National Academies Press: OpenBook
« Previous: 3 Uncombusted Fuels and Combustion Products: Background Information
Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

4
CANCER

Cancer is a group of diseases characterized by uncontrolled growth and spread of abnormal cells (ACS 2003g; WHO 2003). Cancer can affect almost any tissue of the body. Known causes include external factors (such as chemicals, radiation, and infectious agents) and internal factors (such as mutations, hormones, and immune conditions). Such factors may act together or in sequence to initiate or promote carcinogenesis (ACS 2003g). In adults, a latent period of 10 years or more may elapse between exposure or mutation and the detection of cancer.

Cancer is the second-leading cause of death in the United States, exceeded only by heart disease. Among the member states of the World Health Organization (WHO), cancer is the third-leading cause of death, after heart disease and infectious or parasitic diseases (WHO 2003). In the United States in 2000, lung cancer was the leading cause of cancer deaths among both men and women, followed by prostatic cancer in men and breast cancer in women (Jemal et al. 2003). Each year, cancer leads to 12% of deaths worldwide, equivalent to about 6 million deaths (WHO 2003). Among men, lung and stomach cancers are the most common worldwide; among women, breast and cervical cancers are the most common (WHO 2003).

This chapter summarizes the results of epidemiologic studies of cancer outcomes related to exposure to fuels and combustion products. The committee considered the findings of those investigations as a means of determining what types of cancers Gulf War veterans might be at increased risk for as a consequence of exposure to fuels in the course of using military equipment, to tent heater fumes, and to smoke from oil-well fires. Because only a dozen years have passed since the Gulf War, studies of the Gulf War veterans themselves for cancer outcomes, which are characterized by considerable latent periods, would not yet be expected to be informative. Chapter 3 presented a general introduction on fuels and combustion products and a summary of toxicologic information on them. Appendix D contains tables that describe studies of populations exposed to relevant agents; many of the studies are referred to repeatedly in this chapter because their findings are related to several cancers of the specific anatomic sites and tissues reviewed.

In this chapter, the section on each type of cancer contains pertinent findings from cohort studies and then from case-control studies, first for fuels and then for combustion products, followed by the committee’s conclusions regarding the relationship between cancer of the specific type and exposure to fuels or combustion products. The tables included at the end of this chapter contain results from the primary studies on which the committee bases its conclusions.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Those tables are presented in reverse chronological order by each type of study design. The committee reviewed over 500 epidemiologic studies on cancer related to exposure to fuels and combustion products and selected studies that met its inclusion criteria for more thorough evaluation. Briefly, the studies had to appear in peer-reviewed publications, identify exposure relevant to the committee’s charge, and identify a specific health outcome (for example, the study must specify a type of cancer as opposed to considering all cancers together). Chapter 2 discusses the committee’s inclusion criteria in more detail.

This chapter reviews epidemiologic studies of cancer in adults, which would be pertinent to the occurrence of cancer in Gulf War veterans themselves; studies of childhood cancer are reviewed in Chapter 7, on reproductive and developmental effects, because the committee was concerned with such outcomes in the offspring of Gulf War veterans as a possible result of parental exposure. Epidemiologic studies assessing gender-specific cancers (for example, female breast cancer and prostate cancer) are included in the committee’s review. Seven percent of the 697,000 US military personnel sent to the Persian Gulf were women.

For the combustion products of crude oil and petroleum-derived fuels, the epidemiologic data complement the vast amount of toxicologic information on poly cyclic aromatic hydrocarbons (PAHs) (particularly benzo[a]pyrene), other combustion products, and soot. There are numerous studies of occupational cohorts heavily exposed to PAHs (for example, from coal tar and asphalt), usually in combination with other products of combusted petroleum-derived fuels (for example, exhausts from various sources and metals) and soot. The conclusions from that large, complex body of information have been addressed by several expert bodies, including the International Agency for Research on Cancer (IARC 1985), which (IARC 1984a, 1984b) have been virtually unanimous in judging that PAHs and soot are most probably human carcinogens, particularly for skin after dermal exposure.

Urban firefighter studies were not included in the committee’s review. The committee agreed that urban fire fighters are likely exposed to a number of compounds that are not found in combustion products produced from oil-well fires, tent heaters, and vehicles (for example, plastics, asbestos, and PCBs). It would not be possible for the committee to distinguish between health effects in urban firefighters attributable to those compounds versus combustion products as were experienced in the Gulf War. Therefore, the committee made a decision not to include urban firefighter studies in this report.

Cancer sites or types are addressed in this chapter largely according to the ninth revision of the International Classification of Disease (ICD-9).1 That approach is taken in an effort to organize the multitude of site-specific evidence presented in the chapter. In many cases, the findings by various investigators do not follow the strict categorization of the ICD-9.

CANCERS OF THE ORAL CAVITY AND OROPHARYNX

The cancers reviewed in this section include those of the oral cavity, that is, the lips, the lining of the lips and cheeks, the teeth, the gums, the tongue, the floor and roof of the mouth, and the area behind the wisdom teeth) (ICD-9 140–145); and the oropharynx and hypopharynx, the parts of the throat just behind the mouth (ICD-9 146 and 148, respectively). With cancers of the

1  

ICD codes are revised and updated by WHO. Although ICD-10 codes have been published, ICD-9 codes remain the most widely recognized and used. ICD codes were established by WHO to promote international comparability in the collection, processing, classification, and presentation of mortality statistics. The codes group cancers according to their organ or tissue of origin and their histologic features.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

nasopharynx (ICD-9 147) and of the nasal cavity and paranasal sinuses (ICD-9 160)—the next section is on cancers of the nasal cavity and nasopharynx—these cancers were formerly denoted “head and neck cancers”. Recently, some cancer epidemiologists (for example, Berrino et al. 2003; Boffetta et al. 2003) have chosen to consider the hypopharynx with the larynx (ICD-9 161), which it is next to, when assessing risks at that site associated with occupational exposure. In discussing the epidemiologic literature on cancer of the oral, nasal, and upper respiratory tissues, the committee has decided to specify exactly which sites individual researchers were reporting on. The committee has opted to draw conclusions related to the separate tissues that would be exposed during inhalation: along the oral pathway, along the nasal pathway, and their juncture near the larynx.

As for all head and neck cancers, the most important risk factor for cancers of the oral cavity and oropharynx is tobacco use, particularly cigarette-smoking (ACS 2003b, 2003c, 2003d). Additional risk factors for this site are alcohol consumption, vitamin A deficiency, exposure to ultraviolet radiation (sunlight), and increasing age. Some genetic factors, a weakened immune system, chronic irritation, and infection with human papillomavirus also may contribute to the occurrence of oral cancers.

In 2000, there were 10.6 new cases of cancer of the oral cavity and oropharynx per 100,000 people (15.9 among men and 6.2 among women) and 2.7 deaths per 100,000 (4.1 among men and 1.6 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.1 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and cancers of the oral cavity and oropharynx.

Cohort Studies

With the assistance of industrial hygienists and others familiar with a uranium-processing facility in Fernald, Ohio, Ritz (1999) conducted a secondary exposure assessment by using available data on the exposure of 4,128 male workers to kerosene (and to cutting fluids and trichloroethylene). The potential confounding effects of smoking were assessed by examining whether smoking habits were related to chemical exposure in a subset of 20% of the workers on whom smoking information was available in recent personnel files. There was no clear pattern of smoking behavior and exposure, so differences in smoking habits among exposure groups probably did not explain study results. Compared with the US population, the risk of death from oropharyngeal cancers (ICD-8 140–149) in the entire cohort was not notably increased (standardized mortality ratio [SMR] 1.05, 95% confidence interval [CI] 0.48–1.99). Among workers exposed to kerosene, the risks of oropharyngeal cancers derived with conditional logistic regression adjusted for pay status, time of hire, and cumulative radiation dose increased with exposure, but were imprecise (low kerosene exposure relative risk [RR] 1.85, 95% CI 0.37–9.36; moderate kerosene exposure RR 2.87, 95% CI 0.43–19.2; no workers had been categorized with heavy kerosene exposure).

Lagorio et al. (1994) tracked the mortality experience of 2,308 men through 1992; the men had been managers of Italian service stations in 1980. The effort complemented a detailed assessment of exposure at service stations in which 111 attendants were monitored in 1992 (Lagorio et al. 1993). Observation of only a single death from oropharyngeal cancer (ICD-9 140–

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

149) (SMR 0.38, 90% CI 0.02–1.79) during the follow-up period rendered this study uninformative.

Jarvholm et al. (1997) investigated cancer morbidity in a cohort of 4,128 male Swedish workers found by reviewing personnel files of 26 different refineries, distribution companies, lubrication-oil manufacturing industries, tank-cleaning companies, and companies that handled fuel. Exposure was determined from job titles combined with a retrospective review of air monitoring of work areas and personal exposure. The cohort was linked to the Swedish cancer and mortality registers. When the full array of oropharyngeal cancers were grouped (ICD-9 140–149), only six cases were identified in the cohort, so the somewhat increased risk estimates could not be distinguished from no effect, even for the subgroup of distribution workers with long duration and latency (standardized incidence ratio [SIR] 2.5, 95% CI 0.44–7.9).

Case-Control Studies

Zheng et al. (1996) conducted a case-control study to investigate the risk of salivary gland cancer (ICD-9 142) among residents of urban Shanghai. Cases were ascertained from 1988 and 1990. A total of 44 cases and 414 controls (frequency matched by sex and age) were interviewed to determine use of specific cooking fuels. Self-reported use of kerosene was associated with the risk of salivary gland cancer in models adjusted for sex, age, and income (odds ratio [OR] 3.5, 95% CI 1.6–7.4). Similar associations were found in multivariate models that included other possible risk factors but were not adjusted for smoking, which had not been found to be associated with salivary gland cancer.

A large population-based case-control study relying on several cancer registries was conducted in New Jersey, Los Angeles, Atlanta, and Santa Clara and San Mateo Counties, California, in January 1984–April 1985 (Huebner et al. 1992). Data were obtained on 1,114 cases (762 men and 352 women) of histologically confirmed primary oral and pharyngeal cancers (ICD-9 141, 143–146, 148, 149) diagnosed in January 1984–April 1985. The results for men working in selected industries were imprecise with some suggestion of an increased risk in petroleum-industry workers (OR 1.79, 95% CI 0.75–4.25). Employment history was obtained through interviews, and exposure was determined by job category. Smoking did not have an effect on the results.

Combustion Products

Table 4.2 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and cancers of the oral cavity and oropharynx.

Case-Control Studies

Gustavsson et al. (1998) identified incident cancers of the oral cavity (ICD-9 141, 143–145) or of the oropharynx or hypopharynx (ICD-9 146, 148) diagnosed in Sweden in 1988–1991. A total of 545 cases in men were identified by monitoring weekly medical reports and verified with the regional cancer registry; 641 controls were matched to the cases by region and age. Each case and control was interviewed by a trained interviewer according to a standardized protocol. Work histories were reviewed by industrial hygienists blinded to case status to estimate occupational exposure to 17 agents, including PAHs in a job-exposure matrix (JEM) approach. After adjustment for region, age, alcohol consumption, and smoking, high PAH exposure was

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

associated with both cancer at all sites (RR 1.48, 95% CI 1.09–2.01) and pharyngeal cancer (RR 1.52, 95% CI 0.94–2.45). No dose-response relationship, however, was evident in the case of PAH exposure.

Pintos et al. (1998) conducted a study in Sao Paulo, Curitiba, and Loiania, Brazil, to examine the risk of oral cancer (ICD-9 140–145) or pharyngeal cancer (ICD-9 146–149) in relation to the use of wood stoves. The researchers identified 784 incident cases of cancer of the pharynx and mouth from local hospitals and selected two controls for each case from among other hospital inpatients (patients with other cancers or mental disorders were excluded), matching them to cases on age, sex, and trimester of hospital admission. Exposure to products of wood stoves was ascertained with a single yes-no question. After adjustment for lifetime cumulative tobacco use (pack-years), alcohol consumption (kilograms of ethanol), sociodemographic variables, diet, and history of employment in specific industries, the reported use of wood stoves was associated with an increased risk of cancer of the mouth (OR 2.73, 95% CI 1.76–4.24) and pharynx (OR 3.82, 95% CI 1.96–7.42).

The Shanghai case-control study of salivary gland cancer (ICD-9 142) (Zheng et al. 1996) found self-reported use of kerosene for cooking was associated with risk of salivary gland cancer in models adjusted for sex, age, and income (OR 3.5, 95% CI 1.6–7.4). The increases in risk associated with use of coal, gas, and wood for cooking were not as precise.

Dietz et al. (1995) identified incident cancers of the oral cavity (ICD-9 141–145), oropharynx (ICD-9 146), or hypopharynx (ICD-9 148) in Heidelberg, Germany, and evaluated the effects of using fossil-fuel stoves for heating and cooking. Cases were ascertained in 1989–1992 from all patients seeking treatment at the Otorhinolaryngology Department at the University of Heidelberg within 3 years after first diagnosis. They identified 100 and 105 cases of oral cavity and pharyngeal cancer, respectively. Controls were recruited from the same medical center and general outpatient department and matched to cases on sex, age, and size of place of residence. All subjects were interviewed to ascertain risk-factor information, including alcohol consumption (grams/day), smoking (tobacco-years), and use of fossil-fuel stoves and cookers (coal, briquette, coke, peat, gas, and oil). After adjustment for tobacco and alcohol, use of fossil-fuel single stove heating units for more than 40 years vs 0–20 years was associated with pharyngeal cancer (OR 3.3, 95% CI 1.43–7.55). Fossil-fuel stove use for cooking in kitchen units also increased risk (OR 2.5, 95% CI 1.03–6.30) for more than 40 years of stove use (compared with 0–20 years). The OR for oral cavity cancer (ICD-9 141–145), adjusted for tobacco and alcohol, was 2.4 (95% CI 1.26–4.40) for more than 40 years (compared with 0–20 years) of exposure to fossil-fuel heating units. For exposure to kitchen cooking units, the OR for oral cavity cancer, adjusted for tobacco and alcohol, was 1.6 (95% CI 0.90–2.97) for more than 40 years of stove use (compared with 0–20 years).

Pukkala (1994) examined cancer incidence in Finland in 1971–1985 in 2,369 men and 809 women employed in various occupations. Occupation was ascertained by linking cancer-registry data with occupational and social-class data from the 1970 Finnish Population Census. SIRs were calculated from sex, age, site, and calendar-year-specific rates in the general Finnish population and were adjusted for social class. Risks of various cancers among men employed in transport and communications (lip cancer SIR 0.91, 95% CI 0.73–1.12; tongue cancer SIR 1.17, 95% CI 0.74–1.76; oral cavity cancer SIR 1.17, 95% CI 0.71–1.81; pharyngeal cancer SIR 0.97, 95% CI 0.66–1.39) were not found to be increased; there was a small increase in motor vehicle drivers (tongue cancer SIR 1.56, 95% CI 0.94–2.44), but the 95% CI contained the null.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

In the study by Huebner et al. (1992) of fuel exposures described above, the relationship between job categories potentially involving exposure to combustion products and primary cancers of the oral cavity and pharyngeal region were also assessed. The effect estimates derived for boiler or furnace and heavy-equipment operators (OR 1.50, 95% CI 0.68–3.34), heavy-equipment operators only (OR 1.25, 95% CI 0.78–2.01), motor-vehicle operators (OR 1.01, 95% CI 0.75–1.35), railroad transport workers (OR 1.00, 95% CI 0.30–3.35), mechanics or repairers (OR 0.86, 95% CI 0.66–1.12), and firefighters (OR 0.65, 95% CI 0.23–1.85) were imprecise and suggested no increases for these jobs. Similarly, the results for men working in selected industries were imprecise; no increased risk was observed for transportation workers (OR 1.07, 95% 0.74–1.56) and trucking or warehousing workers (OR 0.86, 95% CI 0.56–1.31). Among women, point estimates exceeded 1 for oral or pharyngeal cancer in association with employment as a motor-vehicle operator, but results were imprecise because of the small numbers (OR for motor-vehicle operator 2.80, 95% CI 0.61–12.9).

Incident oral and pharyngeal cancer was assessed in Shanghai, China (Zheng et al. 1992). A total of 204 cases 20–75 years old were ascertained in 1988–1990 and matched to 414 controls on age and sex. Exposure to potential risk factors was ascertained by interview. The prevalence of men using kerosene stoves among cases was reported to be 27.0% compared with 14.1% of controls (p≤0.01), but no difference was reported for women.

Merletti et al. (1991) conducted a population-based case-control study in Turin, Italy. From July 1982 to December 1984, 103 incident male cases of oral cavity or oropharyngeal cancer were identified. The questionnaire included a detailed occupational history. That information was reviewed by industrial hygienists and physicians experienced in occupational medicine, who determined the probability and intensity of exposure to 16 agents, including PAHs. After adjustment for age, education, geographic region of birth, tobacco-smoking, and alcohol consumption, probable or definite exposure to PAHs was not associated with cancer risk (OR 0.6; study authors stated that confidence interval included 1).

Patient records were abstracted for a case-control study at the Roswell Park Memorial Institute in Buffalo, New York (Decoufle and Stanislawczyk 1977; Viadana et al. 1976). All persons referred to the Institute in 1956–1965 were asked to report their lifetime occupational history and job activities. Their risk of cancer of the buccal cavity and pharynx was compared with that of noncancer controls according to jobs they had ever held or had held for 5 years or more. Data were analyzed and stratified by age at diagnosis (with the cut point at 60 years), and the results were adjusted for smoking. Compared with the risk in clerical workers, no increased risk of cancer of the buccal cavity and pharynx was reported for bus, taxicab, and truck drivers; deliverymen and routemen; locomotive engineers and firemen; mechanics and repairmen; or mine operatives and laborers.

Conclusion

The three cohort mortality studies that assessed the relationship between cancer of the oral cavity and oropharynx and fuels (Jarvholm et al. 1997; Lagorio et al. 1994; Ritz 1999) had limited statistical power and therefore were mostly uninformative. The case-control studies failed to report any consistent relationships between occupational or other self-reported potential exposures to fuels and cancer of the oral cavity and oropharynx (Huebner et al. 1992; Zheng et al. 1996).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels and cancers of the oral cavity and oropharynx.

All available studies of exposure to combustion products and cancer of the oral cavity and oropharynx were of the case-control design, and all were adjusted for cigarette-smoking and other confounders. Results of several studies suggest an association between cancers of the oral cavity and oropharynx and exposure to combustion products. Pintos et al. (1998) demonstrated an association between wood-stove use and cancers of the upper aerodigestive tract, and there were supportive findings from Dietz et al. (1995) that were based on exposure from fossil-fuel stove use in Germany, from Gustavsson et al. (1998) on PAH exposure in Sweden, and from Zheng et al. (1992, 1996) on kerosene-stove use in China.

The committee concludes, from its assessment of the epidemiologic literature, that there is limited/suggestive evidence of an association between exposure to combustion products and cancers of the oral cavity and oropharynx.

CANCERS OF THE NASAL CAVITY AND NASOPHARYNX

Cancers of the nasopharynx (ICD-9 147) or of the nasal cavity and paranasal sinuses (ICD-9 160), which previously have been grouped with oral cancers as “head and neck cancers” were considered as a separate group by the committee. The tissues of the nasal cavity and nasopharynx are subject to exposures that may be somewhat different from those of the tissues of the oral cavity. Nasopharyngeal carcinoma (NPC) is the most frequent malignant tumor of the nasopharynx.

As for other cancers of the head and neck, the most important risk factor for cancers of the nasal cavity and nasopharynx is smoking (ACS 2003b, 2003c, 2003d). Others include diets high in salt-cured fish and meats and infection with the Epstein-Barr virus. Cancers of the nasal cavities and sinuses have been found to be associated with occupational exposures, such as to dusts from wood, textiles, leather, and metals; glues; formaldehyde; solvents used in furniture and shoe production; mustard gas; isopropyl alcohol; and radium.

In 2000, there were 1.4 new cases of cancers of the nasal cavity and nasopharynx per 100,000 in the US (1.9 among men and 0.9 among women) and 0.4 deaths per 100,000 (0.5 among men and 0.2 among women) (Ries et al. 2004). NPC is rare in most parts of the world, with incidences generally less than 1 per 100,000 persons per year (Muir et al. 1987). The highest incidence is observed among southern Chinese (30–50 per 100,000 person-years); it might be attributable to the consumption of salted fish and preserved foods early in life. Several studies of NPC focused on Chinese populations because of the large number of available cases.

Fuels

Table 4.3 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and cancers of the nasal cavity and nasopharynx.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Case-Control Studies

Increased risk of nasopharyngeal cancer posed by 20 occupational exposures was assessed in a case-control study (Armstrong et al. 2000). During a 2-year period, 530 subjects with histologically confirmed NPC were identified from four hospitals in Malaysia and 282 cases underwent interviews that included occupational history and work exposure. Each case was matched by sex and age to a general population control without a history of cancer of the head, neck, or respiratory system. Exposure to motor fuel or oil, assigned according to type of job, was associated with a greater risk of NPC in a crude analysis (OR 1.79, 95% CI 1.16–2.82). However, after adjustment for smoking, passive smoke exposure, and diet, the association was largely reduced, and the CI suggested no effect (OR 1.33, 95% CI 0.81–2.20). A case-control study conducted by Teschke (1997) had only four cases, and no increased risk of sinonasal cancer was observed.

Combustion Products

Table 4.4 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and cancers of the nasal cavity and nasopharynx.

Case-Control Studies

A case-control study of NPC was conducted in rural Zangwu County in the middle 1980s (Zheng et al. 1994). Beginning in 1986, 88 cases of NPC were recruited with 176 controls matched on neighborhood, sex, and age. Subjects were interviewed to determine their use of wood fuels in the year before diagnosis. Use of wood fuels was associated with an increased risk of NPC in unadjusted models (OR 3.7, p=0.02), models adjusted for a sociodemographic confounder score (OR 6.4, p=0.003), and models adjusted for the confounder score, childhood consumption of salted fish, and consumption of herbal tea in the year before diagnosis (OR 5.4, 95% CI 1.5–19.8). There was some evidence that the risk of NPC was increased by household factors that can affect fume concentrations; specifically, the observed wood-fuel association increased in households with lower ventilation (for example, households that had no windows or no windows in the kitchen).

A hospital-based case-control study was conducted in Guangzhou City, China, in March 1983-August 1985 (Yu et al. 1990). Because Chinese living in the area have a high risk of NPC because of dietary factors, the researchers evaluated diet in detail and controlled for its influence in their analyses. There were 306 histologically confirmed incident cases of cancer of the nasal cavity and nasopharynx in subjects who were all less than 50 years old, and 306 controls were selected from the index cases’ neighborhoods of residence and matched on age, sex, and neighborhood. Exposure was based on self-reporting of occupation and exposure to specific risk factors. Subjects were interviewed in a standardized fashion that included inquiries about lifetime occupational history; exposure to dust, smoke, and chemical fumes; use of specific cooking fuels; and exposure to smoke from incense or mosquito coils. When associations were found for a self-reported occupational exposure, an occupational-medicine specialist reviewed the occupational information (job title and activity in job and industry) blindly to determine exposure status independently of case status. Exposure to smoke in a job held for at least 6 months (ever vs never) was found to be associated with an increased risk of NPC (RR adjusted for dietary risk factors in childhood based on self-reported exposure ever 2.4; 1–9 years of

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

exposure 1.6; 10 years or more of exposure 7.6). No confidence intervals were reported, but the authors stated that the results for “ever” exposed or exposed for 10 years or more had a “2-sided p value for the adjusted RR of less than 0.05”. Those results were attenuated when exposure was based on the specialist’s assessment (RR with 1–9 years of exposure 1.6, 95% 1.1–2.5; RR with 10 years or more of exposure 2.7, 95% CI 1.4–5.5). There was no association with domestic exposure to cooking fire, burning incense, or antimosquito coils.

Two case-control studies that collected occupational and environmental risk-factor information for NPC were conducted in Malaysia and included subjects of Chinese origin. The first was a hospital-based study (Armstrong and Armstrong 1983) of 117 histologically confirmed cases diagnosed in 1973–1980 and treated at the only radiotherapy center for NPC in Malaysia. In addition, the researchers interviewed 200 population controls (matched on neighborhood, sex, and ethnicity) to determine risk-factor information, including exposure to smoke and dust in the workplace. Exposure to both smoke and dust was associated with an increased risk of NPC among Chinese participants (RR for smoke exposure 6.0, p=0.006; RR for dust exposure 4.0, p<0.001). There was some evidence of an increased risk for Malays and Indians associated with smoke exposure, but the number of exposed cases was too small (four) to reach any conclusions. Smoke exposure was generated from the burning of wood, paper, grass, and oil and tar from in such occupations as rubber-tapping (wood-smoke exposure) and street-hawking. Some of the jobs seemed to involve more than one type of exposure; from the description of the analyses, the estimates do not seem to have been adjusted for multiple exposure or for risk factors other than age and sex.

The same researchers conducted a second case-control study on histologically confirmed cases of squamous-cell NPC which is described above (Armstrong et al. 2000). There were 119 prevalent cases (diagnosed before 1990) and 163 incident cases (diagnosed in 1990–1992). Exposures to inhalants were coded in a JEM approach by one of the authors blinded to case-control status. Cases and controls did not differ with respect to exposure amount or to median number of hours exposed to engine exhaust after adjustment for smoking and diet. No increased risk of NPC was associated with exposure to engine exhaust.

In addition to those largely Asian studies of wood-burning and other cooking fuels, the committee considered a large pooled reanalysis of cancers of nasal cavities and paranasal sinuses (ICD-9 160). Leclerc et al. (1997) assembled data from 12 previous studies in seven countries that included occupational-exposure information—the same set of 12 studies as reviewed in Demers et al. (1995) and in Luce et al. (2002). The dates of cancer diagnosis spanned 1968–1990. The inclusion criteria for the selected studies were histologic confirmation of cases; age, sex, and smoking information available on both cases and controls; and occupational histories of cases and controls obtained by interviews or questionnaires given to subjects or survivors (proxies). In 10 of the 12 studies, there were a total of 680 male cases of sinonasal cancer (330 squamous-cell carcinomas, 169 adenocarcinomas, 156 cases of other histologic types, and 25 cases of unknown histology) and 250 female cases of sinonasal cancer (102 squamous-cell carcinomas, 26 adenocarcinomas, 104 cases of other histologic types, and 18 cases of unknown histology). For the pooled analysis, subjects’ self-reported occupational information was recoded with the one-or two-digit International Standard Classification of Occupations. The researchers state that they controlled only for study and age category because they found that “introduction of cigarette smoking into the models, in addition to age and study, had no appreciable effect” on the risk of squamous-cell carcinoma. Among men, employment as a motor-vehicle driver was associated with an increased risk of adenocarcinoma (OR adjusted for study and age 2.50, 95% CI 1.03–

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

6.10) but not of squamous-cell carcinoma (OR 1.13, 95% CI 0.78–1.63). Duration of exposure did not change the point estimates for squamous-cell carcinoma, and only shorter exposure duration was linked to the increased risk of adenocarcinomas (OR<10 years 3.29, no CI; OR ≥10 years 0.80, no CI given). Also among men, employment as a cook was associated with an increased risk of squamous-cell carcinoma only in the shorter-duration group (OR adjusted 1.99, 95% CI 1.04–3.83; OR<10 years 2.72; OR≥10 years 1.25). Among women, however, employment as a cook was associated with a suggestion of a decreased risk of squamous-cell carcinoma (based on only three exposed cases; OR ever employed vs never employed 0.51, 95% CI 0.15–1.77; OR<10 years 0.27; OR≥10 years 0.69).

A population-based case-control study of incident nasal cavity and sinus cancer was conducted in British Columbia, Canada (Teschke et al. 1997). There were 48 cases identified in 1990–1992, and 159 population controls were frequency matched to cases on sex and age. Exposure was ascertained in interviews that included occupational history and items on individual exposures. Occupation-disease associations were estimated in models adjusted for age, sex, and tobacco use. There was no association with any particular occupation.

The incidence of pharyngeal, sinonasal, and oropharyngeal or hypopharyngeal cancer was assessed in a population-based case-control study from the Washington state cancer registry (Vaughan 1989). There were 231 cases aged 20–74 years diagnosed in 1979–1983 (sinonasal cancer) or 1980–1983 (pharyngeal cancer) and 552 population controls frequency matched to cases on age and sex (also de facto matched on telephone prefix because they were recruited with random-digit dialing). Histories for all jobs held at least 6 months were taken in interviews with subjects or their proxies. People were classified into 31 industrial and 59 occupational groups, and duration of employment was calculated on the basis of the start and end dates of employment. Models for oropharyngeal, hypopharyngeal, or sinonasal cancer were adjusted for age, sex, tobacco use, and alcohol consumption. Models for NPC were adjusted for age, sex, and race. Employment as a motor vehicle operator or other transportation worker was not associated with sinonasal cancer or pharyngeal cancer.

Conclusion

Little information is available on exposure to fuels and cancers of the nasal cavity and nasopharynx. The two studies reviewed by the committee did not report convincingly positive findings (Armstrong et al. 2000; Teschke et al. 1997).

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels and cancers of the nasal cavity and nasopharynx.

Overall, the results of studies of the relationship between combustion products and cancers of the nasal cavity and nasopharynx are inconsistent, and indirect methods were used to assess exposure. However, positive associations were reported by studies conducted in China (Yu et al. 1990; Zheng et al. 1994) between combustion products (particularly wood smoke) and cancer of the nasopharynx. Those findings are supported by the work of Leclerc et al. (1997) and Armstrong and Armstrong (1983). The committee believes that the evidence is strong enough to suggest an association between combustion products and cancers of the nasal cavity and nasopharynx.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The committee concludes, from its assessment of the epidemiologic literature, that there is limited/suggestive evidence of an association between exposure to combustion products and cancers of the nasal cavity and nasopharynx.

ESOPHAGEAL CANCER

This review focuses on esophageal cancer (ICD-9 150). Risk factors for that cancer are increasing age, sex, ethnicity, dietary habits, chronic reflux esophagitis, alcohol and tobacco use, and work exposure (ACS 2004q, 2004w).

In the United States in 2000, there were 4.7 new cases of esophageal cancer per 100,000 people (7.9 among men and 2.1 among women) and 4.4 deaths per 100,000 (7.7 among men and 1.8 among women) (Ries et al. 2004).

Fuels

Table 4.5 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and esophageal cancer.

Cohort Studies

A cohort of 3,814 male uranium-processing workers in Ohio in 1951–1989 was used to assess the potential relationship between esophageal and stomach cancers and kerosene exposure (Ritz 1999). Kerosene exposure based on a detailed industrial hygiene assessment (light or medium exposure) was associated with an increased risk of esophageal and stomach cancers; those cancers were analyzed together. Light kerosene exposure for 2 years or more with a 15-year lag before disease onset was associated with an RR of 3.46 (95% CI 1.22–9.80); medium kerosene exposure of the same duration and lag were also associated with an increased risk of esophageal and stomach cancer (RR 7.71, 95% CI 2.04–29.1).

Mortality was assessed in a cohort of 15,032 men with 5 years or more of work in 1964–1973 at Imperial Oil Limited refinery in Canada (Hanis et al. 1979). No specific industrial-hygiene assessment was available. There was an additional 11-year update that included 34,597 workers, including those hired in 1964–1983 (Lewis et al. 2000b; Schnatter et al. 1992). In followup through 1973, potential daily exposure to petroleum was associated with a greater risk of combined esophageal and stomach cancers (RR 3.25, p<0.05); the increase was greater with increasing years of employment (Hanis et al. 1979). However, the risk was not consistently observed in later followup studies of the cohort (Lewis et al. 2000b; Schnatter et al. 1992).

In a nationwide survey of gasoline-station attendants in Italy, the SMR for esophageal cancer was 2.34 (90% CI 0.80–5.35) (Lagorio et al. 1994). In small stations, where there were higher sales per employee, the risk of esophageal cancer was greater (SMR 3.42, 90% CI 1.17–7.82). Workers in small stations with higher sales of super-premium gasoline may have experienced higher exposure.

Combustion Products

Table 4.6 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and esophageal cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Cohort Studies

Mortality and cancer incidence were determined in a retrospective cohort study conducted in Stockholm in relation to diesel exhaust and asbestos exposure in bus-garage workers (Gustavsson et al. 1990). Although the primary focus of the work was lung-cancer incidence, esophageal-cancer incidence was also included. An increase was observed in the incidence of esophageal cancer in bus-garage workers compared with a local reference population, but the CI included the null (SMR 3.27, 95% CI 0.89–8.37).

The incidence of esophageal cancer among occupational groups in Sweden as recorded in the 1960 census was evaluated (Chow et al. 1995). SIRs were not increased for “transportation and communication” as a major industrial or occupational category. Of the reported occupations that might involve exposure to exhausts, only locomotive and traffic workers showed any increase in risk (SIR 1.1; p>0.05). The database did not permit adjustment for possible confounders.

Case-Control Studies

In a case-control study of 99 cases of confirmed esophageal cancer and age-matched population controls, no associations were reported between esophageal cancer and exposure to a number of combustion products, including nitrogen oxides (NOx), gasoline emissions, carbon monoxide, PAHs from any source, and mononuclear aromatic hydrocarbons (Parent et al. 2000b). Increases in the risk of esophageal cancer were reported in association with exposures to benzo[a]pyrene, PAHs specifically derived from coal, and PAHs specifically derived from petroleum, but the CIs all included the null. Exposure information was obtained with detailed questionnaires that were analyzed by chemists and industrial hygienists.

The relationship between PAH exposure and squamous-cell carcinoma of the esophagus was determined in a case-control study conducted in Sweden (Gustavsson et al. 1998). Participants were administered a questionnaire to determine exposure information, and work histories were reviewed and coded by an occupational hygienist. Occupational exposure to PAHs was associated with an increased risk of esophageal cancer. Estimated RR attributable to low PAH exposure was 2.01 (95% CI 1.16–3.48), and the RR of high PAH exposure was 1.87 (95% CI 1.11–3.16).

Another occupational case-control study evaluated the relationship between exposure to several types of engine exhausts and combustion products and esophageal cancer (Siemiatycki et al. 1988). Increased risk of esophageal cancer was associated with exposure to wood combustion products (OR 2.3, 90% CI 1.2–4.5) on the basis of only eight cases, but no increased risk was associated with exposure to gasoline exhaust, diesel exhaust, jet-fuel exhaust, propane exhaust, or products of combustion of propane, natural gas, liquid fuel, coal, or coke.

Conclusion

Studies of an association between fuel exposure and esophageal cancer are few, and their results are inconsistent. Two of the studies (Hanis et al. 1979; Ritz 1999) analyzed esophageal and stomach cancer together, so the committee cannot determine which cancer type may have been associated with exposure. Despite the larger number of studies of combustion products and esophageal cancer, no consistent association was observed.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and esophageal cancer.

STOMACH CANCER

This review focuses on gastric cancer (commonly known as stomach cancer) (ICD-9 151). Risk factors for stomach cancers are increasing age, sex, ethnicity, family history, dietary habits, and tobacco and alcohol use (ACS 2004q, 2004w). Heliobacter pylori infection is also a known cause of stomach cancer.

In 2000, there were 8.0 new cases of stomach cancer per 100,000 people (11.6 among men and 5.3 among women) and 4.6 deaths per 100,000 (6.4 among men and 3.2 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.7 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and stomach cancer.

Cohort Studies

Two cohort studies assessed the risk of combined esophageal and stomach cancer posed by exposure to fuels (Hanis et al. 1979; Ritz 1999). The results are presented in the previous section on esophageal cancer. No increased risk of stomach cancer was found in a nationwide survey of gasoline-station attendants in Italy (Lagorio et al. 1994). Amoco Oil Company employees in 1970–1980 who worked in operations were at increased risk for stomach cancer, with an SMR of 2.06 (Nelson et al. 1987), but those working in administration (and presumably were not exposed to petroleum products) had an SMR of 1.80.

Case-Control Studies

A case-control study of 3,726 cases of cancer in men in 19 Montreal hospitals was conducted to determine whether there was an association exists between exposure to fuels and various cancers, including stomach cancer (Siemiatycki et al. 1987a). Exposure was assessed with an industrial-hygiene assessment of exposure based on occupational history. The authors presented 90% confidence intervals and reported borderline increased risks for participants with automotive-gasoline exposure (OR 1.5, 90% CI 1.2–1.9). The risk of stomach cancer also was increased after exposure to kerosene (OR 1.7, 90% CI 1.2–2.3).

Combustion Products

Table 4.8 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and stomach cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Cohort Studies

Several epidemiologic studies examined rates of stomach cancer in defined populations compared with the general population. Occupational information on persons with stomach cancer as reported to the Shanghai Cancer Registry in 1980–1984 was assessed (Kneller et al. 1990). Increased SIRs were reported for fuel suppliers (p value not given), petroleum-refinery workers (p<0.05), and boiler firemen (p<0.01).

A similar approach was taken to evaluating the incidence of stomach cancer that was linked to occupation as reported in the 1960 Swedish census (Chow et al. 1994). There were 16,872 men diagnosed with stomach cancer in 1961–1979, but no increased risk was observed for “transportation and communication workers” as a major industrial or occupational category. Among machine and engine maintenance workers, crane operators were found to have an increased risk of stomach cancer (SIR 1.5, p<0.01). The database did not permit adjustment for potential confounders.

Case-Control Studies

Wu-Williams et al. (1990b) reported on a case-control study of 137 men with stomach cancer as reported to the Los Angeles County cancer registry. Exposure was determined by interviewing participants (or in some cases, surrogates) about occupational history using a structured questionnaire. No relationship was found between exposure to “smoke/exhaust” and stomach cancer.

A case-control study in Italy examined the relationship between occupational exposures and stomach cancer (Cocco et al. 1994). Men with histologically confirmed stomach cancer were interviewed to determine their occupational history. Controls were randomly selected from community and local health-unit registers of the resident population and matched on sex and age. A relative risk of 1.0 for stomach cancer was reported for men in several job categories in which exposure to combustion products might have occurred (for example, mechanics, repairmen, and railroad workers), but the CIs included the null. A small increase in risk was reported in men occupationally exposed to NOx (OR 1.4, 95% CI 1.0–2.1).

An occupational case-control study evaluated the relationship between exposure to several types of engine exhausts and combustion products and stomach cancer on the basis of an industrial-hygiene assessment of occupational history (Siemiatycki et al. 1988). No increased risk was associated with exposure to gasoline exhaust, diesel exhaust, jet-fuel exhaust, propane exhaust, or products of combustion of propane, natural gas, liquid fuel, coal, or coke.

A case-control study was conducted to assess the relationship between heating and cooking fuel-related exposures and stomach cancer in a coal-mining region of Pennsylvania (Weinberg et al. 1985). Cases were identified from death certificates, and followup interviews were conducted with next of kin to determine exposures. Increases in stomach-cancer risk were reported for coal cooking and heating, but the CIs included the null. No increased risks were found for gas cooking and heating. The results were not adjusted for smoking.

Occupational history was obtained by interviewing patients admitted to Roswell Park Memorial Institute in Buffalo, New York (Viadana et al. 1976). Nonneoplastic controls were selected from the same hospital and matched for age and smoking status. An association between the occupations of bus, taxicab, or truck driver and stomach cancer was reported (RR 1.6, p>0.05).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Conclusion

No consistent association between fuels or combustion products and stomach cancer was observed in the studies reviewed by the committee. Two of the studies of fuel exposure (Hanis et al. 1979; Ritz 1999) analyzed esophageal and stomach cancers together, so the committee cannot determine which cancer type may have been associated with exposure. For combustion-product exposure, two studies reported an increased risk of stomach cancer; however, the method used to assess exposure was limited and there was no adjustment for confounders.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and stomach cancer.

COLON CANCER

This review focuses on cancer of the colon (ICD-9 153). Risk factors for this cancer are family history, increasing age, ethnicity, dietary habits, weight and inactivity, and tobacco and alcohol use (ACS 2004p).

In 2000, there were 38.5 new cases of colon cancer per 100,000 people (43.5 among men and 34.8 among women) and 17.6 deaths per 100,000 (21.1 among men and 15.2 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.9 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of and association between exposure to fuels and colon cancer.

Cohort Studies

In a cohort of 3,814 uranium-processing workers, kerosene exposure was associated with a greater risk of death from large intestine cancer, but the CIs included the null (Ritz 1999). Kerosene exposure was assessed with a detailed industrial hygiene survey.

A cohort of 10,763 workers employed at an Amoco Corporation oil refinery in 1970–1980 was retrospectively studied (Nelson et al. 1985, 1987). Exposure was classified by an industrial hygienist on the basis of job type and “rough exposure categories”. Increased SMRs were found for jobs associated with exposure to light aromatic hydrocarbons for all digestive tract cancers (SMR 1.39). Occasional exposure to heavy oils was also associated with increased SMRs for digestive tract cancers (SMR 1.79). No specific digestive-cancer sites were included in the analysis. On the basis of data from the National Cancer Institute, the age-adjusted incidences of gastrointestinal cancers indicate that men are much more likely to be diagnosed with colorectal cancer than esophageal and stomach cancers (Ries et al. 2003). Therefore, it is likely that the increase in digestive tract cancers in the Amoco Corporation oil-refinery workers is due to colorectal cancer.

A cohort of workers at Imperial Oil Limited in Canada was evaluated for cancer outcomes, including colorectal cancer (Hanis et al. 1979; Lewis et al. 2000b; Schnatter et al. 1993). Moderate exposure (defined as less than daily contact with petroleum or its products) and daily exposure were not associated with a greater risk of cancers of the intestines and rectum (the

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

result of the analysis of these cancers was presented together) (Hanis et al. 1979). In a subcohort of 6,672 male marketing and distribution workers who probably were exposed to finished products (gasoline and diesel fuel), the risk of large intestinal cancer was increased (SMR 1.50, 95% CI 0.97–2.21) (Schnatter et al. 1993); employment for 30 years or more (with a 10-year latency) was associated with a greater risk of large intestine cancer than shorter employment (SMR 2.33, p<0.05). In another subcohort of 25,292 Imperial Oil Limited workers, an increased risk of large intestine cancer was found, but the CIs included the null (Lewis et al. 2003).

Case-Control Studies

A population-based case-control study conducted in Sweden assessed the potential for an association between occupational exposure and colon cancer (Gerhardsson de Verdier et al. 1992). Exposure to specific substances was assessed by self-report; although an association was noted with fuels or with work in automotive repair or gasoline stations, the CI included the null (OR 1.8, 95% CI 0.6–5.2).

In a case-control study in Montreal, exposures to multiple fuels were examined with industrial-hygienist review of occupational history, and exposure intensity and probability were estimated (Siemiatycki et al. 1987a). The ORs were not increased (as judged with 90% CIs) for automotive gasoline, aviation gasoline, kerosene, jet fuel, or diesel fuel.

In a case-control study of occupational exposure and colorectal cancer that used a subset of the National Cancer Survey database and the National Occupational Hazard Survey, the probability of exposure to specific substances was assigned with a JEM based on occupational history (Spiegelman and Wegman 1985). A medium-high cumulative-exposure probability score, which summed the products of probability of exposure in each job and duration of exposure, was associated with a greater risk of colorectal cancer (OR 1.53, p=0.01) and colon cancer alone (OR 1.61; p=0.02) among men exposed to fuel oil. Among women exposed to fuel oil, there was no clear increase in risk of colorectal cancer (OR 1.24, p=0.21).

Combustion Products

Table 4.10 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of and association between exposure to combustion products and colon cancer.

Case-Control Studies

(Goldberg et al. 2001) conducted a case-control study of 497 Canadian men with colon cancer and two sets of controls, one population-based and the other a cancer case-control group. Exposure was based on a review of occupational history and expert opinion. Increasing risks were related to a number of occupational exposures, among them diesel-engine emissions (adjusted OR 1.6, 95% CI 1.0–2.5 for those with “substantial” exposure, and adjusted OR 1.2, 95% CI 0.8–1.8 for those with “nonsubstantial” exposure). There was no evidence of increasing risk with duration of exposure.

In the study in Sweden, the potential association between occupational exposure to combustion products and colon cancer was assessed (Gerhardsson de Verdier et al. 1992). Small increases in risk of colon cancer were associated with exposure to combustion gases from coal, coke, wood, and soot and with railway work, but the 95% CIs included the null. Exposure to tar and asphalt did not increase the risk of colon cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

An occupational case-control study evaluated the relationship between exposure to several types of engine exhausts and combustion products, based on an industrial-hygiene assessment of occupational history, and colon cancer (Siemiatycki et al. 1988). Exposure to diesel exhaust led to an increased risk of colon cancer (OR 1.3, 90% CI 1.1–1.6). Exposure to jet-fuel exhaust and products of wood combustion also led to increases in colon-cancer risk, but the 90% CI included the null. No increased risk was reported in association with exposure to gasoline exhaust, propane exhaust, or products of combustion of propane, natural gas, liquid fuel, coal, or coke.

Conclusion

No consistent association was observed in the studies of fuels and colon cancer reviewed by the committee. Three of the studies analyzed colon cancer and rectal cancer together (Hanis et al. 1979; Nelson et al. 1987; Ritz 1999), so the committee could not determine whether exposure to fuels may have been associated with a specific type of cancer. Although the three studies of exposure to combustion products and colon cancer reported positive associations (Gerhardsson de Verdier et al. 1992; Goldberg et al. 2001; Siemiatycki et al. 1988), the committee believes that the evidence of an association is inadequate because of the small number of studies available.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and colon cancer.

RECTAL CANCER

This review focuses on cancers of the rectum, rectosigmoid junction, and anus (ICD-9 154). Risk factors for those cancers are family history, increasing age, ethnicity, dietary habits, weight and inactivity, and tobacco and alcohol use (ACS 2004p).

In 2000, there were 14.6 new cases of rectal cancer per 100,000 people in the US (19.0 among men and 11.1 among women), and 3.0 deaths per 100,000 (4.0 among men and 2.3 among women) (Ries et al. 2004).

Fuels

Table 4.11 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and rectal cancer.

Cohort Studies

As described in the colon-cancer section, a cohort of 10,763 workers at an Amoco Corporation oil refinery employed in 1970–1980 was retrospectively studied (Nelson et al. 1985, 1987). No specific digestive-tract cancer sites were included in analysis. Increased SMRs were found for jobs associated with light aromatic hydrocarbon exposure and occasional exposure to heavy oils. For the reasons described above, it is likely that most of the digestive-tract cancers were colorectal cancers.

In a cohort of workers at Imperial Oil Limited in Canada, increased risk of colorectal cancer was assessed (Hanis et al. 1979; Lewis et al. 2000b; Schnatter et al. 1993). Colon and

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

rectal cancers were analyzed together, and the results are summarized in the colon-cancer section above. No increased risk of colorectal cancer was found in the workers. In another followup study, no increased risk of cancer of the rectum and rectosigmoid junction was found in the cohort (Lewis et al. 2003). However, cancer of the rectum was increased in the marine-operating subgroup employed for 35 years or more (SMR 2.75, 95% CI 1.19–5.41) (Lewis et al. 2000b). Exposure to fuels in this subgroup most likely occurred during loading and unloading operations.

Case-Control Studies

A hospital-based case-control study conducted in Sweden assessed the potential association between occupational exposure and rectal cancer (Gerhardsson de Verdier et al. 1992). Exposure to specific substances was assessed by self-reporting, and no association was noted between fuels and rectal cancer. There was a suggestion of an increased risk in automotive-repair or gas-station workers (OR 1.5, 95% CI 0.4–5.6), but the effect estimates were imprecise.

In a case-control study in Montreal, exposure to multiple fuels was examined with an industrial-hygiene review of occupational history, and exposure intensity and probability were estimated (Siemiatycki et al. 1987a). An increased risk of rectal cancer was reported for exposure to automotive gasoline, aviation gasoline, jet fuel, diesel fuel, and heating oil, but the 90% CIs included the null. No increased risk of rectal cancer was found in association with kerosene.

In a case-control study of occupational exposure and colorectal cancer that used a subset of the National Cancer Survey database and the National Occupational Hazard Survey, a medium-high cumulative exposure probability score was associated with a greater risk of colorectal cancer (OR 1.53, p=0.01) among men exposed to fuel oil (Spiegelman and Wegman 1985). Among women exposed to fuel oil, there was no clear increase in risk of colorectal cancer (OR 1.24, p=0.21). There was no evidence of an exposure-response relationship.

Combustion Products

Table 4.12 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and rectal cancer.

Case-Control Studies

No evidence of increased rectal cancer risk posed by occupational exposure to soot, wood, and coal combustion products was reported in a population-based case-control study of 257 Canadians (Dumas et al. 2000).

In the study in Sweden, the potential association between occupational exposure to combustion products and rectal cancer was assessed (Gerhardsson de Verdier et al. 1992). Exposure was self-reported. Associations were found between rectal cancer and exposure to combustion gases from coal, coke, and wood (OR 2.1, 95% CI 1.0–4.6), soot (OR 2.7, 95% CI 1.2–5.7), and tar and asphalt (OR 1.0, 95% CI 0.3–2.8).

A case-control study conducted in Montreal did not report an increased risk of rectal or rectosigmoid cancer in people exposed to diesel exhaust, jet-fuel exhaust, propane exhaust, or products of combustion from natural gas, liquid fuel, coal, or coke (Siemiatycki et al. 1988). A JEM classification of exposure was used. There was a slight increase in rectal cancer in people

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

exposed to gasoline exhaust (OR 1.2, 90% CI 1.0–1.5), particularly in a subgroup exposed at high concentrations for more than 10 years (OR 1.6, 90% CI 1.1–2.3).

Conclusion

Although some studies reported positive associations between fuels or combustion products and rectal cancer, the results were not consistent, and the number of studies was small. The positive studies failed to include at least one high-quality study supported by an adequate exposure assessment.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and rectal cancer.

LIVER CANCER

This review focuses on hepatic cancer (more commonly referred to as liver cancer) (ICD-9 155) (ACS 2003f). Known risk factors include infection with hepatitis B and hepatitis C viruses; cirrhosis caused by alcohol abuse, hepatitis B and hepatitis C, and excess iron in the liver (from hemochromatosis); aflatoxin; vinyl chloride; thorium dioxide; tobacco use; anabolic steroids; and arsenic.

In 2000, there were 5.3 new cases of liver cancer per 100,000 people (8.1 among men and 3.0 among women) and 4.7 deaths per 100,000 (6.8 among men and 2.9 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.13 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and hepatic cancer.

Case-Control Studies

A case-control study reported on cases of primary liver cancer that occurred in New Jersey in 1975–1980 (Stemhagen et al. 1983). A questionnaire was used to determine occupational history and other factors, such as lifetime residence, medical history, smoking habits, and alcohol consumption. Of many occupations and industries that were examined in the study, the authors report results only for agricultural occupations (the primary purpose of the study) and for occupations and industries for which the relative risks were greater than 2.0 or for which the relative risks were greater than 1.0 and the CIs did not include the null. A total of 11 occupation and industries were in the latter category, including one that is possibly relevant to the Gulf War fuel exposures—gasoline service stations (RR 2.88, 95% CI 1.20–6.88).

Combustion Products

Table 4.14 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and hepatic cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Case-Control Study

The study by Stemhagen et al. included two occupations that are possibly relevant to the Gulf War combustion exposures—road-building (RR 2.60, 95% CI 0.83–8.19) and bus lines (RR 2.80, 95% CI 0.93–8.40) (Stemhagen et al. 1983). Employment at gasoline service stations (RR 2.88, 95% CI 1.20–6.88) probably involved exposure to a mixture of fuel and combustion products.

Conclusion

Only one relevant study that evaluated exposure to fuels or combustion products and hepatic cancer was identified (Stemhagen et al. 1983). Although associations were noted for some occupations, there were few cases with relevant exposure, and the study did not consider all pertinent risk factors.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and hepatic cancer.

PANCREATIC CANCER

This review focuses on pancreatic cancer (ICD-9 157), for which tobacco use stands out as the most import risk factor among increasing age, ethnicity, sex, dietary habits, alcohol use, family history, and some occupational exposures (ACS 2004m, 2004p, 2004q). In addition, diabetes and pancreatitis are risk factors for pancreatic cancer.

In 2000, there were 10.9 new cases of pancreatic cancer per 100,000 people (12.8 among men and 9.4 among women) and 10.6 deaths per 100,000 (12.2 among men and 9.3 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.15 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and pancreatic cancer.

Cohort Studies

In a cohort of 3,814 uranium-processing workers, the relationship between kerosene exposure and pancreatic cancer was imprecise (Ritz 1999). For example, moderate exposure to kerosene for more than 5 years with a 15-year lag yielded an RR of 2.78 (95% CI 0.51–15.2). Kerosene exposure was based on a detailed industrial-hygiene assessment.

No increased risk of pancreatic cancer was found in a cohort of petroleum-refinery workers at Imperial Oil Limited in Canada (Lewis et al. 2003, 2000b; Schnatter et al. 1993), but only a small number of cases were noted. No specific industrial-hygiene assessment was available.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Combustion Products

Table 4.16 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and pancreatic cancer.

Case-Control Studies

Alguacil et al. (2000) conducted a case-control study in Spain to assess the relationship between occupational exposure and pancreatic cancer. A JEM was used to determine exposures to chemicals on the basis of jobs reported in the occupational history. There were few cases of exposure to combustion products, and the relationships reported were imprecise.

A population-based case-control study in Finland examined the relationship between a number of occupational exposures and pancreatic cancer (Kauppinen et al. 1995). Exposure analysis included a JEM. No association was found with exposure to engine exhaust (OR 0.89, 95% CI 0.51–1.53). An increased risk of pancreatic cancer was reported in association with exposure to PAHs, but the CI included the null (OR 1.33, 95% CI 0.69–2.57).

The association between occupational exposure and pancreatic cancers was evaluated in a case-control study in France (Pietri et al. 1990). Overall, transportation workers did not show an increased risk of pancreatic cancer (RR 0.87, 95% CI 0.42–1.77, on the basis of 13 exposed cases).

A case-control study in Montreal reported that exposure to coal combustion products was associated with an increased risk of pancreatic cancer (OR 2.3, 90% CI 1.4–4.0) (Siemiatycki et al. 1988). A JEM classification of exposure was used. The same study did not find increased risk of pancreatic cancer associated with exposures to gasoline exhaust, diesel exhaust, jet fuel exhaust, propane exhaust, or products of combustion of propane, natural gas, liquid fuel, wood, or coke.

Conclusion

Information on the risk of pancreatic cancer posed by fuel exposure is limited. One study reported an association between kerosene exposure and pancreatic cancer (Ritz 1999), but the results were imprecise. One of the four reviewed studies of combustion-product exposure and pancreatic cancer reported a positive finding (Siemiatycki et al. 1988). It found an association between exposure to coal combustion products and increased risk of pancreatic cancer, but it did not find a link between nine other types of combustion products and pancreatic cancer.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and pancreatic cancer.

LARYNGEAL CANCER

This section addresses cancer of the larynx (ICD 161). The most important risk factor for this cancer is tobacco use, particularly cigarette-smoking (ACS 2003a). Additional risk factors include alcohol consumption (which has a synergistic effect with tobacco), dietary habits, vitamin deficiency, exposure to ultraviolet (UV) radiation (sunlight), increasing age, a weak

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

immune system, genetic factors, and some occupational exposures, such as to dusts from wood, textiles, and leather; glues; formaldehyde; solvents used in furniture and shoe production; mustard gas; isopropyl alcohol; and radium.

In 2000, there were 4.0 new cases of laryngeal cancer per 100,000 people in the US (7.2 among men and 1.4 among women), and 1.4 deaths per 100,000 (2.6 among men and 0.5 among women) (Ries et al. 2004).

Fuels

Table 4.17 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and laryngeal cancer.

Cohort Study

Lagorio et al. (1994) tracked the mortality experience of 2,308 men who had been managers of Italian service stations in 1980–1992. Laryngeal cancer was not associated with employment at the service stations (on the basis of three cases). Managers at all stations had an OR of 1.05 (90% CI 0.29–2.72); the OR for managers at small stations was 1.53 (90% CI 0.42–3.96).

Case-Control Studies

De Stefani et al. (1998) conducted a case-control study in Montevideo, Uruguay in 1993–1995. They enrolled 112 incident, histologically confirmed laryngeal-cancer cases in males and 509 controls (from among all other cancer cases, excluding those in sites proximal to the larynx or related to tobacco or alcohol consumption). Exposure was based on self-reported occupational history and exposure, and risks were adjusted for age, smoking, alcohol consumption, residence, education, and income. Cases with self-reported exposure to gasoline who worked as gasoline fillers had increased RRs, but the CIs included the null. There is a suggestion of a dose-response relationship with increased years of exposure to gasoline.

Wortley et al. (1992) conducted a population-based case-control study in western Washington from September 1983 to February 1987 and enrolled 235 incident cases of laryngeal cancer identified from a population-based Surveillance, Epidemiology, and End Results (SEER) registry. They matched 547 population controls to cases by sex and age with random-digit dialing. Subjects were interviewed with a questionnaire that included items on job titles, description of tasks, and nature of industry for each job held 6 months or longer. Job title and industry were coded according to 1980 US census codes. Some 505 individual occupations were collapsed into 62 categories. Job titles were categorized according to duration of exposure: less than 10 years vs 10 years or longer. Furthermore, industrial hygienists created a JEM for some chemical categories, such as diesel fumes, but not for combustion-product exposure. Risks were not found to be increased in vehicle mechanics (OR 1.2, 95% CI 0.6–2.1) or for garage- and gas-station-related work (OR 0.8, 95% CI 0.4–1.8). All results were adjusted for age, education, smoking, and alcohol.

Ahrens et al. (1991) conducted a case-control study to investigate the relationship between occupational factors and laryngeal cancer. A hospital in Bremen, Germany, was the source of 55 men newly diagnosed with histologically confirmed primary laryngeal cancer and 30 more who had been diagnosed during the previous 2 years. Each case was matched to a male patient from the same hospital without a history of cancer or other smoking-related diseases by

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

age. Standardized interviews were conducted and included occupational history with an exposure checklist and questions on smoking and drinking behavior. Broad industrial and occupational categories related to transportation showed moderately increased risks. Increased ORs were associated with self-reported exposure to diesel oil (OR 1.7, 95% CI 0.8–3.5) and gasoline (OR 2.8, 95% CI 1.0–7.7). Self-reported exposure to gasoline was most frequent in mechanics and drivers; diesel-oil exposure was most prevalent among those employed in shipping and in drivers and train operators.

Brown et al. (1988) enrolled 183 histologically confirmed incident cases of squamous-cell carcinoma of the larynx from 56 Texas hospitals in 1975–1980 and obtained 250 population controls (ascertained from death records, driver’s license files, and Health Care Financing Administration [HCFA] records) that were matched to cases on age, vital status, ethnicity, and county of residence. Subjects working in petroleum refining or chemical manufacturing had an adjusted RR of 0.93 (95% CI 0.59–1.46). For subjects with self-reported exposure to diesel and gasoline fumes, they calculated an adjusted RR of 1.50 (95% CI 1.00–2.26). However, it is not apparent from the description in the article whether diesel and gasoline fumes were the uncombusted and combusted form of those agents.

Combustion Products

Table 4.18 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and laryngeal cancer.

Case-Control Studies

A large IARC study assessed 1,010 male cases of laryngeal or hypopharyngeal cancer and 2,176 male population controls identified and interviewed at six European centers in 1980–1983 (for example, Berrino et al. 2003; Boffetta et al. 2003). Exposure in a specific industry or job was defined as working for at least 1 year since 1945 (Boffetta et al. 2003). Results were controlled for study area, age, smoking, and alcohol consumption. Employment in the railway-transport industry was associated with an increased risk of laryngeal cancer overall (OR 1.52, 95% CI 0.97–2.39) and showed a trend with duration of employment (p=0.02). Other transportation-related occupations with the potential for exhaust exposure (motor-vehicle mechanics, other mechanics, railway vehicle loaders, lorry drivers, local and long-distance lorry drivers, and other motor-vehicle drivers) were also associated with increased risks, but the CIs were wide and included the null value. PAHs were among 16 industrial agents for which duration, intensity, and likelihood of exposure were estimated with a JEM (Berrino et al. 2003). For the 695 cases 55 years old or older, the adjusted risk associated with PAH exposure was essentially unity (95% CI 0.7–1.3); for the younger 315 cases, the adjusted OR for PAH exposure for those who had been exposed 20 years or more was 1.1 (95% CI 0.5–2.4).

In a hospital-based case-control study of incident laryngeal cancer in Istanbul, Turkey, 940 cases were identified in 1979–1984, and 1,519 other patients with diagnoses not thought to share risk factors with laryngeal cancer were selected as controls (Elci et al. 2003). Cases and controls were interviewed on admission to the hospital with a questionnaire that included occupational history. After adjustment for age, smoking, and alcohol consumption, Elci et al. (2001) found that drivers had increased risks of laryngeal cancer (OR 1.7, 95% CI 1.1–2.4), but mechanics did not (OR 0.8, 95% CI 0.5–1.3). Elci et al. (2003) described the results of applying a

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

JEM to work histories to estimate specific exposures. The adjusted risks of laryngeal cancer were positively associated with ever having been exposed to diesel exhaust (OR 1.5, 95% CI 1.3–1.9), to gasoline exhaust (OR 1.6, 95% CI 1.3–2.0), or to PAHs (OR 1.3, 95% CI 1.1–1.6), and these associations were similar and generally positive for all the specific sites (supraglottic, glottic, and others). The risk of laryngeal cancer was similar for all three intensity and probability levels of exposure.

In the previously described case-control study of laryngeal cancer in Montevideo, Uruguay (De Stefani et al. 1998) truck drivers had an increased risk of cancer in the glottic area (OR 2.7, 95% CI 0.7–10.7) but not the supraglottic area (OR 0.6, 95% CI 0.1–2.9). When stratified by subsite, self-reported exposure to diesel or gasoline exhaust was positively associated with glottic cancer (diesel exhaust OR 1.9, 95% CI 0.6–5.8; gasoline exhaust OR 1.8, 95% CI 0.6–5.7) but not supraglottic cancer (diesel exhaust OR 0.7, 95% CI 0.2–1.9; gasoline exhaust OR 0.8, 95% CI 0.3–2.1). The authors speculated that the glottic area is potentially more affected by smaller particles, such as those found in exhaust, than the supraglottic area.

Gustavsson et al. (1998) identified incident cancers of the oral cavity (ICD-9 141, 143–145) or the oropharynx or hypopharynx (ICD-9 146, 148) diagnosed in Sweden in 1988–1991. In models adjusted for region, age, alcohol consumption, and smoking, high PAH exposures (categorized on the basis of an industrial-hygiene assessment of the participant’s occupational history) were associated with laryngeal cancer (RR 1.47, 95% CI 0.96–2.24).

Pintos et al. (1998) conducted a case-control study in Sao Paulo, Curitiba, and Loiania, Brazil, that examined the risk of laryngeal cancer in relation to the use of stoves. Reported use of wood stoves was associated with an increased risk of laryngeal cancer (OR adjusted for tobacco and alcohol consumption 2.34, 95% CI 1.17–4.67). In men, the OR was 2.03 (95% CI 1.12–3.67); in women, 16.24 (95% CI 2.66–99.1). Relatively few women were included in the study.

Goldberg et al. (1997) conducted a hospital-based case-control study in France that assessed the risk of laryngeal and hypopharyngeal cancer associated with occupation. The study included 528 male cases and 305 male controls with various other types of cancer. Interviews were conducted to obtain information on occupational history, demographic characteristics, alcohol consumption, and tobacco use. After adjustment only for age, alcohol use, and smoking, an increase in laryngeal-cancer and hypopharyngeal-cancer risk was reported for transportation-equipment operators (OR 1.5, 95% CI 1.0–2.5). When education was included in the adjustments, there was still an increase in risk, but the CI included the null (OR 1.4, 95% CI 0.9–2.3). With full adjustment, small increases in laryngeal-cancer and hypopharyngeal-cancer risk were reported for motor-vehicle mechanics (OR 1.2, 95% CI 0.5–2.5) and for workers employed in railway transportation (OR 1.4, 95% CI 0.6–3.1) and road transportation (OR 1.0, 95% CI 0.4–2.1), but all the CIs included the null.

Pollan and Lopez-Abente (1995) identified 50 incident laryngeal-cancer cases in Madrid in 1982–1985 and recruited 46 population controls matched on age, sex, and residential area and 45 hospital controls (excluding patients with alcohol- or tobacco-related diseases) matched on age, sex, and admission date. Exposure was based on self-reported job history and occupational codes; the researchers collected detailed data on year of hire and termination, occupational activity, and unit. Employment as a transport driver was associated with an increased risk of laryngeal cancer, but this result was based on only eight cases (OR adjusted for age, tobacco use, and alcohol use 2.71, 95% CI 0.85–8.64) and thus imprecise.

Muscat and Wynder (1995) reported on 235 white men with primary laryngeal cancer age-matched to 205 hospital controls with a variety of non-tobacco-related diseases. Although

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

there was an association between laryngeal cancer and self-reported exposure to diesel exhaust, an independent review and coding of occupations and industries potentially exposed to diesel exhaust found no association with diesel-exhaust-exposed jobs.

Researchers at the University of Heidelberg, Germany, conducted a hospital-based case-control study with 164 male incident laryngeal-cancer cases ascertained in February–June 1988 and November 1988–May 1989 (Dietz et al. 1995; Maier and Tisch 1997). They recruited 656 male controls from the same outpatient clinic (excluding those with evidence of cancer) matched to cases on age and residential area. Exposure was based on occupation and exposure to environmental factors such as heating or cooking with fossil-fuel stoves. After adjustment for alcohol consumption and tobacco use, heating and cooking with fossil-fuel stoves were associated with laryngeal cancer (OR for heating with fossil-fuel stoves over 40 years 2.11, 95% CI 1.43–3.12; OR for cooking with fossil fuel stoves over 20 years 1.47, 95% CI 0.92–2.33; and OR for heating with coal, briquettes, or coke 1.52, 95% CI 0.94–2.47). ORs and 95% CIs were calculated by the committee with standard methods from the observed numbers presented in the original paper.

Wortley et al. (1992) conducted a population-based case-control study in western Washington. On the basis of on occupational titles, some jobs with the potential for diesel-exhaust exposure were associated with a small increased risk of laryngeal cancer but all estimates were imprecise: motor-vehicle operator (OR ever employed 1.3, 95% CI 0.8–2.1; OR less than 10 years 1.6; OR at least 10 years 0.8); work in motor-vehicle transportation (OR ever employed 1.3, 95% CI 0.6–2.8; OR less than 10 years 1.1; OR at least 10 years 0.8); cook (OR 1.3, 95% CI 0.4–4.1); and firefighter (OR 0.5, 95% CI 0.0–2.9). All those results were adjusted for age, education, smoking, and alcohol.

In the previously described case-control study investigating the relationship between laryngeal cancer and occupational factors conducted by Ahrens et al. (1991) in Bremen, Germany, a modest increase in risk was observed in subjects employed in the transport and communication industry (OR 1.3, 95% CI 0.64–2.59) and subjects employed as transportation and store workers (OR 1.83, 95% CI 0.94–3.56). Self-reported exposure to fumes and smoke was not associated with an increase in risk (OR 0.7, 95% CI 0.3–1.4).

In the case-control study of incident squamous cell carcinoma conducted by Brown et al. (1988) in Texas, self-reporting ever having been employed in transportation or as a driver was associated with an increase in risk of laryngeal cancer, but the CIs were wide and included the null (OR transportation 1.42, 95% CI 0.86–2.36; OR driver 1.69, 95% CI 0.75–3.83).

An early case-control study by Decoufle and Stanislawczyk (1977) was the result of abstracting the records of about 14,000 white cancer patients at the Roswell Park Memorial Institute in Buffalo, New York. Occupational histories routinely constituted a section of the medical charts. Patients with laryngeal cancer were compared with noncancer controls with respect to jobs ever held or held for 5 years or more. Analyses were stratified by age at diagnosis (with the cutpoint at 60 years), and the results for this type of cancer were adjusted for smoking. Of several exhaust-associated occupations, drivers most consistently showed increased RRs of about 1.5, adjusted for smoking, but the CIs were wide.

Conclusion

Overall, the results regarding exposure to fuels and laryngeal cancer are inconsistent. Two studies reviewed by the committee reported a modest increase in the risk of laryngeal cancer associated with exposure to fuels; however, the exposures in both studies were self-

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

reported (Ahrens et al. 1991; Brown et al. 1988). Wortley et al. (1992) used a JEM and reported an increased risk of laryngeal cancer in vehicle mechanics that was imprecise but no increase in garage and gasoline-station workers.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels and laryngeal cancer.

A number of studies evaluated the potential link between exposure to combustion products and laryngeal cancer. Several studies reported positive findings, including Maier and Tisch (1997) and Dietz et al. (1995) regarding exposure to the emissions of fossil-fuel stoves and Pintos et al. (1998) regarding exposure to wood-stove emissions. There was supportive evidence from Gustavsson et al. (1998) and Elci et al. (2003). Several studies reported small increases in laryngeal-cancer risk for some exposures: Boffetta et al. (2003), De Stefani et al. (1998), Goldberg et al. (1997), Muscat and Wynder (1995), Pollan and Lopez-Abente (1995), Wortley et al. (1992), Ahrens et al. (1991) and Brown et al. (1988); however, the overall results are inconsistent.

The committee concludes, from its assessment of the epidemiologic literature, that there is limited/suggestive evidence for an association between combustion products and laryngeal cancer.

LUNG CANCER

Lung cancer (ICD-9 162) is the leading cause of cancer death of both men and women (ACS 2004s); its incidence is higher in men than in women. The major risk factor for lung cancer is tobacco-smoking. Smoking is believed to be responsible for about 80% of lung-cancer cases. Other risk factors are exposure to environmental tobacco smoke, ionizing radiation (including radon gas), arsenic, PAHs (particularly benzo[a]pyrene), asbestos, chromium, and silica; tuberculosis and some forms of pneumonia that leave scars in the lungs; family history; and some aspects of diet.

Lung cancer is classified into two main types based on the appearance of the cells (ACS 2004n). Small cell lung cancer (SCLC), also called oat cell cancer, accounts for 20% of all cancers and is almost exclusively associated with smoking. It originates primarily in the bronchi or central portion of the lungs, though it can spread rapidly throughout the body. The second type of lung cancer, non-small cell lung cancer, makes up the remaining 80% and is divided into 3 subtypes: squamous cell carcinoma, 25–30% of all lung cancers; adenocarcinoma, 40%; and large-cell undifferentiated carcinoma, 10–15%. Squamous cell carcinoma is also associated with smoking and is found centrally within the lungs. Adenocarcinoma appears in the outer regions of the lungs, and large-cell undifferentiated carcinoma can be found anywhere and can metastasize quickly.

In 2000, there were 62.3 new cases of lung cancer per 100,000 people (79.8 among men and 49.8 among women) and 56.1 deaths per 100,000 (76.9 among men and 41.2 among women), in the United States (Ries et al. 2004).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Fuels

Table 4.19 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and lung cancer.

In a hospital-based case-control study conducted in Montreal in which exposures to multiple fuels were examined with an industrial-hygiene review of occupational history, exposure to kerosene was associated with an increased risk of squamous-cell lung cancer (OR 1.4, 90% CI 1.0–1.9) and adenocarcinoma of the lung (OR 1.5, 90% CI 1.0–2.3) (Siemiatycki et al. 1987a); exposure to heating oil, an increased risk of oat-cell lung cancer (OR 1.7, 90% CI 1.0–2.7); exposure to crude oil, an increased risk of squamous-cell lung cancer (OR 2.8, 90% CI 1.0–7.6); and exposure to diesel fuel, an increased risk of nonadenocarcinoma lung cancer (OR 1.6, 90% CI 1.1–2.4) with indications of an exposure-response relationship. In multivariate analyses, those risks were also increased.

Mortality was assessed in workers at an Imperial Oil Limited refinery in Canada (Hanis et al. 1979; Lewis et al. 2000b; Schnatter et al. 1993). In followup through 1973 of a subcohort with daily exposure to petroleum products, there was an increase in lung cancer (RR 1.89) and a greater increase with increasing years of employment (Hanis et al. 1979). However, there was a greater rate of lung cancer in office workers with no occupational exposure to petroleum products. The study did not control for smoking, and it is possible that the office workers smoked at a higher rate than the refinery workers.

A retrospective cohort study assessed mortality in 7,119 workers at a petroleum refinery in Beaumont, Texas, who worked at least 1 year in 1945–1978 (Raabe et al. 1998). The refinery produced a variety of fuels, including gasoline and jet fuel, and feedstocks for the petrochemical industry. The study did not control for smoking. No association between refinery employment and lung cancer was observed except for an increased SMR in maintenance-craft workers (SMR 1.20, 95% CI 0.99–1.45). A followup of the Beaumont cohort that extended enrollment and vital-status followup to December 31, 1996, found no association between refinery employment and lung cancer (Wong et al. 2001). A nested case-control study was conducted to assess the relationship between lung cancer and employment in the Beaumont, Texas refinery (Rosamilia et al. 1999). There were 112 cases of lung-cancer death that were matched to controls (n=490) by birth date and race. The analysis compared each job category (an indication of the likelihood of fuel exposure) and included some individual information on smoking. No increased risk of lung cancer was found in connection with any specific job category.

Combustion Products

Table 4.20 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and lung cancer. Figures 4.1, 4.2, and 4.3 present results of some of the studies on exhaust-exposed workers, indoor air pollution, and ambient air pollution, respectively. Results were sorted in descending order of relative risk, except for studies in which there are multiple results, which are contiguous entries.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

FIGURE 4.1 Lung cancer and occupations with exposure to combustion products

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

FIGURE 4.2 Lung cancer and indoor air pollution from combustion of fuels

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

FIGURE 4.3 Lung cancer and ambient air pollution from combustion of fuels

Ambient Air-Pollution Studies

Five cohort studies have investigated the association between ambient air pollution and lung cancer: the Harvard Six Cities study (Krewski et al. 2000), the American Cancer Society (ACS) CPS-II Study (Pope et al. 1995, 2002), the California Seventh-Day Adventist Study (Abbey et al. 1999; Beeson et al. 1998), a Dutch study of diet and health (Hoek et al. 2002), and a cardiovascular cohort study in Norway (Nafstad et al. 2003). All except the Norwegian study used fixed-site air-pollution monitors to estimate exposure to ambient air pollution. In the Harvard Six Cities and ACS studies, subjects were assigned at the time of entry into the studies

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

average annual exposure in their city of residence. In the Seventh-Day Adventist study, geostatistical interpolation of annual fixed-site monitor data was assigned to the address of each subject. In the Dutch study, a complicated exposure-assessment algorithm based on regional and local pollution and accounting for land use was used to estimate annual exposure of each subject. In the Norwegian study, dispersion models coupled with emission data were used to estimate annual sulfur dioxide (SO2) and NOx at each subject’s residence.

Table 4.20 shows that in the Harvard Six Cities study and in the Dutch study, RRs of lung cancer were increased with exposure to fine particulate matter (PM) although the CIs included the null. In the Harvard Six Cities study, the RR associated with an increase in PM2.5 of 18.5 µg/m3 was 1.17 (95% CI 0.67–2.04). In the Dutch study, the RRs associated with an increase in air pollution from the 5th to the 95th percentile were 1.1 (95% CI 0.4–2.6) for black smoke (a measure of PM2.5 and elemental carbon) and 1.3 (95% CI 0.4–3.7) for nitrogen dioxide (NO2). In the ACS study, the RR associated with a 10 µg/m3 increase in fine PM was 1.14 (95% CI 1.04–1.23). In the Seventh-Day Adventist Study, RRs increased with increasing exposure: with an increase in PM10 of 24.8 µg/m3, men had an RR of 3.36 (95% CI 1.57–7.19), and women an RR of 1.33 (95% CI 0.60–2.96); with an increase in SO2 of 3.72 ppb, men had an RR of 1.99 (95% CI 1.24–3.20), and women an RR of 3.01 (95% CI 1.88–4.84); and with an increase in NO2 of 19.78 ppb, men had an RR of 1.82 (95% CI 0.93–3.57), and women an RR of 2.81 (95% CI 1.15–6.89). Finally, an association with NOx (a marker of traffic-related combustion products) was found in the Norwegian study: for an increase in NOx of 10 µg/m3, the adjusted hazard ratio was 1.08 (95% CI 1.02–1.15). No associations were found with exposure to SO2.

A number of case-control studies of the association between lung cancer and air pollution have been published. After adjusting for age, smoking, and occupation, Vena (1982) found a 26% increase in risk in persons living for at least 50 years in areas of Erie County, New York, that had high concentrations of total suspended particles (TSP). In a mortality case-control study in Krakow, Poland, an association was found among men living in areas with high concentrations of SO2 and TSP (OR 1.46, 95% CI 1.06–1.99), but no excess risks were found among women (OR for medium or high concentrations 1.17) (Jedrychowski et al. 1990). A study in Trieste, Italy, found lung cancer to be associated with high estimated concentrations of deposited of particles (OR for over 0.298 g/m2 per day compared with less than 0.175 g/m2 per day 1.4, 95% CI 1.1–1.8) (Barbone et al. 1995; Biggeri et al. 1996). Lung-cancer risk was also associated with living near an incinerator in the city (OR 2.6, 95% CI 1.3–5.1). (Alarie et al. 1972; Xu et al. 1996a) found that in northeastern China ORs in men increased with perceived smokiness in the outdoor environment (OR for “somewhat smoky” 1.5, 95% CI 1.2–2.0; OR for “smoky” 2.3, 95% CI 1.7–2.9). Similar results were found in women. In a small study among women living in Athens, Katsouyanni et al. (1991) investigated the association between lung cancer and ambient concentrations of soot. They did not find an association among nonsmoking women (smoking-adjusted, comparing highest quartile with lowest quartile OR 0.7), but found a strong association among women who smoked for long durations (30 years of smoking, comparing highest quartile with lowest quartile OR of 2.23). Nyberg et al. (2000) conducted a case-control study of lung cancer and traffic-related pollution in Stockholm, Sweden, and estimated traffic-related concentrations of NO2 and heating-related concentrations of SO2, using patterns of traffic density and dispersion models. After adjusting for smoking, occupation, exposure to radon, and other risk factors, they found little evidence of an association when they used exposures to NO2 averaged over a 30-year period. A trend in risk was observed with a 20-

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

year lagged metric for exposure (NO2 over 29 µg/m3 OR 1.4, 95% CI 1.1–2.0; SO2 over 129 µg/m3 OR 1.2, 95% CI 0.9–1.7). There was no evidence of an interaction with smoking status.

Some studies were methodologically less rigorous. Yang (1999) conducted a death-certificate case-control study in Taiwan and found an association between lung cancer and a petrochemical air-pollution index defined by the proportion of employees working in the petrochemical industry in each municipality. In a study of occupational and environmental exposures, Jockel et al. (1992) did not find an association with air pollution. Theirs was a small study, and the exposure assessments were not well described; no associations were found with estimated concentrations of SO2. In an ecologic study in Rome, Italy, Michelozzi et al. (1998) did not find an association with distance from incinerators and oil refineries.

Occupational Exposure to Engine Exhaust

This review includes findings from occupational studies in which subjects were considered to be probably exposed to exhaust emissions from internal-combustion engines. The studies have been organized according to whether job titles were used to infer such exposures (that is, no specific estimates of exposure to exhaust-related pollutants were provided) or exposure to exhausts, diesel fumes, or PAHs was estimated with generally accepted principles of occupational epidemiology (that is, use of JEMs, detailed assessment of exposure by experts, and actual measurement of exposures). For studies that used job titles, the assumption regarding exposure was based on general industrial-hygiene information about job types, not necessarily on details of subjects’ occupations. Consequently, one would expect misclassification of exposure to weaken RR estimates.

General Exposure to Exhaust

Three incidence and two mortality cohort studies of bus drivers yielded no strong indication of an association of exposure to exhaust with lung cancer (Alfredsson et al. 1993; Netterstrom 1988; Paradis et al. 1989; Soll-Johanning et al. 2003). An increased relative risk was found in a study of commercial drivers (truck, taxi, and bus drivers) in Geneva (RR 1.61, 95% CI 1.29–1.98) (Guberan et al. 1992), a study of truck drivers in Denmark (RR 1.60, 95% CI 1.28–1.98) (Hansen 1993), a study of truck drivers in Sweden (RR 1.29, 95% CI 0.99–1.65) (Jarvholm and Silverman 2003), and a study of truck drivers in Iceland (RR 2.14, 95% CI 1.37–3.18) (Rafnsson and Gunaarsdottir 1991). There was little evidence that risk increased with duration of employment in the few studies in which the risk was reported.

Almost all the case-control studies showed an increased risk of lung cancer among professional drivers (Bruske-Hohlfeld et al. 1999, 2000; Buiatti et al. 1985; Hansen et al. 1998; Hayes et al. 1989; Menvielle et al. 2003; Muscat et al. 1998; Steenland et al. 1990; Swanson et al. 1993; Zahm et al. 1989). RRs were found to increase with duration of employment in a Danish study (Hansen et al. 1998); a US study of teamster-union drivers (especially diesel- and gasoline-truck drivers) (Steenland et al. 1990); a study in Detroit, Michigan (Swanson et al. 1993); a study in New Caledonia (Menvielle et al. 2003); and a pooled analysis of three American case-control studies (Hayes et al. 1989). No clear-cut trends for professional drivers were found in a German study (Bruske-Hohlfeld et al. 1999, 2000) and in a Swedish study (Damber and Larsson 1985, 1987). Other professions that probably entail exposure to engine exhaust were found to be associated with increased risks of lung cancer, including dock work

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

(Jockel et al. 1998; Menvielle et al. 2003), transportation-equipment management (Menvielle et al. 2003), and shipyard work (Bovenzi et al. 1993; Hayes et al. 1989; Zahm et al. 1989).

Estimated Exposure to Compounds in Exhaust Fumes

The Dutch dietary cohort study (van Loon et al. 1997), from which exposure to ambient air pollution was estimated, was also used to estimate the incidence of occupational exposure to PAHs. Adjustment only for age showed that the RR of lung cancer increased by tertile of exposure. However, adjustment for age, smoking, and other occupational exposures led to risks that declined with increasing exposure. It is possible that the latter analysis, in which other occupational agents were included, was overmatched. An analysis adjusting for all potential confounding factors, excluding the occupational ones, was not conducted. A large census cohort study in Sweden that used a JEM to assign probability and intensity of exposure to diesel emissions on the basis of job titles collected in the 1960 and 1970 censuses showed that risks of lung cancer increased with the magnitude and probability of exposure among men but not among women (Boffetta et al. 2001). A study from the Kaiser Permanente Medical Care Program in California showed a 61% excess of lung cancer among bus-garage mechanics and showed that RRs increased by an index of cumulative diesel exhaust (Van Den Eeden and Friedman 1993).

Male firefighters in Philadelphia did not appear to be at increased risk for death from lung cancer when a surrogate index of exposure to diesel exhaust was used, although the power of the study was likely to be low because of small numbers of cases and a surrogate index of exposure (Baris et al. 2001). In a cohort of US railroad workers, younger diesel-exposed workers were found to be at higher risk, presumably because the younger workers had the greatest duration of exposure (Garshick et al. 1987, 1988; Larkin et al. 2000). Potash miners in Germany were at higher RR for death, but the 95% CIs included the null (Saverin et al. 1999). Risk was found to increase with cumulative exposure to diesel among a cohort of bus garage mechanics in Stockholm (Gustavsson et al. 1990). In the ACS CPS II study, it was reported that risk increased with duration of self-reported exposure to diesel exhaust (Boffetta et al. 1988); a limitation of this analysis is the use of self-reported exposure.

Risk estimates associated with exposure to specific components of exhaust have come out of a number of studies. Increased lung-cancer risk with cumulative exposure to diesel exhaust and to PAHs from occupational sources was found in a previously cited case-control study in Germany (Bruske-Hohlfeld et al. 2000) and in a study in Stockholm (Gustavsson et al. 2000). However, exposure to diesel exhaust from occupational sources was not found to increase the risk of lung cancer in a US hospital-based case-control study (Boffetta et al. 1990) and exposure to PAHs was not found to increase the risk in occupational studies in Buenos Aires (Matos et al. 2000), Montreal (Nadon et al. 1995), and Sweden (Emmelin et al. 1993).

Many studies have assessed the association of indoor air quality related to types of heating or cooking fuels, such as coal, with lung cancer. Studies of exposure to heating or cooking fuels included a number from China (Alarie et al. 1972; Dai et al. 1996; Du et al. 1996; He et al. 1991; Huang et al. 1992; Kleinerman et al. 2002; Lan et al. 1993; Lei et al. 1996; Liu et al. 1993; Metayer et al. 2002; Shen et al. 1996; Wu-Williams et al. 1990a, 1993; Xu et al. 1989, 1996a; Zhong et al. 1999), Hong Kong (Koo et al. 1983) and Taiwan (Chen et al. 1990; Ko et al. 1997), many of which showed such associations. Only one cohort study investigated the effects of indoor air pollution on the risk of developing lung cancer (Lan et al. 2002). That population-based interview study included more than 31,000 farmers living in Xuanwei, China, who were followed from 1976 to 1992. The analysis was based on subjects whose parents used unvented firepits and smoky coal throughout their lives. It showed that changing to stoves that had

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

chimneys decreased the risk of lung cancer (RR<0.6) and that the reduction occurred more than 10 years after the change.

Gulf War Veteran Study

A study assessing respiratory cancer in Gulf War Veterans was identified and reviewed by the committee (Smith et al. 2002). The study population consisted of 405,142 regular active-duty US military personnel who were in the gulf region during the Kuwaiti oil-well fires. Hospitalization records were examined from Department of Defense military treatment facilities from August 1, 1991 until hospitalization, separation from active-duty service, or July 31, 1999. Modeling was used to estimate troop exposure to oil-well fire smoke. The risk of malignant neoplasms of the respiratory and intrathoracic organs was modestly increased in the exposed group, but the CI included the null (adjusted RR 1.10, 95% CI 0.56–2.17). The relatively short observation period (8 years) is a limitation of this study for assessing cancer risk.

Conclusion

Results of studies of fuel exposure and lung-cancer risk are inconsistent. Siemiatycki et al. reported an association between kerosene and crude-oil exposure and squamous-cell lung cancer, between diesel-fuel exposure and nonadenocarcinoma, and between heating-oil exposure and oat-cell lung cancer (Siemiatycki et al. 1987a). Two cohort studies (Imperial Oil and Texas cohorts) did not find an association in workers most likely to have been exposed to fuels.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels and lung cancer.

There was evidence of associations between exposure to ambient air pollution, engine exhausts, and heating sources (coal) and lung cancer. Cohort and case-control studies showed consistently that risks increased with increasing ambient air pollution. There was evidence from both cohort and case-control studies that increasing exposure to engine exhausts and to its components (such as PAHs) increased the risk of lung cancer. As supporting evidence, the case-control studies of heating with coal showed increasing risks of developing lung cancer, and the risks were independent of active smoking and exposure to environmental tobacco smoke. The one cohort study showed that reduction in combustion products from smoky coal, by use of stoves that had chimneys, decreased the risk of lung cancer. Experimental evidence has shown that many compounds present in combustion products of oils in Chinese-style cooking are carcinogenic and some mutagenic. There is additional evidence that kerosene and soot are genotoxic in in vivo and in vitro models and that diesel exhaust is mutagenic. Toxicologic studies found that rats and mice exposed to coal and wood smoke had an increased incidence of lung cancer.

The committee concludes, from its assessment of the epidemiologic literature, that there is sufficient evidence of an association between exposure to combustion products and lung cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

MALIGNANT MELANOMA OF THE SKIN

Skin cancer is the most common type of cancer (ACS 2004u). There are two forms of skin cancer: melanoma (ICD-9 172) and nonmelanoma (basal-cell and squamous-cell carcinomas) (ICD-9 173). The major risk factor for both types of skin cancer is exposure to the sun and other sources of UV radiation. Family history, fair skin, moles, male sex, the inherited disease xeroderma pigmentosum, and immune suppression also play a role in the development of melanoma skin cancers. Melanoma is a much less common form of skin cancer than basal-cell or squamous-cell carcinoma, but is much more serious, usually being fatal if not treated in its early stages (ACS 2004u).

In 2000, malignant melanoma accounted for 17.7 new cases per 100,000 people (22.5 among men and 14.4 among women) and 2.7 deaths per 100,000 (3.8 among men and 1.8 among women) in the United States (Ries et al. 2004).

Most of the studies identified by the committee as meeting its criteria for assessing a potential association between exposure to fuels or their combustion products and skin cancer focused on melanoma. Mortality studies are likely to focus on melanoma, because it is unusual to die of other forms of skin cancer. Several studies of ocular melanoma or uveal melanoma (ICD-8 191) were identified; these are considered separately from cutaneous melanoma because they are classified as malignant neoplasms of nervous tissue. Mortality and morbidity studies of cutaneous melanoma are presented below; information gathered on malignant nonmelanoma skin cancer is summarized in the next section.

Fuels

Table 4.21 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and melanoma skin cancer.

Cohort Studies

In the large Amoco cohort of oil refinery workers, Nelson et al. (1987) found the risk of skin-cancer death among white men increased (SMR 2.01, 95% CI 1.00–3.60), even though the usual “healthy-worker effect” of reduced overall mortality and cancer mortality was seen. The study combined both melanoma and nonmelanoma skin cancer in its definition of skin cancer; however, because death was the end point, virtually all the skin-cancer cases were probably melanoma. When latent period was considered, the mortality was greatest among those with less than 15 years since first exposure (SMR 5.24, p<0.05). The work histories of the people in the cohort were reviewed by industrial hygienists and classified for specific exposures (Nelson et al. 1985). There were 11 skin cancers, and the excess appeared to be concentrated among the maintenance workers (SMR 3.78, p<0.05). Routine exposure to refinery processes was also associated with skin-cancer mortality (SMR 2.68, p<0.05). The failure to adjust for possible confounders (particularly sun exposure) reduces confidence in the finding.

Jarvholm et al. (1997) investigated the incidence of melanoma and nonmelanoma skin cancer separately in a cohort of Swedish refinery workers. A minor increase in the risk of melanoma (SIR 1.1, 90% CI 0.49–2.0) was based on seven exposed cases, who had all received their diagnoses within 20 years of their first exposure.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Mortality has been assessed periodically among workers of Imperial Oil Limited of Canada (Hanis et al. 1979; Lewis et al. 2000b; Schnatter et al. 1992). In comparison with the Canadian general population, the estimated risk of death from melanoma in 1964–1994 among men working in all sectors of the industry was somewhat increased (SMR 1.32, 95% CI 0.83–2.00). Risks had diminished from the previous followup period (1964–1983) for upstream operations and for the marketing and distribution segment (Schnatter et al. 1992). Only two additional cases were observed among marketing and distribution workers in the 11 years of followup through 1994, so SMR was reduced from 2.56 (95% CI 0.94–5.57) to 1.59 (95% CI 0.69–3.14). The occurrence of only a single new case among exploration, drilling, production, or pipeline (upstream) workers reduced the risk considerably from 6.00 (95% CI 2.19–13.06) to 2.82 (95% CI 1.13–5.81). The increased risk of melanoma among the upstream subcohort may be related to exposure to crude oil, but it also coincides with the greater potential for sun exposure of these workers, who spend a considerable amount of time outdoors.

From the Imperial Oil cohort followup through 1994, Lewis et al. (2003) defined an “inception cohort” of younger workers with exposure experience only under more recent occupational-hygiene standards; the cohort consisted of those hired in 1964–1993. In these younger men, the estimated risk of melanoma (SIR 1.25, 95% CI 0.82–1.83) was effectively equivalent to the estimated mortality among the men in the full cohort; it must be noted that the incident cases include many (if not all) of the instances of melanoma deaths in the first analysis. The contrast between mortality among the (few) older female workers and incidence among the newly hired women, however, was considerable, although the CIs overlap slightly. The women’s estimated risk of melanoma (SIR 1.46, 95% CI 0.83–2.37) was similar to both findings for men. The Imperial Oil findings present a fairly coherent picture of an association of exposure to petroleum and derived fuels with melanoma, except that there is no suggestion of a dose-response relationship by duration of employment or intensity of exposure to fuel-related hydrocarbons. Additional uncertainty arises from the lack of control for possible confounders, especially exposure to sunlight.

Case-Control Studies

In an analysis of the data from the multicancer case-control study conducted in Montreal that specifically addressed association between work in various occupations and melanoma, Fritschi and Siemiatycki (1996a) found no association with any industry, occupation, or industrial-hygiene-determined substance related to petroleum. Those substances included C1-C4 alkanes, C5-C17 alkanes, C18+ alkanes, and leaded gasoline with 10, 22, 18, and six exposed cases, respectively. The study has the advantages of addressing incident cases, having full exposure histories interpreted by industrial hygienists, and incorporating information on many potential confounding variables, but the researchers note the study’s limited power to exclude definitively the possibility of an association with any particular exposure.

Combustion Products

Table 4.22 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and melanoma skin cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Cohort Studies

Boffetta et al. (1988) analyzed the mortality experience of the prospective cohort established by the ACS in 1982 from the perspective of occupational exposure to diesel exhaust. The sample was defined as men found in followup who had been 40–79 years old at enrollment and had provided usable information on both smoking and exposure to diesel exhaust (and extensive other information gathered in the self-administered questionnaire). Although only 2 years of followup had accumulated, 7,499 deaths had occurred in the sample of 369,943 people (the above data constraints and loss to followup reduced the original sample by about 25%). When adjusted for age, smoking, and other occupational exposures, an imprecise estimate of risk of melanoma death of those exposed to diesel exhaust was increased (RR 1.67, p>0.05).

Pion et al. (1995) conducted a nested case-control study on the data generated on the ACS cohort at the time of its 6-year followup. The set of 2,780 melanoma cases was made up (in unspecified proportions) of men and women of all ages who had already been so diagnosed at the time of enrollment or were newly diagnosed during the followup period. Three controls without melanoma—matched by sex, age, race, and area of residence—were drawn from the cohort for each case. Although their accuracy might still be questioned, self-reports of exposure would not be expected to be subject to recall bias in a (strictly) prospective study. Several exposures of interest to the committee were addressed (coal tar, pitch, or asphalt; diesel engine exhaust; and gasoline exhaust), but none had an OR exceeding unity. Occupations were analyzed separately by sex. The risk of melanoma was less than 1 for truck drivers (OR 0.72, 95% CI 0.40–1.30); the most marked increase for a male occupation was for (generic) firefighters (OR 2.29, 95% CI 0.85–6.16). No adjustments were made by using information on possible confounders, whose availability is one of the primary merits of the ACS prospective cohort. The commingling of cross-sectional and prospective designs in this analysis raises the possibility that exposure occurred after the cancer event for the subset of original prevalent cases; no mention was made of screening the exposures to account for this temporal issue.

From the Swedish Cancer Environmental Registry updated to include 1970 census results and cancers diagnosed in 1971–1989, Boffetta et al. (2001) partitioned the cohort of men and women who were actively employed according to both the 1960 and 1970 censuses into sets with and without occupational exposure to diesel exhaust. Individual exposure to diesel emissions was assessed with a matrix of probabilities and intensities expected for each job title. Having found a high concordance between the occupations reported on the two censuses, the researchers opted to encode diesel exposure from the occupation specified on the 1960 census. Melanoma was not among the cancer sites for which there was any suggestion of an increased risk for either men or women who had held a job in 1960 with any likelihood of exposure to diesel emissions. Roughly half the male cases in the sample would intersect with those in the case-control study below, which was based on melanomas diagnosed in Sweden from 1961 to 1979 (Linet et al. 1995).

Case-Control Studies

In a case-control study of patients at Roswell Park Memorial Institute in 1956–1965, the occupational histories of about 14,000 white cancer cases were compared with those of noncancer patients (Decoufle and Stanislawczyk 1977; Viadana et al. 1976). For all the several exhaust-associated occupations of interest, the estimates of association with melanoma were increased, some of them markedly (for example, locomotive engineers and firemen and excavating, grading, and road-machinery operators). Because no more than four men with melanoma had held any of those occupations and only three who had held only clerical positions

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

were available for comparison, all the associated confidence intervals included unity, and no adjustment for possible confounders was attempted.

Siemiatycki et al. (1988) addressed the 121 interviewed melanoma cases in the Montreal multicancer case-control study in seeking associations with various exhaust and combustion-product exposures. They reported that melanoma was associated (p<0.05) with propane exhaust (OR 3.3, 90% CI 1.2–9.0) and there were suggestive increases with jet-fuel exhaust (OR 1.8, 90% CI 0.5–6.4) and liquid-fuel combustion (OR 1.8, 90% CI 0.9–3.4). In a more recent publication on this dataset, Fritschi and Siemiatycki (1996a) found no compelling association between any combustion-products exposure and melanoma. For an exposure to be analyzed (with adjustment for age, education, and ethnicity), there had to be at least four exposed cases. Estimates hovering about unity were reported for gasoline-engine emissions, PAHs from petroleum, and carbon monoxide. An appendix listed several additional combustion-products-related exposures for which the lower 95% confidence limit of the estimate of association with melanoma did not exceed 0.9: the air-transport, motor-transport, and railway-transport industries, with six, eight, and five exposed cases, respectively; the occupations of mechanic or motor-transport worker, with five and 10 exposed cases, respectively; and diesel-engine emissions, liquid-fuel combustion products, PAHs from any source, and pyrolysis fumes not classified elsewhere, with 10, six, 57, and 11 exposed cases, respectively. There was no explanation of why the number of interviewed melanoma cases considered had decreased by 18 (15%); only inconsequential perturbations resulted in the statistics for the exposures common to the two publications (assuming that “petrol engine emissions” equates to “gasoline exhaust”).

Nelemans et al. (1993) compared 140 melanoma cases with 181 controls who had other malignancies, all gathered from a cancer registry in the mideastern part of the Netherlands. After adjustment for age, sex, education, pigmentation factors, and exposure to sunlight, those who had ever worked in the “transport and communications” industry had a greater risk of melanoma than those who had not (OR 1.70, 95% CI 0.84–3.46); the difference was intensified by contrasting this group with those who had never worked in any of 10 hypothetically high-risk industries (OR 1.92, 95% CI 0.84–4.35). For the transportation and communications workers, tar products, cutting oils or coolants, and lubricating oils were among the exposures that were self-reported more frequently than by other workers, but no formal analysis was presented, and there was no attempt at a JEM approach to evaluating exposure. This elementary study provides some weak support for the possibility that exposure to combustion products is associated with the occurrence of cutaneous melanoma.

Using the Swedish Cancer Environment Registry database, Linet et al. (1995) identified 3,850 men diagnosed with melanoma in 1961–1979 who had been listed as occupationally active in the 1960 census; this sample would be expected to coincide with the male portion of the set of 5,003 melanoma cases studied in somewhat less detail by Vagero et al. (1990). The global categories for the “transport and communications” industry or occupation showed no indication of a relationship between melanoma and employment that might involve exposure to vehicle exhaust. Occupational subcategories in this sector, however, did show some increases; the risk for “traffic administration” was 1.6 (p<0.05, adjusted for age and region), and the risk for the more specific classification “traffic enforcement or railroad work” even more pronounced (SIR 3.1, p<0.01).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Conclusion

Data from studies of the large Amoco (Nelson et al. 1985, 1987) and Imperial Oil (Hanis et al. 1979) cohorts suggest an increased risk of death from melanoma with exposure to petroleum-derived materials (such as heavy oils and crude oil), and the smaller study (Jarvholm et al. 1997) is not inconsistent with such an association (Lewis et al. 2000b). In all the fuel cohort studies, the peak of melanoma risk occurred with latency (Jarvholm et al. 1997; Nelson et al. 1987) or duration (Lewis et al. 2000b) of around 15 years, rather than showing a dose-response relationship. The studies, however, did not adjust for sun exposure, a major risk factor for melanoma, and the workers—particularly the exploration, drilling, and pipeline workers—may have received considerable sun exposure while performing their jobs. But the one case-control study with fairly reliable exposure analysis (Fritschi and Siemiatycki 1996a) did not support such a conclusion.

The studies addressing exposure to combustion products also failed to adjust for exposure to sunlight. The Montreal case-control study, which had the best exposure assessment, and the record-linkage study of Linet et al. (1995) found isolated effects of specific exposures (propane exhaust and being a traffic administrator, respectively) that were not among the major ones considered by the present committee in evaluating possible effects of combustion products.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and cutaneous malignant melanoma.

NON-MELANOMA SKIN CANCERS

Nonmelanoma skin cancers (basal-cell and squamous-cell carcinomas) (ICD-9 173) are common but rarely fatal. As for cutaneous malignant melanoma, the main risk factor for nonmelanoma skin cancers is exposure to the sun and other sources of UV radiation. Other risk factors for nonmelanoma skin cancer are sex, family history, some specific inherited conditions (for example, basal-cell nevus syndrome and xeroderma pigmentosum), radiation treatment, exposure to some chemicals (such as arsenic, industrial tar, coal, paraffin, and oils), chronic or severe skin problems (such as burns), a weakened immune system, viral infection, and smoking (ACS 2004u).

Fuels

Table 4.23 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and nonmelanoma skin cancers.

Cohort Studies

Jarvholm et al. (1997) investigated cancer incidence in a cohort of workers assembled from the personnel records of 26 companies in the Swedish petroleum industry. The study generated a somewhat stronger, yet still weak, estimate of an association of exposure to fuels with nonmelanoma skin cancers (SIR 1.3, 90% CI 0.61–2.4) than it had with melanoma.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Retrieval from the cancer registry generated an identical number (seven) of melanoma and nonmelanoma skin cancers; this suggests that records were incomplete.

Case-Control Studies

Kubasiewicz and Starzynski (1989) described the methods used to identify and interview men who had incident cases of skin cancer from 1982 through 1988 in the registry for Lodz, Poland, and age-matched sets of population and hospital controls. With selection for “all skin lesions suspected to have a neoplastic origin”, they reported that the 374 interviewed cases (of the 520 identified cases born after 1900) had 278 basal-cell carcinomas, 70 “carcinomas planoepitheliale”, 13 cases of Bowen’s disease (carcinoma in situ, precancerous), and 13 cases of Arning’s carcinoid (superficial basal-cell carcinoma). In an analysis that focused on occupational exposure to PAHs, Kubasiewicz et al. (1991) reported on a final set of 376 cases and their 752 population and 752 hospital controls. Full work histories were gathered, but the exposures analyzed consisted of self-reports on each of 17 possible agents, three of which would fit in our fuels classification: petroleum, petrol, and gasoline. The distinction between petrol and gasoline was not stated, but the results for gasoline exposure were said to be too sparse to analyze. Minor increases in crude risks of skin cancer were reported for both petroleum and petrol (ORs 1.17 and 1.30, respectively). The definition of skin cancer used in the study clearly excluded melanoma, but it did encompass some skin lesions that do not coincide with the malignant classification the present committee intended to address.

Using the Alberta (Canada) Cancer Registry, Gallagher et al. (1996) identified all men with squamous-cell carcinoma and basal-cell carcinoma who were 20–79 years old when diagnosed in 1983–1984. All 225 subjects with a first primary squamous-cell carcinoma were eligible. For first primary basal-cell carcinoma, only one-fourth of the cases with the lesion on the head or neck were retained, but all subjects with such a tumor elsewhere were kept; the total was 314 eligible cases of basal-cell carcinoma. During home interviews, 180 men with squamous-cell carcinoma and 226 men with basal-cell carcinoma completed a questionnaire that included a complete work history and a list of specific exposures of interest. Of 573 age-matched controls with no history of skin cancer drawn from men in the Alberta Health Care Insurance Plan (which includes everyone resident in the province for more than 3 months), 406 men completed the same questionnaire. The sets of squamous-cell carcinoma and basal-cell carcinoma cases were compared with the controls separately. With adjustment for age, skin pigmentation (for example, fair skin), ethnicity, and exposure to sunlight, but not smoking, the risk of squamous-cell carcinoma after exposure to petroleum products (specified to be gasoline and oil) was increased (OR 1.3, 95% CI 1.0–2.0), but not the risk of basal-cell carcinoma (OR 0.9, 95% CI 0.6–1.3). Although the response rates were low and the exposure data came only from self-reports, the results of the study suggest a potential relationship between exposure to petroleum-derived fuels and squamous-cell carcinoma.

Combustion Products

Table 4.24 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and nonmelanoma skin cancers.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Case-Control Studies

Hannukesela-Svahn et al. (1999) analyzed basal-cell carcinomas and other nonmelanoma skin cancers assembled in the Finnish Cancer Registry over 4 decades by linking to subjects’ stated occupations in the 1970 census. For “transportation and communication” (a nonspecific, self-reported occupation code at a single time), the estimated risk of basal-cell carcinoma among men was 1.0 (95% CI 1.0–1.1), which suggests that the risk was not increased. There were only about 15% as many cases of the other types of nonmelanoma skin cancers (predominantly squamous-cell carcinomas) as there were cases of basal-cell carcinoma, so the confidence intervals associated with the estimated risks were wider; the estimated risk in men was slightly below unity, that in women somewhat above.

As discussed above in connection with exposures to fuels, Kubasiewicz et al. (1991) presented an analysis of occupational exposure of 376 skin-cancer subjects (primarily with basal-cell carcinoma) and their 752 population and 752 hospital controls to PAHs. Full work histories had been gathered, but exposure to PAHs was determined on the basis of self-reported exposure to each of 17 possible sources of PAHs. The cases were compared independently (apparently without adjustment beyond matching on age) with both the population and hospital controls, and the results were virtually identical (the statistics related to the population controls are reported here). There was a slight increase in risk in those who had ever been exposed to any of the sources of PAHs (OR 1.15, 95% CI 0.90–1.51; 95% CI calculated with standard methods from the observed and expected numbers presented in the original paper), but there was no indication of a dose-response relationship with duration of exposure.

In the case-control study of squamous-cell carcinoma and basal-cell carcinoma from the Alberta Cancer Registry, Gallagher et al. (1996) presented the results related to several potentially PAH-containing exposures (determined by self-report) in addition to the fuel-related exposure reported above. After adjustment for age, pigmentation, ethnicity, and exposure to sunlight, but not smoking, the risk of squamous-cell carcinoma after exposure to diesel fumes was increased (OR 1.7, 95% CI 1.1–2.5). For basal-cell carcinoma, only a modestly increased PAH-related risk was seen after exposure to diesel fumes (OR 1.1, 95% CI 0.8–1.6).

Conclusion

The committee recognizes that PAHs (present in soot and numerous similar complex mixtures, mostly originating from combustion processes) have long been accepted to be skin carcinogens in animals and humans (ATSDR 1995; IARC 1985). IARC limited the scope of its consideration to chimney soot, so the subjects of the studies on which its conclusions were based were all chimney sweeps. Similarly, the epidemiologic bases of the Agency for Toxic Substances and Disease Registry (ATSDR) conclusions were shale-oil workers and chimney sweeps. As explained in Chapter 2, in planning its approach to its task, the committee considered those types of exposure as too dissimilar to the exposure scenarios in the Persian Gulf to base its conclusions on combustion products and nonmelanoma skin cancers on the conclusions of IARC and ATSDR.

Dermal application of individual PAHs to animals has been shown to lead to skin tumors in a number of studies. Mechanistic studies have demonstrated that some PAHs are genotoxic and can act as initiators, promoters, and complete carcinogens. Despite strong evidence that PAHs are carcinogenic in animal models and that they are genotoxic, the committee used

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

toxicologic information only in a supportive role; that is, it did not base its conclusions solely on toxicologic information.

Of the available epidemiologic studies that met the committee’s criteria, the Gallagher et al. (1996) study appears to be the most reliable and well conducted. It reported one borderline association between fuel exposure and squamous-cell carcinoma and several somewhat stronger associations between self-reported exposure to PAH-containing agents and squamous-cell carcinoma. For the more common type of nonmelanoma skin cancer (basal cell carcinoma), however, the findings with both types of exposure were largely negative. The committee concluded that without inclusion of epidemiologic studies regarding occupations that it had determined were dissimilar to exposures experienced during the Gulf War, a consistent and convincing picture of increased risk of nonmelanoma skin cancer did not emerge.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels and non-melanoma skin cancers.

FEMALE BREAST CANCER

Breast cancer (ICD-9 174) is about 100 times more common among women than men (ACS 2003e). There are a number of known or suspected risk factors for breast cancer in women; some major ones are aging, genetics, family history, years of ovulation, parity, and use of hormone-replacement therapy after menopause.

In 2000, there were 135.1 new cases of breast cancer per 100,000 women in the US and 26.7 deaths per 100,000 (Ries et al. 2004).

Fuels

Table 4.25 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and breast cancer in women.

Cohort Studies

The risk of breast cancer in women was assessed in two cohorts of workers at petroleum companies (Divine et al. 1999b; Lewis et al. 2000b, 2003). Exposure was assessed in the Texaco cohort by using work histories in company records (Divine et al. 1999b), and no increased risk of breast cancer in women was found (SMR 0.71, 95% CI 0.40–1.18). In the Canadian Imperial Oil cohort, exposure was assessed in the same way (Lewis et al. 2000b, 2003), and neither breast cancer mortality nor incidence was markedly increased (SMR 1.08, 95% CI 0.66–1.67; SIR 1.02, 95% CI 0.80–1.28, respectively).

Lagorio et al. (1994) tracked the mortality experience of 357 women who had been managers of Italian service stations in 1980–1992. The effort complemented a detailed assessment of exposure at service stations by monitoring 111 attendants in 1992 (Lagorio et al. 1993). The OR for female breast-cancer risk was 1.04 (90% CI 0.18–3.28) on the basis of two cases.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Combustion Products

Table 4.26 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and breast cancer in women.

Case-Control Studies

Petralia et al. (1999) conducted a case-control study of 301 women who had premenopausal breast cancer and 316 controls from western New York state. JEMs were used to determine occupational exposure to PAHs and benzene, and the researchers endeavored to separate their individual roles. In comparison with women exposed to neither PAHs nor benene, the adjusted risk of breast cancer among women exposed only to PAHs (six cases) was neutral (RR 1.01, 95% CI 0.55–3.45), whereas the equivalent result for the 24 cases exposed only to benzene was 1.70 (95% CI 1.17–2.92). Smaller numbers of subjects in subcategories made separating the agent-specific contributions to estrogen-receptor-positive or -negative types of breast cancer more ambiguous.

Lewis-Michl et al. (1996) conducted a case-control study of the relationship between breast cancer and residence near industry or traffic on Long Island, New York. Among the women who had premenopausal breast cancer (93 in Nassau County and 70 in Suffolk County), the crude risks gave no indication of a relationship with those factors. In the larger group of 627 postmenopausal cases, crude ORs for high traffic density were not impressively increased among Nassau county subjects and did not exceed unity among women in Suffolk County; these results were essentially unchanged by adjustment for proximity to industry. The relevance to Gulf War exposures of the metric—distance from potential sources of vehicular combustion products—is limited by the ecologic nature of the exposure assessment.

Conclusion

No increased risk of breast cancer was found in a cohort of female petroleum workers (Divine et al. 1999b), and only a modest increase was found in another cohort of female petroleum workers (Lewis et al. 2003) and a cohort of female service-station workers in Italy (Lagorio et al. 1994). The two studies of breast cancer and exposure to combustion products evaluated by the committee (Lewis-Michl et al. 1996, Petralia et al. 1999) had essentially negative results.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and breast cancer in women.

MALE BREAST CANCER

Breast cancer occurs in men (ICD-9 175), although only rarely. Breast cancer accounts for about 0.22% of cancer deaths among men (ACS 2004t). Risk factors include aging, family history, radiation exposure, liver disease, estrogen treatment, physical inactivity, obesity, and the congenital condition Klinefelter syndrome.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

In 2000, there were 1.3 new cases of breast cancer per 100,000 men and 0.4 deaths per 100,000 in the United States (Ries et al. 2004).

Fuels

Table 4.27 presents the findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and breast cancer in men.

Case-Control Study

A population-based study of male breast cancer was conducted by using the Danish cancer registry (Hansen 2000). Exposure to gasoline vapors and PAHs (combined) was estimated from trade codes; blue-collar workers who had at least 3 months of employment in service stations, vehicle maintenance, the wholesale gasoline trade, or car-repair shops were classified as exposed. Exposure defined in that way was associated with a greater risk of breast cancer (OR 2.2, 95% CI 1.4–3.6), when birth year and socioeconomic status were controlled for. When a 10-year exposure lag was considered, the risk of breast cancer remained increased (OR 2.5, 95% CI 1.3–4.5). Men younger than 40 years old at the time of first employment had an OR of 3.7, which increased to 5.4 (95% CI 2.4–11.9) for a lag time of 10 years. The study did not control for smoking.

Combustion Products

Table 4.28 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and breast cancer in men.

Case-Control Studies

There are few case-control studies of the relationship between male breast cancer and exposure to combustion products or occupations with combustion-product exposure. As discussed in the previous section on fuel exposure, Hansen (2000) explored the relationship between male breast cancer and occupational exposure to gasoline and vehicle combustion products, including PAHs. Because the exposures were combined in the analysis, it is not possible to determine whether the increased risk of male breast cancer was due to fuel or combustion-product exposure.

Cocco et al. (1998) conducted a case-control study of male breast cancer and occupational exposure. They assessed occupational exposure to PAHs, high temperature, electromagnetic fields, herbicides, pesticides, and organic solvents by applying JEMs based on US census occupation and industry codes. The OR for the association between exposure to PAHs and male breast cancer was not increased. Taxicab drivers and subjects working in the motor-vehicle and equipment industry were at increased risk for male breast cancer.

Conclusion

Hansen (2000) reported positive findings regarding exposure to fuels and combustion products and male breast cancer; but the method used to assess exposure in that study is limited. Cocco et al. (1998) used a JEM and did not find an association between PAH exposure and male breast cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and male breast cancer.

FEMALE GENITAL CANCERS (CERVICAL, ENDOMETRIAL, UTERINE, AND OVARIAN)

Cervical cancer (ICD-9 180) has a number of risk factors, the most important of which is human papillomavirus infection. Other important factors are smoking, immune deficiency, and poor nutrition. Oral contraceptives, multiple pregnancies, maternal diethylstilbestrol use, family history, and socioeconomic status may also contribute to risk (ACS 2004b).

Endometrial cancer (ICD-9 182), which occurs in the inner lining of the uterus, is the most common female reproductive system cancer. A major risk factor in endometrial cancer is exposure to estrogen, which may occur because of early menarche or late menopause (both can lead to longer span of menstruation), infertility, nulliparity, obesity, use of tamoxifen or estrogen-replacement therapy, and some ovarian diseases. Other risk factors are diabetes, diet, age, family history, breast or ovarian cancer, hereditary nonpolyposis colorectal cancer syndrome, and pelvic radiation therapy (ACS 2004c).

Ovarian cancer (ICD-9 183) leads to more deaths than any other cancer of the female reproductive system. Risk factors include age, reproductive history, fertility drugs, estrogen or hormone-replacement therapy, family history, and history of breast cancer. Mutations in the BRCA1 and BRCA2 genes may also increase risk of ovarian cancer (ACS 2004i).

In 2000, there were 32.0 new cases of cervical, uterine, and endometrial cancer per 100,000 women and 6.9 deaths per 100,000 in the US (Ries et al. 2004), and there were 16.3 new cases of ovarian cancer per 100,000 and 8.9 deaths per 100,000 in the US (Ries et al. 2004).

Fuels

Table 4.29 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and cervical, endometrial, uterine, and ovarian cancers.

Cohort Studies

Exposure to fuels did not lead to increased risk of cervical cancer or of uterine cancer in either the Texaco or Imperial Oil cohort of petroleum workers (Divine et al. 1999b; Lewis et al. 2003). There was no notable increase in the risk of cancer of the ovary, fallopian tube, or broad ligaments among the women in the Imperial Oil cohort (SMR 1.74, 95% CI 0.70–3.58) (Lewis et al. 2003), but no information on ovarian cancer was available on female workers in the Texaco cohort (Divine et al. 1999b).

A Finnish study used a record-linkage approach to explore the relationship between occupational exposures and ovarian cancer (Vasama-Neuvonen et al. 1999). Gasoline exposure, derived with a JEM based on job titles, was found not to be associated with ovarian cancer.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Combustion Products

Table 4.30 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and cervical, endometrial, uterine, and ovarian cancers.

Case-Control Studies

A study explored the relationship between burning wood in the kitchen and the risk of cervical neoplasia in women in Honduras (Velema et al. 2002). Although the risk of grade III cervical intraepithelial neoplasia (CIN) showed a trend with increasing duration of exposure to wood smoke (p=0.022), women with low or intermediate exposure (up to 25 years) had lower estimated risks than the reference group of women, who said they had never used wood in the kitchen.

Finnish women born in 1906–1945 and active according to the 1970 census were linked to the national cancer registry for 1971–1995. The Finnish JEM applied to the census job titles was used to define exposure to selected occupational agents. Even in the highest exposure categories, Weiderpass et al. (2001) did not find the risk of cervical cancer to be increased by exposure to exhaust from diesel engines (RR 1.7, 95% CI 0.4–6.8) or gasoline engines (RR 1.3, 95% CI 0.7–2.2) or by exposure to PAHs (RR 1.2, 95% CI 0.3–4.8). Weiderpass et al. (2001) found no increase in risk of endometrial cancer associated with any of those three agents. Vasama-Neuvonen et al. (1999) reported on the results of this record-linkage study for ovarian cancer. Risk was not found to be increased by overall exposure to gasoline-engine exhaust, diesel-engine exhaust, or PAHs; there were, however, suggestions of dose-response trends related to exposure to each type of engine exhaust.

Conclusion

Overall, the studies provide inadequate support for an association between exposure to fuels or combustion products and cervical, endometrial, uterine, or ovarian cancer (Divine et al. 1999b; Lewis et al. 2003; Vasama-Neuvonen et al. 1999; Velema et al. 2002; Weiderpass et al. 2001). The association between wood smoke and CIN observed in the study by Velema et al. should be explored further.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and cervical, endometrial, uterine, or ovarian cancer.

MALE GENITAL CANCERS (PROSTATIC OR TESTICULAR)

Factors that increase the risk of developing prostatic cancer (ICD-9 185) include increasing age, race, family history, diet, and physical inactivity (ACS 2004v). Testicular cancer (ICD-9 186) is an uncommon but highly treatable cancer. Known or suspected risk factors include cryptorchidism, family history, some occupational exposures, multiple atypical nevi, HIV infection, race and ethnicity, body size, and maternal hormone use during pregnancy (ACS 2004j).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

In 2000, there were 176.9 new cases of prostatic cancer per 100,000 men and 30.6 deaths per 100,000 in the US (Ries et al. 2004), and there were 5.7 new cases of testicular cancer per 100,000 men and 0.2 death per 100,000 (Ries et al. 2004).

Fuels

Table 4.31 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and prostatic cancer and the only results on testicular cancer.

Cohort Studies

In a cohort of uranium-processing workers, an industrial-hygiene survey was done to categorize exposure to kerosene (Ritz 1999). At the lowest kerosene exposure, there was no hint of an association with prostatic cancer mortality; while for the six cases with moderate kerosene exposure, the most extreme risk estimate occurred for more than 5 years of exposure and no lag (RR 3.69, 95% CI 0.91–15.0). No workers had the highest exposure to kerosene.

Risk of prostaticcancer was not increased in two cohorts of petroleum-company workers: the Canadian Imperial Oil cohort (Hanis et al. 1979) and the Texaco cohort (Divine et al. 1999b; Lewis et al. 2000b, 2003; Schnatter et al. 1993). Only Lewis et al. (2003) reported on testicular cancer, again with negative findings.

Case-Control Study

In a hospital-based case-control study conducted in Montreal, there was an association between diesel-fuel exposures, as assessed with a JEM based on occupational history, and prostatic cancer (OR 1.7, 90% CI 1.2–2.5) (Siemiatycki et al. 1987a). Modest increases were in association with kerosene, heating oil, and crude oil. No increased risk of prostatic cancer was observed after exposure to automotive gasoline, aviation gasoline, or jet fuel.

Combustion Products

Table 4.32 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and prostatic cancer.

Case-Control Studies

A case-control study in Germany assessed the relationship between occupational factors and prostatic cancer (Seidler et al. 1998). Two JEMs were used to estimate dose-years of exposure, the product of estimates of intensity, duration, and probability of exposure. One of the JEMs yielded estimates of exposure to exhaust fumes that resulted in an increased risk for more than 25 dose-years vs no exposure (OR 2.4, 95% CI 1.2–4.7) but a less pronounced risk for 25 dose-years or fewer (OR 1.2, 95% CI 0.8–1.9). A similar pattern was seen when the other matrix was used to estimate exposure to diesel fumes and fuel; for more than 25 dose-years, the risk was clearly increased (OR 3.7, 95% CI 1.4–9.8), but for 25 dose-years or fewer, the OR was only 1.1 (95% CI 0.7–1.8). For PAH exposure, results contrary to a dose-response relationship were seen: exposure for 25 dose-years or fewer led to an OR of 1.6 (95% CI 1.0–2.4), but for more than 25 dose-years, the OR was 1.4 (95% CI 0.4–4.7).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

In another case-control study, the relationship between occupational risk factors and prostatic cancer was examined on the basis of job history obtained by interview (Krstev et al. 1998). An increased risk of prostatic cancer was reported in firefighters (OR 3.34, 95% CI 1.13–9.91) and railroad-transportation workers (OR 1.66, 95% CI 1.13–2.44). A dose-response trend was observed in the railroad-transportation workers; those working in railroad transportation for fewer than 5 years had an OR of 1.47, for 5–19 years an OR of 1.43, and for 20 years or more an OR of 6.47. No increased risk was associated with other transportation or trucking workers.

Aronson et al. (1996) conducted a population-based case-control study of occupational risk factors in a study that included 449 prostatic cancer cases and over 2,000 controls. A JEM was applied to occupational histories to categorize various exposures as “substantial” or “nonsubstantial” for comparison with nonexposed subjects. Nonsubstantial exposure to diesel-engine emissions or to PAHs from coal was associated with increased risks of prostatic cancer; substantial exposure to the agents was associated with less pronounced risks. Substantial exposure to combustion products of liquid fuel or to PAHs from any source was associated with higher estimated risks than nonsubstantial exposure, but still not markedly increased.

A population-based case-control study in Montreal evaluated exposure to combustion products and prostatic cancer. An industrial hygienist reviewed occupational histories to derive the intensity and probability of various exposures (Siemiatycki et al. 1988). The report presented 90% CIs. Increased risks of prostatic cancer were reported for exposure to combustion products of liquid fuel (OR 1.6, 90% CI 1.2–2.1) or of coal (OR 1.6, 90% CI 1.2–2.2) but not of propane, natural gas, or wood or to any of the types of exhaust considered.

Conclusion

There were not enough relevant data to draw any sort of conclusion about exposure to fuels or their combustion products and testicular cancer.

The evidence is inconsistent regarding an association between fuel exposure and prostatic cancer. Only one study (Siemiatycki et al. 1987a) of the several reviewed by the committee reported a positive association between a fuel-related exposure and prostatic cancer.

Although the studies by Siemiatycki et al. (1987a), Aronson et al. (1996), Seidler et al. (1998), and Krstev et al. (1998) reported several positive associations between occupations having potential for exposure to combustion products or PAHs or having more rigorously derived estimates of exposure to such agents and prostatic cancer, the committee noted that the results were not consistently positive.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and prostatic cancer.

NERVOUS SYSTEM CANCERS

This section summarizes what is known about the relationship between exposure to fuels and combustion products and cancers of the nervous system (ICD-9 191–192). Most nervous system tumors are not associated with known risk factors. The few known risk factors associated with those cancers are radiation, immune system disorders, and family history.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

In 2000, there were 6.6 new cases of brain and other nervous system cancers per 100,000 people (8.0 among men and 5.4 among women) and 4.5 deaths per 100,000 (5.6 among men and 3.7 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.33 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and brain cancer.

Cohort Studies

A series of studies described mortality among refinery, petrochemical, and research workers employed by Texaco (Divine and Barron 1986; Divine et al. 1985, 1999a, 1999b) for at least 5 years, with at least 1 day falling after January 1, 1947. Two of the studies presented the results of followup through 1977 for about 19,077 workers in the three categories, of when only refinery workers were of primary interest to the present committee (Divine and Barron 1986; Divine et al. 1985). With followup through 1993, addition of employees at three refineries acquired after 1977, and the inclusion of workers hired after 1977 (all with at least 5 years of employment), Divine et al. (1999b) considered a total of 28,840 workers, 738,454 person-years, and 9,575 deaths. Mortality from brain tumors (benign or malignant) was somewhat increased among white men who were ever employed in the motor-oil unit (SMR 1.78, 95% CI 0.88–3.19) and more intensely among those who had been employed in that unit for at least 5 years (SMR 3.26, 95% CI 1.40–6.43).

A cohort mortality study of 1,583 workers employed in an oil refinery near Milan, Italy that converted crude oil into a variety of hydrocarbons (solvents, fuel, and lubricants) was conducted to evaluate cancer risk (Bertazzi et al. 1989). Followup on workers was originally from 1949 to 1982; Consonni et al. (1999) extended it to 1991, for a total of 39,857 person-years and 352 deaths. There was an increased mortality from brain cancers, but the number of cases observed was low (five) and the 95% CI included the null value (SMR 2.08, 95% CI 0.67–4.85). Furthermore, the cancer excess was confined to workers employed for less than 15 years and no cases were observed among longer-term workers. Similarly, the largest increase was among workers who died from brain cancer within 9 years of first employment and no cases were observed 20 years after initial employment. No information was provided on brain cancer by type of job. The authors also reported that internal comparisons (using Poisson regression) were noninformative.

A series of cohort studies of petroleum workers at Imperial Oil Limited in Canada was conducted to assess cancer risk (Lewis et al. 2000b, 2003; Schnatter et al. 1993). There wase no notable increase in deaths from nervous system tumors overall (benign or malignant) through 1994 among workers employed in 1964–1983. Minor increases in the estimated risks of malignant brain tumors were found in refinery, upstream (exploration, drilling, production or pipeline), and office workers, whereas the estimate of this risk was less than unity (SMR 0.68, 95% CI 0.31–1.29) among marketing and distribution workers, who would be expected to have greater exposures to petroleum products than workers in other operating segments (Lewis et al. 2000b). In a study of workers hired more recently (1964–1994) by Imperial, the estimated incidence of or mortality from brain malignancies or any other nervoius system tumors were not remarkably increased in either men or women (Lewis et al. 2003).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

A nationwide survey of service stations in Italy defined a cohort of 2,665 managers alive at the beginning of 1981. Following their mortality through 1992, the study authors found 250 deaths among the 2,308 men and 20 among the 357 women (Lagorio et al. 1994). The SMR for nervous system cancer was 2.14 (90% CI 0.93–4.21) for all service-station mangers. For small stations, where sales per employee (a surrogate of exposure) were greater, the risk (based on five observed cases, one of them a woman) was more pronounced (SMR 2.66, 90% CI 1.05–5.59); the observation of a single case among men who attended large service stations just exceeded the expectation of 0.9. The authors note that the station employees were exposed to a combination of hydrocarbons from fuels and combustion products.

Case-Control Studies

A case-control study was conducted to evaluation the relationship between glioma and occupational title (De Roos et al. 2003). Four hundred and eighty-nine cases were identified at three US hospitals. Controls were selected from patients admitted to the same hospitals for nonmalignant conditions. Exposure was assessed by questionnaire which was reviewed by an industrial hygienist. No increased risk was observed for gasoline station attendants (ever worked: OR 0.5, 95% CI 0.3–0.9).

A population-based case-control study of glioma, the most common form of primary malignant brain tumor in adults, was conducted in the San Francisco Bay area (Carozza et al. 2000). The estimated risk of glioma in those ever employed as petroleum or gas workers was 4.9 (95% CI 0.6–42.2), but no increase in the risk of glioma was found in service-station workers. For both occupational categories, the estimated ORs and CIs were virtually identical with or without a 10-year latent period.

Combustion Products

Table 4.34 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and brain cancer.

Case-Control Studies

The study by De Roos et al. (2003) that evaluated the relationship between glioma and occupational title did not find increased risk for car and light truck drivers (ever worked: OR 0.9, 95% CI 0.5–1.7) and heavy truck drivers (ever worked: OR 0.7, 95% CI 0.4–1.1). Railroad workers had a slight increase in risk (ever worked: OR 1.1, 95% CI 0.4–3.3), but the CI included the null and there were only 6 cases.

The case-control study of glioma in the San Francisco Bay area discussed above (Carozza et al. 2000) had similar negative findings for mechanics and motor-vehicle operators whether or not provision was made for a 10-year latent period. For occupations that entailed exposure to combustion products that the committee considered relevant to Gulf War veterans, the most increased risk estimate was for motor-vehicle operators exposed for at least 10 years without the requirement of a latent period (OR 2.1, 95% CI 0.7–6.2). Firefighters had an OR of 2.7 (95% CI 0.3–26.1), but they were presumed to be largely urban.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Conclusion

Several studies reported sporadic associations between fuel exposure and brain cancer (Carozza et al. 2000; Consonni et al. 1999; Divine et al. 1999a; Lagorio et al. 1994; Lewis et al. 2003), but none could be considered a high-quality study supported by an adequate exposure assessment. Data on combustion products and brain cancer were too sparse to determine whether an association exists.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and nervous system cancers.

OCULAR MELANOMA

Ocular melanoma (ICD-9 190) is a rare type of eye cancer that usually develops in the choroids, although 10% occur within the iris. Risk factors include coloration (particularly blue eye color), such inherited conditions as dysplastic nevus syndrome and oculodermal melanocytosis (nevus of Ota), and possibly sun exposure (ACS 2004d).

In 2000, there were 0.7 new case of ocular melanoma per 100,000 people (0.9 among men and 0.6 among women) and 0.1 death per 100,000 (0.1 among men and 0.1 among women) in the United States (Ries et al. 2004).

Fuels

The committee did not identify any relevant studies that examined the relationship between exposure to fuels and ocular melanoma.

Combustion Products

Table 4.35 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and ocular melanoma.

Case-Control Studies

Three studies of ocular melanoma were reviewed by the committee (Ajani et al. 1992; Guenel et al. 2001; Monarrez-Espino et al. 2002) all assessed exposure by using self-reported job titles. Each of these modest-size studies reported that occupations or industries involving exposure to combustion products were associated with increased risks of ocular melanoma with imprecise confidence intervals. Ajani et al. (1992) reported an OR of 1.23 (95% CI 0.55–2.74) in transportation, communications, and other public-utilities workers. Guenel et al. (2001) reported an OR of 1.4 (95% CI 0.5–3.8) in male transport equipment operators. Monarrez-Espino et al. (2002) reported an OR of 1.5 (95% CI 0.66–3.23) in male transportation-equipment operators and an OR of 2.5 (95% CI 0.94–6.58) in female station, engine, heavy equipment operators, and freight handlers.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Conclusion

No studies were identified that evaluated exposure to fuels and an association between increased risk of ocular melanoma. Three studies, all lacking adequately specific exposure assessment, reported increased, but imprecise risks of ocular melanoma in occupations related to transportation (Ajani et al. 1992; Guenel et al. 2001; Monarrez-Espino et al. 2002).

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to combustion products and ocular melanoma.

BLADDER CANCER

The bladder is lined with transitional and squamous cells. More than 90% of bladder cancers (ICD-9 188) arise in transitional cells, and squamous-cell carcinomas represent only about 8% (NCI, 2002). Because cells that line the renal pelvis and ureter are histologically similar to bladder epithelial cells, tumors of the renal pelvis (ICD-9 189.1), ureters (ICD-9 189.2), and urethra (ICD-9 189.3) are considered urothelial-cell tumors, as are primary tumors of the bladder. Renal-cell carcinomas are histologically distinct from tumors that arise from urothelial tissues, such as those in the renal pelvis. This section, however, addresses only bladder cancer (ICD-9 188) unless it is otherwise stated.

The major risk factor for bladder cancer is smoking. Demographic factors that have some influence on the occurrence of bladder cancer are race (highest in whites, lowest in Asians), increasing age, sex (males at higher risk), and family history. Chronic bladder inflammation due to infections, bladder or kidney stones, or parasites has been associated with bladder cancer. Known occupational risk factors include use of the drug cyclophosphamide, and exposure to aromatic amines, arsenic, and organic chemicals associated with manufacture of rubber, leather, textiles, and paint (ACS 2004o).

In 2000, there were 21.3 new cases of bladder cancer per 100,000 people (37.8 among men and 9.4 among women) and 4.3 deaths per 100,000 (7.6 among men and 2.3 among women) in the United States (Ries et al. 2004).

Fuels

Table 4.36 presents the most relevant findings reviewed by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and bladder cancer. In the text below, the studies are discussed in roughly chronologic order, whereas more recent studies (with publications on a given study population grouped) are generally presented first in the table.

Case-Control Studies

In the earliest case-control study of bladder cancer reviewed by the committee (Cole et al. 1972), an industrial-hygiene-derived group of occupations associated with petroleum products was among 13 occupational classes into which subjects were categorized. When smoking and age were controlled for, no excess risk was observed among men ever or usually working in this petroleum-product group. Some later case-control studies of bladder cancer considered

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

employment in petroleum-related industries or occupations, as abstracted from work histories, among the possible exposures analyzed. Several found associations between petroleum-related entries in work histories and the occurrence of bladder cancer (Howe et al. 1980; Iscovich et al. 1987; Najem et al. 1982), and two found only imprecise indications of increased risks related to such industrial or occupational categories (Risch et al. 1988; Teschke et al. 1997). Using exposures to specific agents as recalled and reported by the subjects themselves, Mommsen and Aagard (1984) reported an increase associated with “work with oil or gasoline” (RR 2.71, 95% CI 1.18–6.17) but a less compelling increase associated with “work with kerosene or asphalt” (RR 3.12, 95% CI 0.88–11.00) (Mommsen et al. 1982). The committee did not regard the exposure assessments conducted in these case-control studies as sufficiently specific or reliable to be indicative themselves of an association between exposure to fuels and bladder cancer. The committee gave more weight to larger studies that included exposure analyses guided by expert understanding of the potential exposures involved in various industrial processes, as described below.

Participants in the National Cancer Institute (NCI) National Bladder Cancer Study (NBCS) were drawn from 10 centers and studied with a common protocol (Hartge et al. 1984). There were several reports from the individual centers—for example, Detroit (Silverman et al. 1983), New Jersey (Schoenberg et al. 1984), and Utah (Schumacher et al. 1989)—or from specific perspectives, such as mechanics (Smith et al. 1985), motor-vehicle exhaust-related occupations (Silverman et al. 1986), the chemical industry (Zahm et al. 1987), and racial differences (Schairer et al. 1988). The summary reports by (Silverman et al. 1989a, 1990, 1989b) on nonwhite men, white women, and white men are focused on analyses by industry and occupational titles with adjustment for smoking.

The analyses of the multicenter dataset with the most power were conducted on the large subsample of 2,100 white males who had bladder cancer and 3,874 white male population controls (Silverman et al. 1989b). For those ever employed in petroleum-processing, the increase in risk was marginally increased (RR 1.3, 95% CI 1.0–1.8); the largest risk was associated with the extraction of crude oil (RR 2.4, 95% CI 1.1–5.5). Three center-specific publications presented reanalyses involving about 62% of the white male subjects; the results were consistent with those above for petroleum-processing and a suggestion of increased risk among service-station workers, a category that had not been reported on in (Silverman et al. 1989b). In the publication on the 332 white men from the Utah study center (Schumacher et al. 1989), a slightly more increased risk of bladder cancer was reported for those who had worked at least 10 years in the “fuel industry”, but the estimate for this small subset was imprecise. In the New Jersey subset of 658 white men, Schoenberg et al. (1984) found virtually the same risk as seen in the full NBCS sample for refinery workers (again imprecise in the smaller sample) and a strong association in garage or gas-station workers (OR 2.35, 95% CI 1.47–3.78). In analyzing the Detroit subset of 303 white men, Silverman et al. (1983) found a large refining-related risk based on only six exposed subjects who had bladder cancer (RR 6.0, 95% CI 0.7–49.8) and a more modest increase in those ever employed in gasoline service (RR 1.6, 95% CI 0.8–3.5), which was further reduced when adjusted for smoking (RR 1.3).

In the NBCS subsample of 126 bladder-cancer cases in nonwhite men and their 383 controls, the risk of bladder cancer among petroleum workers (RR 2.1, 95% CI 0.5–9.2) or service-station workers (RR 1.6, 95% CI 0.5–4.9) was somewhat higher (Silverman et al. 1989a). Although the subsample of white women was larger (652 cases and 1,266 controls), there were

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

too few subjects employed in fuel-related jobs for analysis (Silverman et al. 1990). All the results from the NBCS are less compelling because the analyses were based only on industry or job title.

In a study of several types of cancer, Siemiatycki et al. (1987b) investigated the relationship between exposure to various fuels and 486 incident cases of bladder cancer in 29 Montreal hospitals and 2,196 other-cancer controls (excluding lung and kidney cancer). Exposure was categorized on the basis of industrial-hygiene review of occupational history. No increased risk of bladder cancer was attributable to exposure to aviation gasoline, jet fuel, diesel fuel, or heating oil; the risks posed by exposure to automotive gasoline, kerosene, and crude oil were only slightly above unity (Siemiatycki et al. 1987a). In a later analysis of this dataset focused specifically on bladder cancer (Siemiatycki et al. 1994), employment in the petroleum or coal-products industry was the sole exposure pertinent to fuels, and it showed no association with bladder cancer.

Using a record-linkage approach on the Swedish Cancer Environment Registry, Steineck et al. (1989) constructed a case-control study by identifying the 10,123 male bladder-cancer cases diagnosed in 1961–1979 that had also been listed in the 1960 national census. Controls were selected from that census. A JEM was applied to the job reported by each subject in that census with a note on the likelihood that the given job was predictive of exposure to a particular agent. There was no association of exposure with the risk of bladder cancer in men who had reported jobs deemed moderately or highly likely to involve exposure to gasoline.

Steineck et al. (1990a) gathered cases of cancer of the lower urinary tract (bladder, ureter, renal pelvis, and urethra) diagnosed in Stockholm County from 1985 to 1987 and population-based controls. Gasoline and various combustion products were among the occupational exposures assessed by an industrial hygienist on the basis of each participant’s work history. The 254 male subjects who had urothelial cancer were compared with 287 age-matched controls (Steineck et al. 1990b). After adjustment for age and smoking, the estimated risk of cancer associated with having worked with gasoline was somewhat increased (RR 1.4, 95% CI 0.7–2.9); the effect on risk apparently was more concentrated in the group considered to have the highest exposure (RR 2.5, 95% CI 0.8–7.5).

Of 1,716 incident bladder-cancer cases in people 40–85 years old identified in the State Health Registry of Iowa in 1986–1989, 1,135 men and 317 women were interviewed (Zheng et al. 2002); next of kin completed the questionnaires for 156 cases. Interviews were completed with 2,434 controls frequency-matched by sex and age group. A detailed occupational history was obtained for each job held for 5 years or more. Analyses were adjusted for smoking status, age, and family history of bladder cancer. On the basis of only seven exposed cases, men who had worked in the petroleum and coal-products industry or the petroleum-refining industry did not have an increased risk of bladder cancer. The estimated risk in garage and service-station workers was a bit high (OR 1.7, 95% CI 0.9–3.1). It was not possible to include the women in the analysis because of a lack of occupational exposure.

The European Merged Bladder Cancer Study (EMBCS) was a large effort that pooled data gathered at 11 European centers according to similar protocols in 1976–1996. There was one center each in Denmark and Greece; two each in France, Italy, and Spain; and three in Germany. Nine centers generated individual publications that considered employment in petroleum-related fields: Denmark (Jensen et al. 1987), France (1) (Clavel et al. 1994; Cordier et al. 1993), France (2) (Hours et al. 1994), Germany (1) (Claude et al. 1988; Kunze et al. 1992), Germany (3) (Pesch et al. 2000b), Greece (Rebelakos et al. 1985), Italy (1) (Vineis and Magnani 1985), Italy (2) (Porru et al. 1996), and Spain (1) (Gonzalez et al. 1989). The overall EMBCS sample was made

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

more uniform by limiting it to people 30–79 years old and to incident cases (only cases in those interviewed within 2 years of their diagnosis); this resulted in the exclusion of 755 male cases and 525 of their controls (Kogevinas et al. 2003) and 253 female cases and 357 of their controls (‘t Mannetje et al. 1999). To take advantage of the greater statistical power provided by the overall large sample, the committee decided to consider the findings in the pooled analyses on the 3,346 male cases and their 6,840 controls (Kogevinas et al. 2003) and on the 700 female cases and their 2,425 controls (‘t Mannetje et al. 1999). Occupations were recoded from the original work histories, and a JEM was applied to assess exposures to specific substances. On the basis of three exposed cases, the risk of bladder cancer was found not to be increased (when adjusted for age, smoking, and study center) in men ever employed as petroleum-refining workers (OR 0.52, 95% CI 0.10–2.69), the only fuel-related exposure presented; for the women, not even this occupational category was reported.

Reporting on fuel-related exposures was also quite sparse in publications concerning the individual study centers. In reporting on 658 male cases at the first French center (of whom 97 were excluded from the pooled study), Cordier et al. (1993) reported a heightened risk estimate (OR 4.04, 95% CI 0.78–21.03, based on seven exposed cases, p=0.10) in men who had worked in the petroleum-refining industry. In the first German center’s sample of 531 male cases (of whom 168 were excluded from the pooled study), Claude et al. (1988) reported no association between bladder cancer and having ever been employed as a service-station attendant (OR 0.33, 95% CI 0.04–2.87, based on one exposed case) or as an oil refinery worker (OR 1.50, 95% CI 0.25–8.87, based on three exposed cases). In the same German sample, however, Kunze et al. (1992) found that exposure to petroleum in the workplace (self-reported by 26% of the sample) was associated with an increased crude risk (OR 1.4, 95% CI 1.1–1.9); when adjusted for smoking, the risk increased (p for trend=0.01), although not monotonically, over duration of exposure. Despite its overall size, the pooled EMBCS sample analyzed by (Kogevinas et al. 2003) for men and by ‘t Mannetje et al. (1999) for women actually contains little information on fuel-related occupational exposures and bladder cancer.

Combustion Products

Table 4.37 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and bladder cancer.

Cohort Studies

Boffetta et al. (2001) used the Swedish Cancer Environment Registry developed by linking the Swedish Cancer Registry of all cancers diagnosed in 1971–1989 with records from the 1960 and 1970 Swedish censuses. According to industrial-hygiene criteria, each occupation and industry listed on the 1960 census was categorized for intensity and likelihood of exposure to diesel emissions. The rates of several cancers (including bladder cancer) in 1971–1989 in the diesel-exhaust cohort thus defined were compared with their incidences in the complementary non-diesel-exposed cohort. In both men and women, there was no indication of an association with bladder cancer.

The Netherlands Cohort Study was established to obtain detailed information on cancer risk factors, including occupational history, in a set of 58,279 men 55–69 years old in 1986 assembled from 204 municipal registries. Zeegers et al. (2001) conducted a nested case-control

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

study of this cohort by linking with cancer registries to identify 532 cohort members diagnosed through 1992 with urothelial cancer (including cancers of the ureter, renal pelvis, and urethra along with bladder cancers). A randomly selected subset of 1,630 men from the cohort (after elimination of those who had cancer other than skin cancer in 1986) served as controls. Industrial hygienists reviewed the occupational histories and derived cumulative estimates of exposure to PAHs and diesel exhaust and exposure to paints and aromatic amines. In the highest tertiles for both PAH and diesel-exhaust exposure, there were small increased risks (RR 1.18, 95% CI 0.62–2.24; RR 1.17, 95% CI 0.74–1.84, respectively). There was no evidence of dose-response relationships for bladder-cancer risk over the tertiles of PAH and diesel-exhaust exposure.

Case-Control Studies

As for fuel-related exposure, a number of case-control studies investigated the possibility of an association between exposure to combustion products and bladder or urothelial-cell cancers. Studies that approached exposure assessment only by addressing industry or occupation reported some suggestive findings, but they are too nonspecific as to exposure agent to be relied on in drawing a conclusion about a possible association with bladder cancer. For the transport occupations and industries related to driving and the railroads, there were reports of increased risk (Hoar and Hoover 1985; Howe et al. 1980; Iscovich et al. 1987) and reports of possibly less stable estimates of increased risk (Brownson et al. 1987; Decoufle and Stanislawczyk 1977; Notani et al. 1993; Risch et al. 1988). This situation was similar for mechanics. Brownson et al. (1987) had a positive finding, and several other researchers reported less certain increased risks of bladder cancer in those working in this field (Decoufle and Stanislawczyk 1977; Teschke et al. 1997). The findings on even these “core” combustion product occupations were not consistent across all studies, and they were found among a clutter of results for diverse types of ill-defined employment that might involve some exposure to combustion products.

NCI’s NBCS included 2,100 white men with bladder cancer at 10 US Surveillance, Epidemiology, and End Results centers whose exposure was categorized by industry and occupation (Silverman et al. 1989b). There were modest increases in risk among those who had ever been railroad workers (RR 1.3, 95% CI 0.9–2.0), mechanics (RR 1.2, 95% CI 1.0–1.4), or drivers (RR 1.2, 95% CI 1.1–1.4). In comparing the professional drivers among the white men with those who had never held an exhaust-related job, Silverman et al. (1986) found stronger results when the occupation was the subject’s “usual” occupation rather than merely one that he had “ever” held; dose-response relationships by duration of employment were demonstrated in those who had ever worked as truck or taxi drivers. Among nonwhite men (Silverman et al. 1989a), neither mechanics nor drivers had an increased risk, but an increased risk was observed in automobile mechanics (RR 1.4, 95% CI 0.4–4.4) and taxicab drivers or chauffers (RR 1.3, 95% CI 0.5–3.2). For white women (Silverman et al. 1990), the only occupation with exposure to combustion products that had enough subjects for analysis was motor-vehicle driver, for which the estimated risk was not meaningfully increased (RR 1.1, 95% CI 0.4–3.0). The size of this multicenter study compensates in part for its being limited to an analysis of occupations.

In the Iowa case-control study of bladder cancer (Zheng et al. 2002), men who had ever worked in 10 years or more in the railroad-transportation industry showed an increased risk of bladder cancer (OR 1.7, 95% CI 1.0–3.1). The “general automotive repair shops” industry and the occupation of mechanic or repairer also showed associations with bladder cancer in men, especially for those employed for more than 10 years (OR 3.4, 95% CI 1.3–9.0; OR 1.4, 95% CI 1.0–1.8, respectively). The estimated risks observed in men who had worked in the

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

transportation-services industry (OR 2.8, 95% CI 0.7–11.8) or held the job of driver (OR 1.3, 95% CI 0.9–1.8) were less certain. As for fuel exposure, there were no industries or occupations likely to have combustion-product exposure in which enough women had worked to merit analysis.

Considered from the perspective of individual industries or occupations, many classifications could involve exposure to combustion products, so a standard unifying approach for converting an entire work history into exposure (such as a JEM) would be extremely useful. Several studies reported equivocal increases in the risk of bladder cancer posed by exposure to diesel exhaust, either determined by self-reports (Howe et al. 1980; Risch et al. 1988) or derived from self-reports of “usual” occupation (Iyer et al. 1990). More-reliable risk estimates were generated in several studies, described below, that used more-comprehensive exposure-assessment methods informed by industrial hygienists’ understanding of the likelihood of exposure to particular substances in various jobs at different times.

Bonassi et al. (1989) contrasted the work histories of 121 men in Italy’s Bromida Valley who had been diagnosed with bladder cancer over a 10-year period with those of 342 age-matched community controls. Eleven occupations were categorized as having high risk on the basis of previous studies, and at least a year of work was required to be regarded as exposed. After adjustment for smoking, auto mechanic and truck driver had almost identical increased but imprecise risks. A JEM was used to partition the subjects into “definite”, “possible”, and unexposed classes for PAHs. After adjustment for smoking, the risk of bladder cancer was increased with possible PAH exposure (OR 1.63, 95% CI 0.95–2.83) and even more certain with definite PAH exposure (OR 2.20, 95% CI 1.12–4.38). Simultaneous adjustment for exposure to aromatic amines, however, slightly reduced the risk estimate for definite PAH exposure and diminished its precision (OR 2.14, 95% CI 0.82–5.60).

Steineck et al. (1990b) compared 254 Swedish men who had urothelial cancer (including an unspecified number of tumors of the renal epithelium) with 287 age-matched controls. An industrial hygienist estimated eight types of combustion product exposures on the basis of the subjects’ work histories: diesel or petrol exhaust and soot or combustion gases from coal, oil, or wood. The results (adjusted for age and smoking) were most suggestive for exposure to diesel exhaust (RR 1.7, 95% CI 0.9–3.3). There were patterns of increasing risk with greater exposure to petrol and diesel exhausts individually, and the effect appeared to be concentrated in the seven people who had been exposed moderately or highly to both agents (RR 7.1, 95% CI 0.9–58.8).

In the Montreal multicancer case-control study Siemiatycki et al. (1988) found no association between exposure to products of combustion of gasoline, diesel, jet fuels, coke, or liquid fuel and bladder cancer, but an association was seen between products of combustion of natural gas bladder cancer (OR 1.6, p<0.05). More-detailed analyses (Siemiatycki et al. 1994) yielded similar findings. For natural-gas combustion, all partitions of exposure scales (except duration) suggested a dose-response relationship, and the effect was most specifically associated with higher concentration. For the other industrial-hygiene-coded occupational exposures (to diesel exhaust, benzo[a]pyrene, or coal tar and pitch), no relationships with bladder cancer were found. The strongest association was for the motor-transport industry, but the effect was about the same whether employment had been for less than 10 years (OR 1.9, 95% CI 1.2–2.8) or for more than 10 years (OR 1.7, 95% CI 1.2–2.5).

The Swedish Cancer-Environment Registry (CER), which was used by Boffetta et al. (2001) as described above to define a diesel-exposure cohort, has also been the source for developing case-control studies. People employed in specific occupations as of the 1960 census

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

(and later the 1970 census) have been linked with incident cancers reported in the national cancer registry for 1961–1979 (and later for 1971–1989). Steineck et al. (1989) applied a JEM to 10,000 men whose bladder cancer was newly diagnosed in 1961–1979 and who had been reported as employed in 1960 census. They also reported results of 714 cancers of the renal epithelial tissues (cancers of the renal pelvis) separately, which would be expected to be included in the set of 824 cancers of the renal pelvis that McLaughlin et al. (1987) reported on by occupation in an article on this version of the CER, which also presented separate analyses on a set of renal-cell carcinomas (RCCs).

The original work histories from the 11 studies pooled in the EMBCS were coded according to a Finnish JEM aimed at estimating occupational exposures to PAHs and diesel exhaust. In contrasting the 3,346 male cases with their 6,840 controls, Kogevinas et al. (2003) found that each of the tertiles for PAH exposure had ORs greater than 1, demonstrating a dose-response relationship, culminating in an estimate of 1.23 for the high-exposure category (95% CI 1.07–1.40). Those who had ever worked as motor-vehicle drivers or mechanics showed slightly increased risks (adjusted for age, smoking, and study center); a subgroup of automobile mechanics had a more definitively increased risk (OR 1.38, 95% CI 1.02–1.87). The findings from the considerably smaller set of 700 female bladder-cancer subjects and their 2,425 controls were more limited (‘t Mannetje et al. 1999) because occupational exposures were minimal. Kogevinas et al. (2003) noted that higher risks were found in the earlier studies, perhaps implying an improvement in occupational exposure conditions in 1976–1996. The combination of those datasets into a unified analysis supports an association between exposure to the products combustion of petroleum-derived fuels and bladder cancer.

Some of the urothelial cases presented in Pesch et al. (2000b) were incorporated into the EMBCS (Kogevinas et al. 2003; ‘t Mannetje et al. 1999). However, nearly 30% of the 704 male cases and 50% of the 331 female cases had been excluded from the pooled analyses because they were prevalent cases or fell outside the age range. The publication was of particular interest to the committee because Pesch et al. used British, German, and their own task-defined JEMs to conduct detailed extractions of occupational exposure to combustion products. Again, there were not enough occupationally exposed women for the results to be precise. Only for “tar, pitch, and related products” among the men were all three JEMs applied; the British JEM appeared to be most liberal in attributing exposure (as was also the case for PAHs), and the task-based JEM most conservative. Nonetheless, the resulting risk estimates were similar in the two systems; the highest exposure category in each showed an effect to be likely (British JEM OR 1.6, 95% CI 1.1–2.3; task-based JEM OR 1.8, 95% CI 1.0–3.4). For PAHs, the British JEM, but not the task-based JEM, yielded increased risk estimates with a dose-response relationship, but for exhaust there was no clear indication of increased risk of bladder cancer. Where a contrast could be shown, the JEMs applied to subjects’ work histories inferred more exposure than did the self-assessments.

Similarly, with 97 case-control pairs eliminated for not meeting the eligibility criteria for the pooled study, the dataset of Clavel et al. (1994) was one of the two French studies incorporated into the EMBCS (Kogevinas et al. 2003). An extensive analysis of PAH exposure was performed for all 658 pairs of male bladder-cancer cases and hospital controls assembled and interviewed in 1984–1987. With adjustment for age, hospital, residence, ethnicity, smoking, and coffee consumption, exposure to PAHs was associated overall with bladder cancer (OR 1.3, 95% CI 1.0–1.7). Dose-response relationships were evident over most classifications of PAH exposure. For cumulative PAH exposure, however, the lowest category (containing almost 50%

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

of the exposed cases) had an increased risk (OR 1.7, 95% CI 1.2–2.4), almost as large as that for the much smaller highest exposure group (OR 1.8, 95% CI 0.8–3.9), whereas the intermediate cumulative exposure groups had lower estimated risks.

Additional Studies on Cancers of the Renal Pelvis

Of the analyses discussed above, Steineck et al. (1990b), Pesch et al. (2000b), and Zeegers et al. (2001) merged cancers of the renal epithelium with bladder cancers, and the others addressed only bladder cancers. The two studies discussed below conducted separate analyses of sets of cancers of the renal epithelium.

In conjunction with the bladder-cancer cases in Denmark that were pooled into the EMBCS (Kogevinas et al. 2003; ‘t Mannetje et al. 1999), Jensen et al. (1988) gathered a set of 96 cancers of the renal pelvis or ureter (none of the urethra) and hospital controls. A broad grouping (chemical, petrochemical, or plastics industry or exposure to gasoline or petroleum), which was related at least in part to fuel exposures, showed a risk for men and women combined. For the 60 male cases, increased smoking-adjusted risks were obtained for occupational exposure to coke or coal (RR 4.0, 95% CI 1.2–13.6) or to asphalt or tar (RR 5.5, 95% CI 1.6–19.6). Similarly, during the process of assembling the set of RCCs in Australia that were incorporated into the International Renal-Cell Cancer Study (IRCCS), McCredie and Stewart (1993) gathered 58 male and 89 female cases of renal pelvic cancer. With the exception of the PAH-related exposure related to having ever been employed in blast-furnace or coke-oven work, the exposures of interest are fuel-related. The age- and sex-adjusted risk associated with having been exposed to gasoline was increased, and the risks in two more categories were also increased: having ever been employed in the petroleum-refining industry (RR 2.97, 95% CI 1.10–8.02) and having ever been occupationally exposed to “other petroleum products”, which were said to include jet fuel, heating oil, kerosene, or diesel fuel (RR 2.16, 95% CI 1.44–4.08).

McLaughlin et al. (1987) and Steineck et al. (1989) reported on renal pelvic cancer cases diagnosed in 1961–1979 among men reported as employed in the 1960 census, as culled from the Swedish CER. The analysis of McLaughlin et al. (1987) was limited to industries and occupations listed in 1960 for 821 cases, whereas Steineck et al. (1989) performed a more rigorous exposure analysis by applying a JEM to this information for a set of about 100 fewer cases (probably because of the constraint that they were required to have been 20–64 years old in 1960). Some modest increases in risk were reported by both these record-linkage studies in association with exposures related to combustion products. Compared with the risk estimates associated with the same JEM-derived exposures on the much larger set of only bladder cancers (Steineck et al. 1989), only the estimates of an association with soot from oil, coal, or wood were somewhat higher (but still uncertain) for renal pelvic cancer; the risks of renal pelvic cancers posed by exposure to combustion gases from oil or coal were less than unity, whereas those of bladder cancer had suggested possible increases.

Conclusion

Although there was a suggestion of a relationship between fuel exposure and bladder cancer in some of the case-control studies that categorized exposure on the basis of expert review of job history, the relationship was not consistently increased in any study with detailed and specific exposure assessment. Associations were reported with work in petroleum-related industries and occupations in the NBCS (Silverman et al. 1989b), but these results were based on

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

broad occupational and industrial categories, and the true exposure to fuels is unknown. The large EMBCS, which pooled data from 11 European study centers, overall yielded a negative finding for petroleum refining as an occupation (Kogevinas et al. 2003). The individual northern German center (Kunze et al. 1992) reported an association with petroleum as an occupational exposure, but 32% of the cases in the sample were excluded from the pooled EMBCS because they were prevalent cases or fell outside the permissible age range. The committee concludes that the available epidemiologic data on exposure to fuels, which may be difficult to segregate entirely from exposure to combustion products in some studies, and bladder cancer is insufficient to conclude that there is a relationship between fuels and the occurrence of bladder cancer.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between fuels and bladder cancer.

Studies assessing the relationship between exposure to combustion products and bladder cancer have also not been consistently positive, and no studies assessed exposure on the basis of measurements. However, a pooled analysis of occupation and bladder cancer in western Europe (Kogevinas et al. 2003) included 3,346 men who had bladder cancer and 6,840 controls. With adjustment for smoking, questionably increased risks were noted in exhaust-related occupations. A JEM was applied to the occupational histories to derive measures of exposure to PAHs, benzopyrene, and diesel-engine exhaust. Exposure was evaluated as the product of the prevalence of exposure and the average magnitude of exposure in each occupation. The risk increased with higher exposures to PAHs (OR 1.23, 95% CI 1.07–1.4 in the highest tertile and benzopyrene (OR 1.27, 95% CI 1.04–1.54) in the highest tertile. A slightly increased risk was observed for diesel exhaust. In a related study by Pesch et al. (2000b), which included some of the cases pooled in the European study, similar findings were noted with some of the JEM-derived exposures to exhausts and PAHs. Clavel et al. (1994), in cases also included in the pooled European study, carried out a more detailed assessment of PAH exposures based on expert review of work-history information and found apparently stable associations with average and cumulative PAH exposures and total duration of PAH exposures. The results taken together constitute limited or suggestive evidence of an association between combustion products and bladder cancer, but the lack of exposure measurements and the heterogeneity of results precludes classifying this association as sufficient.

The committee concludes, from its assessment of the epidemiologic literature, that there is limited/suggestive evidence to determine whether an association exists between combustion products and bladder cancer.

KIDNEY CANCER

Over 90% of kidney cancers (ICD-9 189) in adults are renal-cell carcinomas (RCCs) or adenocarcinomas (ICD-9 189.0) (ACS 2004r). Most other malignant kidney tumors are transitional-cell carcinomas that arise in the renal pelvis (ICD-9 189.1), ureter (ICD-9 189.2), or urethra (ICD-9 189.3); these are jointly referred to as urothelial carcinomas or cancers of the renal pelvis. Because they resemble bladder tumors in their behavior and microscopically, the studies that assess risk factors for cancers of the renal pelvis separately were considered with bladder cancers in the previous section. In this section, if researchers specifically addressed only

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

RCCs, that is indicated; otherwise, cancers of the renal pelvis were included in a more global class of kidney cancers that would have been dominated by RCCs.

Smoking and obesity are the major risk factors for kidney cancer. Other risk factors are diet, increasing age, male sex, some hereditary conditions (such as Von Hippel-Lindau disease and hereditary papillary renal-cell carcinoma), and dialysis treatment for kidney disease. Such medications as phenacetin and diuretics (or the high blood pressure that they are used to treat) have also been associated with RCC, as has occupational exposure to asbestos, cadmium, and some organic solvents (ACS 2004r).

In 2000, there were 12.1 new cases of cancer of the renal pelvis and kidney per 100,000 (16.9 among men and 8.3 among women) and 4.2 deaths per 100,000 (6.2 among men and 2.8 among women) in the United States (Ries et al. 2004).

During the 1980s, it was noted that kidney tumors in male rats occurred after chronic gasoline exposure by inhalation. Considerable attention was directed toward monitoring human populations exposed to gasoline for similar tumors. Detailed toxicologic investigations later determined that the carcinogenicity in male rats was attributable to a protein peculiar to them, so gasoline would not be a human renal carcinogen, at least by the same mechanism.

Fuels

Table 4.38 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and kidney cancer. In the text below, the studies are discussed in roughly chronologic order, whereas more recent studies (with publications on a given population grouped) are generally presented first in the table’s sections by study design.

Cohort Studies

Wong et al. (1993) conducted a cohort mortality study in 18,135 US petroleum-distribution workers potentially exposed to gasoline. The assessment of exposure to hydrocarbons (HCs), both aromatic and aliphatic, was calibrated in parts per million for each job category in four periods (Smith et al. 1993). Several exposure variables (duration of exposure, time since first employment, job category, cumulative exposure, cumulative frequency of peak exposure, and year of first exposure) were analyzed separately for land-based and marine workers. The risk of kidney cancer was not increased overall (RR 0.73, 95% CI 0.47–1.09) nor did it show any sort of a dose-response relationship. A nested case-control study was conducted to investigate further whether this cancer might be associated with gasoline exposure in the land-based subpopulation of these distribution workers (Wong et al. 1999). The 12 cases of kidney cancer were each matched by age, sex, and company with up to five controls selected from the remainder of the cohort. Again, no relationship was found between kidney cancer and specific jobs ever held, duration of employment, duration of exposure, cumulative exposure, cumulative frequency of peak exposure, or year of first exposure.

A nested case-control study of whether kidney cancer is associated with exposure to HCs was conducted by identifying 100 cases of primary RCC in six petroleum-company cohorts at 36 refinery locations (Poole et al. 1993). The cases were identified by review of all 18,323 death certificates associated with the cohorts. For each case, four refinery-worker controls were selected who were alive and kidney-cancer-free at the time of the case’s diagnosis and who matched by employer, refinery location, and decade of birth. A team of industrial hygienists

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

assessed exposure by evaluating every job with respect to five classes of HCs associated with various refinery processes; they used three-point scales for intensity and for frequency (daily, weekly, and monthly) of exposure to each HC class. By apply the scales to the individual work-history records, they derived cumulative exposure scores for each HC class for every subject. RCC was not associated with having ever been exposed to any of the five specific petroleum-refinery HC classes or with cumulative exposure score or exposure duration.

Suggestive findings on kidney cancer in a cohort mortality study of workers at Exxon refineries in Louisiana, New Jersey, and Texas (Shallenberger et al. 1992) motivated conduct of a nested case-control study of kidney cancer diagnosed by 1990 among workers with at least 1 month of service in 1970–1982 (Gamble et al. 1996). A total or 37 confirmed RCCs or renal-cell adenocarcinomas occurred in 1970–1990; 32 of the patients had died by the time the study began. (Nine of the 32 cases, diagnosed at the Louisiana refinery in 1979–1992 and first reported by Hanis et al. (1982), were also included in the analysis of Poole et al. (1993) discussed above. Four controls per case were selected from among cohort members “at risk” (alive and cancer-free) on the date of the case’s diagnosis and frequency-matched on sex, race, and dates of birth and hire. A JEM constructed with the assistance of plant industrial hygienists was used to assess the potential intensity and duration of exposure to various classes of petroleum products, defined in terms of chain length, saturation, and whether aromatic. The cases had slightly higher mean cumulative exposure scores for the various HC categories. Workers exposed at the highest level had a 4- to 5-fold increase in risk of developing kidney cancer, but all the 95% CIs included the null value (with the lower limits ranging from 0.65 to 0.88 and the upper limits from 20.8 to 32.9 for various HC groups), and none of the HC categories showed a consistent dose-response relatoinship over quartiles of cumulative exposure. When analyses were adjusted for other risk factors—such as smoking, blood pressure, and body-mass index (BMI)—the risk estimates for each HC class generally increased, but the confidence intervals also widened. The risk associated with long tenure (over 38 years) was substantially increased but imprecise; the authors asserted that the trend over tenure was “close to significance” but gave no statistics.

Lewis et al. (2000b) found that the 41 kidney-cancer deaths observed in 1964–1994 among men who work for Imperial Oil in Canada any time in 1964–1983 were associated overall with a neutral SMR of 0.96 (95% CI 0.69–1.30); a somewhat increased risk was observed among distribution workers (SMR 1.14, 95% CI 0.64–1.88). Similarly, the 15 incident cases of kidney cancer diagnosed in 1969–1994 in the younger cohort of men who were hired in all sectors of the company in 1964–1994 (Lewis et al. 2003) corresponded closely to the incidence expected on the basis of rates in the general Canadian population (SIR 1.00, 95% CI 0.56–1.65).

Case-Control Studies

A population-based case-control study included 495 white residents (313 men and 182 women) of the Minneapolis-St. Paul standard metropolitan statistical area diagnosed with RCC in 1974–1979 and 714 controls selected from the general population matched on age and sex. Only exposures experienced before 1973 were analyzed. For 251 cases, the questionnaire eliciting occupational history and other possible risk characteristics had to be administered to proxies, so the “self-reports” of 25 suspected occupational exposures might be unreliable. McLaughlin et al. (1984) reported that exposure to “petroleum, tar, or pitch products” was the only one of them that showed an apparent association with renal cancer when adjusted only for age and smoking; the association remained high (OR for men 1.6, 95% CI 0.9–2.7; OR for women 4.6, 95% CI 0.4–51) when adjusted for all other risk factors considered (demographic,

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

medical, and dietary). In a later publication, McLaughlin et al. (1985) addressed petroleum-related occupational risks of the male subjects (313 cases and 428 controls) in greater detail. They controlled for potential confounding by age, smoking, BMI, drinking, country of ancestry, and phenacetin use. No trend in RCC was seen by years of employment in the petroleum occupations as a whole. In the subgroup of gasoline-station attendants, however, a modest trend with years of employment was observed.

In the Montreal multicancer case-control study, 181 men who had kidney cancer of any kind (ICD 189) were compared with the other 2,196 cancer cases remaining after the exclusion of lung and bladder cancers (Siemiatycki et al. 1987b). Siemiatycki et al. (1987a) addressed occupational exposures to petroleum-derived fuels, as coded from work histories. Adjusting for age, socioeconomic status, ethnicity, smoking, and blue- or white-collar job history led to imprecise estimates of increased risk posed by automotive gasoline, kerosene, diesel fuel, heating oil, and crude oil. Statistically significant screening results motivated more detailed logistic analysis of aviation gasoline (OR 3.1, 90% CI 1.5–6.5) and of jet fuel (OR 3.1, 90% CI 1.5–6.6); both showed increased risk, concentrated among six persons who had “substantial” exposure. The subset of 142 men who had RCC specifically was analyzed more intensely with respect to those two fuel exposures in comparison with population and cancer controls (Parent et al. 2000a). The adjusted ORs for RCC were 3.5 (95% CI 1.4–8.7) for jet fuel and 3.5 (95% CI 1.4–8.6) for aviation gasoline; and they were slightly higher for RCC than for kidney cancer in general.

Sharpe et al. (1989) conducted a case-control study of 164 male and female RCC cases diagnosed from January 1982 to June 1987 in nine hospitals in Montreal; they probably included many of the 181 male kidney-cancer (142 RCC) cases in the Montreal multicancer case-control study (Parent et al. 2000a, Siemiatycki et al. 1987a, Siemiatycki et al. 1988). In an effort to minimize recall bias, the control subjects were selected from patients who had hematuria but in whom urinary cancer had been ruled out. Having found no association between smoking history and RCC, the authors did not adjust other analyses for this factor. Self-reported occupational exposure to gasoline was associated with a slightly increased estimated risk of RCC (OR 1.09, 95% CI 0.55–2.15), and the risk estimate for kerosene exposure was higher but still imprecise (OR 2.04, 95% CI 0.69–6.28). Both findings are concordant with the results of Siemiatycki et al. (1987a) for those fuels.

Kadamani et al. (1989) gathered 210 RCC cases (142 male and 68 female) from 29 hospitals in Oklahoma City and Tulsa, Oklahoma that were newly diagnosed from July 1981 to August 1983. Controls matched by age and sex were recruited from the general population by random-digit dialing. Work histories were reviewed by two industrial hygienists who scored each job for its likely degree of HC exposure. From the job scores, a lifetime time-weighted average HC exposure index was derived for each subject; those who had any HC exposure were subdivided into low, moderate, and high groups. With adjustment only for weight, the women showed no increase in risk related to HC exposure (OR 0.7, 95% CI 0.3–1.4). An apparent increase in RCC (adjusted for weight and education) was found in the moderately exposed men (OR 2.7, 95% CI 1.2–6.5), but a clear increase was not observed among the highly exposed men (OR 1.6, 95% CI 0.7–3.6). Of the moderately exposed men, those under 60 years old were at the greatest risk for RCC (OR 3.0, p<0.05).

Asal et al. (1988a, 1988b) drew from the same source population with an additional six hospitals and 1 more year of observation to study 315 RCC cases. The cases were matched to both hospital and population controls. An analysis by industry with adjustment for age, smoking,

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

and weight found increases in the risk of RCC in both men (OR 4.3, 95% CI 1.7–10.9) and women (OR 1.6, 95% CI 0.4–6.5) who worked at least a year in petroleum refining and distribution (Asal et al. 1988a). In the full logistic model for RCC in males, the second-most predictive factor (after weight) was “petroleum work” (OR 6.6, 95% CI 2.3–19.2).

Partanen et al. (1991) interviewed 338 ever-employed people who had newly diagnosed RCC (the partition between males and females was not stated) in the Finnish Cancer Registry in 1977 and 1978. Each case had been matched by year of birth, sex, and survival status at the time of data-gathering to one or two responding controls drawn from the Population Register Centre, for a total of 484 controls. Job histories and other risk factors were obtained with mailed questionnaires completed by subjects or next of kin for deceased subjects. For both cases and controls, only about one-fourth of the subjects were alive. Industry and occupation were coded by international and Nordic standards, respectively, for each year from 1920 to 1968 (allowing a 10-year latent period). From that information, an industrial hygienist derived duration, magnitude, and cumulative amount of exposure to gasoline, diesel fuel (and other distilled fuel oils), PAHs, and six other agents. Exposed status was defined as at least 5 years of low-level exposure or at least 1 year of high-level exposure. There was good control for potential confounding by age, sex, obesity, smoking, and caffeine consumption. An increased OR for men and women combined was reported for gasoline (OR 1.72, 95% CI 1.03–2.87), but the increased estimate for “diesel fuels and other distilled fuel oils” was more uncertain (OR 1.20, 95% CI 0.63–2.27). ORs for men alone were slightly lower (1.63 and 1.15 for gasoline and diesel, respectively); 95% confidence limits included the null. The highest OR for men was for those exposed to gasoline without exposure to diesel (OR 2.05, 95% CI 1.05–3.98); the risk in men exposed to both categories of fuel was lower (OR 1.29), and the risk in men exposed only to diesel was not increased (OR 0.68). The authors concluded that they had found evidence of an association between exposure to gasoline and RCC.

Mandel et al. (1995) presented the overall results on occupational exposures and RCC from the IRCCS. This multicenter case-control study involved 1,732 RCC cases in men and women diagnosed in 1989–1991 in five countries. Controls were all population-based, but they were identified in different ways in the separate centers. The questionnaires used by the different study centers to determine exposures also varied somewhat in level of detail. The pooled analysis considered only occupations and exposures that were common to all centers, and it controlled for age, smoking, BMI, study center, and subjects’ level of education. Mandel et al. (1995) reported comprehensive adjusted RRs among all the men in the IRCCS (1,050 cases and 1,429 controls) for several types of exposure to petroleum: oil-refinery workers (RR 1.3, 95% CI 0.6–2.4), gas-station attendants (RR 1.3, 95% CI 0.9–1.9), and those who had self-reported exposure to gasoline (RR 1.6, 95% CI 1.2–2.0) or to jet fuel, heating oil, kerosene, or diesel fuel (RR 1.6, 95% CI 1.3–2.1, and a positive dose-response relationship). The risk of RCC was not associated with number of years worked in occupations involving gasoline exposure, and the gasoline-related risks were substantially diminished by adjusting for the other petroleum products noted above, so the authors concluded that the study was negative for an association between gasoline and RCC.

Pesch et al. (2000a) identified RCC cases newly diagnosed in 1991–1995 in five regions of Germany. The sample consisted of 570 male cases and 365 female cases interviewed in the hospital (representing a response rate of 88%) and 2,650 male and 1,648 female controls (71% response rate) matched on sex, age, and region and interviewed at home. In addition to applying both the British and German JEMs, the researchers derived a third matrix for exposure

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

estimation that incorporated information on tasks performed within jobs. For every subject, separate exposure indexes were derived for the exposure agents of interest on the basis of duration, probability, and intensity as characterized with each of the three JEMs. For analysis, the cases’ indexes were partitioned into four groups at the 30th, 60th, and 90th percentiles of the exposed controls’ distribution. For the occupational category “production and use of petroleum products”, which was said to include “transport and use of mineral oils and fuel”, the males were found to have slight increases in risk associated with the higher two categories of duration (OR 1.1, 95% CI 0.3–1.1; and OR 1.3, 95% CI 0.6–2.9, respectively).

Combustion Products

Table 4.39 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and kidney cancer.

Cohort Studies

As described in connection with bladder cancer, Boffetta et al. (2001) used the Swedish CER of all cancers diagnosed in 1971–1989 and records from the 1960 and 1970 Swedish censuses and applied industrial-hygiene criteria to occupations and industries listed on the 1960 census to define a cohort occupationally exposed to diesel emissions. The incidences of several cancers (including kidney cancer) in 1971–1989 in this cohort were compared with incidences in the complementary non-diesel-exposed cohort. In the small set (1,479) of women exposed to diesel exhaust, there was no indication of an association with the occurrence of kidney cancer. In the large sample (54,404) of men exposed to diesel exhaust, however, there was a slight but seemingly real increase in the risk of kidney cancer (SIR 1.06, 95% CI 1.02–1.11, adjusted only for age).

Case-Control Studies

McLaughlin et al. (1987) also used the Swedish CER to define a cohort of men employed in 1960. A search for them in the Swedish tumor registry in 1961–1979 revealed 7,405 RCC cases. There was an increased crude risk of RCC among men employed in the automobile-transportation industry in 1960 (OR 1.33, 95% CI 1.03–1.70, calculated by the committee from numbers in the article), but the overall ORs for industries and occupations with the potential for combustion-product exposure were not consistently increased.

Associations with several occupational exposures involving exhaust and combustion products were also investigated in the Montreal multicancer case-control dataset (Siemiatycki et al. 1988). No definitive associations were found between any of the agents and kidney cancer (in general). The most common exposure (to gasoline exhaust) yielded the strongest indication of increased risks (OR 1.2, 90% CI 0.9–1.4), but the pattern observed in connection with duration and intensity did not suggest a dose-response relationship. Jet-fuel exhaust showed the strongest association (OR 1.4, 90% CI 0.5–3.9). On the basis of the same four cases (which probably also constitute a subset of the people exposed to aviation gasoline and jet fuel), the reanalysis of the dataset for RCC specifically (Parent et al. 2000a) found a somewhat more intense risk for jet-fuel exhaust (OR 2.7, 95% CI 0.9–8.1). Parent et al. (2000a) reported that there was little to suggest an association between motor transport as an industry and RCC (OR 1.0, 95% CI 0.6–1.8) or between motor transport as an occupation and RCC (OR 1.1, 95% CI 0.7–1.8).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The study of 164 RCC cases in Montreal reported by Sharpe et al. (1989) appears to overlap with the Montreal multicancer study discussed above (Parent et al. 2000a; Siemiatycki et al. 1987a, 1988). There was a strong signal for exposure to PAH-laden tar or pitch (OR 9.29, 95% CI 1.16–74.20, p<0.02). The risks posed by occupational exposure to burning coke (OR 2.0, 95% CI 0.49–8.14) or burning coal (OR 2.54, 95% CI 0.96–6.99) were somewhat increased. When integrated with domestic exposure, however, a fairly strong dose-response relationship (p<0.025) was found for exposure to burning coal.

The Finnish population-based case-control study of RCCs of Partanen et al. (1991) reported mildly increased risks among men who worked for more than 5 years in the “transportation and storage” industry (OR 1.13, 95% CI 0.63–2.02) or in transportation occupations (OR 1.09, 95% CI 0.59–2.00) compared with those who never held such jobs in 1920–1968. A small increase in risk was observed among men who had been occupationally exposed to PAHs (OR 1.21, 95% CI 0.43–3.45). Those results provide some slight evidence of a combustion-product effect in association with RCC.

The finding in male gas-station attendants (RR 1.3, 95% CI 0.9–1.9) reported by Mandel et al. (1995) for the entire IRCCS, as discussed above in conjunction with fuels, could also be interpreted as evidence of a relationship between motor-vehicle exhaust and RCC. And the finding in men who had worked in the blast-furnace and coke-oven industry (RR 1.7, 95% CI 1.1–2.7) is evidence of a relationship between PAH exposure and RCC.

Mellemgaard et al. (1994) reported on occupational risk factors in the Danish subpopulation of the IRCCS, which comprised 368 RCC cases (226 men and 142 women) diagnosed in 1989–1992. Controls matched on sex and age in 5-year intervals were drawn from the Central Population Register. Exposures were determined with an interview conducted in a subject’s home; occupation was encoded according to the International Standard Classification of Occupation and Industry (ISCOI) by the Standard Industrial Classification. The analysis controlled for age, BMI, and smoking history. In addition to occupational risks similar to those reported on by Mandel et al. (1995) for the entire IRCCS, Danish men who were truck drivers for at least a year (at least 10 years before being interviewed) had an increased risk of RCC (OR 3.1, 95% CI 1.3–7.7).

Schlehofer et al. (1995) reported on the occupational risks observed for the 185 male RCC cases and 192 controls selected from among the residents of Heidelberg, Germany, the site of another participating center in the IRCCS. In addition to the exposure reported on by Mandel et al. (1995), this group reported an apparent increase in the risk of RCC in men who worked in jobs with exposure to exhaust gas for at least 5 years (RR 1.82, 95% CI 1.03–3.22).

The population-based case-control study of RCC in five regions of Germany (Pesch et al. 2000a) used three JEMs (including one that went down to the level of tasks within jobs) to generate estimates of considerably more combustion-related exposures than the single result reported above in the section on exposures to fuels. The grouping “railway brakemen, signalmen, and shunters” was among the seven occupations that had positive results (OR 6.2, 95% CI 1.6–23.4), but, in contrast with the other six positive occupations, too few women had the jobs to permit a comparison that might be informative about the stability of the result. The findings for the occupation of motor-vehicle driver were essentially negative for both men and women. There was a suggestion of a dose-response relationship for exposure of both men and women to “tar, pitch, mineral oil” in the British JEM, but only for women at the highest exposure was the increased risk marginally stable (OR 2.1, 95% CI 1.0–4.5). For PAHs, the British JEM indicated somewhat increased risks of RCC in both men and women but task-JEM-generated risk

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

estimates uniformly less than 1.0 in men (no results were presented for women). In several additional case-control studies, the risk of kidney cancer (or specifically RCC) was examined in relation to various exposures associated with combustion products without addressing in isolation any possible risks posed by fuels.

Several other case-control studies reviewed by the committee contained results of interest but were considered less reliable, largely because of the self-reported or less agent-specific nature of their exposure assessments. Exposures to a melange of possibly PAH-containing substances were reported by two of these additional case-control studies.

McLaughlin et al. (1984) conducted a traditional case-control study of RCC cases gathered in the Minneapolis-St. Paul area of Minnesota. They reported increased age- and smoking-adjusted risks in both women (OR 4.6, 95% CI 0.4–125.3) and men (OR 1.7, 95% CI 1.0–2.9) who had self-reported exposure to the rather nonspecific agent “petroleum, tar, and pitch products”. When adjusted for other potential confounding factors, the estimated risk in men exposed to the PAH-containing products was reduced (OR 1.6, 95% CI 0.9–2.7). The sample of working men was large enough to permit analysis by duration of exposure, which showed the increased risk to be concentrated among those who had 20 years or more of self-reported exposure (OR 2.6, 95% CI 1.2–5.7).

As part of Canada’s National Enhanced Cancer Surveillance System (NECSS), Hu et al. (2002) obtained completed questionnaires from 1,279 people (691 men and 588 women) who had newly diagnosed RCC in 1994–1997 in eight provinces (excluding Quebec). A total of 5,380 cancer-free controls (2,704 men and 2,676 women) were assembled from the provinces, frequency matched on age and sex to the overall distribution in the NECSS database of 18 cancer types. The grouping “coal tar, soot, pitch, creosote, asphalt” was one of 17 occupational exposures about which the respondents were asked. Adjusted for age, province, education, BMI, tobacco and alcohol use, and meat consumption, an association between this collection of PAH-bearing agents and RCC was evident in men (OR 1.4, 95% CI 1.1–1.8); in women, an estimated increase of similar magnitude was not nearly as certain (OR 1.3, 95% CI 0.7–2.3).

Four of the additional studies reported risks of renal cancer in association with having worked as a vehicle driver.

The male RCC subjects in the Oklahoma case-control study were also contrasted with their population controls by longest-held occupation with adjustment for age, smoking, and weight. Having a longest-held occupation of “operative”, a category noted to have a high prevalence of truck drivers, was found to be associated with the occurrence of RCC (OR 2.0, 95% CI 1.1–3.5) (Asal et al. 1988a).

Hospitals that report newly diagnosed cancers to the Missouri Cancer Registry follow a standardized protocol to gather information on smoking, alcohol use, and occupational history, but the details of the procedure were not reported. Brownson (1988) assembled the 205 white male RCC cases diagnosed from July 1984 to June 1986 and compared them with the 615 white men reported to have cancers not related to smoking (that is, excluding oral, pancreatic, laryngeal, lung, and bladder cancers). Analyses of usual occupation vs all other occupations—adjusted for age, smoking, and alcohol use—found increased risks in automobile mechanics (OR 1.8, 95% CI 0.4–2.0), machinists (OR 2.2, 95% CI 0.5–10.3), and drivers of heavy trucks (OR 3.1, 95% CI 1.1–8.5). The available data apparently were not detailed enough to permit investigation of a full spectrum of demographic and medical factors in conjunction with occupation or to pursue dose-response issues.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Auperin et al. (1994) matched 138 male RCC cases diagnosed in 10 French hospitals in 1987–1991 with 107 cancer controls and 128 noncancer controls, matching by sex, age, hospital, and interviewer. Complete occupational histories were blindly converted into ISCOI codes. After further adjustment for education, smoking, and the Quetelet index of height to weight before diagnosis, no increase in risk (OR 0.5, 95% CI 0.2–1.6) was found in men who had ever been transport-equipment operators, the only combustion-product-related occupation reported on, compared with the pooled cancer and noncancer hospital controls.

Delahunt et al. (1995) compared 710 men who had RCC entered with an active occupation code into the New Zealand Cancer Registry in 1978–1986 with 12,756 employed men who had non-urinary-tract cancers diagnosed during the same period. A crude increase in risk of RCC in firefighters (urban or forest application not specified) was intensified by adjusting for age and smoking status (RR 4.69, 95% CI 2.47–8.93). In transportation-equipment operators, however, there was no observed increase in risk (RR 0.91, 95% CI 0.63–1.32, unadjusted).

Conclusion

No key study that was positive for an association between exposure to fuels and kidney cancer was identified, but the uniformly negative results of the nested case-control study (Poole et al. 1993) of a comprehensive sample of RCC cases in the petroleum industry with excellent exposure assessment were compelling.

Although some studies of exposure to combustion products and kidney cancer suggested a possible association based on job title, the results were not consistently positive.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between fuels or combustion products and kidney cancer.

NON-HODGKIN’S LYMPHOMA

Non-Hodgkin’s lymphoma (NHL, ICD-9 200, 202) is a cancer originating in the B cells or, less frequently, the T cells of the lymphatic tissue (ACS 2004k). It encompasses the many types of lymphoma that remain after the exclusion of the B-cell lymphoma known as Hodgkin’s disease (HD) and is characterized by Reed-Sternberg cells (ICD-9 201). Within the evolving classification systems for lymphohematopoietic cancers overall, there have been a series of systems just for NHL. In ICD-9 coding, those cancers are subdivided within ICD-9 200 and ICD-9 202. The nonconstancy in the terminology and coding for reporting diagnoses of or deaths from this family of diseases complicates epidemiologic research on their etiology. Many risk factors have been identified for NHL: genetic or acquired defects of the immune system; infection by HIV, related T-cell viruses, or Epstein-Barr virus or by some bacteria (such as Heliobacter pylori); aging; obesity, the only recognized “lifestyle” factor; radiation; chemotherapy drugs; and possibly some chemicals, with benzene, herbicides, and insecticides most often implicated.

In 2000, there were 19.0 new cases of non-Hodgkin’s lymphoma per 100,000 (23.4 among men and 15.4 among women) and 8.2 deaths per 100,000 (10.3 among men and 6.7 among women) in the US (Ries et al., 2004).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Fuels

Table 4.40 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and NHL, presented in reverse chronologic order within type of study design.

Cohort Studies

In conjunction with a detailed exposure assessment of service-station attendants in Italy, Lagorio et al. (1993, 1994) tracked the mortality experience through 1992 of 2,308 men who had been service-station managers in 1980. With only three cases of death due to NHL, the estimated increase in mortality risk was imprecise (SMR 1.73, 90% CI 0.47–4.48).

Jarvholm et al. (1997) investigated cancer morbidity and mortality in a cohort of 4,128 male Swedish petroleum-industry workers by screening the Swedish cancer registry and death certificates. Statistics broken down by site were not presented for cancer mortality, and the incidence information was grouped for all lymphomas (ICD-9 200–202). Exposures were determined from job titles and supported by limited air monitoring of work areas and personal-exposure monitoring, which used benzene concentration as an index rather than a more global measure of HC concentrations. With nine lymphomas observed, the risk was not increased overall (SIR 0.93, 90% CI 0.48–1.6), and there was no suggestion of an increase with duration and latency.

A large cohort of Canadian workers for Imperial Oil did not show an increase in overall mortality from NHL through 1994 among men or women who had worked there any time from 1964 to 1983 (Lewis et al. 2000b). A slight increase among men in the marketing and distribution sector (SMR 1.12, 95% CI 0.65–1.79) showed no relationship to cumulative exposure to total HCs in a nested case-control study (Schnatter et al. 1996). Similarly, among the younger, overlapping cohort of 17,230 men in the entire company who had been first hired in 1964–1994, the incidence of NHL in 1969–1994 was not increased on the basis of 20 incident cases (Lewis et al. 2003).

On the basis of a positive finding for lymphohaematopoietic malignancies in a cohort-mortality study, Lewis et al. (2000a) and Huebner et al. (2000) followed up the incidence of such cancers among 8,942 employees who worked at an Exxon Corporation facility (a combined refinery and chemical plant) in Baton Rouge, Louisiana, in 1970–1992. Work histories from company payroll records were reviewed to assign workers to the occupation and unit where each spent most of his working time. Information on potential confounding factors was not consistently available. Cancers newly diagnosed in 1983–1994 were ascertained from the Louisiana cancer registry. There were 22 cases of NHL in men (SIR 1.06, 95% CI 0.67–1.61) with an indication that any excess occurred among those first employed before 1950 (SIR 1.44, 95% CI 0.85–2.27).

Case-Control Studies

In the Montreal multicancer case-control study of 206 men who had NHL, Siemiatiycki et al. (1987a) found no relationship between exposure to any of seven petroleum-derived fuels and this form of cancer.

Francheschi et al. (1989) interviewed 208 patients who had histologically confirmed NHL and were treated at a hospital in northeast Italy from June 1985 to March 1988. Their self-reported exposures were compared with those of 401 hospital controls who had acute conditions.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Adjusted for age and sex, the estimated risk of NHL in “petrochemical workers” (“chemical workers” were reported on separately) was somewhat increased (RR 1.83, 95% CI 0.87–3.84).

Blair et al. (1993) assembled white, male NHL cases diagnosed in 1980–1983 from the Iowa State Health Registry and from a surveillance network of hospitals in Minnesota. They interviewed 622 of them (or next of kin of 13% who were deceased) and 820 population controls (or next of kin) who were free of lymphohematopoietic cancer and matched by state, age, and year of death. With adjustment for age, state, smoking, family history of lymphoproliferative diseases, exposure to agricultural pesticides, hair-dye use, and response by next of kin, the risk associated with having worked in the petroleum-refining industry for at least 1 year was questionable (SIR 1.6, 95% CI 0.5–5.8).

Combustion Products

Table 4.41 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and NHL, presented in reverse chronologic order within type of study design.

Cohort Study

Boffetta et al. (1988) tracked the vital status of participants in the ACS II prospective cohort 2 years after enrollment, when detailed information that included a detailed occupational history had been gathered. With adjustment for age, smoking, and other occupational exposure, lymphomas considered together (ICD-9 200–202) showed no increase in risk (RR 0.92) associated with self-reported exposure to diesel-engine exhaust in men 40–79 years old at the time of enrollment.

Case-Control Studies

An early occupational case-control study was constructed by reviewing medical records of patients in 1956–1965 at Roswell Park Memorial Institute (Decoufle and Stanislawczyk 1977; Viadana et al. 1976). When all lymphomas were considered together, locomotive engineer or fireman was the only exhaust-related occupation that showed an intensified (but still imprecise) risk when limited to those with at least 5 years of exposure (RR 2.13).

The Montreal multicancer case-control study’s analysis of 206 men who had NHL showed only a modest increased risk posed by exposure to jet-fuel exhaust (OR 1.7, 95% CI 0.5–5.2) on the basis of four exposed cases (Siemiatycki et al. 1988).

The case-control study of NHL in Iowa and Minnesota (Blair et al. 1993) focused on agricultural exposure, but it also reported on some exposures related to combustion products derived by application of a JEM to each subject’s work history. The risk estimates for NHL associated with exposure to gasoline or diesel exhaust, to asphalt or creosote, or to oils or greases showed little deviation from unity; for each of these, the risk was slightly greater for high-intensity exposure.

Using the NECSS, Mao et al. (2000) identified histologically confirmed cases of NHL diagnosed in eight Canadian provinces in 1994–1997. Completed mailed questionnaires were received from 764 male cases, 705 female cases, and 5,073 cancer-free population controls that were frequency-matched for age and sex. The analyses of the self-reported exposures were adjusted for age, province, and BMI. Both the men (OR 1.2, 95% CI 0.9–1.5) and the women (OR 1.3, 95% CI 0.7–2.3) had somewhat increased estimated risks of NHL in association with

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

self-reported exposure to coal tar, soot, pitch, creosote, or asphalt. Exposure to mineral, cutting, or lubricating oils (which are often considered vehicles for PAH exposure) posed a marginally increased risk among the men (OR 1.3, 95% CI 1.0–1.5), but not among the women (OR 0.8, 95% CI 0.4–1.4).

In the study of 12 areas of Italy, the only exposure reported by Costantini et al. (2001) of possible relevance to a relationship of exposure to combustion products with NHL was having worked as a transport operator. The occurrence of NHL, also including chronic lymphocytic leukemia (CLL), in men was not related to having worked in this capacity (OR 0.9, 95% CI 0.7–1.3).

Several studies linking national censuses with tumor registries have given some suggestion of a relationship between the transport occupations reported in the census and NHL. In Sweden, the risk in male truck drivers was increased (SIR 1.4, p<0.05) (Linet et al. 1993), but the risk in women employed in the transport industry was not increased (Linet et al. 1994). A similar study in Denmark (Skov and Lynge 1991) found less-certain increases in both men (RR 1.12, 95% CI 0.91–1.38) and women (RR 1.24, 95% CI 0.40–2.90) who were unskilled transport workers.

Conclusion

Although some risk estimates were greater than unity, the reasonably well-conducted studies on NHL had no firmly positive findings. In the petroleum-refining and -distribution cohort with the most thorough quantitative exposure assessment, a nested case-control study found no indication of a dose-response relationship (Schnatter et al. 1996). Similarly, in the case-control study with the most objective exposure assessment, there was no indication of an association with any of the fuels (Siemiatycki et al. 1987a) or their combustion products (Siemiatycki et al. 1988).

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and NHL.

HODGKIN’S DISEASE

Hodgkin’s disease (HD, also called Hodgkin’s lymphoma) (ICD-9 201) is a cancer that originates in the lymphatic tissue (ACS 2004k). HD is a B-cell lymphoma characterized by microscopically identifiable Reed-Sternberg cells; all other cancers of the lymphatic tissues are called non-Hodgkin’s lymphoma. The only known risk factors for HD are infectious mononucleosis (caused by the Epstein-Barr virus) and lowered immunity. HD has not been associated with family history, diet, or environmental exposure.

In 2000, there were 2.7 new cases of Hodgkin’s disease per 100,000 people (3.3 among men and 2.3 among women) and 0.5 death per 100,000 (0.6 among men and 0.4 among women) in the United States (Ries et al. 2004).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Fuels

Table 4.42 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and HD, in reverse chronologic order within type of study design.

Cohort Studies

As noted above in the section on NHL, in their investigation of cancer morbidity among Swedish petroleum industry workers, Jarvholm et al. (1997) found no association between exposure to fuels and the incidence of all types of lymphoma combined.

A cohort study of mortality in 1949–1982 was conducted to evaluate cancer risk among 1,583 workers employed in an oil refinery near Milan, Italy, that converted crude oil into a variety of HCs (solvents, fuel, and lubricants) (Bertazzi et al. 1989). A later study extended followup to 1991 (for a total of 39,857 person-years and 352 deaths) (Consonni et al. 1999). The observation of two deaths from HD in 1949–1991 resulted in an increased but imprecise estimate of risk (SMR 1.51, 95% CI 0.17–5.44).

On the basis of a finding of lymphohematopoietic malignancies in a cohort mortality study (Lewis et al. 2000a), Huebner et al. (2000) followed up the incidence of such cancers among 8,942 employees working at an Exxon Corporation facility (a combined refinery and chemical plant) in Baton Rouge, Louisiana, in 1970–1992. Work histories from company payroll records were reviewed to assign each worker to the occupation and unit where he spent most of his working time. Information on potential confounding factors was not consistently available. A search of the Louisiana cancer registry for newly diagnosed cancers in 1983–1994 found four cases of HD among the men in this cohort (SIR 1.54, 95% CI 0.42–3.95), without any clear relationship to job title or year first employed.

Mortality and morbidity associated with HD have been assessed among the petroleum workers of Imperial Oil Limited in Canada (Hanis et al. 1979; Lewis et al. 2000b; Schnatter et al. 1993). In considering mortality from 1964 to 1994, Lewis et al. (2000b) found no increase in the risk of HD for men or women (SMR 0.68, 95% CI 0.28–1.41; SMR 0.79, 95% CI 0.02–4.42; respectively). There was no association between working in any particular division of the industry and HD. Lewis et al. (2003) studied cancer incidence in a younger, overlapping cohort of 17,230 men who had been first hired into any sector of the company in 1964–1994; on the basis of 11 incident cases, the incidence of HD in 1969–1994 was not notably increased (SIR 1.05, 95% CI 0.52–1.88).

Case-Control Study

In the Canadian multicancer case-control study, Fritschi and Siemiatiycki (1996b) reported no fuel-related results related to the small set of 54 HD cases.

Combustion Products

Table 4.43 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and HD.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×
Cohort Study

As noted above in the section on NHL, the 2-year followup of the ACS prospective cohort yielded negative findings on an association between self-reported diesel exposure and all lymphomas; a nonspecified number of HD deaths were included in that category (Boffetta et al. 1988).

Case-Control Studies

Similarly, the negative results related to combustion-product-related occupations of the case-control study of patients at the Roswell Park Memorial Institute concerned all lymphomas analyzed together (Decoufle and Stanislawczyk 1977; Viadana et al. 1976).

With only 54 male HD cases, the only finding from the Canadian multicancer case-control study that was at all related to combustion products was an imprecise increased risk posed by exposure to cooking fumes (Fritschi and Siemiatycki 1996b).

In the study of 12 areas of Italy (Costantini et al. 2001), the only reported exposure of possible relevance to the influence of combustion products on HD was the occupation of transport operator. The finding in men was entirely negative (OR 0.8, 95% CI 0.4–1.7).

Conclusion

Overall, the studies described here are limited by their small numbers of cases and the nonspecificity of their exposure assessments. The studies of workers at the Exxon chemical and refinery plant in Baton Rouge (Huebner et al. 2000) and of Italian refinery workers (Consonni et al. 1999) reported similar 50% increases in the estimated risk of HD after exposure to petroleum-related products, but both were imprecise, and neither showed a clear relationship to a specific job or to duration of employment.

In the case of combustion products, there is virtually no information beyond what is shared with NHL in analyses of all lymphomas together.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to fuels or combustion products and Hodgkin’s disease.

MULTIPLE MYELOMA

Multiple myeloma (ICD-9 203) is a type of cancer formed by malignant plasma cells (ACS 20041). The overgrowth of plasma cells can produce tumors in several sites, including the soft interior of the bone marrow. So far, few risk factors for multiple myeloma have been identified. Known risk factors include increased age, race, radiation exposure, family history, some plasma-cell diseases, and some occupational exposures.

In 2000, there were 5.5 new cases of multiple myeloma per 100,000 people (6.8 among men and 4.5 among women) and 3.8 deaths per 100,000 (4.7 among men and 3.3 among women) in the United States (Ries et al. 2004).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Fuels

Table 4.44 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and multiple myeloma (in reverse chronologic order within type of study design).

Cohort Studies

A retrospective cohort study examined mortality through 1989 in 9,026 petroleum-product distribution workers who worked for at least 1 year in land-based terminals for four US companies in 1946–1985 (Wong et al. 1993). Using available industrial-hygiene data, Smith et al. (1993) derived a detailed JEM in terms of total HCs for the tasks performed by distribution workers throughout this period and used it to estimate cumulative lifetime and peak exposures for the people in the cohort. Cases of myeloma were not noted separately, but there was no relationship to total HC exposure for the 18 deaths from cancers of “other lymphatic tissues” (ICD-8 203–203, 208) among land-based distribution workers. The 11 multiple-myeloma deaths in the group were carried into a nested case-control study (Wong et al. 1999), in which they were matched with as many as five controls who were still alive at the time of the cases’ deaths and did not die from myeloma, leukemia, or kidney cancer (the subjects of simultaneous nested investigations). The quantitative exposure information was used only to compare the mean values without measures of variance. The authors noted that the cases’ averages were “similar or slightly lower” compared with those of the controls; that was the case for duration of employment, duration of exposure, and cumulative exposure, but the cases’ mean peak exposure was 28% higher than the controls’. When specific jobs ever held were analyzed, the risk was highest in foremen and supervisors (OR 1.92, 95% CI 0.43–8.59).

On the basis of a positive finding for lymphohematopoietic malignancies in a cohort mortality study (Huebner et al. 2000; Lewis et al. 2000a) followed up the incidence of such cancers among 8,942 employees who worked at an Exxon Corporation facility (a combined refinery and chemical plant) in Baton Rouge in 1970–1992. Work histories from company payroll records were reviewed to assign each worker to the occupation and unit where he spent most of his working time. Information on potential confounding factors was not consistently available. Newly diagnosed cancers in 1983–1994 were ascertained from the Louisiana cancer registry. There were nine cases of multiple myeloma (SIR 1.39, 95% CI 0.64–2.64) without any obvious pattern related to length of work or job title. A formal case-control study was not done.

A series of epidemiologic studies (Hanis et al. 1979; Lewis et al. 2000b, 2003; Schnatter et al. 1992, 1993) have been conducted on workers for Imperial Oil in Canada. Schnatter et al. (1996) conducted a nested case-control study on seven deaths from multiple myeloma among male fuel-distribution workers in the Imperial Oil cohort with followup through 1983 (Schnatter et al. 1993). Four controls from the cohort were selected who were alive after a respective case’s death and matched on decade of birth. A panel of industrial hygienists estimated quantitative HC and benzene exposures by considering loading and unloading technology at each location during various periods, types of materials handled, typical tasks performed by workers and job title, typical environmental conditions, and historical industrial-hygiene surveys at some of the sites. There was no association between exposure to HCs or benzene and multiple myeloma.

The most recent update on the Canadian Imperial Oil cohort checked vital status through 1994 (Lewis et al. 2000b). All-causes mortality was lower than that in the general population, and there was no association between working in refineries and multiple myeloma (SMR 0.70,

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

95% CI 0.30–1.37). In marketing and distribution workers, the SMR was 1.94 (95% CI 1.11–3.15), with a more pronounced risk in workers who had 25–34 years of employment (SMR 3.06, 95% CI 1.47–5.63). In the younger, overlapping cohort of 17,230 men in the entire company who had been first hired in 1964–1994, the incidence of multiple myeloma between 1969 and 1994 was not increased, on the basis of three incident cases (Lewis et al. 2003).

Case-Control Studies

Incident cases of multiple myeloma in 1977–1981 were identified in NCI’s Surveillance, Epidemiology, and End Results tumor registries in four geographic areas (Washington state, Utah, Atlanta, and Detroit) (Demers et al. 1993; Morris et al. 1986). Controls were selected with random-digit dialing or similar methods. Interviews were conducted to gather information on 698 cases and 1,683 controls; for 221 deceased or severely ill cases the interview was completed with a proxy, whereas proxies were needed for only 1% of the controls. Respondent-selected chemical exposures were grouped into exposure categories by a toxicologist. After adjustment for sex, age, race, and geographic area, the risk of multiple myeloma did not exceed unity in those exposed to aliphatic HCs (including gasoline, diesel, and kerosene exposures), whether or not proxy responses were excluded from the sample. Demers et al. (1993) revisited the dataset’s work histories to determine risks associated with various occupations and industries. There was no association with multiple myeloma among those who had worked as gas-station attendants (OR 0.8, 95% CI 0.4–1.5). The only suggestion of a relationship with fuel exposure was seen in those said to have worked in the petroleum-refining and coal-product manufacturing industries (OR 1.2, 95% CI 0.4–3.1), but the increase vanished when the analysis was limited to self-respondents.

Linet et al. (1987) interviewed 100 cases of multiple myeloma diagnosed at seven Baltimore hospitals in 1975–1982. They were individually paired with hospital controls by sex, age, year of diagnosis, and hospital. Inquiries about occupational and environmental exposures were part of the telephone interviews conducted with each subject or next of kin. According to discordant-pair analysis with adjustment for whether the respondent was a subject or a proxy, exposure to petroleum was most strongly related with multiple myeloma (OR 3.7, 95% CI 1.3–10.3) of all the risk factors considered in the study (medical and pharmaceutical, as well as occupational). In a study this small, discordant-pair analysis is unwieldy for adjusting for confounders.

Men diagnosed with multiple myeloma (1,098 cases) in Denmark in 1970–1984 as listed in the Danish Cancer Registry were each matched to four controls (who were alive in the year of diagnosis) on the basis of sex and year of birth (Heineman et al. 1992). Work history since 1964 was obtained by linkage to the Danish Supplemental Pension Fund, and this information was abstracted by industrial hygienists to derive whether (and how long) each subject had been experienced to a variety of workplace exposures. Workers ever employed in the fuel, oil, or gas industry did not have an increased risk of multiple myeloma (OR 0.8, 95% CI 0.4–1.6). Among workers considered exposed to gasoline or oil products, the estimated ORs were consistently slightly increased. The same protocol generated data on 1,010 Danish women who had multiple myeloma and 4,040 population controls, but the analyses of interest were limited to the 363 cases and 1,517 controls who had been in the workforce at any time after 1964. Exposure to coal or oil products was the only exposure pertaining to fuels reported for the women, and there was no suggestion that it was related to multiple myeloma.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Combustion Products

Table 4.45 presents the findings considered most relevant by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and multiple myeloma, presented in reverse chronologic order within type of study design.

Nested Case-Control Studies

On the basis of self-reported exposure to diesel-engine exhaust, Boffetta et al. (1988) defined a cohort of men who were 40–79 years old at the time of enrollment in the ACS prospective cohort study. Persons who had an extant cancer diagnosis at the time of enrollment were not excluded. After 2 years of followup, comparison of multiple-myeloma mortality in the diesel-exhaust-exposed cohort with that in the complementary nonexposed cohort yielded a mildly increased estimated risk (RR 1.21). After 4 years of followup on the ACS cohort, Boffetta et al. (1989) conducted a more conventionally designed nested case-control study that considered only people who had been cancer-free at the time of enrollment. Exposure to combustion products was addressed in self-reports and the even less specific surrogate of main occupation. The risks posed by the three categories of combustion-product-related agents (coal tar, pitch, or asphalt; diesel exhaust; and gasoline exhaust) showed no consistent pattern. Fully adjusted risks of multiple myeloma were increased but imprecise in truck drivers (OR 2.8, 95% CI 0.5–16.1) but more emphatically increased in railroad workers (OR 7.1, 95% CI 1.2–43.6), although both estimates were based on only three exposed cases.

Wong et al. (1999) matched each of 11 multiple-myeloma deaths observed among 9,026 land-based petroleum-distribution workers with up to five controls on sex, year of birth, and company. None of the occupational groups (mechanics, drivers, and loaders) that might be expected to have job-related exposure to diesel exhaust had an increased rate of multiple myeloma. None of the available surrogates of dose (cumulative and peak HC exposures, duration of employment or of exposure, and year of first exposure) was found to be associated with multiple myeloma in logistic regressions.

Lee et al. (2003) contrasted the occupational exposure of 446 people who had primary incident multiple-myeloma cases in 1971–1999 in a cohort of 365,424 male Swedish construction workers with that of the remainder of the cohort. A JEM was developed on the basis of exposure monitoring conducted from 1971 to 1976 and used to determine each worker’s exposure to several substances, including diesel exhaust and asphalt. There was no association between asphalt exposure and multiple myeloma. With adjustment for age, BMI, and the other occupational exposures, the risk associated with diesel exhaust exposure was slightly increased (OR 1.3, 95% CI 1.00–1.77), but a dose-response relationship was not apparent over the three exposure levels used.

Case-Control Studies

In the set of 698 multiple-myeloma cases gathered from the SEER system in 1977–1981, working as a vehicle mechanic, the occupation most closely associated with combustion-product exposure, showed no association with multiple myeloma, with or without the proxy cases included (Demers et al. 1993). Self-reported exposures to diesel, jet-fuel, or automobile exhaust; coal fumes; and smoke were grouped as “carbon monoxide” (Morris et al. 1986). After adjustment for age, sex, race, and study site, the adjusted OR for all cases was 1.8 (95% CI 1.0–3.2); the reliability of this finding was increased by the fact that the result was a bit stronger

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

when proxy respondents were excluded from the analysis (OR 1.9, 95% CI 1.1–3.2). Williams et al. (1989) focused specifically on 69 “light-chain” multiple-myeloma cases in this case group, 46 of whom had completed the interview themselves. Adjusted for age, sex, race, residence, and educational attainment, the RR for all light-chain respondents was 2.9 (95% CI 1.0–8.4); for self-respondents, the relationship was considerably stronger (RR 6.1, 95% CI 2.0–18.2). When the light-chain multiple-myeloma cases were removed, the risk for all other multiple-myeloma cases became uncertain (OR 1.3, 95% CI 0.7–2.7) and led the authors to conclude that diesel exhaust was most strongly associated with the light-chain variety of multiple myeloma. However, there apparently has been no more research into this intriguing possibility.

Flodin et al. (1987) enrolled 131 cases (75 men and 56 women) of multiple myeloma identified at hospitals and clinics in six Swedish cities. They had been diagnosed in 1973–1983, but interviewing did not start until 1981, so short-term survivors were underrepresented; cross-checking with cancer registers suggested that the researchers had enlisted only about one-third of the cases occurring in the nominal study period. The 431 population-based controls, selected with no matching criteria beyond residence in the catchment areas, were the same people used in the authors’ case-control study of CLL (Flodin et al. 1988). After adjustment for smoking and other potential confounders in a Mantel-Haenszel analysis, the risk of multiple myeloma in men exposed to engine exhaust remained fairly firmly established (OR 2.1, 95% CI 1.2–3.9) on the basis of 35 exposed cases.

Heineman et al. (1992) identified 1,098 men diagnosed with multiple myeloma in 1970–1984 in the Danish Cancer Registry and drew 4,169 age- and sex-matched population controls from the Danish Central Population Registry. Work histories since 1964 were accessed from the Danish Supplementary Pension Fund, and from them industrial hygienists determined possible or probable exposure to various agents and duration of exposure. Ever having worked in the transportation industry (OR 1.3, 95% CI 1.0–1.6) was associated with an increased risk of multiple myeloma, but there was little evidence of a dose-response relationship (p for trend=0.24) when those who worked less than 5 years were compared with those who worked 5 years or more. No increased risk was associated with exposure to tar, asphalt, or soot. Exposure to engine exhaust was associated with a modest increase in estimated risk (OR for “possible” exposure 1.3, 95% CI 1.0–1.6; OR for “probable” exposure 1.2, 95% CI 0.9–1.6). Possible exposure to engine exhaust was the one exposure that retained borderline statistical significance when included in a logistic model simultaneously adjusted for exposure to gasoline, phthalates, and vinyl chloride. The exhaust findings are less convincing because “possible” exposure showed a stronger effect than “probable” exposure, and the results for duration also were contrary to dose-response expectations. A companion study of Danish women who had multiple myeloma analyzed 363 cases and 1,517 controls on whom work information was available from the files of the pension fund (Pottern et al. 1992). The increases in the risk of multiple myeloma associated with possible or probable exhaust exposure (OR 1.4, 95% CI 0.6–3.2; OR 1.6, 95% CI 0.4–5.5, respectively) were imprecise in this smaller sample.

In the study of 12 areas of Italy, the only exposure reported by Costantini et al. (2001) of possible relevance with respect to an association between exposure to combustion products and multiple myeloma was the occupation of transport operator. In men, the estimated risk was totally negative (OR 0.5, 95% CI 0.2–1.1).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Conclusion

No consistent relationship between exposure to fuels and multiple myeloma was detected in the studies described above; most studies reported no association. On the basis of small numbers of cases considered to have fuel exposure and the imprecision of the exposure assessments accompanying the case-control and cohort studies considered, there was a limited ability to detect an association between exposure to fuels and multiple myeloma.

Sizable studies of Swedish construction workers (Lee et al. 2003) and of Danish men (Heineman et al. 1992) both used industrial-hygiene methods to derive exposures from work histories, but their estimated risks of multiple myeloma after exposure to exhaust were just barely suggestive. Furthermore, the magnitude of the risks was fairly uniform, showing indications of a dose-response relationship with intensity, duration, or likelihood of exposure. The risk posed by carbon monoxide in the US SEER study (Morris et al. 1986), in which industrial hygienists grouped self-reports, was similarly marginal. Imprecise estimates of increased multiple-myeloma risk in association with exhaust exposure have been reported among Danish working women (Pottern et al. 1992) and in a study in which exposure was declared prospectively (Boffetta et al. 1989), whereas the most stable positive result, reported in a small study of Swedish men (Flodin et al. 1987), was based on self-reported exposure.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between fuels or their combustion products and multiple myeloma.

LEUKEMIAS

Leukemias [ICD-9 204–208] are malignant diseases that arise from precursor cells of white blood cells. As for the lymphomas, characterizing cases gathered retrospectively for epidemiologic studies and integrating the results of studies conducted over several decades is particularly challenging because a succession of diagnostic criteria, with corresponding groupings and nomenclature, have been used. Individual leukemias may have unique etiologic factors (for example, T-cell leukemia is caused by the virus HTLV-I), but the recognized risk factors for leukemias in general include exposure to radiation or certain chemicals (for instance, occupational exposure to benzene or chemotherapy with alkylating agents), some genetic conditions (such as some chromosomal abnormalities including, Down syndrome), and particular acquired blood diseases (for example, myelodysplastic syndromes may develop into acute myeloid leukemia) (NCI 2004).

Although all leukemias originate in the bone marrow, there are 4 main types, classified by the type and developmental stage of the cells involved. Leukemias can be either acute—in which the cells grow rapidly and are not able to mature—or chronic—in which the cells grow and accumulate slowly, and look mature. Classification also depends on which cell type is affected. Lymphocytic leukemias affect the lymphocytes, a type of white blood cell that makes up lymphoid tissue; myeloid leukemias affect granulocytes or monocytes, both of which are white blood cells that circulate and protect the body against infection (ACS 2004n). Acute lymphocytic leukemia (ALL) affects children more frequently than adults, while chronic lymphocytic leukemia (CLL) affects only adults, mostly over the age of 40 (ACS 2004e, 2004g). Acute myeloid leukemia (AML), also called acute non-lymphocytic leukemia (ANLL) is the

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

most common leukemia and usually affects adults, particularly men, although it can occur in children (ACS 2004f). Chronic myeloid leukemia (CML) mostly affects adults and is rare in children (ACS 2004h). These 4 types of leukemias can be further divided into sub-types, based on progression of the cancer and cell sub-types.

In 2000, there were 11.9 new cases of leukemia per 100,000 people (15.2 among men and 9.4 among women) and 12.4 deaths per 100,000 (15.9 among men and 9.8 among women) in the US (Ries et al. 2004).

Fuels

The second Gulf War committee found sufficient evidence to conclude that benzene is associated with leukemia, AML in particular. Benzene is a component of all the petroleum-derived fuels under consideration by the current committee and of their exhaust. For example, gasoline contains 0.5–2.5% benzene, and kerosene and the related jet fuels contain 0.1–1% benzene (Ritchie et al. 2003). This committee, therefore, did not revisit the literature related to the possibility of an association between fuels and leukemia.

Combustion Products

Table 4.46 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and leukemia, in reverse chronologic order within type of study design.

Cohort Study

Boffetta et al. (1988) tracked the vital status of participants in the ACS II prospective cohort 2 years after enrollment, when information that included detailed occupational histories had been gathered. After adjustment for age, smoking, and other occupational exposures, all leukemias considered together (ICD-9 204–208) had an imprecisely estimated risk (RR 1.29) associated with self-reported exposure to diesel-engine exhaust among men who were 40–79 years old at the time of enrollment. The deaths analyzed include those of persons newly diagnosed during the followup period and any occurring among those already diagnosed with leukemia at the time of the cohort’s inception; this introduces the possibility of recall bias into the exposure factor defining the cohort in this nominally prospective study.

Nested Case-Control Study

Wong et al. (1999) conducted a nested case-control study of leukemia in a cohort of US land-based petroleum-distribution workers. The entire group of 35 leukemias and the subgroup of 13 AMLs were analyzed for association with several relevant occupational subcategories (mechanics, drivers, and loaders). None of those occupational groups was associated with increased rates of leukemia or AML.

Case-Control Studies

The early case-control study of patients at the Roswell Park Memorial Institute yielded no clearly positive findings on all leukemias in association with occupations that had likely exposure to combustion products (Decoufle and Stanislawczyk 1977). The most increased risk estimates were in small sets of taxi drivers and mine workers (RR 2.08 and 2.16, respectively),

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

but when analysis was limited to those with at least 5 years in these jobs the risks were below unity.

Flodin et al. (1988) conducted a case-control study of CLL (ICD-9 204.15) by identifying cases at hospitals and clinics in five Swedish cities. The resulting 71 male and 40 female cases were compared with 431 population-based controls that were selected on the basis of no matching criteria beyond residence in the catchment areas. Most of the cases were diagnosed in 1975–1984, but some were survivors who had been diagnosed as early as 1964. There was a detailed assessment of radiation exposure, but it and smoking were found to have no influence on the occurrence of CLL. After adjustment for age, sex, farm work, and exposure to horses, DDT, fresh wood, or solvents, the authors found an association between self-reported occupational exposure to engine exhaust and CLL (Mantel-Haenszel incidence rate ratio 2.2, 95% CI 1.2–4.2). The authors noted that benzene is present in exhaust, but they speculated that other leukemogenic agents were involved because the benzene concentratios would be very low. Methodologic weaknesses in sample assembly and in assessment of exposures of interest to this committee reduce the confidence that can be placed in the result.

Lindquist et al. (1991) interviewed 125 patients (76 men and 49 women) who had acute leukemia, classified according to the French-American-British (FAB) system. They were compared with 125 neighborhood controls matched for age and sex. The data were analyzed as exposure-discordant matched pairs, thus forcing accounting for other possible confounders to be indirect. Professional drivers had an increased risk of acute leukemia (OR 3.0, 95% CI 1.1–9.2). The same effect was apparent in a smaller number of recreational drivers and was diminished only modestly when drivers who had also worked as painters were excluded. There was evidence of a dose-response relationship in that the risk was higher in those who had a greater duration of exposure (OR 5.0, p<0.05). The authors suggested that this finding might be attributable to the presence of benzene at about 5% in Swedish fuels.

The case-control study conducted by Costantini et al. (2001) interviewed 383 men and 269 women in 12 areas of Italy who were newly diagnosed with leukemia (ICD-9 204–208) in 1991–1993. The 1,779 controls were randomly selected from each of the regions and frequency-matched for sex and age. Work as a transport operator (as abstracted from the work history) was the only reported exposure possibly related to the influence of the combustion products on the occurrence of leukemia. Among men, the risk of that occupation was only modestly increased (OR 1.1, 95% CI 0.7–1.7). The sample was of substantial size, and the interview’s coverage was considerable, but there was no attempt to use the available information to adjust occupational exposures for possible confounders.

Conclusion

The studies address leukemias overall and subtypes (AML and CLL), which represent substantial heterogeneity in health outcome in a fairly limited set of evidence. The apparent associations with exposure are related to separate types of leukemia, and the authors of the studies (Flodin et al. 1988; Lindquist et al. 1991) note that any increase in leukemia risk is difficult to attribute specifically to exhaust because of concurrent exposure to fuels and benzene. The exposure assessments in all the studies are based on information from sources of questionable reliability (personal interviews or medical records) or have a low degree of specificity for combustion products.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between exposure to combustion products and leukemia.

MYELODYSPLASTIC SYNDROMES

Myelodysplastic syndromes (MDSs) are hematopoietic disorders of uncertain etiology in which the production of blood cells is compromised (ACS 2004a). Some of the syndromes are marked by high frequency of progression to AML. The FAB system of classification has recently been superseded by one from WHO. MDSs are fairly common sequelae of chemotherapy or radiation therapy, in which case they are regarded as secondary MDSs. Other risk factors thought to be related to the development of MDSs included aging, some genetic conditions, environmental or occupational exposure to radiation or chemicals (particularly benzene), and smoking.

Fuels

Table 4.47 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to fuels and MDSs, in reverse chronologic order.

Case-Control Studies

In a pilot study, Farrow et al. (1989) interviewed 39 men and 24 women who had an MDS diagnosed in a hospital in Wales from October 1985 through September 1986. Age- and sex-matched controls were recruited from the outpatient clinics on the same day. The researchers used an approach modeled after that used by Siemiatycki’s team in the Montreal multicancer case-control study (Gerin et al. 1985) to derive occupational exposures from the interview data. Preliminary adjusted analyses reported associations (p<0.01) with exposure for at least 6 months to “petrol diesel liquids” or to “petrol diesel fumes” (combustion exhaust may have been intended, rather than volatized fuel).

The same group of researchers (West et al. 1995) used the same recruitment procedures and exposure-assessment technique to gather 400 case-control pairs from three areas of the UK, which probably included the 63 pairs described in Farrow et al. (1989). Of 635 newly diagnosed primary MDS cases identified, only 400 were interviewed and paired with noncancer-patient controls matched for age, sex, residence, hospital, and year of diagnosis. The low ascertainment, primarily due to failure to interview subjects before it was no longer possible, is potentially biasing if an etiologic factor influenced the progression of the disease. Analysis was based on discordant pairs, and this limited adjustment to the matching factors. Two of the five classes of “petroleum products” (“diesels and petrols” and “oils and greases”) might be considered to fit within the exposures that the present committee is considering. The risk related to “oils and greases” was modestly increased (OR 1.29, 95% CI 0.88–1.89), and neither exposure showed any suggestion of a dose-response relationship with duration or intensity of exposure.

Another pilot study (Nisse et al. 1995) reported on the first 100 cases of MDS diagnosed at the University Hospital of Lille (from September 1991 through July 1993) and suggested associations both with “oils and greases” and with exhaust gases. Collection continued through

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

February 1996 for the full set of 204 newly diagnosed cases and 204 sex- and age-matched population controls (Nisse et al. 2001). Cases and controls were questioned by a trained interviewer to gather extensive demographic, medical, and occupational information. The resulting data were reviewed and distilled by a group of experts in occupational exposure to estimate lifetime cumulative exposure to various agents according to the approach of the Montreal multicancer case-control study (Gerin et al. 1985). The exposures were classified as exposed or nonexposed and regarded from the perspective of duration and frequency. The risk factors analyzed included gasoline, “oil”, exhaust gases, and PAHs. Cases were more likely to have been occupationally exposed to oil (OR 4.2, 95% CI 2.0–9.9), but the risk posed by petrol exposure was less precise (OR 2.5, 95% CI 0.9–7.7). Beyond control provided by the matching variables in a Mantel-Haenzel test, there was apparently no adjustment for possible confounders despite their availability in the detailed database developed.

Several other case-control studies of MDSs (Goldberg et al. 1990; Ido et al. 1996; Nagata et al. 1999; Rigolin et al. 1998) contained no analyses specifically of fuels (or combustion products) but might be considered indirectly related to fuels by virtue of their focus on solvents, primarily benzene, and were reviewed by the second Gulf War committee (IOM 2003). With the exception of the results of Goldberg et al. (1990), the risks associated with solvent exposures (all based on self-reports) were increased.

Combustion Products

Table 4.48 presents the most relevant findings considered by the committee in drawing its conclusion on the possibility of an association between exposure to combustion products and MDSs, in reverse chronologic order.

Case-Control Studies

For their 400 case-control pairs, West et al. (1995) reported that exposure to coal tar (a PAH-containing substance) showed no association with the occurrence of MDSs. The risks posed by exposure to exhaust gases showed minor increases, and there was not a monotonic increase in risk with increasing duration of exposure. For the only occupational grouping possibly characterized by exposure to combustion products (“transport operating, material moving and storing and related”), there was no association with MDSs.

In their case-control study of French subjects, Nisse et al. (2001) found that MDS cases were as likely as controls to have been exposed to exhaust gases (OR 1.0, 95% CI 0.5–1.9), but their estimated risk related to exposure to PAHs was only modestly increased (OR 1.8, 95% CI 0.7–4.6). Machine operators, an occupational group potentially exposed to combustion products, manifested a risk of MDS (OR 2.8, 95% CI 1.3–6.4). Logistic regressions were said to have been performed on variables that were associated with MDSs in univariate comparisons, but the fully adjusted results were not presented systematically.

Conclusion

Nisse et al. (2001) obtained positive findings for exposure to petroleum-related substances. Given the benzene content of fuels, one might have expected a clear picture to emerge, but it did not in the work of West et al. (1995).

For combustion products, only the increased risk estimate for the not particularly substance-specific occupational category of machine operator appeared stable (Nisse et al. 2001).

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Despite the apparent rigor of the data-collection methods used, the researchers’ analyses were rudimentary, and failed even to adjust for possible confounders when the information was at hand.

The committee concludes, from its assessment of the epidemiologic literature, that there is inadequate/insufficient evidence to determine whether an association exists between fuels or combustion products and myelodysplastic syndromes.

SUMMARY OF CONCLUSIONS

The committee’s conclusions about the strength of the evidence between fuels and combustion products and various types of cancers are summarized in Box 4.1.

BOX 4.1
Summary of Findings Regarding the Association Between Specific Cancers and Exposure to Fuels and Combustion Products

Sufficient Evidence of a Causal Relationship

No conclusions

Sufficient Evidence of an Association

Combustion products and lung cancer

Limited/Suggestive Evidence of an Association

Combustion products and:

Cancers of the oral cavity and oropharynx

Cancers of the nasal cavity and nasopharynx

Laryngeal cancer

Bladder cancer

Inadequate/Insufficient Evidence of an Association

Fuels and:

Cancers of the oral cavity and oropharynx

Cancers of the nasal cavity and nasopharynx

Esophageal cancer

Stomach cancer

Colon cancer

Rectal cancer

Hepatic cancer

Pancreatic cancer

Laryngeal cancer

Lung cancer

Melanoma

Non-melanoma skin cancer

Female breast cancer

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Male breast cancer

Female genital cancers (cervical, endometrial, uterine, and ovarian cancers)

Prostate cancer

Testicular cancer

Nervous system cancers

Kidney cancer

Bladder cancer

Hodgkin’s disease

Non-Hodgkin’s lymphoma

Multiple myeloma

Myelodysplastic syndromes

Combustion products and:

Esophageal cancer

Stomach cancer

Colon cancer

Rectal cancer

Hepatic cancer

Pancreatic cancer

Melanoma

Female breast cancer

Male breast cancer

Female genital cancers (cervical, endometrial, uterine, and ovarian cancers)

Prostate cancer

Testicular cancer

Nervous system cancers

Ocular melanoma

Kidney cancer

Non-Hodgkin’s lymphoma

Hodgkin’s disease

Multiple myeloma

Leukemia

Myelodysplastic syndromes

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLES

TABLE 4.1 Cancers of the Oral Cavity and Oropharynx and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Ritz 1999

Uranium-processing workers in Fernald, Ohio (ICD-8 140–149)

9

1.05 (0.48–1.99)

 

Kerosene exposure (industrial-hygiene reconstruction)

 

 

 

Low exposure (>2 yr, 15-yr lag)

na

1.85 (0.37–9.36)

 

Moderate exposure (>2 yr, 15-yr lag)

na

2.87 (0.43–19.2)

Lagorio et al. 1994

Filling-station attendants in Italy (exposure reconstruction using monitoring)

1

0.38 (0.02–1.79)a

Cohort Study—Incidence

Jarvholm et al. 1997

Petroleum-industry workers in Sweden (ICD-9 140–149) (qualitative industrial-hygiene-interpretation of personnel records)

 

 

 

≥1-yr duration; ≥1-yr latency

6

1.2 (0.54–2.5)a

≥1-yr duration; ≥20-yr latency

5

2.0 (0.79–4.2)a

≥10-yr duration; ≥20-yr latency

5

2.2 (0.90–4.8)a

Refinery operators

 

 

≥1-yr duration; ≥1-yr latency

1

0.58 (0.03–2.8)a

≥1-yr duration; ≥20-yr latency

1

1.1 (0.06–5.3)a

≥10 yr-duration; ≥20-yr latency

1

1.3 (0.06–5.9)a

Distribution workers

 

 

≥1-yr duration; ≥1-yr latency

3

1.8 (0.48–4.6)a

≥1-yr duration; ≥20-yr latency

2

2.3 (0.40–7.0)a

≥10-yr duration; ≥20-yr latency

2

2.5 (0.44–7.9)a

Case-Control Studies

Zheng et al. 1996

41 salivary gland cancer (ICD-9 142) cases among residents of Shanghai, China; self-reported agents (not smoking adjusted)

 

 

 

Occupational exposure to petroleum products

14

1.8 (0.8–3.8)

Kerosene exposure from cooking

13

3.5 (1.6–7.4)

With multivariate analysis

13

3.0 (1.4–6.8)

Gas exposure from cooking

28

1.3 (0.6–2.6)

Huebner et al. 1992

762 cancers of oral cavity and pharynx (ICD-9 141, 143–146, 148, 149) among male residents of California, Georgia, and New Jersey; self-reported occupation (smoking adjusted)

 

 

 

Petroleum-industry workers

16

1.79 (0.75–4.25)

Cancer of tongue

8

3.2 (1.15–8.9)

Cancer of mouth

1

0.41 (0.05–3.46)

Cancer of pharynx

7

2.31 (0.75–7.15)

Service-station workers

48

0.85 (0.53–1.36)

NOTE: na=not available.

a90% CIs reported in this paper.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.2 Cancers of the Oral Cavity and Oropharynx and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Studies—Adjusted for smoking, unless otherwise noted

Gustavsson et al. 1998

545 cancer cases among male residents of two regions in Sweden; industrial-hygiene-derived agents

 

 

 

Oral cavity

 

 

PAHs (low)

25

0.99 (0.57–1.73)

PAHs (high)

41

1.39 (0.86–2.25)

Pharynx

 

 

PAHs (low)

28

1.06 (0.61–1.82)

PAHs (high)

44

1.52 (0.94–2.45)

Pintos et al. 1998

784 pharyngeal-cancer cases among residents of three cities in southern Brazil; self-reports

 

 

 

Pharynx—use of wood stove

na

3.82 (1.96–7.42)

Males

 

2.82 (1.63–4.86)

Females

 

5.78 (0.52–64.3)

Mouth—use of wood stove

na

2.73 (1.76–4.24)

Males

 

2.52 (1.69–3.76)

Females

 

2.77 (1.09–7.02)

Zheng et al. 1996

41 salivary gland cancer cases among residents of urban Shanghai, China; self-reported agents (not smoking adjusted)

 

 

 

Fuel used for cooking

 

 

Kerosene

13

3.5 (1.6–7.4)

Coal

38

1.6 (0.5–5.6)

Gas

28

1.3 (0.6–2.6)

Wood or straw

6

1.6 (0.6–4.4)

Dietz et al. 1995

Incident cancer cases among residents of Heidelberg, Germany; self-reported agents

 

 

 

105 pharyngeal cancer cases

 

 

Air pollution on job (>20 yr)

21

0.94 (0.53–1.65)a

Traffic jams on way to work (>20 yr)

20

0.85 (0.48–1.50)a

High traffic emissions, residential (>20 yr)

23

0.74 (0.43–1.27)a

Outdoor air pollution, residential (>20 yr)

8

0.48 (0.20–1.07)a

Heating, fossil-fuel stoves (>40 yr)

33

2.60 (1.54–4.37)a

Cooking, fossil-fuel stoves (>20 yr)

24

1.41 (0.81–2.44)a

100 oral cavity cancer cases

 

 

Air pollution on the job (>20 yr)

22

1.09 (0.62–1.92)a

Traffic jams on way to work (>20 yr)

21

1.03 (0.58–1.82)a

High traffic emissions, residential (>20 yr)

34

1.41 (0.86–2.31)a

Outdoor air pollution, residential (>20 yr)

14

0.67 (0.35–1.29)a

Heating, fossil-fuel stoves (>40 yr)

33

1.60 (0.97–2.65)a

Cooking, fossil-fuel stoves (>20 yr)

34

2.09 (1.26–3.48)a

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Huebner et al. 1992

1,114 oral- and pharyngeal-cancer cases among residents of five US locations; self-reported job titles

 

 

 

762 men

 

 

Boiler or furnace operators

20

1.50 (0.68–3.34)

Heavy-equipment operators

55

1.25 (0.78–2.01)

Motor-vehicle operators

157

1.01 (0.75–1.35)

Railroad-transport workers

8

1.00 (0.30–3.35)

Cooks or other food service workers

83

1.00 (0.67–1.47)

Mechanics or repairers

207

0.86 (0.66–1.12)

Transportation worker (industry)

86

1.07 (0.74–1.56)

Trucking or warehousing workers (industry)

62

0.86 (0.56–1.31)

352 women

 

 

Motor-vehicle workers

7

2.80 (0.61–12.9)

Cooks or other food service workers

44

1.34 (0.78–2.28)

Zheng et al. 1992

204 oral- and pharyngeal-cancer cases among residents of urban Shanghai, China; self-reports (not smoking adjusted)

 

 

 

Men

 

 

Petroleum products, occupational exposure

53

1.76 (1.10–2.82)a

Use of kerosene stove

31

2.24 (1.27–3.97)a

Women

 

 

Petroleum products, occupational exposure

15

1.63 (0.72–3.72)a

Use of kerosene stove

13

1.01 (0.45–2.26)a

Merletti et al. 1991

86 oropharyngeal-cancer cases among residents of Turin, Italy

 

 

 

Transportation and communication (occupation) PAHs (JEM-derived agent)

8

0.7 (0.9–1.5)

Any exposure

56

1.0

Probable or definite exposure

20

0.6

Decoufle and Stanislawczyk 1977

Buccal-cavity and pharyngeal cancer cases among male patients at Roswell Park Memorial Institute in Buffalo, New York (job history, including durations, from medical charts)

 

 

 

Occupations (ever)

 

 

Deliverymen and routemen

na

1.16, ns

Machinists

na

1.36, ns

Mechanics and repairmen

na

0.98

Bus, taxicab, and truck drivers

na

1.37, ns

Locomotive engineers and firemen

na

1.64, ns

Viadana et al. 1976

Bus, taxicab, and truck drivers (not smoking adjusted)

89

1.44, ns

Exposed 5+ yr

68

1.33, ns

 

Locomotive engineers and firemen

18

1.64, ns

 

Exposed 5+ yr

17

1.97, ns

NOTE: na=not available; ns=not statistically significant (p<0.5) for a risk estimate above unity.

aUnadjusted ORs and 95% CI calculated with standard methods from observed numbers presented in original paper.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.3 Cancers of the Nasal Cavity and Nasopharynx and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Studies

Armstrong et al. 2000

282 nasopharyngeal-cancer cases among Chinese residents of Selangor and Federal Territory, Malaysia; self-reported exposures (smoking adjusted)

 

 

 

Motor fuel and oil

83

1.33 (0.81–2.20)

Teschke et al. 1997

48 sinonasal-cancer cases among residents of British Columbia, Canada; self-reported exposures (smoking adjusted)

 

 

 

Service-station attendants and managers

4

0.8 (0.2–2.8)

 

20-yr latency

4

1.0 (0.2–3.7)

TABLE 4.4 Cancers of the Nasal Cavity and Nasopharynx and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk

Case-Control Studies

Leclerc et al. 1997

930 sinonasal-cancer cases pooled from 12 studies in seven countries; job titles (not smoking-adjusted)

 

 

 

Motor-vehicle driver, males

41

1.13 (0.78–1.63)

 

<10 yr

 

1.15

 

≥10 yr

 

1.22

Armstrong et al. 2000

282 nasopharyngeal-cancer cases among Chinese residents of Selangor and Federal Territory, Malaysia; self-reported agents (smoking-adjusted)

 

 

 

Cooking fumes

19

0.93 (0.38–2.27)

10-fold exposure increase

 

0.97 (0.78–1.21)

Engine exhaust fumes

59

1.05 (0.61–1.79)

10-fold exposure increase

 

1.00 (0.88–1.14)

Wood fumes

27

1.65 (0.69–3.92)

10 fold exposure increase

 

1.08 (0.84–1.38)

Other fumes

29

1.46 (0.66–3.23)

10-fold exposure increase

 

1.12 (0.91–1.39)

Zheng et al. 1994

88 nasopharyngeal-cancer cases among residents of Zangwu County, China; self-reported agents (not smoking-adjusted)

 

 

 

Use of wood fuel (adjusted)

80

5.4 (1.5–19.8)

Use of wood fuel (crude, matched)

80

3.7 (p=0.02)

With windows in home

73

3.6 (p=0.03)

Without windows in home

7

7.8 (p=0.009)

With good ventilation in home

24

3.1 (p=0.07)

With poor ventilation in home

56

4.7 (p=0.01)

With kitchen in home

66

3.4 (p=0.04)

With kitchen outside in shack

14

5.9 (p=0.01)

With windows in kitchen

59

3.4 (p=0.08)

Without windows in home

7

5.4 (p=0.06)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk

Yu et al. 1990

306 nasopharyngeal cancer cases among residents of Guangzhou City, China; self-reported agents (not smoking-adjusted)

 

 

 

Combustion products (univariate)

63

2.4 (1.4–4.2)

1–9 yr

32

1.6 (0.9–2.9)

10+ yr

31

7.1 (2.5–20.6)

Combustion products (multivariate)

 

 

1–9 yr

32

1.8 (0.9–3.6)

10+ yr

31

9.0 (2.8–28.8)

TABLE 4.5 Esophageal Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Lewis et al. 2000b

Refinery workers in Toronto, Canada

 

 

 

Males

42

0.96 (0.69–1.29)

Refinery segment

20

1.10 (0.67–1.70)

Marketing or distribution segment

14

1.04 (0.57–1.74)

Females

3

1.54 (0.32–4.50)a

Ritz 1999

Uranium-processing workers in Fernald, Ohio—esophageal and stomach cancers (analyzed together)

 

 

 

Kerosene, light exposure

 

 

>2-yr duration, no lag

10

1.98 (0.77–5.09)

>2-yr duration, 15-yr lag

9

3.46 (1.22–9.80)

>5-yr duration, no lag

5

0.96 (0.32–2.94)

>5-yr duration, 15-yr lag

3

1.26 (0.31–5.15)

Kerosene, moderate exposure

 

 

>2-yr duration, no lag

5

3.00 (0.81–11.2)

>2-yr duration, 15-yr lag

5

7.71 (2.04–29.1)

>5-yr duration, no lag

4

2.86 (0.60–13.6)

>5-yr duration, 15-yr lag

4

10.7 (2.26–50.7)

Lagorio et al. 1994

Filling-station attendants in Italy (exposure reconstruction using monitoring)

4

2.34 (0.80–5.35)b

 

Men

4

2.41 (0.82–5.51)b

 

At smaller stations

4

3.51 (1.20–8.03)b

 

Women

0

0.0 (0.0–36.9)a

aRisk estimates and 95% CIs calculated with standard methods from observed and expected numbers presented in original paper.

b90% CIs presented.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.6 Esophageal Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Study—Mortality

Gustavsson et al. 1990

Male bus-garage workers in Stockholm, Sweden

 

 

 

Stockholm referent

4

1.93 (0.53–4.94)

 

Sweden referent

4

3.27 (0.89–8.37)

Cohort Study—Incidence

Chow et al. 1995

Esophageal-cancer cases among male residents of Sweden

 

 

 

Transport Industry

143

1.0 (0.84–1.18)a

Locomotive/traffic workers

118

1.1 (0.91–1.32)a

Machine, engine maintenance

46

1.0 (0.74–1.34)a

Case-Control Studies

Parent et al. 2000b

99 esophageal-cancer cases among male residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Nitrogen oxides (any)

21

0.9 (0.6–1.6)

Gasoline-engine emissions (any)

41

0.9 (0.6–1.5)

Carbon monoxide (any)

45

0.7 (0.4–1.1)

PAHs from any source (any)

64

0.9 (0.5–1.5)

Benzo[a]pyrene (any)

24

1.1 (0.7–1.9)

Benzo[a]pyrene (nonsubstantial)

19

1.0 (0.5–1.7)

Benzo[a]pyrene (substantial)

5

2.3 (0.8–6.5)

PAHs from coal (any)

10

1.2 (0.6–2.5)

PAHs from coal (nonsubstantial)

4

0.7 (0.2–2.1)

PAHs from coal (substantial)

6

2.0 (0.8–5.3)

PAHs from petroleum (any)

64

1.0 (0.6–1.6)

PAHs from other sources (any)

16

0.7 (0.4–1.2)

Mononuclear aromatic hydrocarbons (any)

29

0.8 (0.5–1.3)

Gustavsson et al. 1998

122 esophageal-cancer cases among male residents of two regions in Sweden; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

PAHs (low)

32

2.01 (1.16–3.48)

PAHs (high)

37

1.87 (1.11–3.16)

Siemiatycki et al. 1988

107 esophageal-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Gasoline exhaust

47

1.0 (0.8–1.4)b

Diesel exhaust

12

0.6 (0.4–0.9)b

Jet-fuel exhaust

2

2.5 (0.4–14.8)b

Propane exhaust

0

0.0 (0.0–1.7)b

Propane combustion

4

1.2 (0.5–3.0)b

Natural-gas combustion

3

0.8 (0.3–2.0)b

Liquid-fuel combustion

5

0.7 (0.3–1.4)b

Coal combustion

4

0.9 (0.4–1.9)b

Coke combustion

2

2.1 (0.5–9.8)b

Wood combustion

8

2.3 (1.2–4.5)

a95% CIs calculated with standard methods from risks and observed numbers presented in original paper.

b90% CIs reported in this paper.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.7 Stomach Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Lewis et al. 2000b

Refinery workers in Toronto, Canada

 

 

 

Males

83

0.70 (0.56–0.87)

Refinery segment

42

0.80 (0.57–1.08)

Marketing/distribution segment

31

0.86 (0.58–1.22)

Females

4

0.54 (0.15–1.38)

Ritz 1999

Uranium-processing workers in Fernald, Ohio—esophageal and stomach cancers (analyzed together)

 

 

 

Kerosene, light exposure

 

 

>2-yr duration, no lag

10

1.98 (0.77–5.09)

>2-yr duration, 15-yr lag

9

3.46 (1.22–9.80)

>5-yr duration, no lag

5

0.96 (0.32–2.94)

>5-yr duration, 15-yr lag

3

1.26 (0.31–5.15)

Kerosene, moderate exposure

 

 

>2-yr duration, no lag

5

3.00 (0.81–11.2)

>2-yr duration, 15-yr lag

5

7.71 (2.04–29.1)

>5-yr duration, no lag

4

2.86 (0.60–13.6)

>5-yr duration, 15-yr lag

4

10.7 (2.26–50.7)

Lagorio et al. 1994

Filling-station attendants in Italy (exposure reconstruction using monitoring)

6

0.60 (0.26–1.18)b

 

Men

6

0.64 (0.28–1.27)b

Women

0

0.0 (0.0–5.27)a

Nelson et al. 1987

Amoco Oil refinery cohort

 

 

 

Maintenance jobs

4

0.86

Operations jobs

9

2.06

Case-Control Study

Siemiatycki et al. 1987a

250 stomach cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Automotive gasoline

44

1.5 (1.2–1.9)b

Aviation gasoline

3

0.8 (0.3–2.7)b

Kerosene

24

1.7 (1.2–2.3)b

Jet fuel

1

0.2 (0.0–1.7)b

Diesel fuel

10

1.0 (0.6–1.6)b

Heating oil

15

1.4 (0.9–2.1)b

Crude oil

3

1.4 (0.4–5.0)b

aRisk estimates and 95% CIs calculated with standard methods from observed and expected numbers presented in original paper.

b90% CIs presented.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.8 Stomach Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Incidence

Chow et al. 1994

Male residents of Sweden

 

 

 

Industries

 

 

Transportation and communication

1162

1.0

Occupations

 

 

Transport and communication

1062

1.0

Machine and engine maintenance

348

1.0 (p<0.05)

Crane operators

66

1.5 (p<0.01)

Forklift operators

102

1.0

Dock workers

134

1.3 (p<0.05)

Warehouse workers

443

1.1 (p<0.05)

Kneller et al. 1990

Residents of Shanghai, China

 

 

 

Fuel suppliers

23

1.33

Petroleum-refinery workers

5

4.39 (p<0.05)

Boiler firemen

118

1.59 (p<0.01)

Case-Control Studies

Cocco et al. 1994

640 stomach-cancer cases among residents of Italy; occupation (not smoking-adjusted)

 

 

 

Mechanics, repairmen and allied

 

 

Ever worked

36

1.0 (0.6–1.5)

Worked 21+ yr

19

0.8 (0.5–1.5)

Railroad workers, drivers, and allied

 

 

Ever worked

51

1.0 (0.7–1.5)

Worked 21+ yr

30

1.3 (0.8–2.2)

Nitrogen oxides

 

 

Ever exposed

453

1.4 (1.0–2.1)

Exposed 21+ yr

na

0.7 (0.3–1.5)

Siemiatycki et al. 1988

250 stomach-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Gasoline exhaust

111

1.1 (0.9–1.3)a

Diesel exhaust

41

0.9 (0.7–1.1)a

Jet-fuel exhaust

2

0.8 (0.2–3.5)a

Propane exhaust

7

1.2 (0.6–2.2)a

Propane combustion

12

1.5 (0.9–2.4)a

Natural-gas combustion

11

1.3 (0.8–2.2)a

Liquid-fuel combustion

16

0.9 (0.6–1.3)a

Coal combustion

14

1.3 (0.8–2.0)a

Coke combustion

3

1.5 (0.4–5.8)a

Weinberg et al. 1985

178 stomach-cancer cases among residents of four counties in Pennsylvania; next-of-kin reports (not smoking-adjusted)

 

 

 

Coal heating

162

 

Digestive-cancer controls

 

1.69 (0.85–3.36)

ASHD controls

 

1.08 (0.49–2.37)

Neighborhood controls

 

1.20 (0.52–2.78)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Coal cooking

90

 

Digestive-cancer controls

 

1.00 (0.63–1.57)

ASHD controls

0.84 (0.55–1.27)

Neighborhood controls

0.97 (0.59–1.59)

Gas heating

125

 

Digestive-cancer controls

 

0.61 (0.36–1.02, p<0.08)

ASHD controls

0.62 (0.36–1.06)

Neighborhood controls

0.64 (0.33–1.24)

Gas cooking

129

 

Digestive-cancer controls

 

0.67 (0.40–1.12)

ASHD controls

0.49 (0.29–0.83, p<0.005)

Neighborhood controls

0.54 (0.28–1.06)

Decoufle and Stanislawczyk 1977

Stomach-cancer cases among patients at Roswell Park Memorial Institute in Buffalo, New York

 

 

 

Bus drivers

1

2.06

Taxicab drivers and chauffeurs

1

2.15

Truck and tractor drivers

8

1.56

Bus, taxicab, and truck drivers

5

0.89

Locomotive engineers and firemen

2

1.14

Machinists

2

0.92

Mechanics and repairmen

15

1.60

NOTE: na=not available; ASHD=arteriosclerotic heart disease.

a90% CIs reported in this paper

TABLE 4.9 Colon Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Lewis et al. 2000b

Imperial Oil workers in Canada—cancer of the large intestine except rectum (ICD 153)

 

 

 

Males

183

1.06 (0.91–1.22)

Refinery segment

65

0.88 (0.68–1.12)

Marketing-distribution segment

63

1.20 (0.92–1.53)

Females

19

0.88 (0.53–1.37)

Ritz 1999

Uranium-processing workers in Fernald, Ohio—colon and rectal cancers (analyzed together)

 

 

 

Kerosene, light exposure

 

 

>2-yr duration, no lag

10

1.13 (0.49–2.60)

>2-yr duration, 15-yr lag

9

1.20 (0.50–2.91)

>5-yr duration, no lag

9

1.26 (0.52–3.01)

>5-yr duration, 15-yr lag

7

1.40 (0.52–3.74)

Kerosene, moderate exposure

 

 

>2-yr duration, no lag

8

1.80 (0.64–5.06)

>2-yr duration, 15-yr lag

7

2.11 (0.75–5.97)

>5-yr duration, no lag

5

1.13 (0.31–4.18)

>5-yr duration, 15-yr lag

4

1.91 (0.50–7.27)

Nelson et al.

Amoco Oil refinery cohort—cancer of the large intestine

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

1987

 

 

 

 

Maintenance jobs

12

1.11

 

Operations jobs

12

1.19

Case-Control Studies

Gerhardsson de Verdier et al. 1992

352 colon-cancer cases among residents of Stockholm, Sweden; self-reported occupations and agents (not smoking-adjusted)

 

 

 

Petrol station and automotive repair (occupation)

12

1.8 (0.6–5.2)

Self-reported agents

 

 

Combustion gases from coal, coke, and wood

21

1.3 (0.6–2.5)

Soot

24

1.7 (0.8–3.4)

Tar and asphalt

8

0.7 (0.2–1.9)

Siemiatycki et al. 1987a

233 cases among residents of Montreal, Canada—industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Colon cancer

 

 

Automotive gasoline

39

1.0 (0.7–1.2)a

Aviation gasoline

7

1.7 (0.7–3.6)a

Kerosene

14

0.7 (0.5–1.1)a

Jet fuel

7

2.1 (0.9–5.1)a

Diesel fuel

10

0.7 (0.4–1.1)a

Heating oil

13

0.9 (0.6–1.5)a

Crude oil

3

1.5 (0.3–6.8)a

Rectosigmoid cancer

 

 

Automotive gasoline

25

0.9 (0.7–1.3)a

Aviation gasoline

3

0.8 (0.2–2.7)a

Kerosene

11

0.9 (0.5–1.4)a

Jet fuel

2

0.8 (0.2–3.8)a

Diesel fuel

4

0.4 (0.2–0.9)a

Heating oil

6

0.6 (0.3–1.2)a

Crude oil

0

Spiegelman and Wegman 1985

370 colon cancer cases in seven US metropolitan areas; fuel oil; JEM-derived agents (not smoking-adjusted)

 

 

 

Males: colorectal cancer

na

1.53 (p=0.01)

Colon cancer only

na

1.61 (p=0.02)

Females: colorectal cancer

na

1.24 (p=0.21)

Colon cancer only

na

1.34 (p=0.12)

NOTE: na=not available.

a90% CIs reported in this paper.

TABLE 4.10 Colon Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Studies

Goldberg et al. 2001

497 colon cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agent (smoking-adjusted)

 

 

 

Diesel-engine emissions

 

 

Nonsubstantial

45

1.2 (0.8–1.8)

Substantial

35

1.6 (1.0–2.5)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Gerhardsson de Verdier et al. 1992

352 colon-cancer cases among residents of Stockholm, Sweden; self-reported occupations, agents (not smoking-adjusted)

 

 

 

Railway work

5

1.4 (0.3–6.3)

Combustion gases from coal, coke, and wood

21

1.3 (0.6–2.5)

Soot

24

1.7 (0.8–3.4)

Tar and asphalt

8

0.7 (0.2–1.9)

Siemiatycki et al. 1988

364 colon-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Gasoline exhaust

147

1.0 (0.9–1.2)a

Diesel exhaust

68

1.3 (1.1–1.6)a

Short, low

6

0.7 (0.3–1.4)a

Short, high

5

0.6 (0.3–1.3)a

Long, low

27

1.5 (1.0–2.2)a

Long, high

30

1.7 (1.2–2.5)a

Jet-fuel exhaust

4

1.3 (0.4–4.2)a

Propane exhaust

7

0.9 (0.5–1.8)a

Propane combustion

12

1.0 (0.6–1.6)a

Natural-gas combustion

6

0.5 (0.2–0.9)a

Liquid-fuel combustion

19

0.9 (0.6–1.4)a

Coal combustion

8

0.5 (0.3–0.8)a

Coke combustion

2

0.7 (0.1–3.5)a

a90% CIs reported in this paper.

TABLE 4.11 Rectal Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Lewis et al. 2000b

Refinery workers in Toronto, Canada

 

 

 

Males

55

0.87 (0.65–1.13)

Refinery segment

30

1.09 (0.73–1.55)

Marketing and distribution segment

12

0.62 (0.32–1.08)

Females

3

0.60 (0.12–1.74)

Ritz 1999

Uranium-processing workers in Fernald, Ohio—colon and rectal cancers (analyzed together)

 

 

 

Kerosene, light exposure

 

 

>2-yr duration, no lag

10

1.13 (0.49–2.60)

>2-yr duration, 15-yr lag

9

1.20 (0.50–2.91)

>5-yr duration, no lag

9

1.26 (0.52–3.01)

>5-yr duration, 15-yr lag

7

1.40 (0.52–3.74)

Kerosene, moderate exposure

 

 

>2-yr duration, no lag

8

1.80 (0.64–5.06)

>2-yr duration, 15-yr lag

7

2.11 (0.75–5.97)

>5-yr duration, no lag

5

1.13 (0.31–4.18)

>5-yr duration, 15-yr lag

4

1.91 (0.50–7.27)

Nelson et al.

Amoco Oil refinery cohort

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

1987

 

 

 

 

Maintenance jobs

5

1.67

 

Operations jobs

5

1.78

Case-Control Studies

Gerhardsson de Verdier et al. 1992

217 rectal-cancer cases among residents of Stockholm, Sweden; self-reported agents (not smoking-adjusted)

 

 

 

Petrol station/automotive repair

7

1.5 (0.4–5.6)

Combustion gases from coal, coke, and wood

21

2.1 (1.0–4.6)

Soot

21

2.7 (1.2–5.7)

Tar and asphalt

7

1.0 (0.3–2.8)

Siemiatycki et al. 1987a

190 rectal-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Automotive gasoline

24

1.1 (0.8–1.6)a

Aviation gasoline

4

2.5 (0.6–10.3)a

Kerosene

11

0.9 (0.6–1.6)a

Jet fuel

4

2.1 (0.6–7.4)a

Diesel fuel

11

1.4 (0.8–2.5)a

Heating oil

11

1.5 (0.8–2.6)a

Spiegelman and Wegman 1985

Colon and rectal cancer cases in seven US metropolitan areas

 

 

 

Fuel oil: males; colorectal cancer

na

1.53 (p=0.01)

Colon-cancer only

na

1.61 (p=0.02)

Females: colorectal cancer

na

1.24 (p=0.21)

Colon-cancer only

na

1.34 (p=0.12)

NOTE: na=not available.

a90% CIs presented.

TABLE 4.12 Rectal Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Studies

Dumas et al. 2000

257 rectal-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Soot

 

 

Any

17

0.8 (0.5–1.3)

Substantial

4

0.8 (0.3–2.4)

Coal

 

 

Any

8

0.6 (0.3–1.3)

Substantial

6

1.3 (0.5–3.3)

Wood

 

 

Any

9

0.7 (0.3–1.4)

Substantial

4

1.0 (0.3–3.2)

Gerhardsson de Verdier et al. 1992

217 rectal-cancer cases among residents of Stockholm, Sweden; self-reported agents (not smoking-adjusted)

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Railway work

5

2.3 (0.5–10.4)

Combustion gases from coal, coke, and wood

21

2.1 (1.0–4.6)

Soot

21

2.7 (1.2–5.7)

Tar and asphalt

7

1.0 (0.3–2.8)

Siemiatycki et al. 1988

190 rectal-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Gasoline exhaust

89

1.2 (1.0–1.7)a

Short, low

17

1.3 (0.8–2.1)a

Short, high

9

1.0 (0.5–1.8)a

Long, low

27

1.1 (0.7–1.6)a

Long, high

36

1.6 (1.1–2.3)a

Diesel exhaust

35

1.1 (0.9–1.5)a

Jet-fuel exhaust

3

1.7 (0.4–7.8)a

Propane exhaust

3

0.7 (0.2–2.4)a

Propane combustion

10

1.1 (0.6–1.9)a

Natural-gas combustion

5

0.8 (0.3–1.8)a

Liquid-fuel combustion

14

1.1 (0.7–1.8)a

Coal combustion

7

0.7 (0.4–1.3)a

Coke combustion

0

0.0 (0.0–1.7)a

a90% CIs reported in this paper.

TABLE 4.13 Hepatic Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Study

Stemhagen et al. 1983

265 hepatic-cancer cases among residents of New Jersey; industries (not smoking-adjusted)

 

 

 

Motor vehicles and equipment manufacturing

11

2.20 (0.95–5.07)

Gasoline service stations

12

2.88 (1.20–6.88)

TABLE 4.14 Hepatic Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Study

Stemhagen et al. 1983

265 hepatic cancer cases among residents of New Jersey; industries (not smoking-adjusted)

 

 

 

Bus lines

7

2.80 (0.93–8.40)

Gasoline service stations

12

2.88 (1.20–6.88)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.15 Pancreatic Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies

Lewis et al. 2003

Petroleum workers in Canada

 

 

 

Males, mortality

4

0.51 (0.14–1.31)

Females, mortality

1

0.48 (0.01–2.69)a

Males, incidence

5

0.59 (0.19–1.39)

Females, incidence

0

0.00 (0.00–1.63)a

Ritz et al. 1999

Uranium-processing workers in Fernald, Ohio

 

 

 

Kerosene, light exposure

 

 

>5-yr duration, 15-yr lag

na

1.33 (0.31–5.66)

Kerosene, moderate exposure

 

 

>5-yr duration, 15-yr lag

na

2.78 (0.51–15.2)

NOTE: na=not available.

aUnadjusted risk estimates and 95% CIs calculated with standard methods from observed and expected numbers presented in original paper.

TABLE 4.16 Pancreatic Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Case-Control Studies

Alguacil et al. 2000

185 pancreatic-cancer cases among residents of eastern Spain (smoking-adjusted)

 

 

 

Industrial-hygiene assessment

 

 

PAHs

13

0.81 (0.37–1.76)

Low

10

0.74 (0.31–1.77)

High

3

1.11 (0.24–5.21)

Exposed ≥10-yr, 10 yr before diagnosis

 

 

Low

8

1.08 (0.39–3.00)

High

2

1.73 (0.22–13.8)

Finnish JEM assessment (substantial vs low exposure)

 

 

Benzo[a]pyrene

6

3.10 (0.73–13.2)

Diesel-engine exhaust

4

2.39 (0.50–11.4)

Gasoline-engine exhaust

4

2.42 (0.51–11.6)

PAHs

2

0.78 (0.12–5.18)

Kauppinen et al. 1995

595 pancreatic-cancer cases among residents of Finland; JEM-derived agents (smoking-adjusted)

 

 

 

Engine exhaust

19

0.89 (0.51–1.53)

PAHs

14

1.33 (0.69–2.57)

Siemiatycki et al. 1988

117 pancreatic-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Gasoline exhaust

47

0.9 (0.7–1.2)a

Diesel exhaust

15

0.6 (0.4–0.9)a

Jet-fuel exhaust

2

1.6 (0.3–7.7)a

Propane exhaust

2

0.7 (0.2–2.5)a

Propane combustion

3

0.8 (0.3–2.2)a

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Natural-gas combustion

6

1.5 (0.7–3.1)a

Liquid-fuel combustion

10

1.3 (0.8–2.3)a

Coal combustion

10

2.3 (1.4–4.0)a

Nonsubstantial

2

0.7 (0.2–2.4)a

Substantial

8

3.5 (1.7–7.3)a

Coke combustion

2

2.2 (0.4–10.5)a

NOTE: na=not available.

a90% CIs reported in this paper.

TABLE 4.17 Laryngeal Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Study—Mortality

Lagorio et al. 1994

Male filling-station attendants in Italy

 

 

 

Stations of all sizes

3

1.05 (0.29–2.72)a

Small stations

3

1.53 (0.42–3.96)a

Case-Control Studies

De Stefani et al. 1998

112 laryngeal-cancer cases among residents of Montevideo, Uruguay (smoking-adjusted)

 

 

 

Gasoline fillers (job title)

2

1.4 (0.2–7.7)

Gasoline (self-reported agent)

22

1.4 (0.8–2.6)

1–20 yr

6

1.1 (0.4–2.8)

21+ yr

16

1.7 (0.9–3.5)

Supraglottic

na

1.3 (0.5–3.4)

Glottic

na

2.5 (0.8–8.2)

Wortley et al. 1992

235 laryngeal-cancer cases among residents of 13 counties in western Washington; occupation (smoking-adjusted)

 

 

 

Vehicle mechanics

32

1.2 (0.6–2.1)

Garage and service-station work

12

0.8 (0.4–1.8)

Ahrens et al. 1991

100 laryngeal-cancer cases among residents of Bremen, Germany; self-reported agents (smoking-adjusted)

 

 

 

Diesel oil

na

1.7 (0.8–3.5)

Gasoline

na

2.8 (1.0–7.7)

Brown et al. 1988

183 laryngeal-cancer cases among residents along Gulf Coast of Texas; industry (smoking-adjusted)

 

 

 

Petroleum refining and chemical manufacturing (ever vs never)

47

0.93 (0.59–1.46)

NOTE: na=not available.

a90% CIs reported.

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

TABLE 4.18 Laryngeal Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Study—Incidence

Boffetta et al. 2001

Cohort exposed to diesel exhausted defined on industrial-hygiene coding of occupation on 1960 Swedish census (not smoking-adjusted)

 

 

 

Males

730

1.09 (1.01–1.17)

Probability of exposure

 

 

Low

272

1.1 (0.99–1.27)

Medium

218

0.92 (0.80–1.05)

High

204

1.1 (0.96–1.28)

Intensity of exposure

 

 

Low

473

1.1 (0.95–1.17)

Medium

127

1.1 (0.88–1.26)

High

94

0.99 (0.81–1.22)

Females

5

2.39 (0.78–5.57)

Case-Control Studies

1,010 cases of laryngeal cancer (ICD-9 161) or hypopharyngeal cancer (ICD-9 146.4, 146.5, 148, 149.8) among male residents of France, Italy, Spain, and Switzerland (IARC six-center study) (smoking-adjusted)

 

 

Boffetta et al. 2003

Railway transport (ever vs never) (industry)

44

1.52 (0.97–2.39)

 

Duration

 

p for trend=0.02

1–10 yr

11

1.1

11–20 yr

5

0.3

21+ yr

28

2.2 (p<0.05)

Occupations (ever vs never)

 

 

Motor-vehicle mechanics

17

1.09 (0.55–2.16)

Other mechanic

42

1.39 (0.87–2.23)

Railway-vehicle loaders

14

1.39 (0.61–3.15)

Lorry drivers, local

42

1.14 (0.73–1.79)

Lorry drivers, long-distance

37

1.28 (0.78–2.10)

Other motor-vehicle drivers

12

1.32 (0.58–3.03)

Berrino et al. 2003

695 cases ≥55 yr old

 

 

 

PAHs (JEM-derived agent)

na

1.0 (0.7–1.3)

315 cases <55 yr old

 

 

PAHs (JEM-derived agent)

263

0.7 (0.3–1.4)

Likelihood of exposure

 

 

Possible

107

0.5 (0.2–1.1)

Probable

156

0.8 (0.3–1.8)

Duration of exposure

 

 

<10 yr

na

0.6 (0.2–1.8)

10–19 yr

na

0.9 (0.4–2.1)

≥20 yr

na

1.1 (0.5–2.4)

Elci et al. 2003

940 laryngeal-cancer cases among male residents of Istanbul, Turkey (smoking-adjusted)

 

 

 

Diesel exhaust (JEM-derived agent)

297

1.5 (1.3–1.9)

Intensity of exposure

 

 

Low

161

1.5 (1.1–1.8)

Medium

91

1.7 (1.2–2.3)

High

45

1.6 (1.0–2.4)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Probability of exposure

 

 

Low

92

1.6 (1.2–2.2)

Medium

148

1.5 (1.1–1.9)

High

57

1.6 (1.1–2.4)

Gasoline exhaust (JEM-derived agent)

220

1.6 (1.3–2.0)

Intensity of exposure

 

 

Low

141

1.5 (1.2–2.0)

Medium

78

1.8 (1.3–2.5)

Probability of exposure

 

 

Low

86

1.6 (1.1–2.2)

Medium

131

1.7 (1.3–2.2)

High

3

0.7 (0.2–2.9)

PAHs (JEM-derived agent)

376

1.3 (1.1–1.6)

Intensity of exposure

 

 

Low

189

1.4 (1.1–1.7)

Medium

138

1.3 (1.0–1.6)

High

49

1.5 (1.0–2.2)

Probability of exposure

 

 

Low

106

1.4 (1.0–1.8)

Medium

176

1.4 (1.1–1.7)

High

94

1.3 (1.0–1.7)

Elci et al. 2001

Occupations

 

 

 

Drivers

75

1.7 (1.1–2.4)

Glottis

22

2.2 (1.3–3.8)

Supraglottis

27

1.2 (0.7–1.9)

Other

26

1.8 (1.1–3.1)

Mechanics

28

0.8 (0.5–1.3)

De Stefani et al. 1998

112 laryngeal-cancer cases among male residents of Montevideo, Uruguay (smoking-adjusted)

 

 

 

Automobile mechanic (job title)

8

1.1 (0.5–2.8)

Supraglottic

na

1.3 (0.3–5.9)

Glottic

na

5.5 (1.3–23.5)

Truck driver (job title)

8

0.8 (0.3–1.8)

Supraglottic

na

0.6 (0.1–2.9)

Glottic

na

2.7 (0.7–10.7)

Tractor driver (job title)

7

0.6 (0.2–1.5)

Gasoline fillers (job title)

2

1.4 (0.2–7.7)

Diesel exhausts (self-reported agent)

17

0.8 (0.4–1.4)

1–20 yr

3

0.3 (0.1–0.9)

20+ yr

14

1.4 (0.7–2.8)

Supraglottic

na

0.7 (0.2–1.9)

Glottic

na

1.9 (0.6–5.8)

Gasoline exhausts (self-reported agent)

22

0.9 (0.6–1.8)

1–20 yr

7

0.8 (0.3–1.9)

20+ yr

15

1.2 (0.6–2.3)

Supraglottic

na

0.8 (0.3–2.1)

Glottic

na

1.8 (0.6–5.7)

Gustavsson et al. 1998

Laryngeal-cancer cases among male residents of two regions in Sweden

 

 

 

PAHs (low)

26

0.77 (0.46–1.28)

PAHs (high)

53

1.47 (0.96–2.24)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Pintos et al. 1998

Laryngeal-cancer cases among residents of three cities in southern Brazil

 

 

 

Use of wood stove

na

2.34 (1.17–4.67)

Males

 

2.03 (1.12–3.67)

Females

 

16.24 (2.66–99.1)

Goldberg et al. 1997

Laryngeal cancer cases among male residents of France

 

 

 

Industry

 

 

Railway transportation

30

1.4 (0.6–3.1)

Road transportation

31

1.0 (0.4–2.1)

Other transportation

16

0.8 (0.3–2.2)

Repair of motor vehicles

21

0.9 (0.4–2.0)

Occupation

 

 

Motor-vehicle mechanics

27

1.2 (0.5–2.5)

Transport-equipment operators

118

1.4 (0.9–2.3)

Maier and Tisch 1997

Laryngeal-cancer cases among residents of Heidelberg, Germany

na

 

 

PAH

 

2.7 (1.2–6.1)

Fossil-fuel single stove

 

 

0–20 yr

 

1.0

20–40 yr

 

1.2 (0.8–1.9)

>40 yr

 

2.5 (1.5–4.1)

Dietz et al. 1995

Laryngeal-cancer cases among residents of Heidelberg, Germany

 

 

 

Air pollution on job (>20 yr)

34

1.44 (0.91–2.26)a

Traffic jams on way to work (>20 yr)

23

0.69 (0.41–1.14)a

High traffic emissions, residential (>20 yr)

39

33.8 (13.34–90.85)a

Outdoor air pollution, residential (>20 yr)

21

1.00 (0.58–1.71)a

Heating, fossil-fuel stoves (>40 yr)

57

2.11 (1.43–3.12)a

Cooking, fossil-fuel stoves (>20 yr)

33

1.47 (0.92–2.33)a

Coal, briquette, or coke heating

138

1.52 (0.94–2.47)a

Gas heating

16

0.52 (0.29–0.93)a

Muscat and Wynder 1995

Laryngeal cancer cases among residents of New York, Illinois, Michigan, and Pennsylvania

 

 

 

Truck driver

30

1.22 (0.65–2.28)a

Diesel-exhaust jobs

36

0.96 (0.5–1.8)

Automobile mechanics

13

130 (0.4–4.1)

Diesel exhaust

13

1.47 (0.5–4.1)

Diesel fumes

16

6.4 (1.8–22.6)

Pollan and Lopez-Abente 1995

Laryngeal-cancer cases among male residents of Madrid, Spain

 

 

 

Mechanics and assembly workers

8

1.56 (0.49–4.98)

1–20 yr

4

1.21 (0.25–5.82)

>20 yr

4

3.04 (0.52–17.71)

Transport drivers

8

2.71 (0.85–8.64)

1–20 yr

3

5.76 (0.71–46.48)

>20 yr

4

1.96 (0.43–8.96)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Wortley et al. 1992

Laryngeal-cancer cases among residents of western Washington state

 

 

 

Vehicle mechanics

32

1.2 (0.6–2.1)

<10 yr

 

0.6

≥10 yr

3.2

Motor vehicle operators

54

1.3 (0.8–2.1)

<10 yr

 

1.6

≥10 yr

0.8

Transportation, motor vehicles

19

1.3 (0.6–2.8)

<10 yr

 

1.1

≥10 yr

0.8

Diesel fumes—peak

 

 

None

112

1.0

Low

58

1.2 (0.7–1.9)

Medium

65

1.1 (0.7–1.8)

Diesel fumes—duration

 

 

<1 yr

118

1.0

1–9 yr

70

1.0 (0.7–1.6)

≥10 yr

47

1.0 (0.6–1.8)

Diesel fumes—exposure scores

 

 

<5

158

1.0

5–19

39

1.3 (0.7–2.2)

>20

38

1.0 (0.6–1.7)

Ahrens et al. 1991

Laryngeal-cancer cases among male residents of Bremen, Germany (smoking-adjusted)

 

 

 

Transportation and communication (industry)

24

1.3 (0.64–2.59)

Fumes or smoke (self-reported agent)

na

0.7 (0.3–1.4)

Brown et al. 1988

183 laryngeal-cancer cases among residents along gulf coast of Texas (smoking-adjusted)

 

 

 

Transportation, communication, utilities, and sanitary services (industry)

63

1.62 (1.04–2.51)

Transportation (occupation)

39

1.42 (0.86–2.36)

Specific occupations

 

 

Driver

15

1.69 (0.75–3.83)

Mechanic

33

1.06 (0.63–1.77)

Diesel or gasoline fumes (industrial-hygiene-coded agent)

79

1.50 (1.00–2.26)

<5 yr

32

1.8 (1.0–3.1)

5–14 yr

16

1.3 (0.6–2.7)

≥15 yr

26

1.6 (0.8–2.9)

Decoufle and Stanislawczyk 1977

Laryngeal cancer cases among male patients at Roswell Park Memorial Institute in Buffalo, New York (job history, including durations, from medical charts) (smoking-adjusted)

 

 

 

Occupations (ever)

 

 

Bus, taxicab, and truck drivers

na

0.95

Delivery and routemen

na

0.68

Locomotive engineers and firemen

na

1.07, ns

Machinists

na

1.38, ns

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Mechanics and repairmen

na

0.95

NOTE: na=not available; ns=not statistically significant (p<0.5) for risk estimate above unity; IARC International Agency for Research on Cancer.

aUnadjusted odds ratios and 95% CIs calculated with standard methods from observed numbers presented in original paper.

TABLE 4.19 Lung Cancer and Exposure to Fuels—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Study—Mortality

Lewis et al. 2000b

Refinery workers in Toronto, Canada

 

 

 

Males

478

0.85 (0.77–0.93)

Refinery segment

221

0.96 (0.83–1.09)

Marketing and distribution segment

150

0.87 (0.74–1.02)

Females

34

1.10 (0.76–1.53)

Case-Control Studies

Rosamilia et al. 1999

Lung-cancer cases nested among refinery workers at Mobil in Beaumont, Texas

 

 

 

Process operations, duration of employment

 

 

0 yr

50

1.0

<5 yr

12

0.66 (0.31–1.36)

5–14 yr

9

0.64 (0.28–1.44)

15+ yr

41

1.34 (0.82–2.19)

Maintenance and mechanical operations, duration of employment

 

 

0 yr

18

1.0

<5 yr

26

1.12 (0.55–2.31)

5–14 yr

26

0.70 (0.35–1.42)

15+ yr

42

1.32 (0.68–2.55)

Siemiatycki et al. 1987a

857 lung-cancer cases among residents of Montreal, Canada; industrial-hygiene-derived agents (smoking-adjusted)

 

 

 

Automotive gasoline

 

 

Oat cell

26

1.2 (0.8–1.6)a

Squamous cell

53

1.0 (0.8–1.3)

Adenocarcinoma

26

1.1 (0.8–1.5)

Other cell types

19

0.8 (0.6–1.1)

Aviation gasoline

 

 

Oat cell

1

0.4 (0.1–3.2)

Squamous cell

2

0.4 (0.1–1.6)

Adenocarcinoma

2

0.9 (0.2–3.8)

Other cell types

1

0.4 (0.1–3.1)

Kerosene

 

 

Oat cell

0

0.9 (0.6–1.5)

Squamous cell

34

1.4 (1.0–1.9)

Adenocarcinoma

15

1.5 (1.0–2.3)

Other cell types

13

1.2 (0.8–1.9)

Jet fuel

 

 

Oat cell

2

1.3 (0.2–7.0)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Squamous cell

1

0.2 (0.0–2.4)

Adenocarcinoma

2

1.2 (0.2–6.6)

Other cell types

1

0.6 (0.1–6.0)

Diesel fuel

 

 

Oat cell

10

1.7 (0.9–3.0)

Squamous cell

20

1.5 (1.0–2.2)

Adenocarcinoma

7

1.0 (0.5–1.9)

Other cell types

9

1.4 (0.8–2.5)

Heating oil

 

 

Oat cell

13

1.7 (1.0–2.7)

Squamous cell

25

1.3 (0.9–1.8)

Adenocarcinoma

11

1.3 (0.8–2.1)

Other cell types

5

0.6 (0.3–1.1)

Crude oil

 

 

Oat cell

1

0.7 (0.1–8.0)

Squamous cell

7

2.8 (1.0–7.6)

Adenocarcinoma

1

0.8 (0.1–9.7)

Other cell types

0

a90% CIs calculated.

TABLE 4.20 Lung Cancer and Exposure to Combustion Products—Selected Epidemiologic Studies

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Ambient Air Pollution

Cohort Study—Incidence

Nafstad et al. 2003

Male residents of Oslo, Norway (smoking-adjusted)

 

 

 

NOx

 

 

0–9.99 μg/m3

18

1.00

10–19.99 μg/m3

17

1.02 (0.75–1.39)

20–20.99 μg/m3

22

1.33 (0.87–2.04)

30+ μg/m3

24

2.22 (1.30–3.79)

Per 10 μg/m3

 

1.10 (1.03–1.17)

SO2

 

 

0–9.99 μg/m3

18

1.00

10–19.99 μg/m3

21

0.84 (0.57–1.23)

20–20.99 μg/m3

20

0.78 (0.53–1.16)

30+ μg/m3

18

0.56 (0.33–0.95)

Per 10 μg/m3

 

0.96 (0.88–1.04)

Cohort Studies—Mortality

Hoek et al. 2002

Residents of Netherlands—Netherlands Cohort Study on Diet and Cancer (smoking-adjusted)

 

 

 

Black smoke

na

1.06 (0.43–2.63)

NO2

na

1.25 (0.42–3.72)

Pope et al. 2002

Residents of US—Cancer Prevention Study II (smoking-adjusted)

na

 

 

PM2.5 mean conc.; per 10 μg/m3

 

1.14 (1.04–1.23)

SO4 mean conc.; per 6.5 μg/m3

1.10 (1.04–1.15)a

SO2 mean conc.; per 9.7 ppb

1.03 (0.98–1.07)a

NO2 mean conc.; per 27.9 ppb

0.96 (0.88–1.05)a

CO mean conc.; per 1.7 ppm

0.97 (0.92–1.03)a

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

O3 mean conc.; per 47.9 ppb

 

0.96 (0.86–1.06)a

Krewski et al. 2003

Residents of the US—Harvard Six Cities Study (smoking-adjusted)

na

 

 

PM2.5 mean conc.; per 18.5 μg/m3

 

1.17 (0.67–2.04)

SO4 mean conc.; per 7.5 μg/m3

1.14 (0.66–1.96)

SO2 mean conc.; per 19.8 ppb

1.13 (0.66–1.95)

NO2 mean conc.; per 15.8 ppb

1.15 (0.65–2.04)

Abbey et al. 1999

Seventh-Day Adventist nonsmokers in California (smoking-adjusted)

 

 

 

Males

18 total

 

PM10>100 μg/m3; per 43 days/yr

 

2.38 (1.42–3.97)

PM10 mean conc.; per 24.08 μg/m3

3.36 (1.57–7.19)

SO2 mean conc.; per 3.72 ppb

1.99 (1.24–3.20)

NO2 mean conc.; per 19.78 ppb

1.82 (0.93–3.57)

Females

12 total

 

PM10>100 μg/m3; per 43 days/yr

 

1.08 (0.55–2.13)

PM10 mean conc.; per 24.08 μg/m3

1.33 (0.60–2.96)

SO2 mean conc.; per 3.72 ppb

3.01 (1.88–4.84)

NO2 mean conc.; per 19.78 ppb

2.81 (1.15–6.89)

Beeson et al. 1998

Seventh-Day Adventist male nonsmokers in California—Incidence (smoking-adjusted)

16 total

 

 

PM10, hour in excess of μg/m3

 

 

40 μg/m3

 

4.50 (1.31–15.44)

50 μg/m3

4.96 (1.54–16.00)

60 μg/m3

4.72 (1.69–13.18)

80 μg/m3

3.43 (1.71–6.88)

100 μg/m3

2.95 (1.71–5.09)

PM10 mean conc.; per 24 μg/m3

5.21 (1.94–13.99)

SO2 mean conc.; per 3.7 ppb

2.66 (1.62–4.39)

NO2 mean conc.; per 1.98 ppb

1.45 (0.67–3.14)

Case-Control Studies

Nyberg et al. 2000

1,042 lung-cancer cases among residents of Stockholm, Sweden; residence linked to detailed emission database (smoking-adjusted)

 

 

 

Traffic-related air pollution—NO2; 30-yr averages

 

 

Per 10 μg/m3

 

1.05 (0.93–1.18)

<15.20 μg/m3

242

1.0

15.20–19.84 μg/m3

276

1.18 (0.93–1.49)

19.85–25.05 μg/m3

252

0.90 (0.71–1.14)

25.06–30.54 μg/m3

160

1.05 (0.79–1.40)

≥30.55 μg/m3

112

1.17 (0.84–1.62)

Traffic-related air pollution—NO2; 10-yr averages, 20-yr lag

 

 

Per 10 μg/m3

 

1.10 (0.97–1.23)

<12.78 μg/m3

243

1.0

12.78–17.34 μg/m3

264

1.15 (0.91–1.46)

17.35–23.16 μg/m3

250

1.01 (0.79–1.29)

23.17–29.25 μg/m3

165

1.07 (0.81–1.42)

≥29.26 μg/m3

120

1.44 (1.05–1.99)

Air pollution from heating—SO2; 30-yr averages

 

 

Per 10 μg/m3

 

1.00 (0.96–1.05)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

<41.30 μg/m3

245

1.0

41.30–52.74 μg/m3

254

1.06 (0.83–1.35)

52.75–67.13 μg/m3

272

0.98 (0.77–1.24)

67.14–78.20 μg/m3

152

0.90 (0.68–1.19)

≥78.20 μg/m3

119

1.00 (0.73–1.37)

Air pollution from heating—SO2; 10-yr averages, 20-yr lag

 

 

Per 10 μg/m3

 

1.01 (0.98–1.03)

<66.20 μg/m3

239

1.0

66.20–87.59 μg/m3

270

1.16 (0.91–1.47)

87.60–110.29 μg/m3

259

1.00 (0.79–1.27)

110.30–1 29.09 μg/m3

151

0.92 (0.70–1.21)

≥129.10 μg/m3

123

1.21 (0.89–1.66)

Marsh et al. 1998

142 lung-cancer cases among residents of four smelter towns in Arizona; atmospheric-diffusion modeling (smoking-adjusted)

 

 

 

Highest exposure level

 

 

None and background

na

1.00

Low

na

0.93 (0.51–1.70)

Medium, high, and very high

na

1.00 (0.53–1.89)

Duration above background

 

 

0

na

1.00

0–13 yr

na

1.31 (0.63–2.74)

13–26 yr

na

1.20 (0.65–2.20)

>26 yr

na

0.71 (0.34–1.48)

Cumulative exposure

 

 

0

na

1.00

0–13

na

1.22 (0.56–2.65)

13–35

na

1.19 (0.62–2.27)

>35

na

0.90 (0.48–1.71)

Occupational exposure

 

 

Definite smelter

na

1.73 (0.99–3.01)

Potential smelter

na

2.24 (1.27–3.98)

Barbone et al. 1995

755 lung-cancer cases among residents of Trieste, Italy (smoking-adjusted)

 

 

 

Level of particulate deposition from fixed-site monitoring stations

 

 

<0.175g/m2-day

188

1.0

0.176–0.298 g/m2-day

256

1.1 (0.8–1.5)

>0.298 g/m2-day

311

1.4 (1.1–1.8)

p for trend=0.022

Jockel et al. 1992

194 lung-cancer cases among residents of five cities in Germany (smoking-adjusted)

 

 

 

Emission index of SO2 from energy-consumption statistics

 

 

Low air-pollution level

41

1.0

High air-pollution level

39

1.01 (0.53–1.91)

Semiquantitative index of B[a]P, TSP, and SO2 from variety of sources

 

 

Low air-pollution level

36

1.0

High air-pollution level

44

1.16 (0.64–2.13)

Katsouyanni et al. 1991

101 lung-cancer cases among female residents of Athens, Greece—Cumulative air-pollution exposure from fixed-site

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

monitoring stations according to borough (smoking-adjusted)

 

 

Nonsmokers

 

 

1st quartile (low)

16

1.0

2nd quartile

16

1.20

3rd quartile

8

0.43

4th quartile (high)

8

0.69

p for trend=0.20

Current and past smokers

 

 

1st quartile (low)

6

1.0

2nd quartile

11

2.14

3rd quartile

14

2.72

4th quartile (high)

22

8.56

p for trend=0.007

Jedrychowski et al. 1990

Lung-cancer cases among residents of Cracow, Poland—air pollution index from fixed-site monitoring stations for TSP and SO2 and geographically interpolated (smoking-adjusted)

 

 

 

Males

 

 

Low air pollution

650

1.0

Medium air pollution

129

1.00 (0.75–1.33)

High air pollution

122

1.46 (1.06–1.99)

Females

 

 

Low air pollution

124

1.0

Medium or high air pollution

74

1.17 (0.70–1.96)

Nonsmoking males

 

 

Low air pollution

32

1.0

High air pollution

17

1.45 (0.74–2.87)

Nonsmoking females

 

 

Low air pollution

56

1.0

High air pollution

23

1.16 (0.48–2.80)

Vena 1982

417 lung-cancer cases among white male residents of Erie County, New York (smoking-adjusted)

 

 

 

Years in high or medium air pollution, from 21 fixed-site monitoring stations of TSP, interpolated to town level

 

 

0–29 yr

54

1.00

30–49 yr

114

1.03

>50 yr

249

1.26

Indoor Air Pollution

Cohort Study—Incidence

Lan et al. 2002

Farmers in Xuanwei County, China (smoking-adjusted)

 

 

 

Males

na

 

No stove improvement

 

1.0

Changed to stove with chimney

0.59 (0.49–0.71)

0–9 yr after improvement

1.79 (1.48–2.18)

10–19 yr after

0.25 (0.19–0.31)

≥20 yr

0.07 (0.03–0.17)

Cooked food <20 yr

1.0

Cooked ≥20 yr

1.42 (1.05–1.93)

Females

na

 

No stove improvement

 

1.0

Changed to stove with chimney

0.54 (0.44–0.65)

0–9 yr after improvement

1.41 (1.15–1.73)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

10–19 yr after

 

0.24 (0.19–0.31)

≥20 yr

 

0.17 (0.10–0.31)

Case-Control Studies

Kleinermann et al. 2002

846 lung-cancer cases among residents of Pingliang and Qingyang, China; self-reported agents (smoking-adjusted)

 

 

 

Type of fuel

 

 

Biomass

554

1.00

Coal

278

1.29 (1.03–1.61)

Amount of coal (tertiles)

 

 

0

325

1.00

I

146

1.18 (0.92–1.51)

II

207

1.06 (0.83–1.34)

III

134

1.29 (0.96–1.73)

p for trend=0.1 73

% of time coal used

 

 

No coal use

478

1.00

0.7–56%

102

1.99 (1.46–2.71)

57–100%

262

1.51 (1.20–1.91)

p for trend=0.024

Lan et al. 2001

97 lung-cancer cases among residents of Xuanwei County, China; self-reported agents (smoking-adjusted)

 

 

 

Smoky coal (tons)

 

 

Total

 

 

<130

na

1.0

130–240

na

1.48 (0.73–3.20)

>240

na

3.21 (1.23–9.03)

p for trend=0.01

Males

 

 

<130

na

1.0

130–240

na

1.30 (0.50–3.15)

>240

na

1.88 (0.54–7.09)

P for trend=0.32

Females

 

 

<130

na

1.0

130–240

na

2.21 (0.64–8.14)

>240

na

7.94 (1.46–60.44)

p for trend=0.008

Lan et al. 1993

139 lung-cancer cases among female farmers in Xuanwei County, China; self-reported agents (all subjects smoked) (smoking-adjusted)

 

 

 

Use of smoky coal

74

7.53 (3.31–17.17)

<3 tons/yr

23

8.24 (2.33–29.17)

≥3 tons/yr

51

7.53 (3.03–18.72)

Use of smoky coal from Laibin mine

 

 

Used after 20 yr old

12

1.84 (0.56–6.05)

Used before 20 yr old

10

5.10 (0.97–26.81)

Lifetime use

57

9.89 (3.95–24.75)

Simonato et al. 2000

305 lung-cancer cases among residents of Venice, Italy; self-reported agents (smoking-adjusted)

 

 

 

Venice islands

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Indoor heating

30

1.0

No heating

39

13.7 (4.2–45.3)

No coal

19

1.0

Ever coal

49

1.3 (0.5–3.3)

Venice inland

 

 

Indoor heating

67

1.0

No heating

74

1.4 (0.8–2.2)

No coal

24

1.0

Ever coal

119

0.4 (0.2–0.9)

Zhong et al. 1999

504 lung-cancer cases among nonsmoking female residents of Shanghai, China; self-reported agents (smoking-adjusted)

 

 

 

Fuel used for cooking

 

 

Coal

166

1.00

Coal and gas

123

0.92 (0.63–1.35)

Gas

312

0.90 (0.66–1.23)

Shen et al. 1998

70 lung-cancer cases among nonsmoking female residents of Nanjing, China; self-reported agents (smoking-adjusted)

 

 

 

Single-factor analysis

 

 

Gaseous fuel in home

na

1.51 (0.47–4.78)

Coal stove for heating

na

1.78 (0.79–4.02)

Ko et al. 1997

117 lung-cancer cases among nonsmoking female residents of Kaohsiung, Taiwan; self-reported agents (smoking-adjusted)

 

 

 

No cooking or gas

172

1.0

Coal or anthracite

25

1.3 (0.3–5.8)

Wood or charcoal

113

2.7 (0.9–8.9)

Presence of fume extractor

108

1.0

No fume extractor

160

8.3 (3.1–22.7)

Dai et al. 1996

120 lung-cancer cases among nonsmoking female residents of Harbin, China; self-reported agents (smoking-adjusted)

 

 

 

Period of coal-stove use in bedroom

 

 

1–19 yr

na

4.46 (1.61–12.33)

≥30 yr

na

18.75 (3.94–29.32)

Period of heating with coal

 

 

01–24 yr

na

5.81 (1.67–20.22)

25–34 yr

na

4.70 (1.28–17.18)

Du et al. 1996

849 lung-cancer cases among residents of Guangzhou, China; next-of-kin report (smoking-adjusted)

 

 

 

Males—coal-fumes exposure

na

0.89

Nonsmokers

 

1.50 (0.69–3.27)

Smokers

4.29 (2.33–7.88)

Females—coal-fumes exposure

na

2.21 (1.16–4.21)

Nonsmokers

 

1.56 (0.57–4.25)

Smokers

2.89 (1.09–8.65)

Lei et al. 1996

792 lung-cancer cases among residents of Guangzhou, China; next-of-kin report (smoking-adjusted)

 

 

 

Exposure to coal smoke

 

 

Males

 

 

Infrequent

350

1.0

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Regular

196

1.08 (0.84–1.40)b

Females

 

 

Infrequent

48

1.0

Regular

177

0.90 (0.55–1.45)b

Luo et al. 1996

102 lung-cancer cases among residents of Fuzhou, China; self-reported agents

 

 

 

Air pollution due to indoor burning of coal

na

7.6 (3.7–15.7)

Squamous-cell carcinoma (smoking-adjusted)

na

14.1 (1.37–145.8)c

Adenocarcinoma

na

6.0 (1.96–18.31)c

Shen et al. 1996

263 lung-cancer cases among residents of Nanjing, China; self-reported agents

 

 

 

Squamous-cell carcinoma—coal stove for heating

na

3.72 (0.88–15.71)

Xu et al. 1996a

1,249 lung-cancer cases among residents of Liaoning Province, China; self-reported agents (smoking-adjusted)

 

 

 

Cooking characteristics—males

 

 

Burning kang (coal)

 

 

1–19 yr

na

1.7 (p<0.05)

≥20 yr

na

2.1 (p<0.05)

Gas fuel for cooking

 

 

1–9 yr

na

0.9

≥10 yr

na

0.8

Cooking characteristics—females

 

 

Burning kang (coal)

 

 

1–19 yr

na

1.3

≥20 yr

na

2.3 (p<0.05)

Cooking in bedroom (using coal)

 

 

1–19 yr

na

1.5

≥20 yr

na

1.8 (p<0.05)

Gas fuel for cooking

 

 

1–9 yr

na

0.9

≥10 yr

na

0.8

Indoor air-pollution index (based on cooking fuel, place of cooking, and weighted by duration)—males

 

 

I (low)

na

1.0

II

na

1.1 (0.8–1.4)

III

na

1.2 (0.9–1.6)

IV (high)

na

1.6 (1.2–2.3)

Indoor air pollution index—females

 

 

I (low)

na

1.0

II

na

1.2 (0.9–7.8)

III

na

1.3 (0.9–1.9)

IV (high)

na

1.5 (1.0–2.4)

Liu et al. 1993

316 lung-cancer cases among residents of Guangzhou, China; self-reported agents, interviewer measurements (smoking-adjusted)

 

 

 

Cooking fuel—males

 

 

Coal

200

1.0

Gas

14

0.48 (0.15–1.6)

Wood

8

0.57 (0.11–3.0)

Other

2

0

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

Cooking fuel—females

 

 

Coal

81

1.0

Gas

8

0.90 (0.24–3.3)

Wood

3

0.67 (0.04–11.7)

Other

0

0

Size of kitchen ventilation opening—males

 

 

0.0–0.4 m2

79

1.0

0.5–0.9 m2

58

0.77 (0.36–1.7)

1.0–1.4 m2

48

0.23 (0.10–0.56)

1.5–1.9 m2

19

0.49 (0.16–1.5)

>2.0 m2

20

0.15 (0.05–0.44)

p for trend <0.001

Size of kitchen ventilation opening—females

 

 

0.0–0.4 m2

22

1.0

0.5–0.9 m2

27

0.11 (0.02–0.60)

1.0–1.4 m2

24

0.13 (0.02–0.74)

1.5–1.9 m2

7

0.09 (0.01–0.63)

>2.0 m2

12

0.06 (0.01–0.32)

p for trend <0.001

Huang et al. 1992

Lung-cancer cases among residents of Sichuan, China

 

 

 

Coal-burning indoors

na

1.59 (1.01–2.07)

Chen et al. 1990

135 lung-cancer cases among residents of Taiwan; self-reported agents

 

 

 

Cooking fuels: wood and coal

 

 

Epidermoid carcinoma

na

0.85 (ns)

Small-cell carcinoma

na

1.08 (ns)

Adenocarcinoma

na

1.02 (ns)

Cooking fuels: charcoal, gas, and electricity

 

 

Epidermoid carcinoma

na

1.00 (ns)

Small-cell carcinoma

na

1.00 (ns)

Adenocarcinoma

na

1.00 (ns)

Koo et al. 1983

200 lung-cancer cases among female residents of Hong Kong, China; self-reported agents

na

 

 

Type of cooking fuel

 

 

Kerosene

 

0.75 (0.35–1.58)d

Wood or grass

0.74 (0.37–1.47)d

Liquid-petroleum gas

0.44 (0.29–0.68)d

Gas

1.31 (0.79–2.18)d

Charcoal

0.96 (0.55–1.69)d

Coal

0.32 (0.07–1.32)d

Electricity

3.03 (0.28–76.24)d

Occupational Studies of Engine Exhausts

Cohort Studies—Incidence

Jarvholm and Silverman 2003

Diesel-exposed construction workers in Sweden (smoking-adjusted)

 

 

 

Truck drivers

61

 

SIR reference group: carpenters and electricians

 

1.29 (0.99–1.65)

SIR reference group: population

 

1.14 (0.87–1.46)

Heavy-construction equipment operators

61

 

SIR reference group: carpenters and electricians

 

0.87 (0.66–1.11)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

SIR reference group: population

 

0.76 (0.58–0.97)

Soll-Johanning et al. 2003

Bus drivers or tramway employees in Copenhagen, Denmark—nested case-control study (smoking-adjusted)

 

 

 

No lag time

 

 

Cumulated employment

 

 

<3 months

5

0.74 (0.23–2.39)

3 months-2 yr

29

1.00

2-<10 yr

54

1.26 (0.69–2.28)

10-<20 yr

22

1.39 (0.69–2.81)

20+ yr

43

0.63 (0.32–1.14)

Diesel-exposed in another job

 

 

No

80

1.00

Yes

57

0.85 (0.53–1.36)

Air-pollution index (ordinal scale—0–10—reflecting level of pollution from bus routes used)

 

 

Low

14

1.00

High

41

1.12 (0.40–3.12)

>10-yr lag time

 

 

Cumulated employment

 

 

<3 months

4

0.50 (0.14–1.81)

3 months-2 yr

27

1.00

2-<10 yr

45

1.03 (0.54–1.95)

10-<20 yr

22

1.34 (0.65–2.77)

20+ yr

43

0.54 (0.28–1.03)

Diesel-exposed in another job

 

 

No

74

1.00

Yes

49

0.80 (0.48–1.32)

Air-pollution index

 

 

Low

14

1.00

High

39

0.99 (0.36–2.75)

Soll-Johanning et al. 1998

Bus drivers or tramway employees in Copenhagen, Denmark

 

 

 

Males

 

 

Denmark reference

473

1.6 (1.5–1.8)

Copenhagen reference

473

1.2 (1.1–1.3)

Time since first employment

 

 

0–14 yr

35

1.2

15–29 yr

77

1.55 (p<0.001)

≥30 yr

361

1.7 (p<0.001)

Females

 

 

Denmark reference

15

2.6 (1.5–4.3)

Copenhagen reference

15

2.2 (1.2–3.6)

Time since first employment

 

 

0–14 yr

3

1.3

15–29 yr

10

3.5 (p<0.001)

≥30 yr

2

3.8

Guberan et al. 1992

Male commercial drivers in Geneva, Switzerland

64

1.61 (1.29–1.98)

Netterstrom 1988

Male bus drivers in Denmark

15

0.55 (0.33–0.99)

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

Cohort Studies—Mortality

Jarvholm and Silverman 2003

Diesel-exposed construction workers in Sweden (smoking-adjusted)

 

 

 

Truck drivers

57

 

SMR reference group: carpenters and electricians

 

1.37 (1.04–1.78)

SMR reference group: population

1.18 (0.89–1.53)

Heavy-construction equipment operators

49

 

SMR reference group: carpenters and electricians

 

0.83 (0.61–1.09)

SMR reference group: population

0.70 (0.51–0.92)

Borgia et al. 1994

Male taxi drivers in Rome, Italy

76

1.23 (0.97–1.55)e

 

Age at first enrollment

 

 

18–44 yr

24

1.30 (0.83–1.93)e

45–54 yr

22

1.16 (0.76–1.83)e

55+ yr

30

1.23 (0.83–1.76)e

Duration of membership

 

 

0–9 yr

35

1.18 (0.83–1.66)e

10–19 yr

35

1.50 (1.06–2.10)e

20+ yr

6

0.68 (0.25–1.48)e

Time since first enrollment

 

 

0–9 yr

25

1.37 (0.88–2.02)e

10–19 yr

40

1.44 (1.04–1.98)e

20+ yr

11

0.70 (0.35–1.25)e

Alfredsson et al. 1993

Male bus drivers in five counties in Sweden

64

1.0 (0.8–1.3)

Hansen 1993

Male truck drivers in Denmark (compared with other employed men; SES adjusted)

 

 

 

All respiratory cancers

84

1.60 (1.28–1.98)

Lung cancer

76

1.60 (1.26–2.00)

Guberan et al. 1992

Commercial drivers in Geneva, Switzerland (compared with active unskilled workers)

77

1.50 (1.23–1.81)

 

Time from first exposure

 

 

0–14 yr

2

0.67 (0.08–2.41)e

15–24 yr

11

1.18 (0.59–2.12)e

25–34 yr

24

1.30 (0.83–1.93)e

35–44 yr

21

1.35 (0.84–2.07)e

45+ yr

21

2.59 (1.60–3.96)e

p for trend <0.02

Rafnsson and Gunnarsdottir 1991

Taxi and truck drivers in Reykjavik, Iceland

 

 

 

Truck drivers only (no other occupation)

24

2.14 (1.37–3.18)

Duration of employment (ever worked)

 

 

<2 yr

na

2.70 (0.74–6.92)

2–10 yr

na

2.46 (0.99–5.08)

11–30 yr

na

0.68 (0.01–3.76)

>30 yr

na

2.32 (0.85–5.04)

Any

na

2.09 (1.32–3.13)

Taxi drivers only (no other occupation)

12

1.39 (0.72–2.43)

Duration of employment (ever worked)

 

 

Suggested Citation:"4 Cancer." Institute of Medicine. 2005. Gulf War and Health: Volume 3: Fuels, Combustion Products, and Propellants. Washington, DC: The National Academies Press. doi: 10.17226/11180.
×

Reference

Study Population

Exposed Cases

Estimated Relative Risk (95% CI)

 

<2 yr

na

0

2–10 yr

na

0.59 (0.02–3.30)

11–30 yr

na

0

>30 yr

na

1.60 (0.33–4.66)

Any

na

0.84 (0.23–2.16)

Paradis et al. 1989

Male bus drivers in Montreal, Canada

78

0.92 (0.73–1.14)

 

<30 yr duration

34

1.01 (0.70–1.38)

≥30 yr duration

44

0.85 (0.62–1.13)

Stern et al. 1981

Motor vehicle examiners in New Jersey—respiratory-system cancer

16

1.02 (0.58–1.65)e

Kaplan 1959

Baltimore and Ohio Railroad workers

 

 

 

Directly exposed occupations

49

0.88 (0.65–1.16)f

Less-exposed occupations

67

0.72 (0.56–0.91)f

Rarely exposed occupations

38

0.89 (0.64–1.24)f

Case-Control Studies

Menvielle et al. 2003

228 lung-cancer cases among residents of New Caledonia; self-reported job titles and industrial-hygienist-assigned exposure to specific agents (smoking-adjusted)

 

 

 

Dockers

22

1.1 (0.5–2.3)

<5 yr

10

0.5 (0.2–1.4)

5+ yr

12

3.3 (1.0–10.7)

Transportation-equipment managers

23

1.0 (0.5–2.0)

<5 yr

18

0.8 (0.4–1.6)

5+ yr

5

5.8 (1.1–30.7)

Motor-bus, lorry, and van drivers

13

2.7 (1.1–7.0)

<15 yr

8

2.1 (0.7–22.1)

15+ yr

5