Document K6dNy0zaBVkreM182nMdbYJE0

Mortality Among California Highway Workers Neil Maizlish, PhD, Jay Beaumont, PhD, and Jim Singleton, MS Standardized proportional mortality ratios (PMR) were computed for a population of highway workers. Hazards of highway maintenance work include exposure to solvents, herbicides, asphalt and welding fumes, diesel and auto exhaust, asbestos, abrasive dusts, hazardous material spills, and moving motor vehicles. Underlying cause of death was obtained for 1,570 workers who separated from the California Department of Transportation between 1970 and 1983, and who died in California between 1970 and 1983 (inclusive). Among 1,260 white males, the major findings were statistically significant excesses of cancers of digestive organs (PMR = 128), skin (PMR = 218), lymphopoietic cancer (PMR = 157), benign neoplasms (PMR = 343), motor vehicle accidents (PMR = 14I), and suicide (PMR = 154). Black males (N = 66) experienced nonsignificant excesses of cancer of the digestive organs (PMR = 191) and arteriosclerotic heart disease (PMR = 143). Among 168 white females, deaths from lung cancer (PMR = 189) and suicide (PMR = 215) were elevated. White male retirees, a subgroup with 5 or more years of service, experienced excess mortality due to cancers of the colon (PMR = 245), skin (PMR = 738), brain (PMR = 556), and lymphosarcomas and reticulosarcomas (PMR = 514). Deaths from external causes (PMR = 135) and cirrhosis of the liver (PMR = 229) were elevated among white males with a last job in landscape maintenance. White males whose last job was highway maintenance experienced a deficit in mortality from circulatory diseases (PMR = 83) and excess mortality from emphysema (PMR = 250) and motor vehicle accidents (PMR = 196). Further epidemiologic and industrial hygiene studies are needed to confirm the apparent excess mortality and to quantify occupational and nonoccupational exposures. However, reduction of recogrued hazards among highway maintenance workers is a prudent precautionary measure. Key words: cancer, accidents, herbicides, asphalt, solvents INTRODUCTION In 1985, approximately 9 3 , 6 0 0 workers w e r e employed by state departments of transportation to maintain the 681 , 9 9 9 miles of t h e basic intercity highway network in t h e United States [Transportation Research Board, 1986; Smith, 19861. T h e s e w o r k e r s a r e potentially exposed to n u m e r o u s health and safety hazards, yet chronic health effects among highway maintenance workers have not been widely studied. Recognized hazards of highway maintenance andlor traffic control work include exposure t o asbestos, hot asphalt and asphalt fumes, Portland cement, solvents, Occupational and Envlronmental Health Unit, University of California, Davis Address reprint requests to Neil Maizlish, California Department of Health Services, 2151 Berkeley Way, Room 504, Berkeley, CA 94704. Accepted for publication June 27, 1987 @ 1988 Alan R. Liss, Inc. I@ 364 Maizlish et al. pesticides, welding fumes, ultraviolet radiation, carbon monoxide, hydrogen sulfide, 7ead, automobile and diesel exhaust, silica dust from abrasive blasting, various hazardous materials from highway spills, noise, moving motor vehicles and equipment, excessive heat, cold, and vibration, animal and insect bites, parasites, vectorborne diseases, and toxic plants (Table I). Of particular concern are cancer risks that have been associated with hazardous materials used in highway maintenance work. Lung cancer has been implicated in occupational exposure to asbestos, coal tar among roofers, asphalt production, welding fumes, and diesel exhaust [NIOSH, 1980; Hammond et al, 1976; Silverstein et al, 1985a; Wilson, 1984; Beaumont et al, 1981; Schenker, 1980, NIOSH, 1977al. Cancers of the digestive organs have been associated with asbestos, coal tar pitch volatiles, and cement [Goldsmith, 1982; Hammond et al, 1976; Silverstein et al, 1985a; McDowall, 19841. Leukemia has been linked to benzene, agricultural and landscaping work, and coal tar pitch volatiles [Infante et al, 1977; Fasal et al, 1968; Milham, 1971; Blair and Thomas, 1979; Blair et al, 1985; Burmeister et al, 1983; Cantor and Blair, 1984; Cantor et al, 1985; Pearce et al, 1986; Silverstein et al, 19851. Multiple myeloma has been found among agricultural workers [Pearce et al, 1985; Cantor et al, 19841. Non-Hodgkin's lymphomas and soft tissue cancers have been associated with pesticides, particularly with phenoxyacetic herbicides [Hardell and Sandrom, 1979; Eriksson et al, 1981; Hoar et al, 1986; Schumacher, 1985; Hardell et al, 19811. The present study of the mortality experience of employees of the California Department of Transportation (CalTRANS) was undertaken as a surveillance exercise to address concerns raised by these reports in the literature. Standardized proportional mortality analyses are particularly useful as tools for mortality surveillance and hypothesis generation. The following health outcomes, based on previous reports, were specifically of concern prior to conducting the study: 1) lymphopoietic cancers in landscaping and/or in the use of bituminous materials; 2) soft tissue sarcoma in landscaping involving potential exposure to phenoxyacetic herbicides; 3) lung cancer in workers exposed to bituminous materials; and 4) digestive organ cancers in workers exposed to bituminous materials. While not strictly an a priori hypothesis, examination of mortality from malignant melanoma, which was reported among workers at the CalTRANS headquarters building in a previous incidence study [Austin, 19851, was also preplanned. METHODS Study Site The California Department of Transportation oversees the maintenance of approximately 26,000 miles of primary and secondary roadways, and, in 1985, employed approximately 15,000 permanent workers [Smith, 19861. The workforce is currently organized into 17 divisions, 11 geographic districts, and a headquarters. The districts vary in size from 350 to 3,500 workers, Los Angeles and San Francisco districts being the largest. The Maintenance Division employs 6,000 workers whose activities have been typical of highway maintenance operations (Table I). The Construction and Engineering Division employs approximately 2,000 workers, mostly engineers who inspect the work of contractors. CalTRANS contracts out major construction projects. Approximately 500 workers are employed in equip- I- I 366 Maizlish et al. ment shops, where major vehicle repair and outfitting is done. Routine and minor maintenance of vehicles is done at the approximately 400 maintenance stations. Other work sites include 2 warehouses employing 30 to 40 workers, laboratories for analysis of soils, metals, and asphalt, and a headquarters building where approximately 3,500 administrative and clerical workers are employed. Study Population The study population was defined as all CalTRANS employees who left em- ployment for any reason between 1970 and 1983 (inclusive) and who died in Califor- nia between January 1, 1970, and December 31, 1983. There was no minimum employment duration for inclusion in the study population, which included workers who died while active, retired, quit, or were fired, laid off, or otherwise terminated. The study population included all CalTRANS employees: maintenance workers, materials laboratory technicians, engineers, administrators, office workers, secre- taries, etc. Candidates for study were identified from a cumulative, computerized personnel file known as the separation file. This file listed the name, social security number, birth date, sex, ethnicity, and employment data at separation of former CalTRANS employees who left employment since 1970. The separation file was the sole source ..- f t for ascertaining potential deaths. Pension records and employee records at the district 31 3? level were not available. 7, 7 %I &, Vital status of 27,162 individuals in the separation file was determined by ,L computerized linkage using the social security number as the match key to an I, 5 automated form of the State of California death certificate registry [Arellano et al, -Yi Fs 19841. The linkage yielded 1,653 social security number matches, 1,545 of which I were exact in last name and sex. The other 108 matches were visually scanned for name, sex, race, and date of death, and 25 were judged with confidence to be true matches. The total number of deaths for analyses was 1,570. Out-of-statedeaths were not ascertained. We have no reason to believe that persons dying out of state would have different mortality patterns from those dying in state. Death Certificate Information The underlying cause of death listed on the death certificate was coded by nosologists at the California Department of Health Services according to the revision of the International Classification of Diseases (ICD) in effect at the time of death. Deaths spanned the ICD 8th (1968-1978) and 9th (1979-present) revisions. Causes of death of interest in this study that were affected by differences in the 8th and 9th ICD revisions include leukemia, whose ICD codes in the 9th revision (204-208) were expanded from those of the 8th (204-207). Occupational information The source of occupational information was also the separation file. The following information was available: district, last job classification, date of separation, and reason for separation (including retirement or death). Because year of hire was not coded, latency or duration effects could not be investigated. Subgroup analyses were carried out for retiree deaths and decedents with a last job classification of highway maintenance and landscape maintenance. Eligibility for retirement was at least 5 years of service and 50 years of age throughout the study period. According to CalTRANS Highway Worker Mortality 367 officials, landscape maintenance workers had the greatest potential for pesticide exposure, and highway maintenance workers had the greatest potential for exposure to chemical hazards in pavement and roadside maintenance. Industrial hygiene data of chemical exposures have not been routinely collected. Statistical Analyses Standardized proportional mortality ratios (PMRs) and proportional cancer mortality ratios (PCMRs) were calculated using the Monson [1974] program as adapted by Maizlish [1986], and reference proportional mortalities for the U.S. population through 1980. The PMR program calculated expected number of deaths for up to 57 causes based on the proportions of all deaths in the reference population and the distribution of deaths in the study population in strata of age at death, sex, race, and year of death. Similarly, the PCMR program calculated expected number of deaths for up to 25 cancer causes based on the proportions of cancer deaths in the reference population. The PMR or PCMR is the ratio of observed deaths to the expected deaths, multiplied by 100. A PMR greater than 100 indicates greater than expected mortality, while a PMR of less than 100indicates a deficit in mortality. The Poisson distribution was used to assess statistical significance and to compute 95% confidence intervals. RESULTS Characteristics of the Study Population The 1,570 decedents were mostly male (88%) and white (90%) (Table II). Hispanic decedents represented approximately 3% of whites. Approximately 50% of 118 nonwhite males were black, and 28% were Oriental. Mean age at death was 57 for white males and females and 50 for black males. Resignations accounted for approximately two-thirds of all separations, retirement accounted for 12% ,and deaths accounted for 16%. Standardized Proportional Mortality White males. Statistically significant excess mortality due to all malignant neoplasms (PMR = 118), benign neoplasms (PMR = 343), and external causes (PMR = 122) was observed among 1,260 white males (Table III). Lymphopoietic cancers (PMR = 157) and cancers of the digestive organs (PMR = 128), skin (PMR = 218), and brain (PMR = 162) contributed to the overall cancer excess (Table IV). Nonsignificant elevations in cancer of the large intestine (PMR = 143) and stomach cancer (PMR = 145) accounted for most of the excess cancer of the digestive organs. Of the 14 skin cancers, 11 were malignant melanoma. Lymphosarcomas and reticulosarcomas (PMR = 219) and leukemia (PMR = 162) accounted for most of the excess lymphopoietic cancer excess. More deaths than expected were also observed for testicular cancer (PMR = 217), although this was not statistically significant. Ten of the 12 benign neoplasms were brain tumors. Suicides (PMR = 154) and accidents (PMR = 120), particularly motor vehicle accidents (PMR = 141), were mainly responsible for the overall excess in deaths due to external causes (Table 111). Three soft tissue sarcomas (ICD 171) were observed versus 1.6 expected. One cancer of the pleura (ICD 163.9) was observed. There was a significant deficit in 368 Maizlish et al. TABLE II.Demographic and Occupational Characteristics, Separated CalTRANS Employees, 1WO-1983 Item Total deaths Males Whitea Black Orientalb Other Females White' Black Oriental Other Separation category Resignationslterminations Retirement Death Other Last job classification Landscape maintenance Highway maintenanced Laboratory workers, inspectors, engineers, administrators, and other office workers N 1570 1378 1260 66 3139 192 168 10 11 3 1070 194 258 48 106 327 1137 % 100 88 80 42 1 12 110 1 - 69 13 16 2 271 72 'Includes 47 Hispanic decedents. bIncludes Chinese, Japanese, Korean, and Vietnamese. 'Includes 3 Hispanic decedents. dIncludes highway maintenance, heavy equipment operators, and supervisors. deaths due to all circulatory diseases (PMR = 88) and symptoms, senility and ill- defined conditions (PMR = 12) (Table Ill). Nonwhite males. Black males (N = 66) experienced nonsignificant excess deaths due to digestive organ cancers [PMR = 7/3.67 = 191; CW5%) = 77,3941 and arteriosclerotic heart disease (PMR = 143) (Table HI).Three leukemias were observed among 33 Oriental males (Chinese, Japanese, Korean, or Vietnamese). Based on crude (age unadjusted) U.S. proportional mortalities for Chinese and Japanese males in 1980, only 0.27 deaths would be expected (PMR = 1091,p < .01). Females. Excess deaths due to all malignant neoplasms (PMR = 125) and suicides (PMR = 215) were found among 168white females (Table III). A significant excess of lung cancer (17/9.0, PMR = 189) contributed to most of the cancer excess. Nonsignificant deficits in circulatory disease mortality (PMR = 79) were also ob- served. Among 10 black females, proportional mortality from external causes was elevated (PMR = 203, 95% CI: 65,475). White male retirees. Among 160white male retirees, significant excess deaths were found for all malignant neoplasms (PMR = 152) and nonsignificant excesses for cirrhosis of the liver (PMR = 195) (Table V). A significant deficit in proportional mortality was found for arteriosclerotic heart disease (PMR = 67). Significant excess cancers of the digestive organs (PMR = 182)-mainly cancer of the large intestine i I a$1 cu 2 3 B 3 oe! El Y 3 v a II G 2 TABLE IV.Standardized Proportional Mortality Ratios (PMRs) for Cancers and Standardized Proportional Cancer Mortality Ratios (PCMRs) in CalTRANS White Males, 1970-1983 Cause of death Obs. EXD. PMR 95% CIa PCMR Total cancer deaths Cancer of buccal cavity Cancer of the digestive organs r of the esophagus Cancer of the stomach Cancer of large intestine Cancer of rectum All cancer of liver Cancer of pancreas All cancer of lung Cancer of slun Cancer of soft tissues Cancer of prostate Cancer of testis Cancer of bladder Cancer of kidney Cancer of brain and other CNS All lymphopoietic cancer Lymphosarcoma and reticulosarcoma Leukemia and aleukemia 341 289.27 118b 106,131 8 8.88 90 39,177 90 70.14 128' 103,157 9 7.02 128 59,243 15 10.35 145 81,239 35 24.52 143 99,199 8 6.50 123 53,242 4 4.93 81 31,215 19 14.81 128 77,200 105 107.60 98 80,119 14 6.42 218' 119,366 3 1.55 194 40,587 18 13.92 129 76,204 4 1.84 217 84,563 9 6.50 138 63,272 10 7.55 132 64,273 16 9.88 162 93,263 44 28.05 157b 115,210 11 5.03 219' 110,392 18 11.10 162 96,256 100 78 108 109 123 120 104 68 109 83 193' 103 188 114 114 143 135 191 158 TI = Poisson confidence intervals. bp < .01 Poisson two-tailed test. 'p < .05 Poisson two-tailed test. (PMR = 245), skin (PMR = 738), brain (PMR = 556), and all lymphopoietic tissue (PMR = 227)-contributed to the overall excess cancer mortality. Landscape and highway maintenance. Cancer mortality among white males with a last job of landscape maintenance was unremarkable (Table VI). Landscape maintenance workers experienced a nonsignificant excess of deaths due to cirrhosis of the liver (PMR = 229) and motor vehicle accidents (PMR = 170). Among highway maintenance workers (Table VI), there was a significant excess of'deaths from emphysema (PMR = 250) and all accidents (PMR = 169), including motor vehicle accidents (PMR = 196). There was a nonsignificant elevation of deaths from cancer of the digestive organs (PMR = 151)-primarily cancer of the stomach (PMR = 247) and prostate cancer (N = 226). A significant deficit of deaths due to circulatory disease (PMR = 83) was also observed. Standardized Proportional Cancer Mortality Among white males a significantly elevated PCMR was found for skin cancer (PCMR = 193) (Table IV). Proportional cancer mortality due to lung cancer was significantly less than expected (PCMR = 83). Greater than expected numbers were observed for leukemia (18 vs. 12.9), lymphosarcomas and reticulosarcoma (11 vS. 5.7), digestive organ cancer (90 vs. 83.1), and brain cancer (16 vs. 11.1). Among white females, there was a nonsignificant excess of proportional cancer mortality due to respiratory system cancers (PCMR = 152) and a nonsignificant deficit due to digestive organ cancers (PCMR = 52). 1 41 1/ ( I II .I! ( I( 1 1 I.( . h i I I i \ I I \ I j \ I \ I k Highway Worker Mortality 371 TABLE V. Standardized Proportional Mortality Ratios in CalTRANS White Males, 1970-1983, Retired at SeDaration* Cause of death Obs. Exp. PMR 95% CIa Total deaths All malignant neoplasms Cancer of the digestive organs Cancer of stomach Cancer of large intestine Cancer of rectum All cancer of lung Cancer of skin Cancer of brain and other CNS All lymphopoietic cancer Lymphosarcoma and reticulosarcoma Cancer of other lymphatic tissue Benign neoplasms All circulatory disease . I Arteriosclerotic heart disease All respiratory diseases All diseases of the digestive system Cirrhosis of liver All external causes 160 65 42.74 152' 19 10.46 182b 3 1.47 204 9 3.67 245b 3 .90 333 18 16.97 106 6 .81 738' 7 1.26 556' 8 3.53 227 3 .58 5 14b 4 1.32 302 2 .46 433 54 76.58 71' 36 53.81 67' 10 9.58 104 12 8.29 145 10 5.14 195 10 11.58 86 117,194 110,284 68,614 112,465 67,978 63,167 270,1608 223,1146 98,447 103,1517 76,773 43,1565 53, 91 41, 93 50,191 75,253 94,358 41,158 I *CalTRANS service and disability retirements. 'CI = Poisson confidence intervals. 11 bp < .05 Poisson two-tailed test. 'p < .01 Poisson two-tailed test. DISCUSSION The major findings were excess mortality from neoplasms (malignant and benign), accidents, and suicides. Exposure data were not sufficient to associate the I apparent risks with specific job hazards, job classifications, job duration, or time , since first exposure. Nonetheless, the pattern of mortality was consistent, in part, with previously hypothesized effects, and plausibly related to known occupational hazards and suspect agents. Before discussing the possible etiologic significance of these findings, methodologic limitations must be addressed. Limitations I PMR methodology. The proportional mortality ratio is not a direct measure of absolute risk for a cause of death. The PMR study design has been criticized for nonindependence of cause-specific results; Le., risk estimates for specific causes may be overstated if PMRs for other causes are low. This may occur if there are selection effects related to hiring practices (i.e., healthy worker effect). Empirical evidence suggests, however, that PMRs are frequently accurate representations of the mortality I experience [Beaumont et al, 1981; St. Claire and Butler, 1981; Roman et al, 19831. PMR analyses have also been criticized for selective ascertainment of deaths. In the present study, ascertainment of California deaths was highly complete. Errors in record linkage using the California Automated Mortality Linkage System (CAMLIS) are small; 4-6% of matches have been found to be false-positive or false-negative [Arellano, 19851. Out-of-state deaths were not ascertained, but this would not be 1 expected to substantively alter the findings. We have no specific data to prove or 372 Maizlish et al. TABLE VI. Standardized Proportional Mortality Ratios in CalTRANS White Males, 1970-1983, Highway Maintenance Classifications at Separation Cauce of death Highway maintenance Landscape maintenance Ohs P M R 95% CI+a Obs. PMR 95% C1-c Total deaths 307 87 All malignant neoplasms 81 117 93,146 16 86 Cancer of the digestive 01'gam 25 151 97,223 3 67 Cancer of stomach 6 227 83,495 0 0 All cancer of lung 25 98 63,145 4 60 Cancer of skin 2 122 12,439 1 212 Cancer of prostate 7 226 91,466 1 110 Cancer of brain and other CNS 4 160 40,410 1 142 All lymphopoietic cancer 8 115 50,226 1 50 Benign neoplasms 3 345 69,1011 0 0 All circulatory disease 110 83b 68, 99 33 92 All respiratory diseases Emphysema 15 93 52,153 8 250b 108,492 6 131 1 109 All diseases of the digestive system 15 92 51,152 8 187 Cirrhosis of liver 13 127 68,217 6 229 All external causes 73 145' 114,182 23 135 All accidents 53 169' 127,222 14 162 Motor vehicle accidents 33 196' 134,276 10 170 49,140 13-,197 15-,153 - - 63,130 48-,285 80,368 84,498 85,203 89,272 82,313 +aCI = Poisson coefficient intervals. ! II 2.9) I not sf in thi with I avail I j deati I * nom; canc 1 certi Po*' I this histc I h recr risk I sifie I eve] refute the notion that the frequency of cause-specific out-migration of CalTRANS workers is different from other California residents. California is widely recognized as a desirable state for retirement, and this may minimize potential biases due to out- Multiple comparisons. Multiple comparisons were unavoidable in screening many causes of death, and some findings may be due to chance. Nonetheless, excess lymphopoietic cancers and digestive organ cancers were hypothesized effects, and excess malignant melanoma mortality was confirmed. J- Confounding biases. Adjustments using indirect standardization were made for age at death, sex, race, and year of death. Other determinants of mortality, including socioeconomic status (SES), state of residence, cigarette smoking, and alcohol consumption were not controlled. Cancer of the colon, lymphopoietic tissue, melanoma, and brain cancer are associated with increasing social class [Milham, 1985; Berg and Howell, 1975; Levin et al, 19791. Because higher SES occupations such as engineers, administrators, and other professionals were a large part of the study population, uncontrolled SES may be responsible for part of the elevated mortality. However, as Milham [I9851 pointed out, controlling for SES presents difficulties because SES, occupation, and potential exposures may not be independent risk factors. Based on comparisons of age-adjusted mortality rates (per 100,OOO) for white males, 1970-1979, geographic variations in mortality between the U .S. and California are unlikely to account for observed excesses of colon cancer (U.S., 20.0; California, 18.8), leukemia (U.S., 9.1; California, 9.2), or melanoma (U.S., 2.4; California, I I I I i 41 1 i I 1 I c Yea Eti wo COI Pa' am Mi inc f01 Pr ro M 2, Vt m W 0 Highway Worker Mortality 373 2.9) [EPA/NCI, 19831. Except for lung cancer and emphysema, cigarette smoking is not strongly implicated in the types of cancer deaths observed in excess among males in this study [Surgeon General, 19821. Excessive alcohol consumption is associated with suicides, cirrhosis of the liver, and fatal motor vehcle accidents perkelman et al, 1986; Baker et al, 19841. Data on cigarette or alcohol consumption were not available for the study group. Misclassification errors. Misclassification of underlying cause of death on death certificates is not a likely explanation of findings for colorectal cancer, melanoma, lymphopoietic cancers, or cancers of the stomach, lung, or brain. These cancers generally have greater than 90% detection rates in comparisons of death certificates and hospital records. However, soft-tissue cancers are seriously underreported on death certificates, and this could mask greater proportional mortality for this cause [Percy et al, 19811. An important limitation of this study was the lack of exposure data, work histories, knowledge of previous or concomitant non-CalTRANS employment and recreational activities involving hazardous substances, and data on nonoccupational risk factors. Exposure classification based on last-job classification may have misclassified exposure, depending on the degree of job mobility, which was not known. Lack of hire date precluded investigation of duration or latency effects. However, the subgroup of white male retirees experienced greater proportional mortality at various cancer sites than all white males. This is consistent with a dose-duration effect, because retirees had at least 5 years of employment. Because most cancers I have 10-30 year latency periods, the exposures of etiologic significance may have occurred during the 1950s and 1960s. Current work practices or materials may be substantively different. However, latencies for leukemia may be as short as 5-10 years, so recent exposures may be relevant. Etiologic Significance There have been few previous mortality studies of state highway maintenance workers, or workers exposed to the same chemical and physical hazards under similar conditions. Mortality studies of white male workers, involved in excavating, grading, paving, road machinery operaton and highway maintenance in Minnesota [Bender and Parker, 19871, Washington [Milham, 19831 and California in 1960 [Petersen and Milham, 19801 confirm excess mortality from accidental causes among white males, including motor vehicle accidents. Excess leukemia mortality (SMR = 420) was 1 found among Minnesota highway maintenance workers with 30-39 years of service. Primary malignant neoplasms of the lung were elevated (PMR = 119) in Washington road workers, but not in California (PMR = 92) [Petersen and Milham, 19801, or Minnesota (SMR = 70). Excess prostate cancers were found in workers from Wash- ington (PMR = 155, ages 20-@), and an elevated brain cancer risk (SMR = 294) I was reported among Canadian highway and bridge maintenance workers [Howe and , Lindsay, 19831. I Kovasznay [1984] studied 362 New York State highway sprayers exposed to 2,4,5-T and 2,4-D and a comparison group of age-matched, nonexposed motor vehicle registrants. A small, nonsignificant excess of cancers was reported, but the number of cancer cases was small, follow-up was short, and other maintenance \ workers were not included. In a survey of 15 state hig-hway- commissions or boards of health, Baylor and Weaver [I9681 did not report serious health effects associated 374 Maizlish et al. with exposure to asphalt in highway construction. However, neither death certificates nor medical records were reviewed by the authors. Possible health effects may be suggested from studies of other workers exposed to the same materials, albeit under different conditions. Lymphopoietic cancer. Excess lymphopoietic cancers were found among both white and Oriental males (Table IV).Excess mortality from leukemia has been found among workers with exposure to benzene [IARC, 19821 and possibly other solvents, and among agricultural workers with potential exposures to solvents, pesticides, and viruses, and a subgroup of Minnesota highway workers [Bender and Parker, 19871. Painters, highway line stripers, technicians in materials testing laboratories, asphalt workers, motor vehicle maintenance workers, and other CalTRANS employees are potentially exposed to solvents. Exposure may occur in asphalt operations using cutback asphalts, in which solvents (usually naphtha or other petroleum distillate) blended with asphalt cement evaporate during curing [Asphalt Institute, 19821. Another important exposure may occur when organic solvents, including gasoline, are used for cleaning bitumin contamination from skin [Burgess, 19811. This may also have the effect of solubilizing asphalt in a manner analogous to skin painting. Laboratory testing for asphalts used in highway or street construction have historically relied on benzene. Benzene, carbon disulfide, carbon tetrachloride, and trichloroethylene are currently recommended for laboratory testing of asphalt [ASTM, 1984; AASHTO, 19821. At CalTRANS, asphalt testing for certification of contractors' work requires the use of benzene, trichloroethylene, toluene, xylene, gasoline, naphtha, Stoddard solvent, 1 , 1,1-trichloroethane,and kerosene [CalTRANS, 19781. Hazardous-material spills frequently involve fuels, flammables, and other solvents. In California, 8 reportable spills occur on average each week, with one-third of the incidents involving leaks and drops from moving motor vehicles [Shaw et al, 19861. This suggests hazards of potential high-level exposures from large spills and lower-level exposures from contaminated roadway dusts. The mixing, loading, and application of pesticides, usually from spray trucks, have been tasks shared among members of landscape maintenance crews at CalTRANS. Excess mortality from lymphopoietic cancers was not observed among white males with a last job in landscape maintenance. Phenoxyacetic herbicides were reported to have been used prior to 1970. Since the 1970s, amitrole, triazines, bromocil, cacodylic acid, dichlobenil, dalapon, diquat, diphenamid, diuron, trichlorophenol acetic acid, glyphosate, fosamine ammonium, chlorflurenol, monosodium methylarsonate (MSMA), orazalin, and sodium trichloroacetate (TCA) have been used [CalTRANS, 19811. Since the 1970s, organophosphates and carbamates of relatively low acute toxicity, pyrethrin, bacillus thuriengensis (BT), and petroleum oil, among others, were used as insecticides. Amitrole, glyphosate, and oryzalin are currently considered animal oncogens [USEPA, 1985; USDHHS (NTP), 19831. Skin cancer. Skin cancers, mostly melanoma, were significantly elevated in white males (Table IV)and retirees (Table V). Climatic conditions in California favor year-round outdoor highway maintenance work. Consequently, exposure to solar radiation may occur more frequently than in many other states, and this may explain, in part, excess melanoma in the study group. It has been suggested that cumulative exposure to ultraviolet (UV) radiation is associated with nonmelanoma skm cancers, while blistering overdoses are more closely associated with melanoma [Page and Asire, 19851. Occupational risk factors of nonmelanoma skin cancer include exposure I I I I I 0 I I I I I I I I 1 I I i ! I I I I ( I 4 Highway Worker Mortality 375 to UV radiation or sunlight [Birmingham, 19781, X-radiation, arsenic, and coal tar pitch [NIOSH, 1977bl. Cancer of digestive organs. Colon and stomach cancers were elevated among white and black males (Table IV). Minnesota highway maintenance workers with long latencies experienced an excess of intestinal cancers (SMR = 580). Recent studies have identified elevated risks of digestive organ cancer among asbestos workers [Goldsmith, 19821, roofers exposed to coal tar pitch volatiles [Hammond et al, 1976; Silverstein et al, 1985a1, cement workers [McDowall, 19841, chemists [Hoar and Pell, 19811, rubber workers [Monson and Nakano, 1974; McMichael et al, 19741, non-coke-oven steel workers [Redmond et al, 19811, workers exposed to oil mists and abrasive grinding [Jarvholm et al, 1981; Decouflt, 1978; Park et al, 1984; Silverstein et al, 1985b], pattern/model makers in the automobile industry [Swanson and Belle, 19821, and workers with sedentary jobs [Garabrant et al, 1984; Gerhardsson et al, 1986; Vena et al, 19851. Relevant exposures among CalTRANS workers included asbestos, asphalt fumes or coal tar pitch volatiles containing polyaromatic hydrocarbons, cement and other abrasive dusts, and sedentary jobs. The pattern of excess stomach cancer without a concomitant increase in lung cancer among highway maintenance workers (Tables III and VI) has been observed in cement workers, but not in workers exposed to asbestos or coal tar pitch volatiles. An excess of colon cancers is also consistent with sedentary work or uncontrolled socioeconomic status among professionals, whose colon cancer risk is higher than other workers, and who made up a large percentage of the study group. Nonoccupational factors in the epidemiology of digestive organ cancers have received considerable attention. There is much speculation that dietary factors involv- ing fat, meat, and fiber consumption influence national differences [Schottenfeld and Winawer, 19821. Lung cancer. Excess lung cancers were observed among white females in PMR and PCMR analyses. Historically, females have been underrepresented in maintenance jobs, and it is not likely that exposures in maintenance work were responsible for the observed excess. Most female decedents were office workers, and exposures in the office environment, including cigarette smoke and asbestos [CalOSHA, 19801, may be explanatory. The absence of lung cancer excesses in males does not, without qualification, support a role for inhalation exposures to asphalt fumes or coal tar pitch volatiles. However, if dermal or other routes of exposure affecting other sites were etiologic, expected lung cancer mortality might be observed. The findings do not preclude excess lung cancers for small groups of males whose risk may have been obscured when included in larger groups. A deficit in lung cancer mortality was also observed in Minnesota highway maintenance workers (SMR = 70). Fatal accidents. Excess mortality from motor vehicle accidents was found among white males (Table 111). Road vehicles account for approximately one-third of work-related injury deaths in the U.S. [Baker et al, 19841. Excess mortality due to fatal accidents, particularly motor vehicle accidents, is not surprising in a population of workers exposed to vehicular traffic and heavy equipment. According to CalTRANS officials, motorists currently break through construction barriers on average 5 times per day. Of 98 fatal motor vehicle accidents, 54% were due to a collision with a stopped vehicle or pedestrian on a highway, and 60%separated on the same 376 Maizlish et al. date as death, indicating a death during active status. However, because time of day and day of week were not available in our analysis, the work-relatedness of fatal accidents is not precisely known. Other Findings. White males and females experienced excess suicides (Table m).Unidentified pyschosocial stressors on and off the job may be implicated. The finding of excess cirrhosis deaths among white male retirees (Table V) also suggests a role for excess alcohol consumption, or interactions or misdiagnoses regarding solvents. Arteriosclerotic heart disease among black males, brain cancers and benign neoplasms in white males (nearly all of which were brain tumors), and emphysema and prostate cancer among highway maintenance workers were not hypothesized effects and require confirmation in future studies. CONCLUSIONSAND RECOMMENDATIONS Associations between specific causes of death and employment at the California Department of Transportation have been observed. It is unlikely that the pattern of multisite cancer mortality and fatal accidents could have all arisen by chance, but methodologic limitations and uncontrolled potentially confounding factors may be partly responsible for the associations. Cancer of the digestive organs, lymphopoietic cancers, skin cancers, and motor vehicle accidents are all plausibly related to recognized hazards of highway maintenance work, but also to nonoccupational factors. Further epidemiologic investigations are necessary to c o n f m the apparent excess mortality and to quantify occupational and nonoccupational risk factors and their interactions. State, county, and municipal departments of transportation and the highway and street construction industry are encouraged to evaluate current work practices and materials and to take precautionary steps to reduce hazards. ACKNOWLEDGMENTS The authors acknowledge the Office of Employee Safety and Health, California Department of Transportation, for assistance in carrying out this study. Early discussions with Dr. Mike Stimmann, University of California, Davis, are also gratefully acknowledged. This study was supported by grants from the Public Service Research and Dissemination Program, University of California, Davis, and the National Institute of Environmental Health Sciences (NRSA 7 F32 ES05246-02) (N.M.). REFERENCES American Association of State Highway Transportation Officials (AASHTO) (1982): Materials. 13th ed. Part I. Specifications. Part II. Tests. Washington, D.C. pp 69, 99. American Society for Testing and Materials (ASTM) (1984): "1984 Annual Handbook of ASTM Standards." Section 4. Construction. Road paving materials. Volume 4.03. Philadelphia. Arellano MG, Petersen GR, Petitti DB, Smith RE (1984): The California automated mortality linkage system (CAMLIS). Am J Pub Health 74:1324-1330. Arellano M (1985): Department of Epidemiology and International Health, University of California, San Francisco (personal communication). Asphalt Institute (1982): "Introduction to Asphalt and Some of Its Uses." Manual series No. 5 (MS-5). 8th ed. Austin D (1985): California Department of Health Services, Berkeley, CA. I I I 1 I I I I I ! I I I I I I I ',I I I i I I I I 4 II Highway Worker Mortality 377 Baker SP, O'Neill B, Karpf RS (1984): "The Injury Fact Book." Lexington, MA: Lexington Books, pp 195,238. Baylor CH, Weaver NK (1968): A health survey of petroleum asphalt workers. Arch Environ Health 171210-214. Beaumont J, Leet TL, Okun AH (1981): Occupational data sets appropriate for proportionate mortality ratio analysis. Banbury Report 9: "Quantification of Occupational Cancer." Cold Spring Harbor, NY: Cold Spring Harbor Laboratory, pp 391-411. Bender A, Parker D (1987): "The Minnesota Highway Maintenance Worker Mortality Study, 19451984." Minneapolis: Minnesota Department of Health. Berg JW, Howell HA (1975): Occupation and bowel cancer. J Toxic01 Environ Health 1:75-89. Berkelman R, Ralston M, Herndon J, Gwinn M, Bertolucci D, Dufour M (1986): Patterns of alcohol consumption and alcohol-related morbidity and mortality. CDC Surveillance Summaries 35:1SS5ss. Birmingham D (1978): Dermatoses. In "Occupational Diseases: A Guide to Their Recognition." DHEW (NIOSH) Pub. No. 77-181, Washington, D.C.: USGPO. Blair A, Thomas T (1979): Leukemia among Nebraska farmers: a death certificate study. Am J Epidemiol 110:264-273. Blair A, Everett G , Cantor K, Gibson R, Schuman L, Isaacson P, Blattner W, Van Lier S (1985): Leukemia and farm practices. Am J Epidemiol 122535, 1985. Burgess WA (1981): "Recognition of Health Hazards in Industry." New York: John Wiley & Sons, p 160. Burmeister LF, Everett GD, Van Lier SF, Isaacson P (1983): Selected cancer mortality and farm practices in Iowa. Am J Epidemiol 118:72-77. California Department of Transportation (1978): Testing Manual. Volume XI. Testing and control procedures. 3rd ed. Division of Construction, Sacramento. California Department of Transportation (1981): Vegetation control. A guide for herbicide application. Division of Highway Maintenance, Sacramento, pp 15-18. California Department of Transportation (1985): Code of safe operating practices. Division of Highway Maintenance, Sacramento. CalOSHA (1980): Division of Occupational Safety and Health. Department of Industrial Relations. Inspections W1597036808, R7611011808. Cantor K, Blair A (1984): Farming and mortality from multiple myeloma: a case-control study with the use of death certificates. J Natl Cancer Inst 72:251-255. Cantor K, Everett G , Blair A, Gibson R, Schuman L, Isaacson P (1985): Farming and non-Hodgkin's lymphoma. Am J Epidemiol 122535. DecouflC P (1978): Further analysis of cancer mortality patterns among workers exposed to cutting oil mists. J Natl Cancer Inst 61:1025-1030. EPA/NCI (1983): U.S. cancer mortality rates and trends, 1950-1979. U.S. Environmental Protection Agency and National Cancer Institute. EPA-600/1-83-015a, Washington, D.C. : USGPO. Eriksson M, Hardell L, Berg 0, Moller T, Axelson 0 (1981): Soft-tissue sarcomas and exposure to chemical substances: a case-referent study. Br J Ind Med 38:27-33. Fasal E, Jackson EW, Klauber MR (1968): Leukemia and lymphoma mortality and farm residence. Am J Epidemiol 87:267-274. Garabrant DH, Peters JM, Mack TM, Bernstein L (1984): Job activity and colon cancer risk. Am J Epidemiol 119:1005-1014. Gerhardsson M, Norell SE, Kiviranta H, Pedersen NL, Ahlbom A (1986): Sedentary jobs and colon cancer. Am J Epidemiol 123:775-780. Goldsmith JR (1982): Asbestos as a systemic carcinogen: the evidence from eleven cohorts. Am J Ind Med 3:341-348. Hammond EC, Selikoff U, Lawther PL, Seidman H (1976): Inhalation of benzpyrene and cancer in man. Ann NY Acad Sci 271:116-124. Hardell L, Sandrom A (1979): Case-control study: soft-tissue sarcomas and exposure to phenoxyacetic acids or chlorophenols. Br J Cancer 39:711-719. Hardell L, Eriksson M, Lenner P, Lungren E (1981): Malignant lymphoma and exposure to chemicals especially solvents, chlorophenols and phenoxyacetic acids: a case-control study. Br J Cancer 43: 169-176. Hoar SK, Blair A, Holmes FF, Boysen CD, Robel RI, Hoover R, Fraumeni JF (1986): Agricultural herbicide use and risk of lymphoma and soft-tissue sarcoma. JAMA 256:1141-1147. 378 Maizlish et al. Hoar SK, Pell A (1981): A retrospective cohort study of mortality and cancer incidence among chemists. J Occup Med 23:485-494. Howe GR, Lindsay JP (1983): A follow-up study of ten-percent sample of the Canadian labor force. Cancer mortality in males, 1965-73. JNCI 70:37-44. Infante PF, Rinsky RA, Wagoner JK, Young RJ (1977): Leukemia in benzene workers. Lancet II (8028)176-78. Institute of Transportation Studies (1984): Pavement maintenance and rehabilitation: techniques using asphalt. UCB-ITS-CN-84-1. University of California, Berkeley. International Agency for Research on Cancer (IARC) monographs (1982): Chemicals, industrial pro- cesses and industries associated with cancer in humans. Vol 1-20, Suppl4. Lyons, France. Jarvholm B, Lillienberg L, Stallsten G , Thiringer G, Axelson 0 (1981): Cancer morbidity among men exposed to oil mist in the metal industry. J Occup Med 23:333-337. Key MM, Henschel AF, Butler J, Ligo RN, Tabershaw IR (1978): "Occupational Diseases: A Guide to Their Recognition." DHEW (NIOSH) Pub. No. 77-181, Washington, D.C.: USGPO. Kovasznay BM (1984): "Study of Health Effects of Herbicides in Highway Maintenance Workers." PhD Dissertation. Johns Hopkins University. Dissertation Abstr. Int. UMI-84-14291, Ann Arbor. Levin DL, Devesa SS, Godwin JD, Silverman DT (1979): "Cancer Rates and Risks," 2nd ed. DHEW Pub. NO. (NIH) 79-691. Maizlish N (1986): "A Microcomputer Program for Proportional Mortality Analyses." Keene, CA: National Farm Workers Health Group. McDowall ME (1984): A mortality study of cement workers. Br J Ind Med 41:179-182. McMichael AJ, Spirtas R, Kupper LL (1974): An epidemiologic study of mortality within a cohort of rubber workers, 1964-72. J Occup Med 16:458-464. Milham S (1971): Leukemia and multiple myeloma in farmers. Am J Epidemiol 94:307-310. Milham S (1983): "Occupational Mortality in Washington State, 1950-1979." DHHS (NIOSH) Pub. No. 83-116, Washington, D.C., USGPO. Milham S (1985): Improving occupational standardized proportionate mortality ratio analysis by social class stratification. Am J Epidemiol 121:472475. Monson R (1974): Analysis of relative survival and proportional mortality. Comp Biomed Res 7:325- 332. Monson R, Nakano KK (1984): Mortality among rubber workers. I. White male union employees in Akron, Ohio. Am J Epidemiol 103:284-296. National Institute for Occupational Safety and Health (1977a): "Criteria for a Recommended Standard. Occupational Exposure to Asphalt Fumes." DHEW (NIOSH) Pub. No. 78-106, Washington, D.C.: USGPO. National Institute for Occupational Safety and Health (1977b): "Criteria for a Recommended Standard. Occupational Exposure to Coal Tar Products." DHEW (NIOSH) Pub. No. 78-107, Washington, D.C.: USGPO. National Institute for Occupational Safety and Health (1978): "Health and Safety Guide for Highway and Street Construction." DHEW (NIOSH) Pub. No. 78-196, Washington, D.C.: USGPO. National Institute for Occupational Safety and Health (1980): "Workplace Exposure to Asbestos. " DHHS (NIOSH) Pub. No. 81-103, Washington, D.C.: USGPO. Page HS, Asire AA (1985): "Cancer Rates and Risks," 3rd ed. HHS (NIH) Pub. No. 85-691, Washington, D.C.: USGPO. Park R, Silverstein M, Maizlish N, Mirer F (1984): Mortality among workers exposed to cutting fluids and abrasives: bearing plant I. NIOSH Contract 210-85-5104 (unpublished). Pearce NE, Smith AH, Fisher DO (1985): Malignant lymphoma and multiple myeloma linked with agricultural occupations in a New Zealand cancer registry-based study. Am J Epidemiol 121:225237. Pearce NE, Sheppard RA, Howard JK, Fraser J, Lilley BM (1986): Leukemia among New Zealand agricultural workers: a cancer registry-based study. Am J Epidemiol 124:402-409. Percy C, Stanek E, Gloeckler L (1981): Accuracy of cancer death certificates and its effects on cancer mortality statistics. Am J Pub Health 71:242-250. Petersen G, Milham S (1980): Occupational mortality in the state of California, 1959-1961. DHEW (NIOSH) Pub. No. 80-104, USGPO, Washington. Redmond CK, Weiland HS, Rockette HE, Sass R, Weinberg G (1981): Long-term mortality experience of steelworkers. NIOSH Pub. No. 81-120, USGPO, Washington. i I I I I I I I , I I I I I I I I I 1 i I I I 4 Highway Worker Mortality 379 Roman E, Beral V, Inslup H, McDowall M, Adelstein A (1983): A comparison of standardized and proportional mortality ratios. Stat Med 3:7-14. Schenker MB (1980): Diesel exhaust-an occupational carcinogen? J Occup Med 22:41-46. Schonenfeld D, Fraumeni J (eds) (1982): "Cancer Epidemiology and Prevention. " Philadelphia: WB Saunders, pp. 703-727. Schottenfeld D, Winawer SJ (1982): Large intestine. Chapter 40. Schumacher MC (1985): Farming occupations and mortality from non-Hodgkin's lymphoma in Utah. J. Occup Med 27:580-584. Shaw GM, Windham GC, Leonard A, Neutra RR (1986): Characteristics of hazardous material spills from reporting systems in California. Am J Pub Health 76:540-543. Silverstein M, Maizlish N, Park R, Mirer F (1985a): Mortality among workers exposed to coal tar pitch volatiles and welding emissions: an exercise in epidemiologic triage. Am J Pub Health 75:12831287. Silverstein M, Park R, Maizlish N, Mirer F (1985b): "Mortality Among Workers Exposed to Cutting Fluids and Abrasives: Bearing Plant II." NIOSH Contract 210-81-5104 (in press). Smith FA (1986): "Transportation in America: A Statistical Analysis of Transportation in the United States," 4th ed. Washington: Transportation Policy Associates. St. Claire N, Butler W (1981): "A review of Proportionate Mortality Study Design in Occupational Epidemiology." Paper presented at the American Public Health Association annual meeting, Los Angeles, November 5, 1981. Surgeon General (1982): "The Health Consequences of Smoking: Cancer." Office on Smoking and Health, USDHHS, USGVO. Washington, DC pp 5-8. Swanson GM, Belle S (1982): Cancer morbidity among woodworkers in the US automobile industry. J Occup Med 24:315-319. Transportation Research Board (1986): Progress report on maintenance and operations personnel. Washington, DC: National Research Council. Circular 301. Vena JE, Graham S , Zielezny M, Swanson MK, Barnes RE, Nolan J (1985): Lifetime occupational exercise and colon cancer. Am J Epidemiol 122:357-365. U.S. Department of Health and Human Services (1983): "Third Annual Report on Carcinogens." DHHS (National Toxicology Program) Pub. No. NTP 82-330. U.S. Environmental Protection Agency (1985): "Pesticides Registered for Use on Food Which the Agency Has Determined or Preliminarily Determined to have Oncogenic Effects in Animals. " (Revised 11121/85.) Wilson K (1984): Asphalt production and lung cancer in a Danish village. Lancet II(8398):354.