BY DALSU BARIS, Division of Cancer Epidemiology and Genetics, National Cancer Institute, Rockville, Md.; THOMAS J. GARRITY, Local No. 22 City Firefighters Union, International Association of Fire Fighters, Philadelphia, Pa.; JOEL LEON TELLES, Delaware Valley Healthcare Council of HAP, Philadelphia; ELLEN F. HEINEMAN, Division of Cancer Epidemiology and Genetics; and SHEILA HOAR ZAHM, Division of Cancer Epidemiology and Genetics.
Firefighters are exposed under uncontrolled conditions to a wide variety of toxic chemicals including known and suspected carcinogens, such as benzene and formaldehyde in wood smoke, polycylic aromatic hydrocarbons (PAHs) in soots and tars, arsenic in wood preservatives, asbestos in building insulation, diesel engine exhaust, and dioxins.1-5 The complexity of the firefighters’ exposures increased with the widespread introduction of synthetic building materials in the late 1950s. The combustion of the common plastic polyvinyl chloride (PVC), for example, has been estimated to produce dozens of different chemicals, including cyanide.6
Previous studies of firefighters have re-ported excess risks of malignant melanoma; multiple myeloma; leukemia; and cancers of the bladder, kidney, brain, and colon7,8 Although a large number of studies have been done over the past two decades, the risk estimates for specific cancer sites have been inconsistent, and the degree of occupational cancer risk has not been fully understood. Most of the epidemiological studies examined relatively small populations with too few recent person-years to detect the effects of building material changes. The studies also typically lacked any information on exposure or surrogates of exposures, except duration of employment in a few studies.9-11
In response to the concerns of Philadelphia firefighters, the Philadelphia Fire Department (PFD) collaborated with the Local No. 22 City Firefighters Union, the International Association of Firefighters; and the National Cancer Institute (NCI) in the largest firefighter study to date-a retrospective cohort study of the mortality experience of 7,789 firefighters employed by the PFD any time between January 1, 1925, and December 31, 1986. (This time frame was chosen to avoid a biased sample of “survivors.”)
The project analyzed the mortality patterns and presented known or suspected carcinogens associated with the various types of cancer to which firefighters are exposed through the by-products of combustion.
In general, the firefighters (men) in this study were long-term employees of the PFD. All had spent some part of their career at the PFD fighting fires. The average length of employment was 18 years (see Table 1 on page 108).
The following information was extracted from PFD personnel records [Employee Service Record (ESR) cards]: the firefighter’s name, address, badge number, and phone number; date of birth; social security number (recent cards); the dates of attendance at the Fire Academy; the dates of all fire company assignments; and the types of fire company assignments, including the company number, company type (ladder or engine, for example), position, and start and stop dates for each position. All information pertaining to disciplinary actions and the names of currently employed firefighters were masked before the ESRs were photocopied, in preparation for forwarding to the NCI, who prepared the database.
Vital status as of December 31, 1986, was determined through a variety of sources such as PFD current personnel records, City of Philadelphia Board of Pension records, the National Death Index (NDI), the Veterans Affairs Beneficiary Information Record Locator System, the Pennsylvania Department of Transportation’s driver’s license records, credit bureaus, Union archives, and the personal knowledge of long-term firefighters. The underlying cause of death was obtained from death certificates from the state bureau of vital records. A total of 4,987 (58.6%) firefighters were confirmed as living; 2,220 (26.1%) were determined to be deceased; 582 (6.8%) were lost to follow-up; and 722 (8.5%) were excluded because of missing data, leaving 7,789 firefighters eligible for the analysis (see Table 2 on page 108).
A job profile was created for each firefighter. The ESR cards provided a sequential occupational history for each firefighter based on company assignments (Engine Company 22 or Ladder Company 16, for example) and positions (firefighter or battalion chief, for example). The information on fire company assignments was abstracted and computerized to create a job history file with more than 54,000 work assignment entries.
PFD job titles changed over time, causing different titles to be given to the same or similar positions as they evolved-as an example, the present title of lieutenant evolved from “assistant foreman” and, before that, “fire company supervisor.” To simplify the analysis, job titles were grouped in a job dictionary, making it possible to track changes in job titles for positions over the years (see Table 3).
Ideally, we would have liked to have been able to characterize each firefighter by the number of fires he attended over his career with additional details on the types of fires to which he responded and the kinds of tasks he performed, but the PFD does not keep records of fires attended by an individual firefighter. Therefore, firefighters’ career runs were used to estimate firefighting variables in the analysis. Using PFD, Philadelphia’s Fireman’s Hall (the fire museum), and Union records, a matrix that included the following was created for each subject: the company type (ladder or engine) and the number and estimates of the annual average number of runs for each fire company for the period between 1935 and 1986. For data not available, the information we had was interpolated from the two nearest years, where possible, or runs were factored from the closest known year based on the increase or decrease in total city alarms.
In 1969, the PFD began converting its fleet of fire trucks from gasoline to diesel. Anecdotal reports from firefighters indicate that the first diesel trucks produced exceedingly dirty exhaust. Firefighters reported that these trucks, when started with cold engines, emitted dark clouds of smoke. Some of the fire stations assigned diesel apparatus were built when horses were still used to pull equipment. The curator of Fireman’s Hall, Jack Robrecht, maintained a log of fire apparatus (truck) assignments dating back to the start of the career fire department in 1871. His log was used to create a database of fire truck assignments to fire stations between 1871 and 1986 that included the information of whether the engine was gasoline or diesel.
Philadelphia has a variety of station house designs that affected the degree to which diesel engine emissions reached firefighters’ living quarters. Older stations were typically two or three stories with the apparatus floor, watch desk, and kitchen area on the first floor; offices, bunkroom, and dayrooms(s) were on the upper floor(s). Stairs and poles connected the upper floors to the apparatus floor. These openings permitted diesel emissions to migrate throughout the station. Firefighters have reported that diesel soot was deposited on the walls, ceiling, and other surfaces throughout the older stations.
The newest fire station design separates the apparatus area (garage) from the other areas of the stations. An intermediate station design, constructed during the 1950s and early 1960s, is one story and has the living space on either side of the apparatus floor.
In all three types of stations, the only mechanical exhausts for gas and diesel engine emissions are wall-mounted exhaust fans. None of the stations have diesel emission-control systems.
Every fire station was classified by design: (1) “closed” with a compartmentalized apparatus area, separate from the rest of the station; (2) intermediate style; and (3) “open,” as described above. These design features and assigned fire apparatus engine type were included in the runs matrix.
In this study, firefighter death rates were compared with those of the general population-specifically, U.S. white males of the same age (in this study, the vast majority of firefighters were white men). Standardized mortality ratios (SMRs) and exact Poisson 95% confidence intervals (CI) were computed with the EPICURE program.12 (Also, see Glossary of Study Vocabulary above.) Observed and expected numbers were calculated for five-year age and calendar periods. SMRs were calculated for subgroups of cohorts classified by position (any PFD job, firefighter jobs only), duration of employment (
Second, mortality between groups of firefighters was compared on the basis of number of runs. A firefighter’s cumulative lifetime run and his total runs during the first five years of employment as a firefighter were calculated by multiplying the time (in years) spent at each fire company assignment by the average number of runs for that company for that year (adjusted by the number of platoons).
Because the runs data were available beginning with 1935, 1,312 firefighters hired prior to 1935 were excluded from these analyses, leaving 6,477 firefighters (restricted cohort). To evaluate if the full and restricted cohorts differed substantially in mortality experience, we calculated the cause-specific SMRs for the restricted cohort. For the internal comparisons between subgroups of the restricted cohort, categories of runs were defined as low, medium, and high, based on the values below median, >=median and <75th percentile, or >=75th percentile of the corresponding cutpoints among deceased firefighters.
For additional analysis, we combined medium- and high-exposure groups and defined only two exposure groups as follows: (a) low:
Duration of employment and number of runs were treated as time-dependent variables14 This means that firefighters moved from one exposure category to another over their follow-up period instead of being assigned to a fixed category of exposure. This allowed firefighters to contribute their person-year experience at different ages and calendar years progressively through ordered categories of exposure.
For diesel exposure, mortality rates between firefighters ever exposed to diesel and the U.S. white male population were compared. The time (in years) spent in various companies was multiplied by the relevant proportion of annual number of diesel-exposed runs at that company for that year. For cumulative exposure to diesel, four exposure groups were defined as follows: (a) unexposed, (b) low:
There were 2,220 deaths with known cause identified among the Philadelphia firefighters cohort (see Table 4). In comparison with the U.S. white males, the cohort had similar mortality from all causes of death combined (SMR=0.96, 95% CI=0.92-0.99) and all cancers (SMR=1.10, 95% CI=1.00-1.20).
Statistically significant excess risks were observed for colon cancer (SMR=1.51, 95% CI=1.18-1.93) and ischemic heart disease (SMR=1.09, 95% CI=1.02-1.16).
Elevated, although not statistically significant, risks were found for cancers of the buccal cavity and the pharynx (SMR=1.36, 95% CI=0.87-2.14), bladder cancer (SMR=1.25, 95% CI=0.77-2.00), non-Hodgkin’s lymphoma (SMR=1.41, 95% CI=0.91-2.19), multiple myeloma (SMR=1.68, 95% CI=0.90-3.11), and benign neoplasms, tumors that have not spread to any other parts of the body (SMR=1.65, 95% CI=0.89-3.07).
There were statistically significant deficits of deaths from nervous system diseases, cerebrovascular diseases, respiratory diseases, genitourinary diseases, all accidents, and suicide.
The number of observed cases of most other cancers and nonmalignant causes of death were similar to or slightly lower than expected. In general, SMRs were higher among firefighters 65 years of age and older.
Duration of Service
Analyses by duration of employment in all PFD jobs did not reveal any pattern of increasing mortality with increasing duration. However, significantly elevated SMRs were observed among subjects with fewer than 10 years of employment for cancers of the buccal cavity and pharynx (SMR=4.04, 95% CI=1.52-10.8), lung (SMR=2.08, 95% CI=1.25-3.44), pancreas (SMR=3.03, 95% CI-1.14-8.06), and prostate (SMR=3.61, 95% CI=1.17-11.20). When employment in firefighter positions only was considered in the analyses, the SMRs were slightly lower in the <10 years duration of employment category for these cancers.
The risks of mortality for kidney cancer, non-Hodgkin’s lymphoma, multiple myeloma, and benign neoplasms were highest in firefighters with at least 20 years of service, and the risks tended to increase with duration of employment, except for non-Hodgkin’s lymphoma.
The SMRs for colon cancer were 1.78 (95% CI=1.12-2.82) for firefighters with
Lagging exposure by 10 years to allow for a latent period did not substantially change these results.
Among firefighters who worked only in engine companies, the significantly elevated SMRs were observed for cancer of the buccal cavity and pharynx (SMR=2.00, 95%CI=1.11-2.63), colon (SMR=1.94, 95% CI=1.38-2.73), multiple myeloma (SMR=2.54, 95% CI=1.15-5.67), and benign neoplasms (SMR=2.95, 95% CI=1.41-6.19).
Nonsignificant excess mortality rates were detected for laryngeal cancer (SMR=1.90, 95% CI=0.74-4.56), skin cancer (SMR=1.89 CI=0.84-4.14), and non-Hodgkin’s lymphoma (SMR=1.64, 95% CI=0.85-3.15).
Among firefighters who worked only in ladder companies, there was higher than expected mortality for stomach cancer (SMT=1.85, 95% CI 0.69-4.92), bladder cancer (SMR =1.81, 95% CI=0.45-7.23), non-Hodgkin’s lymphoma (SMR=2.65, 95% CI-0.86-8.23), and leukemia (SMR=2.75, 95% CI=1.03-7.33). Except for leukemia, the SMRs did not reach statistical significance.
Year of Hire
Analysis by year of hire did not reveal any strong or consistent patterns. Among subjects hired before 1935, the only notable excess was for cancer of the buccal cavity and pharynx (SMR=2.11, 95% CI-1.13-3.91). Among firefighters hired between 1935 and 1944, significantly increased mortality was noted for all cancers (SMR=1.21, 95% CI=1.05-1.40), colon cancer (SMR=2.00, 95% CI=1.38-2.90), kidney cancer (SMR =2.11, 95% CI=1.06-4.23), non-Hodgkin’s lymphoma (SMR=2.19, 95% CI=1.18-4.07), and benign neoplasms (SMR=2.63, 95% CI=1.09-6.32). There were 20 observed deaths from colon cancer among firefighters who were hired after 1945, yielding a significantly elevated SMR of 1.60 (95% CI =1.03=2.49). No other cause of death was significantly elevated in this group.
Number of Runs
The data for annual fire company runs was not available for the years prior to 1935. The only strong patterns of increasing mortality associated with an increasing cumulative number of runs in all positions was for pancreas cancer (SMR=1.01, 95% CI=0.48-2.12) for the low category (SMR =1.21, 95% CI=0.51-2.92) for the medium category, and SMR=1.55, 95% CI=0.74-3.26) for the high category.
There was a threefold increased risk of benign neoplasms in the medium and high categories for cumulative runs (SMR=3.32, 95% CI=1.07-10.29 and SMR=3.19, 95% CI=1.03-9.88). The results were similar when the analyses included cumulative runs for a firefighter, a lieutenant, and a captain.
Risks for ischemic heart diseases, cirrhosis of the liver, and suicide decreased with increasing cumulative number of runs.
There were no significant excesses for specific causes of death among firefighters who were never exposed to diesel compared with the U.S. white male population. We also calculated the RRs for major causes of death among diesel-exposed firefighters by cumulative number of runs relative to nonexposed firefighters. The RRs for all causes of death and for all cancer combined were either below unity or close to unity in all exposure categories. Compared with the low-exposed group, the RR was elevated for prostate cancer in all three exposed groups and in the medium exposed group for leukemia and malignant neoplasms of the brain and nervous system; however, none of these excesses were statistically significant. The analyses weighted by station design and number of diesel apparatus per station were similar to the unweighted analyses.
Overall, Philadelphia firefighters experienced slightly lower mortality compared with the general population. This is consistent with the healthy worker effect (see Glossary on Page 110), typically seen in studies of occupational populations. A stronger healthy worker effect might have been expected, given the extreme physical requirements for the job; however, the small healthy worker effect among Philadelphia firefighters was consistent with the weak effect observed in most other studies of firefighters.  Deficits of cardiovascular disease mortality are usually the main contributors to the healthy worker effect15, but Philadelphia firefighters had an SMR of 1.01 (95% CI=0.96-1.07) for circulatory disease and an SMR of 1.09 (95% CI=1.-2-1.16) for ischemic heart disease.
Our findings of statistically significant deficits of deaths from nervous system diseases, cerebrovascular diseases, respiratory diseases, genitourinary diseases, all accidents, and suicide were consistent with the results from other cohort studies.16-18, 
The risk of colon cancer was increased approximately 50 percent but did not increase consistently with duration of employment. Risks for cancers of the buccal cavity and pharynx, kidney, non-Hodgkin’s lymphoma, and multiple myeloma were also elevated.
- Colon cancer. Firefighters are exposed to carcinogens such as asbestos and PAHs that have been linked to colon cancer.19 An increased risk of colon cancer has also been found among workers who were exposed to woods, metals, plastics, fiberglass, and a variety of fumes and solvents.20 Excess colon cancer (SMR=1.83) has been reported in many studies of firefighters. In one study, a significant excess in colon cancer was shown among 1,867 Buffalo, New York, firefighters on the basis of 16 deaths with a suggested dose-response relationship (SMR=4.07) for those with the longest employment.21
STUDY RESULTS FOR SELECTED CANCER SITES
In our study, we observed a significant excess mortality of colon cancer (SMR=1.51) among Philadelphia firefighters. The risk was higher for those hired after 1935 (SMR=1.81). The excess risk did not increase with duration of employment or number of lifetime runs. No information on diet as a potential confounding factor was collected in the present study or any of the previous firefighter studies.
Firefighters are likely to be more occupationally active in comparison with the general population. Despite the observation in the literature that increased physical activity is associated with reduced risk of colon cancer22,23, we observed an increased risk of colon cancer among Philadelphia firefighters.
- Multiple myeloma. A relationship between multiple myeloma and exposure to asbestos, benzene, pesticides, paints and solvents, engine exhaust, and metals has been reported by various researchers.24 Heyer observed an SMR of 2.25 (95% CI=0.47-6.60) for multiple myeloma among Seattle firefighters with a tenfold risk among those with 30 years of service (SMR=9.89, 95% CI=1.0-35.73). 
Similarly, this study found that the risk of multiple myeloma increased with the firefighter’s duration of employment (20+ years; SMR=2.31, 95% CI=1.04-5.16). In a meta-analysis of the published studies, Howe and Burch concluded that there was consistent evidence of an association between multiple myeloma and firefighting.25
- Non-Hodgkin’s lymphoma. Benzene and 1,3-butadiene, which form as combustion products from the burning of plastics and synthetics, have been associated with lymphoma risk.26 Every published study of firefighters indicates increased mortality.
In this study’s cohort, a significant increased mortality was observed among subjects with 20 or more years of employment. Firefighters hired after 1935 had a slightly higher risk. The excess risk was not associated with increasing number of lifetime runs, however. In fact, the SMR was higher in the low category.
- Bladder cancer. There is strong evidence of an association between PAHs, a class of carcinogenic organic substances, and cancer of the bladder 27 PAHs have been detected at fire scenes.28 Guidotti and Vena reported a threefold increase in bladder cancer deaths for firefighters in contrast with the general population mortality rate. ,  A slight nonsignificant elevation of mortality of bladder cancer compared with that of U.S. white men (SMR=1.25) was observed in this study. The risk was highest among those with the higher number of runs during their first five years of employment as firefighters (based on four observed deaths; SMR=2.59, 95%CI=0.64-9.84) and those who were hired before 1935 (SMR=1.71, 95%CI=0.94-3.08).
- Renal (kidney) cancer. This type of cancer is generally not considered an occupational cancer, but there have been reports of excess risk among several occupational groups exposed to asbestos, PAHs, solvents, petroleum products, gasoline, lead, and cadmium.29 In this study’s cohort, a twofold excess mortality rate was found among firefighters with 20 or more years of service (SMR=2.20, 95% CI=1.18-4.08).
- Prostate cancer: Nine of 11 firefighter studies have shown an elevated risk for prostate cancer among firefighters. The majority of the studies showed a 30 percent to 50 percent increase. An incidence study revealed a significantly increased risk of prostate cancer among Seattle and Tacoma firefighters (SMR=1.4, 95% CI=1.1-1.7), but there was no gradient with duration of employment.30
In this study, a significantly elevated SMR for prostate cancer among firefighters with fewer than 10 years of employment was observed (SMR=3.61, 95%=1.17-11.20). The risk did not rise with increasing duration of employment or number of runs.
- Leukemia. The classification for benzene is as an agent that has been proven to be carcinogenic in humans, increasing the risk of leukemia. Benzene usually is present in high concentrations in the fire environment. High levels have been measured at building and car fires.31 Several studies have shown an increased risk of leukemia among firefighters.
Our study, however, provided no consistent evidence of increased mortality for leukemia among Philadelphia firefighters. The only statistically significant finding was an SMR of 2.75 (95% CI=1.03-7.33) among firefighters who worked in ladder companies.
Lung cancer. Lung cancer has been associated with vinyl chloride, asbestos, soots, PAHs, diesel exhaust, and formaldehyde, all of which are common airborne contaminants in fire smoke.32 Interestingly, most of the firefighter studies have not shown significant increases of lung cancer. Howe and Burch obtained a pooled estimate of 1.08 for lung cancer mortality from nine studies. They concluded that there was no evidence of an increased risk of lung cancer associated with firefighting. A few additional studies published after this polled analysis did not find an excess risk of lung cancer among firefighters.,
The Philadelphia cohort also showed no increased risk. No smoking data have been available in any of the firefighter studies, including this study.
We did not have any information pertaining to potential confounding factors such as the use of tobacco and alcohol or diet. Survey data indicate that the smoking prevalence among firefighters is similar to that of people in “blue collar” jobs.33-35 Although SMRs for some of the smoking-related sites such as cancers of the buccal cavity, bladder, and kidney were elevated, the SMR for lung cancer was not elevated, adding to the weight of the evidence from other studies that firefighters are not at excess risk of lung cancer. In addition, SMR for nonmalignant respiratory diseases among Philadelphia firefighters was significantly decreased (SMR=0.67, 95% CI=0.55-0.82).
LIMITATIONS OF THE STUDY
The limitations of the study include the following:
- A lack of detailed individual exposure information. Duration of employment, runs, and station house design were used in place of individual exposure monitoring. These measures would be suitable as a surrogate for exposure if exposures do not differ over a firefighter’s career or among firefighters. However, combustion products and other firefighter exposures have changed over time with the advent of new building materials. Exposures can vary among firefighters at the same fire. As an example, first-in firefighting (the activity of firefighters on the first-arriving engine and ladder companies) often requires working inside a structure when fire conditions are most severe. They, therefore, may have higher exposure to toxic and carcinogenic products of combustion. Auto fires are especially dangerous sources of polyvinyl chloride, polyurethane, and benzene36 and accounted for more than 50 percent of all non-structure fires in Philadelphia from 1945 to 1986. These fires are almost always confined to the auto of origin and are extinguished by the first-in engine company.
Firefighters who work in residential areas may be exposed to products of combustion that differ from those encountered by firefighters responding to fires in industrial or commercial areas. Failure to detect the variations between firefighters at the same fire or the variations across fires might have resulted in misclassification of exposure. Therefore, duration of employment or number of runs, although relatively easily available and readily recorded, may not classify subjects by exposure rank correctly and could result in misclassifications that may dilute the risk estimates. This is a concern, since our study showed several small nonsignificant associations.
In addition, when duration is used as a proxy for exposure, and if there is a survivor effect, a true relationship over the range of duration of employment may be masked.37 This may be the reason for observing higher mortality rates in the shorter duration of employment category for cancer of the pancreas, prostate, and lung.
Small numbers of observed deaths in the categories of duration of employment and cumulative runs resulted in imprecise risk estimates with wide confidence intervals, especially among the restricted cohort with runs data. The study also had a limited power to examine the effect of diesel exposure because of the small number of exposed firefighters and short latency period. Many outcomes were examined by several characteristics of the cohort; therefore, multiple testing may have led to some significant findings by chance alone.
STRENGTHS OF THE STUDY
Most of the previous studies included relatively small populations and lack information about the degree of firefighters’ exposure to toxic substances. This study included nearly 8,000 firefighters with job histories. These data along with the PFD’s tracking of the number of times each fire truck responded to a call allowed us to carry out analysis with more detailed exposure information.
Our study provides no evidence of an association between the occupation of firefighting and increased risk of overall mortality, all cancers combined, or lung cancer. However, we observed increased mortality for colon cancer, non-Hodgkin’s lymphoma, multiple myeloma, and kidney cancer, consistent with several earlier studies. There were deficits of deaths from nervous system disease, respiratory disease, and genitourinary diseases. Deaths from all accidents were fewer than expected, indicating a possible effect of training, supervision, or accident prevention programs.
It is critical to continue to follow up this cohort to acquire enough person-years to examine the effect of diesel exposure. Cross-sectional studies of dietary habits and other lifestyle factors (such as smoking) would be helpful in interpreting the results of this mortality study, since none of these potential confounding factors were collected in the pres-ent study or previous studies. Carrying out cancer incidence instead of mortality studies would be helpful for cancers of low or moderate fatality rates. The future studies of firefighters should include individual exposure assessment. In the meanwhile, firefighters should continue to take precautions to limit and control their exposure to fire contaminants through the proper use of personal protection equipment.
(Editor’s note: Complete protocol and statistical calculation details, as well as a complete list of references, are available from Betsy Dwane, National Cancer Institute, 6120 Executive Blvd. EPS 8098, Rockville, MD 20892-7242).
1. Froines, J.R., W.C. Hinds, R.M. Duffy, et al, “Exposure of firefighters to diesel emissions in fire stations,” Am Ind Hy Assoc J; 1987, 48:202-207.
2. “Overall Evaluation of Carcinogenicity: An update of IARC Monographs Volumes 1-42,” International Agency for Research on Cancer, IARC Monogr Eval Carcinog Risks Hum Suppl; 1987, 7:1-440.
3. Jankovic, J., W. Jones, J. Burkhart, et al, “Environmental study of firefighters,” Ann Occup Hyg; 1991, 35:581-602.
4. McDiarmid, M.A., P.S.J. Lees, J. Agnew, et al, “Reproductive hazards of firefighting; II. Chemical hazards,” Am J Ind Med; 1991, 19:447-472.
5. Siemiyatcki, J. Risk Factors for Cancer in the Workplace. (Boca Raton, Fla.; CRC Press, 1991), Table 1.
6. Terill, J.B., R.R. Montgomery, R.F. Reinhart, “Toxic gases from fires,” Science; 1978, 200:1343-1347.
7. Golden, A.L., S.B. Markowitz, P.J. Landrigan, “The risk of cancer in firefighters,” Occup Med; 1995, 10:803-820.
8. Guidotti, T.L., “Occupational mortality among firefighters: Assessing the association,” J Occup Environ Med; 1995, 37:1348-1356.
9. Aronson, K.J., G. Tomlinson, L. Smith, “Mortality among firefighters in metropolitan Toronto,” Am J Ind Med; 1994, 26:89-101.
10. Guidotti, T.L., “Mortality of urban firefighters in Alberta: 1937-1987,” Am J Ind Med; 1993, 23:921-940.
11. Heyer, N., N.S. Weiss, P. Demers, et al, “Cohort mortality study of Seattle firefighters: 1945-1983,” Am J Ind Med; 1990, 17:493-504.
12. Preston, D.L., J.H. Lubin, D.A. Pierce. EPICURE Users Guide. (Seattle, WA: Hirosoft International Corporation, 1992).
13. Checkoway, H., D.J. Crawford-Brown, N.E. Pearce. Research Methods in Occupational Epidemiology. (Oxford: Oxford University Press, 1989).
14. Swaen, G.M., “Investigating dose response relations in occupational mortality studies: something to keep in mind,” Br J Ind Med; 1987, 44:642-644.
15. McMichael, A.J., “Standardized mortality ratios and the ‘healthy worker effect’: Scratching beneath the surface,” J Occup Med; 1976, 18:165-168.
16. Beaumont, J.J., G.S. Chu, J.R. Jones, et al, “An epidemiologic study of cancer and other causes of mortality in San Francisco firefighters,” Am J Ind Med; 1991, 19:357-372.
17. Demers, P.A., N.J. Heyer, L. Rosenstock, “Mortality among firefighters from northwestern United States cities; (British J. Indus Med; 1992, 49:664-670.
18. Tornling, G., P. Gustavsson, C. Hogstedt,” Mortality and cancer incidence in Stockholm firefighters,” Am J Ind Med; 1994, 25:219-228.
19. Tolbert, P.E., “Oils and cancer,” Cancer Causes and Control; 1997, 8:386-405.
20. Schottenfeld, D., S.J. Winawer, “Cancers of the large intestine,” in Cancer Epidemiology and Prevention (second ed.), D. Schottenfeld and J.F. Fraumeni, eds. (New York: Oxford University Press, 1996), 813-840.
21. Vena, J.E., R.C. Fiedler, “Mortality of a municipal-worker cohort: IV. Firefighters, Am J Ind Med; 1987, 11, 671-684.
22. Colditz, G.A., C.C. Cannuscio, A.L. Frazier, “Physical activity and reduced risk of colon cancer: implications for prevention,” Cancer Causes Control; 1997, 8:649-667.
23. Arbman, G., O. Axelson, M. Frederiksson, et al, “Do Occupational factors influence the risk of colon and rectal cancer in different ways? Cancer; 1993, 72: 2543-2549.
24. Herrinton, L.J., N.S. Weiss, A.F. Olshan, “Multiple myeloma,” Cancer Epidemiology and Prevention, second ed.; 1996, 946-980.
25. Howe, G.R., J.D. Burch, “Firefighters and risk of cancer: an assessment and overview of the epidemiological evidence,” Am J Epidemiol; 1990, 123:1039-1050.
26. Scherr, P. and N. Mueller, “Non-hodgkin’s lymphoma,” Cancer Epidemiology and Prevention, second ed., 920-945.
27. Boffetta, P., N. Jourenkova, P. Gustavsson,” Cancer risk from occupational and environmental exposure to polycyclic aromatic hydrocarbons,” Cancer Causes Control; 1997, 3:444-472.
28. Olshan, A.F., K. Teschke, P.A. Baird, “Birth defects among offspring of firemen,” Am J Epidemiol; 1990, 131:312-321.
29. McLaughlin, J.K., W.J. Blot, S.S. Devesa, et al., Cancer Epidemiology and Prevention, second ed., 1142-1155.
30. Demers, P.A., H. Checkoway, T.L. Vaughan, et al. “Cancer incidence among firefighters in Seattle and Tacoma, Washington (U.S.),” Cancer Causes Control; 1994, 5:129-135.
31. Brant-Rauf, P.W., L.F. Fallon, Jr., T. Tartanitini, et al, “Health hazards of firefighters: exposure assessment,” Br J Ind Med; 1988, 45:606-612.
32. Industrial Disease Standards Panel IDSP Report No. 13, “Report to Workers’ Compensation Board on Cardiovascular Disease and Cancer Among Firefighters,” Toronto, Ont., Canada, 1994.
33. Bates, J.T.,” Coronary artery disease in the Toronto Fire Department,” J Occup Med; 1987, 29:132-135.
34. Sama, S.R., T.R. Martin, L.K. Davis, et al, “Cancer incidence among Massachusetts firefighters, 1982-1986,” Am J Ind Med; 1990, 18:47-54.
35. Sterling, T.D., J.J. Weimkam, “Extent, persistence and consistency of the healthy person effect by all and selected causes of death,” J Occup Med; 1986, 28:348-353.
36. Landrigan, P.J., A.L. Golen, S.B. Markowitz, “Occupational cancer in New York City firefighters.” A report by the Department of Community Medicine, Div. of Environmental and Occupational Medicine, Mount Sinai School of Medicine, City University of New York, 1995.
37. Robins, J., “A graphical approach to the identification and estimation of causal parameters in mortality studies with sustained exposure periods,” J Chronic Dis; 1987, 2:139S-161S.
Glossary of Study Vocabulary
- Cohort study. A study design that traces the mortality or morbidity of a group of people with a common exposure to a potentially hazardous substance-e.g., occupational groups.
- Healthy worker effect. Most occupational epidemiological studies compare workers with the general population because other more appropriate comparison groups are usually unavailable. Since the general population includes people who do not work or cannot work because of illness or disability, a working population is healthier, and a lower mortality rate is expected for most causes of death. A healthy worker effect may conceal a real increase in deaths among workers. Comparisons with similar populations (e.g., another group of workers) are, therefore, more likely to provide accurate estimates of occupational risks.
- Latency period. The time between exposure to a substance(s) and the occurrence of a disease the substance caused.
- Multiple test of significance. When multiple outcomes are examined, some positive associations will occur by chance alone. Caution must be exercised in interpreting the results, especially if there is no prior hypothesis that a particular disease would be elevated in an occupational group.
- Standardized Mortality Ratios (SMR). The SMR is an estimate calculated by comparing the number of deaths observed among firefighters to the number of deaths expected based on a comparison group of the same age and sex during the same time period. The result in this study is measured in terms of SMR. SMR = observed deaths
Some authors multiply the ratio by 100. An SMR greater than 1 (or 100, multiplied by 100) suggests an excess risk of death. An SMR smaller than 1 (or 100) suggests deficit of death. Statistical significance of an SMR is evaluated by using 95% confidence interval (CI), which is the range into which the “true” SMR would fall 95 percent of the time for a given study if the study were to be repeated. If the 95% confidence interval does not include the value 1 (or 100), the SMR is considered statistically significant.
- Survivor effect. Because adequate health is required to continue working and unwell workers may leave the workforce before reaching retirement, it is possible that only the healthiest workers will continue to be employed for a long term. If the duration of exposure is the only available measure of exposure, and if there is a survivor effect, the dose-response may be masked.
Source: Industrial Disease Standards Panel (1994): IDSP Report No. 13, Report to Workers’ Compensation Board on Cardiovascular Disease and Cancer Among Firefighters, Toronto, Ontario.