Document rx2O8Omjx36Rd1dwRLrOB8Ga0

ANNE L. GOLDEN, PhD STEVEN B. MARKOWITZ, MD PHILIP J. LANDRIGAN, MD, MSc THE RISK OF CANCER IN FIREFIGHTERS From the Division of Environmental and Occupational Medicine Department of Community Medicine Reprint requests to: Anne L. Golden, PhD Assistant Professor Box 1057 Mount Sinai School of Medicine One Gustave L. Levy Place New York, NY 10029 Cancer among firefighters has been an area of intensive investigation in occupational medicine for the past two decades. This research has been prompt ed by the recognition that firefighters are exposed in their work to high doses of multiple chemical car cinogens. The full extent of the occupational cancer risk of firefighters is not yet known. It is likely that in the years ahead, additional cancers will be found to be associated with exposures encountered by fire fighters and that additional chemicals to which fire fighters are already known to be exposed will be found to be carcinogenic. Despite the gaps in scien tific knowledge, concern about excess cancer risk has resulted in the provision of disability benefits to firefighters under presumptive occupational cancer legislation in 15 states (Alabama, California, Illinois, Louisiana, Maryland, Massachusetts, Minnesota, Nevada, New Hampshire, North Dakota, Oklahoma, Rhode Island, Tennessee, Texas and Virginia) and in the city of New York. A substantial body of literature now exists on the carcinogenic hazards of firefighting. Of particular concern are cancers that can be plausibly linked with specific toxic and carcinogenic chemical exposures to which firefighters are exposed in the course of their work: leukemia, lymphoma, multiple myeloma, melanoma, and cancers of the respiratory system, di gestive system, genitourinary tract and brain.30'36-45'64 CARCINOGENIC EXPOSURES OF FIREFIGHTERS Firefighters are routinely exposed to complex and dynamic mixtures of chemical substances that OCCUPATIONAL MEDICINE: State of the Art Reviews- Vol. 10, No. 4, October-December 1995. Philadelphia, Hanley & Belfus. Inc. 803 5 PLAINTIFF'S EXHIBIT < 804 GOLDEN, MARKOWITZ, LANDRIGAN are contained in fire smoke and building debris.14 Despite the large numbers of people employed in this occupation, the nature of these exposures is not well defined. Studies that have been completed to date, however, clearly demonstrate the presence of recog nized and suspected human carcinogens in the breathing environment of firefighters at the fire scene. The relative paucity of information about the exposures of firefighters is not sur prising given the complexity of such exposures and the methods by which they are studied. Fires vary greatly in the nature of the materials burned, temperature, size, and ambient weather conditions.14 The nature and concentrations of airborne exposures change at the fire scene over short distances and upon the stage of the fire. The actual exposures received by firefighters further depend on their job tasks at the fire and the type and use of respiratory protection. Finally, measurement of airborne exposures at fires presents formidable technical challenges in sampling methods, equipment, and logistics.37 While studies of firefighters have emphasized the importance of exposures at the fire scene, exposures at the firehouse, where firefighters spend long hours, also may have an impact on their risk of cancer. Diesel exhaust from fire trucks, especially if their en gines are run in closed houses without direct venting to outside air, may lead to high lev els of diesel exhaust emission particulates that are probably carcinogenic.24 Many fire companies are located in old buildings, where deteriorating asbestos-containing insula tion material may produce harmful levels of exposure to resident firefighters. The following sections summarize the available data regarding carcinogenic ex posures in the work environment of firefighters. Benzene Benzene is firmly established as a human carcinogen.36 Numerous studies have shown that benzene is a common airborne contaminant in fire smoke and occurs in con centrations that are considered deleterious in the context of chronic exposures. Treitman, Burgess, and Gold studied ambient environmental levels of a number of air contaminants, including benzene, at more than 200 structural fires in Boston in the mid-1970s.69 Benzene was detected in 181 of 197 (92%) samples taken at fire scenes by air sampling units placed on the chests of firefighters. Half of the samples showed ben zene over 1 part per million (ppm), the current OSHA permissible exposure level. Ap proximately 5% of the samples were above 10 ppm benzene.69 Lowry and colleagues studied firefighters' exposure to benzene at nearly 100 struc tural fires in Dallas in the early 1980s.41 They found benzene at the majority of the fires but did not provide information about the levels measured. They also detected the pres ence of at least 70 organic chemical species regardless of whether synthetic materials were a major part of the materials burned. Brandt-Rauf et al.11 used personal portable sampling devices to measure exposures of 51 firefighters at 14 fires in Buffalo in 1986. The tubes of the sampling devices were attached to the firefighters' turnout gear, thereby representing ambient air outside the mask. Benzene was second only to carbon monoxide as the most common chemical sub stance detected at the fires.11 It was detected in 18 of 26 samples from 12 of 14 fires. When detectable, the concentration of benzene ranged from 8.3 to 250 ppm. In only one sample where benzene was detected was its concentration below 10 ppm. Even when the smoke's intensity was rated as low, benzene was usually present in concentrations ranging from 22 to 54 ppm. The authors noted that respiratory protection was only par tially used or not used at all at the fires judged to be of low smoke intensity.11 Jankovic and colleagues at the National Institute for Occupational Safety and Health (NIOSH) studied benzene and other exposures at 22 fires in the late 1980s, ORIGAN of people J Studies iCOg- lighters at s not suri they are . size, and exposures 'he actual e and the exposures ment, and ires at the may have f their en i high levMany fire ng insula- igenic ex- < have iio *ii con s. lumber of ton in the scenes by >wed benlevel. Ap- 100 struc>f the fires 1 the presmaterials exposures ices were utside the nical subf 14 fires. i only one ven when entrations only par i a and ite 1980s, The Risk of Cancer in Firefighters 805 including 6 training fires, 15 residential fires, and 1 automobile fire.37 Samples were col lected via probes placed inside and outside the masks of working firefighters. In addi tion, industrial hygienists used a variety of sampling devices at the fire scene. Samples were taken separately during the two phases of a fire: knockdown and overhaul. Half of the samples taken during the knockdown phase of the fire showed benzene in concentrations of 1-22 ppm. Of the 29 organic substances analyzed qualitatively by gas chromatography/mass spectrometry, benzene was the most common compound de tected and was the only substance present in all eight samples. To measure the efficacy of respiratory protection, samples for benzene were taken inside and outside the mask.37 Surprisingly, the levels of benzene inside the mask were as high as those taken outside the mask and ranged from nondetectable to 21 ppm. The authors attributed this equivalence in benzene concentrations inside and outside the mask to partial or nonuse of the mask at the fire, especially after the initial phase of fire knockdown. They further suggested that benzene may be present only during the latter part of knockdown.37 During the overhaul phase of the fire, when respiratory protection is frequently re moved, benzene concentrations were low, i.e., less than 1 ppm.37 Asbestos Asbestos is universally recognized as a human carcinogen and has caused an ex cess in risk of a variety of cancers in numerous occupations.36 63 The extent to which a firefighter has potential exposure to asbestos at the fire scene is an interesting and largely unanswered question. Since the building destruction caused by fires and the building de molition actively performed by firefighters during overhaul are likely to dislodge res pirable asbestos fibers, the likelihood that firefighters have exposure to asbestos is high. However, the extent of such exposure is uncertain given intermittent exposure and use of respiratory protection. Markowitz and colleagues at Mount Sinai School of Medicine in New York per formed a cross-sectional study of 212 firefighters who had begun employment in the New York City Fire Department at least 25 years previously.43 All participants had worked principally in ladder companies and, thus, had engaged in overhaul operations frequently. In addition, all participants had worked in locations in New York City where exposure to asbestos-containing materials was considered to be most common: highrise office buildings, warehouses and factories, and poor neighborhoods with high fire activity in the 1960s. Twenty of the 152 (13%) firefighters without prior exposure to asbestos had pleural thickening and/or parenchymal opacities on chest x-ray that represented characteristic sequelae of prior asbestos exposure. All of the chest-ray abnormalities were mild in de gree. Twenty-two of the 60 (37%) firefighters with a history of exposure to asbestos prior to becoming a firefighter showed such radiologic abnormalities. Prevalence of ra diographic abnormalities did not increase with duration of employment as a firefighter or duration from onset of employment, but the study criteria for subject selection as sured a narrow range in these categories. The authors concluded that long-term firefighters in urban areas may have significant exposure to asbestos and are at risk for asbestos-related diseases.43 Although the Mount Sinai study was restricted to pleural and parenchymal fibrosis as outcomes of interest, the results are relevant to the issue of the risk of cancer for firefighters. The finding of excess risk of lung and pleural fibrosis due to asbestos among firefighters indicated that signifi cant asbestos exposure has occurred in this group. Since significant asbestos exposure con fers excess risk for selected cancers, it is reasonable to expect that firefighters have an in creased risk of various cancers as a result of their exposure to asbestos. 806 GOLDEN, MARKOWITZ, LANDRIGAN No environmental study of ambient levels of asbestos at fire scenes has been un dertaken. Jankovic et al. collected airborne fibers on cellulose filters at the scene of structural fires and analyzed these with polarized light microscopy.37 The limit of de tection was 0.4 fibers/ml. Fiber counts were higher during the overhaul phase than the knockdown phase of the fire. No asbestos fibers were detected, but cellulose and glass fibers were obtained. The investigators did not ascertain whether insulation materials were involved in any of the fires. They concluded that their results "do demonstrate the potential for exposures during overhaul when building materials contain asbestos."37 Polycyclic Aromatic Hydrocarbons Polycyclic aromatic hydrocarbons (PAHs) are a class of organic substances that have been implicated as the carcinogenic substances in coal tar pitches, coal tar, and se lected mineral oils.36 They have been associated with excess risk of a variety of cancers, including cancer of the skin, lung, kidney, and bladder.36 Given the combustion of diverse materials at fires, it is likely a priori that fire fighters would be exposed to significant levels of PAHs. Earlier studies of airborne con taminants at fires concentrated on the measurement of acute irritants and asphyxiants, ignoring the presence of PAHs. In their recent study, Jankovic et al. evaluated the pres ence of PAHs at the scene of fires.37 All 14 PAHs measured, including benz(a)pyrene, were present at mean values of 3-63 p,g/m3 during the knockdown phase of the fire. Concentrations of PAHs during overhaul were considerably lower than during knock down and were similar to those seen in ambient air in the absence offire.- Formaldehyde Formaldehyde is considered a probable human carcinogen.36 In animal experi ments, formaldehyde has caused cancer of the nasopharynx and the sinuses. There is also limited evidence that formaldehyde may cause cancer at other organ sites.1-8 The current OSHA permissible exposure level is 0.75 ppm for an 8-hour time-weighted av erage and 2 ppm for a 15-minute short-term exposure. Formaldehyde has been measured at the fire scene by Lowry et al.,41 Brandt-Rauf and colleagues,11 and Jankovic et al.37 Lowry et al. reported combined formaldehyde and acetaldehyde levels, with a mean of 5 ppm and a range of 1 to 15 ppm.41 BrandtRauf and colleagues found aldehydes, including formaldehyde, at 4 of 14 fires at con centrations of 0.1 to 8.3 ppm.11 Jankovic et al. detected formaldehyde at levels up to 8 ppm during knockdown and only 0.4 ppm during overhaul.37 They also reported that airborne concentrations of formaldehyde inside the mask ranged from nondetectable to 0.3 ppm. Diesel Exhaust Considerable experimental and epidemiologic evidence gathered over the past 15 years suggests that constituents of diesel exhaust emissions are carcinogenic and may present a risk to occupations with regular exposure. Firefighters have significant poten tial for exposure to diesel exhaust, because fire trucks with diesel engines are routinely started inside of and backed into firehouses. Froines and colleagues studied the concentration of diesel exhaust particulates in the air inside firehouses in New York, Boston, and Los Angeles in 1985.24 Participat ing firefighters wore personal air samplers throughout the work shift while they were in the firehouse. Unlike studies of air contaminants at the fire scene, the concentrations of airborne diesel particulate measured in this study should accurately reflect the actual exposure of JAN i unf . ^en the glass .`rials e the "37 . that d se vers. fireconlants, presrene, fire, lock- iS 3 The d av- -Rauf ihyde andt: conp to 8 i that ble to ist 15 1 may >otentinely ,tes in cipatere in I ? ui^ of The Risk of Cancer in Firefighters 807 firefighters to diesel emissions. Firefighters obviously do not wear respiratory protection at the firehouse. In addition, firefighters spend much of the work shift inside the firehouse, so that the 8-hour time-weighted average concentration reported by Froines et al. should meaningfully approximate the diesel exhaust exposure of urban firefighters on the job.24 Significant exposure to diesel exhaust particulates was detected.24 Total airborne particulates from diesel exhaust emissions ranged from 170 to 480 (xg/m3. Worst case scenario sampling, during which a very active shift was simulated, detected levels of diesel exhaust particulates in the air of fire houses as high as 748 |xg/m3. The authors conclude that these levels of diesel exhaust emissions may be associated with a signif icant carcinogenic risk and efforts to reduce exposure should be made.24 Unlike expo sures received at the fire scene, diesel exhaust emissions emanate from a specific source that can be controlled with local ventilation attached to the exhaust pipe of the fire truck. Other Agents Although less well studied, there are additional environmental agents to which fire fighters are exposed and for which experimental and/or epidemiologic studies support a relationship between exposure to the agent and the development of cancer. Examples include polychlorinated biphenyls (PCBs), various furans, styrene, and methylene chlo ride. In the studies by Jankovic et al.37 and Lowry et al.41 discussed above, the latter three agents or groups of agents were found in measurable concentrations at multiple fires, but data on actual airborne levels were not provided. Indeed, in the study by Low ery and colleagues, 70 organic agents were repeatedly identified in the smoke at multi ple fires in Dallas.41 Given the large number of chemicals that have been identified as being carcinogenic in the past two decades, at least in rodent tdst systems,55 it is likely that fire smoke contains additional carcinogens beyond those identified to date. Conclusion In conclusion, empirical data are now sufficient to support the notion that fire fighters are exposed to carcinogens in their work environment. The significance of such exposures is still unresolved. The exposures of firefighters are intermittent and variable in intensity. The respiratory protection they use is of uncertain efficacy and limited ac ceptability in the real world. Important exposures such as asbestos and diesel exhaust may occur during overhaul or at the firehouse, when respirators are not typically used. Furthermore, even if the dose of various carcinogens received by firefighters were bet ter known, the residual uncertainty about the degree of risk imparted would be great. Although the fact that firefighters are exposed to carcinogens in their work environment has been established, much additional work remains to be done. Sufficient knowledge exists at present, however, to justify diligent efforts to reduce the exposure of firefight ers to known carcinogenic agents. PREVALENT CANCERS IN FIREFIGHTERS AND ASSOCIATIONS WITH CARCINOGENIC OCCUPATIONAL EXPOSURES The results of 19 epidemiologic studies of cancer in firefighters published in the medical literature are summarized below. The data show that employment as a firefighter increases the risk of developing and dying from certain specific cancers: leukemia, nonHodgkin's lymphoma, multiple myeloma, and cancers of the brain, urinary bladder, and, possibly, prostate, large intestine, and skin. Graphic presentations of data re lated to these specific cancers (Fig. 1-6) include results from all published epidemiologic studies of firefighters that reported on that cancer. (Results for nonspecific organ systems or sites, e.g., digestive system or hematopoietic/lymphatic system, were not included.) For 808 GOLDEN, MARKOWITZ, LANDRIGAN Study Aronson et &L, 1994 (n*I4) Burnett et al., 1994 (n=38) Demers ettl., 1994 (n= 4) ^ vs. police vs. population Tomlins et al., 1994 (n= 5) Guidoni, 1993 (n= 3) Demeisetal., 1992 (n18) ^ vs. police vs. population Beaumont et alH 1991 (n= 5} Grimes et al., 1991 (ns 3) Heyer et al., 1990 (n* 3) Samaetal., 1990 (n 5) ^ vs. police vs. population Vena et at, 1987 (n= 6) Musk et at., 1978 (n 8) 10 201* hraTMTM 103 100 1000 FIGURE 1. Brain cancer risk estimates for firefighters from published epidemiologic studies. Studies listed by first author and publication year (n = observed number of cancers among fire fighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no excess risk) equaling 100 on log10 scale. ^Statistically significant increase in risk ratio (p<0.05>. a given study, the "risk ratio" reported is the measure the authors used to express the as sociation between firefighting and cancer: a standardized mortality ratio (SMR), propor tionate mortality ratio (PMR), standardized incidence ratio. (SIR), or a relative risk, inci dence density ratio or odds ratio multiplied by 100 (RR). The number of cancer cases or deaths observed among firefighters, the risk ratio, and the statistical significance of the re sult are indicated for each study. Unless otherwise stated, the reference group used to cal culate a risk ratio was the general population; certain studies calculated risk ratios for more than one reference group, for example, police officers and the general population. Brain Cancer Chemical exposures that are suspected causes of brain tumors include vinyl chlo ride, benzene, PAHs, PCBs, N-nitroso compounds, triazenes and hydrazines.36'65'71 Re cent epidemiologic studies consistently have found that brain cancer is strongly associ ated with firefighting, as shown in Figure 1. Generally, excess risk was most notable within 15-30 years of exposure, i.e., after a relatively short latency.2'16'68'70 Howe and Burch34 analyzed all cancer mortality studies of firefighters available as of 1989 and concluded that brain cancer fulfilled the criteria indicative of a causal association with firefighting, with a pooled SMR of 143 (95% confidence interval =93-212). A study by Aronson et al.2 of firefighters in metropolitan Toronto reported a sta tistically significant overall SMR of 201 (95% CI=110-337) for brain cancer, with the highest mortality among those with 5-9 years duration of employment as a firefighter (SMR=625, 95% CI= 170-1,600). Demers et al.16 analyzed mortality data from three northwestern cities in the United States and found that firefighters with 10-19 years of employment were at greatest risk (SMR=353, 95% 0=150-700). Although based on only three deaths, an analysis of Honolulu firefighters by Grimes et al.28 found a PMR of 378 (95% 0=122-1,171) for brain and other central nervous system cancers; analy ses by years of employment, were not reported. Tomling et al.68 were unique in finding dose-response relationships between brain cancer incidence and increasing age, dur- GAN ------1 1000 udies. I fire- xcess le asr-->r>es or iereocalmore chlo`Reisociitable e and ? and with a sta:h the ghter three ars of 2d on PMR .r ' - o' , dur The Risk of Cancer in Firefighters 809 ation of employment, and years since hire, and between brain cancer mortality and increasing age, duration of employment, and estimated number of fires fought among Stockholm firefighters who worked during 1931-1983. Cancers of Hematopoietic and Lymphatic Systems Leukemia and lymphoma are associated with environmental and occupational ex posure to benzene and 1,3-butadiene.36-47-49-72 The prevalence of benzene as a solvent, as a component of gasoline, and as a combustion product that forms during the burning of plastics and synthetics, and of 1,3-butadiene, a monomer found in tires and synthetic rubber products, guarantees that firefighters will be exposed to the gases released by these materials as they bum. Chemical exposures that have been associated with multi ple myeloma include benzene and petroleum products. Multiple myleoma risk is also increased in farmers, paper producers, furniture manufacturers, and woodworkers.9 Leukemia As seen in Figure 2, the majority of epidemiologic studies have found that fire fighters are at increased risk of leukemia.2'22'33'50 59 For example, Feuer and Rosenman22 reported a statistically significant PMR of 276 for firefighters compared to police offi cers in New Jersey and an almost twofold increase in mortality compared to the general population in New Jersey and in the United States. Similarly, Sama et al.59 found that firefighters had almost three times the risk of police officers when incident cases re ported to the Massachusetts Cancer Registry from 1982 to 1986 were examined (agestandardized mortality odds ratio=267, 95% 0=62-1,154). Several studies found that the highest risk occurred at older ages, after at least 30 years latency or duration of employment.2-16'33 However, a recent large study from NIOSH12 combining mortality data from 27 states reported excess risk for firefighters younger than 65 (PMR=171, 95% 0=118-240). Study Aronson et al., 1994 (n= 8) Burnett et al., 1994 (n=61) Demers eta]., 1994 (n* 6) Giles et al., 1993 (n= 0) Demers etaL, 1992 (n=l5) Beaumont et al., 1991 (ns 7) Sama etaL, 1990 (n* 6) Heyer et al., 1990 (n= 7) Feuer etal., 1986 (n= 4) Morton et al., 1984 (n= 4) vs. police vs. population vs. police vs. population vs. police * vs. population vs. NJ police vs. NJ population vs. US population 119 100 100 80E3U f 1 127 1267 P112 173 1276 177 1 186 1 346 10 100 1000 Risk Ratio FIGURE 2. Leukemia risk estimates for firefighters from published epidemiologic studies. Studies listed by first author and publication year (n = observed number of cancers among fire fighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no excess risk) equaling 100 on log10 scale. ^Statistically significant increase in risk ratio (p<0.05). 8X0 GOLDEN, MARKOWITZ, LANDRIGAN NonHodgkin's Lymphoma Several studies of firefighters evaluated this group of malignant diseases. Without exception, marked increases in risk were found (data not shown).2-12'15'26-59 The study from the Massachusetts Cancer Registry by Sama et al. found a statistically significant SMOR of 327 (95% 0=119-898) for firefighters relative to police officers.59 Studies by Giles et al.26 from Melbourne, Australia, and Aronson et al.2 from Toronto, Canada, reported that firefighters had twice the risk of non-Hodgkin's lymphoma of males in the general population. Multiple Myeloma Few individual epidemiologic studies of firefighters had sample sizes sufficient to assess risk of multiple myeloma (data not shown). Two of the four published studies that included multiple myeloma found lower than expected risk, based on one2 or two15 cases among firefighters. Two other studies reported increased risk associated with fire fighting.12'33 Although the confidence intervals were wide, the analysis of a cohort of Seattle firefighters by Heyer et al.33 reported an overall SMR of 225 (95% CI=47-660) and, for men with 30 years or more of fire combat duty, a statistically significant SMR of 989 (95% CI=120-3,571). Using the mortality experience for 1984--1990 for fire fighters from 27 states, Burnett et al. found a statistically significant age-adjusted PMR of 148 (95% CI=102-207).12 Howe and Burch34 combined the results of all cancer mor tality studies of firefighters available as of 1989 (including four unpublished reports) and concluded that there was consistent evidence of a causal association between mul tiple myeloma and firefighting (pooled SMR=151,95% CI=91-235). Cancers of Genitourinary System - Bladder Cancer Occupational chemical exposures known to cause bladder cancer include several aromatic amines, solvents, benzidine, PAHs, coal tars and pitches, soot and oils,13-31-36 substances commonly encountered by firefighters, particularly at fires in commercial es tablishments. As seen in Figure 3, the majority of epidemiologic studies found that fire fighting was associated with increased risk for bladder cancer. Guidotti29 and Vena et al.70 both reported a threefold increase in bladder cancer deaths compared to general population rates, with peak risks for firefighters age 60 and older, with latency of 40 or more years. Using incident cases from the Massachusetts Cancer Registry, Sama et al.59 found a statistically significant increased risk for firefighters compared to police officers (SMOR=211, 95% 0=107-414) and to the general population (SMOR=159, 95% 0=102-250). Demers et al.16 reported, based on two deaths, that the rate of bladder can cer was markedly lower than expected in a cohort of firefighters employed at least one year between 1944 and 1979 in Seattle and Tacoma, Washington, and Portland, Oregon (SMR=23, 95% 0=3-83 compared to the general population; age-standardized inci dence density ratio =16, 95% 0=2-124 compared to police officers). However, in a re cent retrospective cohort study among the firefighters from Seattle and Tacoma, the au thors determined that cancer incidence was greater than expected relative to both the general population and the police, based on 18 incident bladder cancer cases among fire fighters reported to a Surveillance, Epidemiology and End Results (SEER) tumor reg istry during 1974-1989.15 Kidney Cancer Occupational exposures that have been implicated as risk factors for renal cell car cinoma include asbestos, PAHs, lead phosphate, dimethyl nitrosamine, coke oven emis- /AN .t :udy cant dies ada, i the it to dies vo15 firert of 560) MR fireMR nor)rts) nul- eral 31,36 lesfirela et leral Oor al.59 cers )5% canone gon ncil re : authe firereg- c. nis- The Risk of Cancer in Firefighters 811 Study Aronson cl al., 1994 (n* 7) Bumeti et al., 1994 (n=37) Demers ei al., I994(n=18) vs. police vs. population Guidotti, 1993 (n- 4) Demers et al., 1992 (n= 2) Beaumont ci al.. 1991 (n= 5) Sama ct al., 1990 (n=26) Vena ct al., 1987 (n= 9) 16[vs. police [ |vs. population 23 j vs. police vs. population to igl 128 99 120 170 ............I 316 571 1211 159 3286 100 Risk Ratio --I 1000 FIGURE 3. Bladder cancer risk estimates for firefighters from published epidemiologic stud ies. Studies listed by first author and publication year (n = observed number of cancers among firefighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no ex cess risk) equaling 100 on log10 scale. * Statistically significant increase in risk ratio (p<0.05). sions, and gasoline.36'56-62 This list clearly includes agents encountered in firefighting; however, the eight epidemiologic studies that assessed kidney cancer in firefighters did not show consistently elevated risk (data not shown). Bumett et al.12 and Guidotti29 did find statistically significant excess mortality among firefighters from 27 states in the United States and from Alberta, Canada, respectively. Guidotti's SMR of 414 (95% CI= 166-853) for kidney and ureter cancer was the highest SMR reported in the study. Risk was greatest after 40-49 years latency and increased with duration of employment as a firefighter and with a calculated index of firefighting exposure opportunity.29 Con versely, a number of studies have reported lower than expected risk among firefight ers.2'6'15-16 Studies from the northwestern United States by Demers and others found lower than expected kidney cancer mortality16 and incidence.15 Although based on only two deaths, the SMR of 27 (95% 0=3-97) for kidney cancer mortality was statistically significant relative to the general population.16 Prostate Cancer High rates of prostate cancer have been reported among workers with cadmium ex posure and in chemists, farmers, loggers, textile workers, painters, and rubber industry workers.20-27'38'48 While no obvious carcinogenic exposure is common to all these groups, occupational risk factors clearly should be considered along with endocrinologic, sexual, and dietary factors in the etiology of prostate cancer. Figure 4 summarizes the data on firefighters' risk for prostate cancer. A 30-50% increase in risk was con sistently found in the majority of studies. Giles et al.26 found that prostate cancer inci dence among firefighters employed in Melbourne, Australia, between 1917 and 1989 occurred at twice the expected rate (SIR=209, 95% 0=67-488). A proportionate mor tality study by Grimes et al.28 from Honolulu found statistically significant increases for prostate cancer in both Caucasian (PMR=370, 95% 0=171-802) and Hawaiian (PMR=335,95% 0=107-1,045) firefighters. On the other hand, Beaumont et al.6 found a statistically significant decrement in prostate cancer mortality (SMR=38, 95% 0=16-75) in a retrospective cohort study of firefighters employed between 1940 and 1979 in San Francisco. 812 Study GOLDEN, MARKOWITZ, LANDRIGAN FIGURE 4. Prostate cancer risk estimates for firefighters from published epidemiologic stud ies. Studies listed by first author and publication year (n = observed number of cancers among firefighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no ex cess risk) equaling 100 on log10 scale. *Statistically significant increase in risk ratio (p<0.05). Testicular Cancer Only two epidemiologic studies specifically addressed testicular cancer in fire fighters.2'26 Giles et al.26 found no association between testicular cancer incidence and employment as a firefighter in Melbourne, Australia, between 1917 and 1989; however, this study was restricted to cancers that occurred between 1980 and 1989, and only two cases were reported. A recent report by Aronson et al.2 found higher than expected mor tality for men employed by the Toronto Fire Department during 1950-1989. Over this 40-year period, three testicular cancer deaths occurred in the cohort when only 1.19 were expected based on the Toronto male population of the same age and calendar pe riod, for an overall SMR of 252 (95% CI=52-737). All three deaths occurred in younger men with less than 15 years as firefighters (SMR=366, 95% 0=75-1,069) and within 20 years of first exposure (SMR=326, 95% 0=67-953). The epidemiologic character istics of testicular cancer show that it occurs most commonly from age 20 to 34, with a white:black ratio of 4:1 and a positive correlation with socioeconomic status.60 The in cidence and mortality rates in men younger than 30 have been increasing over time. Al though occupational risk factors have not been studied well, exposures to solvents and paints have been implicated 23 Testicular cancer risk should be assessed in future stud ies of firefighters. Cancers of the Digestive System Several established occupational exposures increase the risk of cancer of the di gestive system: asbestos, cutting and lubricating oils, dyes, solvents, and metallic com pounds.25,36 It is hypothesized that, once cleared from the airways, inhaled particles and the carcinogens that adhere to them are transferred to the gastrointestinal tract and swal lowed and exert their effect on the digestive epithelium. Cancers of the rectum, colon, liver, pancreas, stomach, and esophagus were assessed in the majority of epidemiologic studies, but too few studies included cancers of the buccal cavity or pharynx for mean ingful discussion. The Risk of Cancer in Firefighters i 813 Large Intestine Of particular relevance to firefighters are the higher than expected rates of colon and rectal cancer observed in workers with exposure to asbestos.63 Figure 5 demon strates that excess rectal cancer has been found consistently in many studies of fire fighters.2'6'12'15'52'5968'70 A similar pattern was evident for colon, colorectal or "intesti nal" cancer,7'15'16'18 26'30'52'70 although the risk ratios tended to be somewhat lower (data not shown). An analysis by Burnett and colleagues12 of mortality data for firefighters from 27 states found a statistically significant excess of rectal cancer, particularly under age 65 (PMR=186, 95% CI=110-294). Orris et al.52 reported significantly higher mor tality in Chicago firefighters during 1940-1988 for both rectal (PMR=164, 95% CI= 114-230) and colon (PMR=131, 95% 0=104-165) cancers. In three other stud ies,2'68'70 rectal cancer mortality among firefighters occurred at twice the expected rate, but these results did not reach statistical significance. Slightly lower than expected mortality was observed in two analyses of firefighters from the northwestern United States.16'33 However, the latest study from this area found that rectal cancer incidence was similar to both the police and the general population, while colon cancer incidence, although not significantly elevated, appeared to increase with duration of employment as a firefighter.15 Liver Cancer Primary liver cancer is rare in the general population of the United States. An giosarcoma of the liver has been associated with occupational and environmental ex posures, including arsenic and vinyl chloride monomer from PVC.21'36 PVC can be as sumed to be present at every structural fire site in recent years involving furniture, electrical wire, and cable insulation and water pipes, and at automobile fires. Five epidemiologic studies reporting results for cancer of the liver (including Study 10 100 1000 Risk Ratio FIGURE 5. Rectal cancer risk estimates for firefighters from published epidemiologic studies. Studies listed by first author and publication year (n = observed number of cancers among fire fighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no excess risk) equaling 100 on log10 scale. ^Statistically significant increase in risk ratio (p<0.05). 814 GOLDEN, MARKOWITZ, LANDRIGAN cancer of the biliary passages and gallbladder) were all based on small numbers of cases observed in firefighters (data not shown). The study with the largest number6 found a twofold excess for liver cancer mortality relative to the United States population among firefighters in San Francisco who were employed between 1940 and 1970 (SMR=191,95% 0=87-363, n=9). Tomling et al.68 found a nonsignificant increase in mortality (SMR=149, 95% 0=41-381, n=4) but a slight decrement in incidence (SMR=85, 95% 0=23-218, n=4) for liver cancer in Stockholm firefighters employed during 1931-1983, relative to regional rates. Three additional studies found no associ ation between firefighting and.liver cancer.2,16,70 Although such an association is bio logically plausible, only a very large study or meta-analysis would have adequate sta tistical power to detect an increase in this rare cancer. Pancreatic Cancer Many occupations and chemical carcinogens have been studied in relation to pan creatic cancer, with little consensus.53 Workers in chemical, petroleum, and metallurgic industries may have particularly high risk from exposures such as benzidine, {3-naphthylamine derivatives, and metal dusts.40,53'54 In general, epidemiologic data suggest that firefighting is not associated with cancer of the pancreas (data not shown). One study found a large but nonsignificant increase in incidence for firefighters compared to police officers (SMOR=319) but not compared to the general population (SMOR=98) in Massachusetts.59 Eight additional investigations assessed pancreatic cancer in fire fighters: one study reported a nonsignificant^ decreased risk (SMR=38),26 three stud ies reported slightly elevated risk,2,6,30 and four studies reported equal risk relative to the general population.15,16,26'68 Stomach and Esophageal Cancer Adenocarcinoma of the stomach and cancer of the esophagus have been associated , with asbestos exposure;10,25,62 as discussed above, asbestos is prevalent at the majority of structural fires. Workers involved in rubber manufacturing, metal working, wood and paper working, and coal mining have also shown high rates of stomach cancer.25 Most of the epidemiologic studies that addressed stomach cancer found a positive association with firefighting,6,15,16,18,33,68,70 but none of the overall results were statisti cally significant (data not shown). Eliopulos et al.18 studied a cohort of firefighters em ployed during 1939-1978 by the Western Australia Fire Brigade. Mortality from stom ach cancer was increased twofold relative to the general population (PMR=202, 95% ksi CI=65-470). A study of firefighters employed in Stockholm during 1931-1983 found a small overall SMR of 121 for stomach cancer mortality;68 however, both incidence and itt. mortality increased with duration of employment and number of fires fought. Although tests for trend did not reach statistical significance, stomach cancer incidence was sig nificantly elevated for firefighters with more than 30 years employment (SMR=289, 95% 0=149-505) or who fought more than 1,000 fires (SMR=264,95% 0=136-461). The data for cancer of the esophagus are more equivocal. Equal numbers of studies found positive6,15,70 and negative2,16,33 associations with firefighting (data not shown). Beaumont et al.6 found that mortality from esophageal cancer occurred at twice the expected rate (SMR=204,95% 0=105-357) in a retrospective cohort study of fire fighters employed between 1940 and 1979 in San Francisco. No increase was demon strated with increasing duration of employment or latency--in fact, the highest rate was seen for those with less than 20 years as a firefighter. The authors postulate that an in 7!: teraction between smoke exposure and alcohol consumption could explain the pattern of cancer mortality in their study population: elevated rates for cancers of the liver, esophagus, buccal cavity, and pharynx. The Risk of Cancer in Firefighters 815 Skin Cancer Skin cancer is a heterogeneous group of diseases, the majority of which are ma lignant melanoma (30,000 new cases in the United States per year) or basal cell or squa mous cell carcinomas (500,000 new cases per year). The most common risk factor for cancers of the skin is prolonged and intense exposure to sunlight. Occupational expo sure to soot and tars, coke oven emissions, arsenic, and cutting oils also have been as sociated with increased risk.19'36 Substances containing carcinogenic agents such as PAHs and PCBs may be absorbed by the skin of exposed body areas, including the hands, arms, face and neck, and other sites when protective clothing is permeated. Con tact with these substances can occur during fire knockdown and overhaul and during the cleaning of clothing or equipment. Figure 6 summarizes the studies that addressed skin cancer risk. (In studies that failed to differentiate melanoma from non-melanoma skin cancer, mortality rates are likely to include only melanoma since other forms of skin cancer are rarely fatal.) Sev eral studies found that firefighters had a statistically significant excess risk of skin can cer compared to the general population.12'22'59 Using deaths reported to a retirement sys tem between 1974 and 1980, Feuer and Rosenman22 found an almost threefold increase in skin cancer mortality for New Jersey firefighters compared to the United States population (PMR=270, p<0.05); firefighters were at somewhat higher risk than the gen eral New Jersey population (PMR=190) but at the same risk as New Jersey police offi cers (PMR=135). Risk among firefighters clearly increased with duration of employ ment and interval since first employment (PMR=388 for more than 25 years duration; PMR=314 for more than 27 years latency); it was not clear which referent pop ulation was used for these comparisons. Sama et al.59 analyzed incident melanoma cases reported during 1982-1986 to the Massachusetts Cancer Registry. They found a statistically significant excess for firefighters in comparison to the state population (SMOR=292, 95% 0=170-503) but no excess in comparison to police officers except in the age group 55-74 years (SMOR=513, 95% 0=150-1,750). Howe and Burch34 Study Aronson et al., 1994 (n= 2) Burnett el al., 1994 (n=38) Demers etal., 1994 (n= 9) Giles et al., 1993 (n= 5) Demers etal., 1992 (n= 6) Beaumont el al., 1991 (n= 7) Sama et al., 1990 (n=l8) Feuer et al., 1986 (ns 4) vs. police vs. population vs. police vs. population vs. police vs. population vs. NJ police vs. NJ population vs. US population 10 73 |------- 163* 100 H 120 1108 1112 98 Ani: " 1169 ....... 1 138 ............. 11292 1 1 135 1m ... / . .J12270 100 Risk Ratio 1000 FIGURE 6. Skin cancer risk estimates for firefighters from published epidemiologic studies. Studies listed by first author and publication year (n = observed number of cancers among fire fighters). Risk ratio expressed by authors as SMR, PMR, SIR, or RR, with null value (no excess risk) equaling 100 on log10 scale. ^Statistically significant increase in risk ratio (p<0.05). 816 GOLDEN, MARKOWITZ, LANDRIGAN combined the results of the studies of cancer in firefighters published through 1989 and determined that there was evidence of a statistically significant increase in risk of melanoma (pooled SMR=173, 95% CI=103-274). However, they concluded that sev eral criteria used to define a causal association were not fulfilled--for example, the abil ity to rule out potential confounders such as sunlight exposure and the limited evidence of a dose-response relationship. Lung Cancer As discussed above, firefighters may be routinely exposed to many known or suspected lung carcinogens, including asbestos, arsenic, PAHs, vinyl chloride and formaldehyde.58 Inhalation exposure can occur during active fire combat as well as dur ing the overhaul phase when protective breathing equipment is usually removed. Accordingly, lung cancer was specified a priori in the majority of epidemiologic studies as an outcome that would be plausibly related to firefighting. Of the 16 published studies that addressed cancer of the respiratory tract, not one found a statistically sig nificant excess risk of lung cancer for firefighters (data not shown). Only two cohort studies29-32 found moderately increased risks: Guidotti from Canada, with an SMR of 142 (95% 0=91-211) for deaths occurring during 1927-1987, and Hansen et al. from Denmark, with an SMR of 163 for deaths occurring during 1970-1980 (95% 0=75-310). A case-control study using Missouri Cancer Registry cases diagnosed be tween 1980 and 1985 found the category that included police, firefighters, and protec tive service occupations had elevated risks for squamous-cell carcinoma, small-cell car cinoma, and other or mixed cell types, but not for adenocarcinoma of the lung.73 These elevated risks were limited to current smokers only. Discussion These epidemiologic studies clearly demonstrate increased risk of several cancers that can be plausibly linked with carcinogenic exposures encountered by firefighters in their work. The data most strongly suggest that firefighters are at increased risk of de veloping and dying from leukemia, nonHodgkin's lymphoma, multiple myeloma, and cancers of the brain and bladder. The majority of studies that examined these cancers found markedly elevated risks for firefighters, and there are no viable alternative hy potheses or strong confounders that could readily explain their increased prevalence. Furthermore, exposure assessment studies have detected substances in the firefighting environment that are known or suspected causes of these cancers. Weaker but still plau sible evidence links firefighting to increased risk of rectal, colon, stomach, and prostate cancers and melanoma. The limitations of the epidemiologic data must be acknowledged. Most of the stud ies examined relatively small populations of firefighters and thus have low statistical power to analyze rare tumors. To increase their sample size, many of the studies ana lyzed deaths occurring over several decades; this technique introduces problems related to (a) trends in diagnoses, (b) differences in exposure over time, since many potential carcinogens, such as chemicals and synthetic materials, were introduced at different times during the relevant exposure periods, and (c) changes in protective equipment and awareness ofhazards. Limited documentation of exposure is also a problem. Some stud ies relied on occupation as recorded on a death certificate or tumor registry, which may reflect the current or most recent job instead of the usual occupation. Recent studies have examined risk in relation to duration of active fire combat duty, latency (years since hire), age at diagnosis (active duty versus retirement), and number of fires fought. How ever, none were able to rank firefighters according to a cumulative index incorporating intensity of exposure. As a result, heavily exposed firefighters are comingled with The Risk of Cancer in Firefighters 817 lightly exposed firefighters, and the risks to the heavily exposed firefighters are diluted out and underestimated by the design of the studies. None of the epidemiologic studies were able to take into account potential con founding variables other than age that could explain the observed associations between firefighting and cancer.42 It is unlikely, however, that increased mortality rates among firefighters can be attributed solely to the personal lifestyle factors--diet, alcohol intake, cigarette smoking--that have been linked with certain cancers. The vast majority of studies found no excess risk of lung cancer, suggesting that firefighters are not more likely to smoke than the general population or other protective service workers. In fact, surveys have found that the proportion of firefighters who smoke is similar to the pro portion of other service and blue collar workers who smoke.5'59'67 In studies of occupa tion and cancer that did collect information on lifestyle factors, most associations re mained unchanged after controlling for cigarette smoking,4'17 and biased attribution of cause of death among smokers compared to nonsmokers has been shown to overesti mate associations between smoking and cancer.66 The latency period for most of the relevant cancers associated with exposure to chemical carcinogens is likely to be at least three or four decades. Therefore, studies to date have not had sufficient follow-up time to detect the full extent of occupational can cer in the firefighters at greatest risk--those who were increasingly exposed to chemi cal carcinogens throughout the 1940s, 1950s, and 1960s without the benefit of modem protective equipment or awareness of hazards. The results of the studies also may be subject to the paradox of the healthy worker and survivor effects.3'35'46 Healthy individuals are more likely than unhealthy persons to seek and gain employment and to remain in theirjobs. This effect is amplified by the strin gent initial screening process and good employment benefits associated with employment as a firefighter, as evidenced by their low all-cause mortality rates. Although the healthy worker effect has less impact on cancer than on other causes of death, the higher than ex pected rates of cancer mortality among firefighters in comparison to the general popula tion and, in particular, to other workers are unsettling. Indeed, the shortcomings of the epi demiologic studies are more likely to dilute or mask associations between occupational exposures of firefighting and cancer than to create falsely positive associations. Few of the results presented reached statistical significance, and the confidence in tervals around the risk ratios were generally wide. Statistical significance is determined by the magnitude of the exposure-disease association, the accuracy or variability of the exposure and outcome measurements, and the size of the study population. Therefore, the small numbers of cancers observed in individual studies contribute to instability in the risk estimates. Future studies that are able to include not just deaths but all incident cancers from large cohorts will benefit from analyzing greater numbers of events. Fig ures 1-6 illustrate the preponderance of evidence implicating certain specific cancers associated with firefighting. Although these cancers warrant particular attention, future investigations should continue to cast a wide net that includes all relevant cancers. The downside of testing many outcomes in relation to a number of exposure variables is that some associations may appear to be statistically significant by chance alone. Because most of the epidemiologic studies used the retrospective cohort study de sign, investigators had access to employer records regarding employment period, work assignments, and vital status, rather than just occupation as recorded on a death certifi cate. Attempts should be made in future studies, particularly those with prospective components, to develop measures of acute and cumulative exposures on an individual basis, although potential misclassification will always be a concern given the nature of the firefighting environment. The techniques of molecular biology increasingly are be ing used to develop biomarkers of exposure in occupational and environmental settings. 818 GOLDEN, MARKOWITZ, LANDRIGAN For example, Liou et al.39 monitored two biomarkers in firefighters: sister chromatid ex change (SCE), a general indicator of genetic damage resulting from exposure to muta gens and carcinogens, and polycyclic aromatic hydrocarbon (PAH)-DNA adducts, which are thought to measure the initiation of carcinogenic changes associated with ex posure to PAHs. After controlling for charcoal-broiled food consumption, cigarette smoking and race, firefighters had a statistically significant fourfold higher risk of de tectable PAH-DNA adduct levels compared to unexposed controls. This association may be specific to urban, structural firefighting; a similar study in wildland firefighters in California found no association between forest fire activity and PAH-DNA adducts.57 The incorporation of biologic markers of exposure, cancer susceptibility, and preclinical effects should be considered in future epidemiologic studies of firefighters. Despite the limitations cited above, the available exposure assessment and epi demiologic studies present convincing and consistent evidence that the toxic exposures encountered in firefighting may increase the risk for certain specific cancers. The rela tively high incidence rates with which some of these cancers occur (prostate, colon, rec tum) and, for rarer cancers, the particularly strong association with firefighting or dismal survival probability (brain, multiple myleoma) underscore the importance of un derstanding and reducing the cancer risks attributable to firefighting. REFERENCES 1. Acheson ED, Barnes HR, Gardner MJ, et al: Formaldehyde in the British chemical industry. Lancet 1:611-616,1984. 2. Aronson KJ, Tomlinson GA, Smith L: Mortality among fire fighters in Metropolitan Toronto. Am J Ind Med 26:89-101,1994. 3. Anighi HM, Hertz-Picciotto I: The evolving concept of the health worker survivor effect. Epidemiology 5:189-196,1994. 4. Axelson O: Aspects of confounding in occupational health. Scand J Work Environ Health 12:486-493, 1978. 5. Bates JT: Coronary artery disease in the Toronto fire department. J Occup Med 29:132-135,1987. 6. Beaumont JJ, Chu GST, Jones JR, et al: An epidemiologic study of cancer and other causes of mortality in San Francisco firefighters. Am J Ind Med 19:357-372,1991. 7. Berg JW, Howell MA: Occupation and bowel cancer. J Toxicol Environ Health 1:75-89,1975. 8. Blair A, Stewart PA, Hoover RN: Mortality from lung cancer among workers employed in formaldehyde industries. Am J Ind Med 17:683--699,1990. 9. Blattner WA: Multiple myeloma and macroglobulinemia. In Schottenfeld D, Fraumeni JF (eds): Cancer Epidemiology and Prevention. Philadelphia, WB Saunders, 1982, pp 795-813. 10. Botha JL, Irwig LM, Strebel PM: Excess mortality form stomach cancer, lung cancer, and asbestosis and/or mesothelioma in crocidolite mining districts in South Africa. AmJEpidemiol 123:30-40,1986. 11. Brandt-Rauf PW, Fallon LF Jr, Tarantini T, et al: Health hazards of fire fighters: Exposure assessment. BrJ Ind Med 45:606-612,1988. 12. Burnett CA, Halperin WE, Lalich NR, Sestito JP: Mortality among fire fighters: A 27 state survey. Am J Ind Med 26:831-833,1994. 13. Cole P, Hoover R, Friedell GH: Occupation and cancer of the lower urinary tract. Cancer 29:1250-1260, 1972. 14. Committee on Fire Toxicology: Fire and Smoke: Understanding the Hazards. Washington, DC, National Academy Press, 1986. 15. Demers PA, Checkoway H, Vaughan TL, et al: Cancer incidence among firefighters in Seattle and Tacoma, Washington (United States). Cancer Causes Control 5:129-135,1994. 16. Demers PA, Heyer NJ, Rosenstock L: Mortality among firefighters from three northwestern United States cities. Br J Ind Med 49:664-670,1992. . 17. Dubrow R, Wegman DH: Setting priorities for occupational cancer research and control: Synthesis of the results of occupational disease surveillance studies. J Natl Cancer Inst 71:1123-1142,1983. 18. Eliopulos E, Armstrong BK, Spickett JT, Heyworth, F: Mortality of fire fighters in Western Australia. Br J Ind Med 41:183-187,1984. 19. Emmett EA: Occupational skin cancers. State Art Rev Occup Med 2:165-177, 1987. 20. Emster VL, Selvin S, Brown SM, et al: Occupation and prostatic cancer. A review and retrospective analysis based on death certificates in two California counties. J Occup Med 21:175-183,1979. f The Risk of Cancer in Firefighters 819 21. Falk H, Caldwell GG, Ishak KG, et al: Arsenic-related hepatic angiosarcoma. Am J Ind Med 2:43-50, 1981. 22. Feuer E, Rosenman K: Mortality in police and firefighters in New Jersey. Am J Ind Med 9:517-527, 1986. 23. Fleming L: Cancers of the reproductive organs. In Rosenstock L, Cullen MR: Textbook of Clinical Oc cupational and Environmental Medicine. Philadelphia, WB Saunders, 1994, pp 591-599. 24. Froines JR, Hinds WC, Duffy RM, et al: Exposure of fire fighters to diesel emissions in fire stations. Am Ind Hyg Assoc J 48:202-207,1987. 25. Frumpkin H: Cancer of the liver and gastrointestinal tract. In Rosenstock L, Cullen MR: Textbook of Clinical Occupational and Environmental Medicine. Philadelphia, WB Saunders, 1994, pp 576-584. 26. Giles G, Staples M, Berry J: Cancer incidence in Melbourne metropolitan fire brigade members, 1980-1989. Health Rep 5:33-38,1993. 27. Greenwald P: Prostate. In Schottenfeld D, Fraumeni JF (eds): Cancer Epidemiology and Prevention. Philadelphia, WB Saunders, 1982, pp 938-946. 28. Grimes G, Hirsch D, Borgeson D: Risk of death among Honolulu fire fighters. Hawaii Med J 50:82-85, 1991. 29. Guidotti TL: Mortality of urban firefighters in Alberta: 1927-1987. Am J Ind Med 23:921-940, 1993. 30. Guidotti TL, Clough VM: Occupational health concerns of firefighting. Annu Rev Public Health 13:151-171, 1992. 31. Gustavsson P, Gustavsson A, Hogstedt C: Excess of cancer in Swedish chimney sweeps. Br J Ind Med 45:777-781, 1988. 32. Hansen ES: A cohort study on the mortality of firefighters. Br J Ind Med 47:805-809,1990. 33. Heyer N, Weiss NS, Demers P, Rosenstock L: Cohort mortality study of Seattle fire fighters: 1945-1983. Am J Ind Med 17:493-504,1990. 34. Howe GR, Burch JD: Fire fighters and risk of cancer: An assessment and overview of the epidemiologic evidence. Am J Epidemiol 132:1039-1050,1990. 35. Howe GR, Chiarelli AM, Lindsay J: Components and modifiers of the healthy worker effect: Evidence from three occupational cohorts and implications for industrial compensation. Am J Epidemiol 128:1364-1375,1988. 36. International Agency for Research on Cancer: IARC Monographs on the Evaluation of Carcinogenic Risks to Humans, Suppl 7, Overall Evaluations of Carcinogenicity: An Updating of IARC Mono graphs. IARC Lyon, France, 1987. 37. Jankovic J, Jones W, Burkhart J, Noonan G: Environmental study of fire fighters. Ann Occup Hyg 35:581-602, 1991. 38. Kipling MD, Waterhouse JAH: Cadmium and prostate cancer. Lancet 1:730-731,1967. 39. Liou SH, Jacobsen-Kram D, Poirier MC, et al: Biological monitoring of fire fighters: Sister chromatid ex change and polycyclic aromatic hydrocarbon-DNA adducts in peripheral blood cells. Cancer Res 49:4929-4935,1989. 40. Lin RS, Kessler II: A multifactorial model for pancreatic cancer in man. JAMA 245:147-152, 1981. 41. Lowry WT, Juarez L, Petty VCS, Roberts B: Studies of toxic gas production during actual structural fires in the Dallas area. J Forensic Sci 30:59-71,1985. 42. Mahaney FX: Studies conflict on fire fighters' risk of cancer. J Natl Cancer Inst 83:908-909, 1991. 43. Markowitz S, Garibaldi K, Lilis R, Landrigan PJ: Asbestos exposure and fire fighting. Ann N Y Acad Sci 643:573-576, 1992. 44. Mastromatteo E: Mortality in city firemen. II. A study of mortality in firemen of a city fire department. Arch Ind Health 20:227-233,1959. 45. McDiarmid MA, Lees PSJ, Agnew J, et al: Reproductive hazards of fire fighting. II. Chemical hazards. Am J Ind Med 19:447-472, 1991. 46. McMichael AJ: Standardized mortality ratios and the "healthy worker effect:" Scratching beneath the sur face. J Occup Med 18:165-168,1976. 47. McMichael AJ, Spirtas R, Kupper LL, Gamble JF: Solvent exposure and leukemia among rubber work ers: An epidemiologic study. J Occup Med 17:234-239, 1975. 48. Monson RR, Fine U: Cancer mortality and morbidity among rubber workers. J Natl Cancer Inst 61:1047-1053,1978. 49. Monson RR, Nakano KK: Mortality among mbber workers. Am J Epidemiol 103:284-296, 1976. 50. Morton W, Marjanovic D: Leukemia incidence by occupation in the Portland-Vancouver metropolitan area. Am J Ind Med 6:185-205, 1984. 51. Musk AW, Monson RR, Peters MJ, Peters RK: Mortality among Boston firefighters, 1915-1975. Br J Ind Med 35:104-108, 1978. ` 52. Orris P, Kahn G, Melius J: Mortality study of Chicago firefighters [abstract]. Revue D'Epidemiologie Et De Sante Publique 40 (Suppl 1 ):S90--91,1992. 53. Partanen T, Kauppinen T, Degerth R, et al: Pancreatic cancer in industrial branches and occupations in Finland. Am J Ind Med 25:851-866,1994. 820 GOLDEN, MARKOWITZ, LANDRIGAN 54. Pietri F, Clavel F: Occupational exposure and cancer of the pancreas. A review. Br J Ind Med 48:583-587, 1991. 55. Rail DP, Hoga MD, Huff JE, et al: Alternatives to using human experience in assessing health risks. Annu Rev Public Health 8:355-3385,1987. . 56. Redmond CK, Ciocco A, Lloyd JW, et al: Longterm mortality study of steel workers. VI. Mortality from malignant neoplasms among coke oven workers. J Occup Med 14:621-629,1972. 57. Rothman N, Correa-Villasenor A, Ford DP, et al: Contribution of occupation and diet to white blood cell polycyclic aromatic hydrocarbon-DNA adducts in wildland firefighters. Cancer Epidemiol Biomark ers Prev 2:341-347,1983. 58. Russi MB, Cone JE: Malignancies of the respiratory tract and pleura. In Rosenstock L, Cullen MR: Text book of Clinical Occupational^ and Environmental Medicine. Philadelphia, WB Saunders, 1994, pp 543-555. ' 59. Sama SR, Martin TR, Davis L, Kriebel D: Cancer incidence among Massachusetts firefighters: 1982-1986. Am J Ind Med 18:47-54, 1990. 60. Schottenfeld D, Warshauer ME: Testis. In Schottenfeld D, Fraumeni JF (eds): Cancer Epidemiology and Prevention. Philadelphia, WB Saunders, 1982, pp 947-957. 61. Selikoff U, Hammond EC: Asbestos-associated disease in United States shipyards. CA Cancer J Clin 28:87-99,1978. 62. Selifkoff XJ, Hammond EC, Seidman H: Mortality experience of insulation workers in the United States and Canada, 1943-1976. Ann N Y Acad Sci 330:91-116,1979. 63. Selikoff IJ, Lee D: Asbestos and Disease. London, Academic Press, 1978. 64. Siemiatycki J: Risk Factors for Cancer in the Workplace. Boca Raton, FL, CRC Press, 1991. 65. Sinks T, Steele G, Smith AB, et al: Mortality among workers exposed to polychlorinated biphenyls. Am J Epidemiol 136:389-398,1992. 66. Sterling TD, Rosenbaum WL, Weinkan JJ: Bias in the attribution of lung cancer as a cause of death and its possible consequences for calculating smoking-related risks. Epidemiology 3:11-16, 1992. 67. Sterling TD, Weinkan JJ: Smoking characteristics by type of employment. J Occup Med 18:743-754, 1976. 68. Tomling G, Gustavsson P, Hogstedt C: Mortality and cancer incidence in Stockholm fire fighters. Am J Ind Med 25:219-228,1994. 69. Treitman RD, Burgess WA, Gold A: Air contaminants encountered by fire fighters. Am Ind Hyg Assoc J 41:796-802,1980. 70. Vena JE, Fiedler RC: Mortality of a municipal-worker cohort: IV. Fire fighters. Am J Ind Med 11:671-684, 1987. 71. Waxweiler RJ, Stringer W, Wagoner JK, et al: Neoplastic risk among workers exposed to vinyl chloride. Ann N Y Acad Sci 271:40-48,1976. 72. Young N: Benzene and lymphoma. Am J Ind Med 15:495-498,1989. 73. Zahm SH, Brownson RC, Chang JC, Davis JR: Study of lung cancer histologic types, occupation and smoking in Missouri. Am J Ind Med 15:565-578,1989. V