagenic activity in the whole fume fraction; however, results of 2 of the 8 samples were inconclusive [Olsen, personal communication].

Two Type III roofing asphalts representing different crude oil sources , one of which was similar to the asphalt air-blown using a ferric chloride catalyst and used by Niemeier et al.[1988] and Sivak et al. [1989]; 18 paving asphalts (representing 14 crude oil sources and various process conditions); and Type I coal tar pitch; and their fume condensates were examined not only for mutagenic activity in a modified Ames assay, but also for PAH content [Machado et al. 1993]. The fume generation temperature of all roofing materials was either 232 or 316°C and that of all paving materials was 163°C (one sample was heated to 221°C). The results of the modified Ames assay are presented in Table 6.

The data indicate that all samples tested exerted mutagenic activity; however, the mutagenic responses of the asphalt fume condensates were approximately 100-fold less than the coal tar pitch samples and weak to moderate in potency [Machado et al. 1993]. Responses for the positive control group were all within the expected ranges.

Results of analyses for PAH content, measured by HPLC fluorescence, of the roofing and paving asphalts, coal tar pitch, and their fume condensates were as follows [Machado et al. 1993]. Concentrations of individual PAHs in samples of asphalt and asphalt fume condensates were less than 50 parts per million by weight (ppm), while most concentrations of individual PAHs in roofing (232°C or 316°C) and all concentrations in paving (163°C, except for one sample at 221°C) asphalts, whole or fumes, were less than 10 ppm and 2 ppm, respectively. Concentrations of individual PAHs in the coal tar pitch samples were 100- to 1000- fold higher than in the roofing and paving samples. Benzo[a]pyrene (BP) was detected in all samples examined; the maximum concentrations of BP in whole asphalt, whole coal tar pitch, asphalt and coal tar pitch fume condensates were approximately 6 ppm, 18,000 ppm, 0.1 - 2.8 ppm, and 250-480 ppm, respectively.

Although PAH content correlated with mutagenicity indices for some samples, for others it did not. The investigators concluded that the data suggest that crude oil source along with processing conditions had some influence on the PAH content of the various materials tested [Machado et al.1993].

Reinke and Swanson [unpublished data 1993] examined the relationship between field-, 146-157°C (295-314°F), and laboratory-generated, 149°C (300°F) and 316°C (600°F), asphalt fume condensates by comparing their chemical content (i.e., PAHs and sulfur heterocyclics) and mutagenic potential. The asphalt tested was a straight run, vacuum distilled 85/100 penetration grade asphalt derived from a blend of Canadian heavy, sour crudes. The field asphalt fume condensates were collected from the head space above an asphalt storage tank, stored between 146-157°C (295-314°F), at a HMA production plant into a cold trap system for about 36 continuous hours. The results of the chemical analyses (GC-MS) for PAHs and sulfur heterocyclics and the modified Ames assay are provided in Table 7 and summarized in Table 8.

The data indicate that field-generated asphalt fume condensates exerted a MI of >0 and < 1, while fumes generated in the laboratory at 149°C (300°F) and 316°C (600°F), exerted MIs of 5.3 and 8.3, respectively.

Chromosomal Aberrations

Condensates of Type I and Type III roofing asphalt fumes generated in the laboratory (same methodology as Sivak et al. 1989) at temperatures (316 ± 10C) similar to actual roofing operations caused a dose-related increase in micronucleus (MN) formation in exponentially growing Chinese hamster lung fibroblasts (V79 cells) [Qian et al. 1995]. The results of immunofluorescent antibody staining showed that both roofing asphalt fume condensates induced mainly kinetochore-positive MN (68-70%). The authors suggested that Type I and Type III roofing asphalt fume condensates are aneuploidogens and possess some clastogenic activities.

Reinke and Swanson [1993] also tested 3 asphalt fume condensates (field and lab-generated) in a chromosomal aberration assay and the results were negative. The authors reported that the absence of positive findings may be explained by the fact that this assay has not as yet been optimized for petroleum asphalt fumes.

Intercellular Communication

The five asphalt roofing fume fractions used by Sivak et al. [1989] were tested for inhibition of intercellular communication, i.e., one of several proposed mechanisms of tumor promotion. 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 in chinese hamster lung fibroblasts (V79) cells in Toraason et al. [1991]. The greatest activity was in fraction 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 concentrations in a concentration dependent fashion.

Carcinogenic Effects

Since publication of the NIOSH criteria document [NIOSH 1977a], there have been reports of carcinogenicity following dermal applications of laboratory-generated asphalt roofing fume condensates [Niemeier et al. 1988; Sivak et al. 1989] and raw roofing asphalt [Sivak et al. 1989]. Additional data from these studies are summarized in detail in Appendix B.

Niemeier et al. [1988] investigated the tumorigenicity of fume condensates generated at 232°C (450°F) and 316°C (601°F) from Types I and III roofing asphalt and Types I and III coal-tar pitch through topical applications to the skin of male CD-1 and C3H/HeJ mice. A total of 48 groups of 50 mice each (1 strain) received applications of cryogenically collected fume condensates singly and in combination (Type III asphalt and Type I coal-tar pitch, both generated at 316°C [601°F]) biweekly for 78 weeks (18 months). Half of each group was exposed to simulated sunlight to determine whether photochemical reactions might alter the carcinogenic activity. Analysis of the skin painting solutions by GC/MS revealed that the solutions containing coal-tar pitch fume condensates had higher concentrations of select PAHs than the solutions containing asphalt fume condensates. The authors report that analysis by nuclear magnetic resonance (NMR) indicated that the asphalt fume condensate was <1% aromatic and >99% aliphatic, whereas the coal-tar pitch condensate was >90% aromatic. BaP was selected as a marker compound based on correlations of BaP concentrations and carcinogenicity.

Tumors were produced by fume condensates of both types of asphalt (see Tables 9 and 10) and both types of coal-tar pitch. The majority of benign tumors were papillomas; the majority of malignant tumors were squamous cell carcinomas. The fume condensates from the coal-tar pitches had slightly greater carcinogenic activity than the fume condensates from the asphalts, but the total amount of select PAHs or BaP needed to produce a 50% tumor incidence was much smaller for the asphalt fume condensates (PAHs, 0.58 to 2.63 mg; BaP, less than or equal to 13.6 mg) than for the coal-tar pitch fume condensates (PAHs, >24.5 to >57.4 mg; BaP, 354 to 405 mg). Tumor response to the coal-tar pitch fume condensates was comparable with that of the BaP controls, based on the total dosage of BaP administered. 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 those generated at 232°C (450°F); a similar increase was noted only for Type III coal-tar pitch fumes. Overall, simulated sunlight inhibited tumorigenic responses. The authors speculated that this inhibition may have resulted from the photo-oxidation or photodestruction of the carcinogenic components of the test materials. Niemeier et al. [1988] concluded that the enhanced carcinogenic activity of the asphalt fume condensates may have been due to their high concentration of aliphatic hydrocarbons, which have cocarcinogenic effects. They also concluded that higher generation temperatures may have further increased that hazard. Finally, Niemeier et al. [1988] concluded that the carcinogenic activity of the coal-tar pitch fume condensates (but not that of the asphalt fume condensates) could be explained by their BaP (or PAH) contents.

Sivak et al. [1989] heated Type III roofing asphalt at 316°C, generated fume condensates, and separated them by high-performance liquid chromatography [Belinky et al.]. The chemical composition of the fractions (A through E) is provided in Table 11. Raw asphalt, neat asphalt (whole or unfractionated condensate) fume, the reconstituted asphalt fume, and the asphalt 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 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 2 groups of Sencar mice (30 male mice per group) for 104 weeks (2 years).

A single initial treatment of BaP followed by individual treatments with fractions A, D, and E was used to test the tumor-promoting activity of the asphalt fume condensate. The cocarcinogenicity of fractions A, D, and E was tested with three different doses of BaP. Fractions A, D, and E were used because they were the fractions Sivak et al. [1989] deemed most likely to exhibit cocarcinogenic or tumor-promoting activity based on their chemical compositions, i.e., primarily long chain alkanes and phenol compounds. Two groups of male Sencar mice were included to allow for possible genetic variation and sensitivity to tumor promotion. One of the two groups of Sencar mice was treated with neat asphalt fume (whole condensate), and the other was used as an unexposed solvent control. The negative control group was treated with cyclohexane and acetone, and the positive control groups were treated with three different concentrations of BaP.
Table 12 presents only the treatment groups which induced histopathologically confirmed carcinomas (malignant tumors), the number of carcinomas per group, the number of mice with histologically confirmed carcinomas, and the average time (in weeks) to carcinoma development. The raw asphalt and neat asphalt fume induced carcinomas (local skin cancers) in 3 of 30 and 20 of 30 C₃H/HeJ mice, respectively. 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 singly. All the combinations of the fractions induced tumors only if they included B or C; combinations A and D; A and E; and A, D, and E failed to induce any tumors. Furthermore, fractions A, D, and E failed to act as either tumor promoters or cocarcinogens. Fourteen of the 30 Sencar mice treated with the asphalt fume condensate developed carcinomas.

As noted previously, only fractions B and C applied singly and in combinations elicited tumor responses. Fractions containing B and C PACs including PAHs, S-PACs, and O-PACs such as alkylated aryl thiophenes, alkylated phenanthrenes, alkylated acetophenones, and alkylated dihydrofuranones. 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 4-ring PACs (refer to previous Table). Sivak et al. [1989] stated the need for additional cocarcinogenesis and tumor-promotion experiments using a wider range of experimental variables, further chemical separation of fractions B and C, more short-term genotoxicity assays, and additional carcinogenicity assays to identify biologically active materials in the roofing asphalt fume condensates.

Table 10 lists the positive tumor responses among the groups of mice studied. The raw asphalt (diluted with a 1:1 solution of cyclohexane and acetone to a final concentration of 0.5 g/ml) produced carcinomas in 3 of 30 C3H/HeJ mice. The neat asphalt fume (diluted with a 1:1 solution of cyclohexane and acetone to a final concentration of 0.5 g/ml) produced carcinomas in 20 of 30 C3H/HeJ mice. Fraction B produced local skin cancers (carcinomas) in 10 of 30 male C3H/HeJ mice, and fraction C produced local skin cancers (carcinomas) in 17 of 30 male C3H/HeJ mice. Fractions A, D, and E failed to produce any carcinomas when applied singly. Of the other combinations of fractions, all produced tumors except the following: A and D; A and E; and A, D, and E. None of the groups of mice with the initiating dose of 200 mg of BaP developed tumors, but 7 of the 9 groups tested for cocarcinogenicity developed carcinomas (see Table 10). Fourteen of the 30 Sencar mice treated with neat asphalt fumes (whole condensate) produced carcinomas, and 1 mouse in the Sencar solvent control group produced 1 tumor (sarcoma). Mice in the C3H/HeJ solvent control group failed to develop tumors, whereas the C3H/HeJ mice in two BaP control groups developed skin tumors (see 0.01% and 0.001% BaP groups in Table 10).


Sivak et al. [1989] observed no tumor responses with the three roofing asphalt fractions (A, D, and E) they considered most likely to exhibit cocarcinogenic or tumor-promoting activities based on their aliphatic hydrocarbon, alcohol, and phenol contents. Treatment with the combined fractions did not produce any synergistic effects. However, tumor responses were elicited by other fractions (B and C) that contained PACs including PAHs, S-PACs, N-PACs, and O-PACs such as alkylated aryl thiophenes, alkylated phenanthrenes, alkylated phenylethanones, and alkylated dihydrofuranones. Fraction B contained most of the S-PACs, and only a few were carried over to fraction C, which contained mainly O-PACs. Because the O-PACs may result from the air-blowing/oxidation refining process common among roofing asphalts, they may be present only in roofing asphalt. If such is the case, the refining process could be altered to eliminate the O-PACs and possibly the carcinogenicity of fraction C. Sivak et al. [1989] stated the need for additional cocarcinogenesis and tumor-promotion experiments using a wider range of experimental variables, further chemical separation of fractions B and C, more short-term genotoxicity assays, and additional carcinogenicity assays to identify biologically active materials in the roofing asphalt fume condensates.

HUMAN HEALTH EFFECTS

ACUTE

Asphalt fumes are irritants to the mucous membranes of the eyes and respiratory tract; hot asphalt can also cause burns of the skin [NIOSH 1977]. It has been reported that irritant effects on the respiratory tract can possibly progress to such nonmalignant lung diseases as bronchitis, emphysema, and asthma [Hansen, 1991; Maizlish et al. 1988]. Workers engaged in road repair and construction reported symptoms of abnormal fatigue, reduced appetite, eye irritation, and laryngeal/pharyngeal irritation [Norseth et al. 1991].

CHRONIC

Considerable data from epidemiological studies on workers exposed to asphalt fumes during paving and roofing operations, and during the production of asphalt, have become available since the publication of the NIOSH criteria document on asphalt [NIOSH 1977]. The mortality experience of Danish mastic asphalt workers [Hansen 1989a; Hansen 1991] and Swedish asphalt road pavers [Engholm 1991] was investigated (see Table 13). Hansen [1989a] reported that the mastic asphalt workers, when compared with the total male Danish population , experienced significantly increased mortality from cancers of the digestive and respiratory systems, with standardized incidence rates (SIR) of 227 (95% confidence interval of 142-344) and 195 ( 95% confidence interval of 236-493), respectively. The SIR for all malignant neoplasms was 195 (95% confidence interval of 153-244). Overall, Hansen [1989a] reported that she observed a three-fold increase in the expected number of lung cancers in the mastic asphalt workers compared with the general Danish population. For an assessment of the induction of primary lung cancer Hansen divided the cohort into subcohorts based on birth year because it was necessary to determine the number of employees potentially exposed to coal tar pitch, which had been added to mastic asphalt during World War II. The SIRs for primary lung cancer were then determined to range from 632 (for employees aged 40 to 54) to about 300(for employees aged 64 to 89). Although smoking histories of the cohort were unknown, an inquiry was made in 1976 into the smoking habits of mastic asphalt workers and a pattern emerged. Based on the approximate rates that were calculated, Hansen suggested that smoking could not account for the three-fold increase she had observed.

When Hansen [1991] updated her cohort and adjusted for smoking and urbanization, she reported that the statistically significant (P<0.01) increase in cancer mortality among mastic asphalt workers remained. The SIR for lung cancer mortality was 224 (95% CI, 145-330). Criticisms by Wong et al. [1992] and Kreich et al.] 1991] of the Hansen studies [1989a; 1991] are provided in the comments section of Table 13 and include the following: possible exposure to coal tar pitch and inadequate adjustment for smoking and urbanization.

Engholm et al. reported [1991] the occurrence of lung (SIR of 207) and stomach cancers (SIR of 207) in Swedish asphalt road pavers (see Appendix C). Data on previous and current smoking histories had been collected and were used in determining the relative risk (RR) for lung cancer. The RR for lung cancer was on the order of 2 before adjustment for smoking, and it was on the order of 3 after adjustment for smoking. Despite the short follow-up period ( an average of 11.5 years) and the very young age (42 years) of the cohort, the authors concluded that this cohort exhibited a slight excess of lung cancer. However, in a later submission to NIOSH [Engholm and Englund 1993], the results of an update based on the inclusion of three additional years of follow-up were reported. Engholm [1993] indicated that: 1) with the additional follow-up, all measures of any cancer risk were not statistically signigicant; 2) the study results may reflect some selection bias; 3) exposure of the cohort is in doubt.

Results of a proportionate mortality study of California highway maintenance workers [Maizlish et al. 1988] and a long-term mortality study of Minnesota highway maintenance workers [Bender et al. 1989] are also presented in Table 13. Maizlish et al.[1988] determined that the increased mortality from all malignant neoplasms for their cohort was not statistically significant. Additionally, exposure measurements and data on tobacco or alcohol consumption of the cohort were unavailable.

Bender et al. [1989] reported that workers with 30 to 39 years of work experience had a statistically significant (P<0.01) SMR of 425 (95% CI, 170-870) for leukemia deaths. The authors concluded, however, that they were unable to relate these findings to asphalt exposure. After additional study of this cohort (case-control studies, cytogenetic studies, updated chhort mortality, and personal air monitoring effort, the Minnesota Department of Health [1993] concluded that it was unlikely that the excess leukemia mortality observed among the highway maintenance workers was job-related.

Only one study is available regarding the mortality experience of roofers [Engholm et al. 1991].

During their investigation of asphalt road pavers in Sweden, these investigators also examined a cohort of roofers (see Appendix C). After adjustment for smoking, the RR for lung cancer in roofers was on the order of 6. The data indicated that though the number of cases was small, there was a lung cancer excess among roofers [Engholm et al. 1991]. Even though the authors acknowledged that the short follow-up period (11.5 years) and the young age (42 years) of the cohort were too short for the normal latency period of a potential carcinogen, they concluded that an excess of lung cancer existed among roofers. In 1993 Engholm and Englund presented to NIOSH information based on their three-year follow-up of Enghom et al. 1991. They concluded that results of the follow-up study " did not permit any final conclusions" regarding health risks of the respective cohorts.

Partanen and Boffetta [1994] conducted a review and meta-analysis of the epidemiologic studies regarding cancer risk in asphalt workers and roofers. They concluded that existing data are insufficient to make a judgment with regard to asphalt. Most epidemiologic studies for lung carcinogenicity (as well as other cancer sites ) are either too non-specific for exposure (e.g., highway maintenance workers , census occupational data), or confounded by coal tar exposure.

In 1987, the International Agency for Research on Cancer (IARC) evaluated the available studies involving asphalt fumes [IARC 1987] and concluded that the carcinogenicity of bitumens (shich include asphalt ) is unclassifiable in humans.

NIOSH investigators (Kyle Steenland) agree with the review of Partanen and Bofetta [1994]. In addition to the studies' deficiencies already enumerated, insufficient latency for workers exposed to asphalt is also noted. Deficiencies of the Hansen [1989, 1991] studies include the unresolved controversy concerning possible exposure to coal tar, possible selection biases, and the appropriate beginning of person-time at risk.

APPENDIX B

REFERENCES