Industrial Disease Standards Panel

In 1985 the Ontario legislature established the Industrial Disease Standards Panel (IDSP) to investigate and identify diseases related to work. The Panel is independent of both the Ministry of Labour and the Workers' Compensation Board. At the end of each fiscal year the WCB reimburses the Ministry for the Panel's expenditures.

The Panel's authority flows from section 95 of the Worker's Compensation Act and its functions are set out as follows:

(8) (a) to investigate possible industrial diseases;

(b) to make findings as to whether a probable connection exists between a disease and an industrial process, trade or occupation in Ontario;

(c) to create, develop and revise criteria for the evaluation of claims respecting industrial diseases; and

(d) to advise on eligibility rules regarding compensation for claims.

Decisions of the Panel are made by its members who represent labour, management, scientific, medical and community interests. Once the Panel makes a finding, the WCB is required to publish the Panel's report in the Ontario Gazette and solicit comments from interested parties. After considering the submissions the WCB Board of Directors decide if the Panel's recommendations are to be implemented, amended or rejected.

To assist with its work the Panel has a small staff of researchers, analysts and support people. In addition to its own staff, the Panel relies heavily on the advice of outside experts in science, medicine and law, as well as input from the parties of interest.

Additional copies of this publication are available in writing:
Industrial Disease Standards Panel
69 Yonge Street, Suite 1004
Toronto, Ontario M5E 1K3
(416) 327-4156

Acknowledgements

We wish to thank all fire department employees who were enroled in this study, and we are grateful for the cooperation of unions and management in this effort. In particular, we thank Chief Joe Gibson, Dr. J.T. Bates, Evelyn Funston, and Jim Pullen. The protocol was initiated by Dr. Lesbia Smith. The data collection phase was greatly assisted by four enthusiastic research assistants, Wendy Findlay, Jani Gendron, Jodi Gendron, and Debbie Morin. We also acknowledge the contributions of Statistics Canada (Martha Fair, Maureen Carpenter, and Pierre Lalonde), the provincial and territorial Registrars of Vital Statistics, the Ontario Cancer Treatment and Research Foundation (Eric Holowaty and Darlene Dale), peer reviewers, and the staff of the Industrial Disease Standards Panel.

Introduction

The Industrial Disease Standards Panel (IDSP) undertook a study of mortality among fire fighters after the unions representing fire fighters made submissions to the Panel concerning excess cardiovascular disease among their members.

Purpose of the Study

There is conflicting evidence of increased risk of cardiovascular disease among fire fighters. Two previous studies among fire fighters in the City of Toronto have shown slight excesses in cardiovascular mortality (Mastromatteo, 1959; Bates, 1987), and one recent study of fire fighters in Hawaii showed a borderline elevation (Grimes et al., 1991). However, numerous other studies carried out in Australia, Canada and the United States have not found an overall increase in cardiovascular mortality (Musk et al., 1978; Eliopulos et al., 1984; Vena and Fiedler, 1987; Beaumont et al., 1991; Demers et al., 1992; Guidotti, 1992). But within some of these studies, sub-groups with either long duration of employment or time since first exposure have exhibited excess risk (Vena and Fiedler, 1987; Demers et al., 1992).

Several studies of fire fighters have demonstrated increased risk for cancer at various sites. A report that combines the results from several studies suggests that fire fighters experience excess mortality from brain cancer and possibly multiple myeloma (Howe and Burch, 1990).

Because of this conflicting evidence, this study was designed to investigate whether there is excess mortality among Metropolitan Toronto fire fighters, with particular attention to cardiovascular disease. Excess mortality is said to occur when there are more deaths in the study population than would be expected based on death rates in a reference population.

This report is intended as a descriptive overview of the main results of the study. The full scientific report ("Scientific Report of the Mortality Study of Fire Fighters in Metropolitan Toronto") is available upon request.

Definition of the Cohort

A retrospective cohort study was conducted, with subjects recruited from all six fire departments within Metropolitan Toronto: the cities of Toronto, North York, Scarborough, Etobicoke and York and the Borough of East York. Data was collected on 5,995 employees of fire departments.

Subsequently excluded were (Table 1):

Following these exclusions, 5414 individuals were included in the analysis. Their mortality experience was analyzed over the period from January 1, 1950 to December 31, 1989.

Methods

Records of employment histories were located in fire department headquarters, personnel offices and archives. All departments maintain paper or card files to some extent, some use microfilm or microfiche and recently some have been storing information on computer. A standard data collection form was developed to facilitate direct entry of information into a computer database. Four data abstractors were briefed on issues of confidentiality and were required to take an oath of secrecy.

The resulting data file containing information on each fire department employee for whom records were available was sent to Statistics Canada (SC). Through record linkage with the National Mortality Data Base (NMDB), SC determined the number of deaths in the cohort, and where applicable, the cause and date of death was added to the file. This system is a very reliable means of determining occurrence and cause of death for residents of Canada when it is provided with the full name and date of birth of an individual (Shannon et al., 1989). SC then removed all personal identifiers from this file and returned it to the IDSP staff for statistical analysis.

Mortality of the cohort specific to age, calendar period, and cause was compared to that of the Ontario male population over the period 1950-1989. Subjects began to contribute person-years at risk on January 1, 1950 or their first day of employment, whichever was later, and ceased to contribute person-years on their date of death, upon reaching age 85 or on December 31, 1989, whichever was earlier. The protocol called for censoring at age 85 on the grounds that certification of cause of death in elderly people is less reliable.

The number of deaths expected from a particular cause was computed by multiplying the person-years in each age and calendar period stratum by the cause-specific Ontario mortality rates for that stratum. The number of deaths observed in the cohort is compared to the number of expected deaths by taking the ratio of observed to expected, and when multiplied by 100 this yields the standardized mortality ratio (SMR):

observed deaths x 100 = SMR
---------------
expected deaths

An SMR greater than 100 may indicate an excess risk of death due to a particular cause. The statistical significance of an apparent excess is judged using a 95% confidence interval (CI) (Breslow and Day, 1987). When the SMR is greater than 100 and the CI excludes 100, there is said to be a statistically significant excess in mortality in the cohort. Likewise, when the SMR is less than 100 and the CI excludes 100, there is said to be a statistically significant deficit in mortality in the cohort. In these situations, chance as an explanation of the results is excluded 19 out of 20 times. Elevated, although not statistically significant (i.e., the 95% CI includes 100), mortality risks are judged in this report as SMRs over 130 with 4 or more deaths.

Information on deaths which were certified to any type of cancer was sent to the Ontario Cancer Treatment and Research Foundation (OCTRF). From records which date back as far as 1964, they returned confirmation of the cause of death and information on cancer incidence, including histological type and date of diagnosis when available.

Results

Descriptive statistics for the 5414 fire department employees are seen in Table 2. At the end of follow-up, 55% of the subjects were classified as current employees, while the remaining 45% were retired or terminated.

Over the period of the study, there were 837 deaths in the cohort and 60 deaths were excluded from the analysis because they occurred after age 85. This left 777 deaths, of which 29% occurred in current workers or within one month of termination of employment, 61% more than two years after termination, and the remaining 10% between these two intervals. The mean age at death was 63 years.

The results below begin by discussing all-cause mortality, and continue with circulatory system, cancer, and respiratory system and digestive system diseases. Summaries of the analyses are seen in the Tables attached, but these are generally not referred to in the text. As well as presenting a summary of the results of this study, we have attempted to evaluate our findings in relation to the body of knowledge available. Salient points are discussed below according to cause.

Overall Mortality

Cause-specific mortality for the total cohort is presented in Table 3. The all-cause SMR is 94 (95% CI: 88-101). As expected, mortality risk among current workers is very low (SMR=56), and that among terminated workers is higher (SMR=119).

Most of the major groups of causes of death, including circulatory system diseases and cancer, have SMRs of about 100, indicating no excess overall in comparison with the general population.

Most of the major groups of causes of death, including circulatory system diseases and cancer, have SMRs of about 100, indicating no excess overall in comparison with the general population. The two exceptions among the major groups of causes of death are infective and parasitic diseases, and endocrine, nutritional and metabolic diseases which exhibit statistically significant deficits in mortality. Among specific causes of death, two causes exhibit statistically significant deficits in mortality: diabetes mellitus and chronic rheumatic heart disease, as these are conditions which may cause an individual not to seek employment as a fire fighter. These will not be discussed further.

Statistically significant excesses are seen for three specific causes of death:

(1) cancer of the brain and other nervous system tumours,

(2) "other" malignant neoplasms, and,

(3) aortic aneurysm.

These will be discussed extensively below.

The Healthy Worker Effect

When this study was designed, there was a concern that the bias known as the healthy worker effect would mask some important risks. This bias can occur when the mortality experience of a working population is compared to that of the general population, which comprises both those who work and those who do not or cannot work. The working population, being healthier, is expected to have a lower mortality rate for many causes, so there is sometimes a problem in determining what exactly constitutes an excess risk for the working population. Fire fighting is unlike most occupations in that to become a fire fighter, one must pass a medical examination, an aptitude test and tests of physical skill. The six fire departments in this study all had approximately the same entrance requirements. Although the requirements were not so stringent in the past, to be hired, one has always had to be physically fit. It does not seem unreasonable to assume that at least at the start of his career, the fire fighter is much healthier than the average person in the general population.

Although the fire fighters experience a lower mortality rate than comparable members of the general population, this is only so for approximately the first ten years since start of employment. After this ten year period, for many causes of death their mortality experience is similar to or higher than that of the general population.

In this study, a strong healthy worker effect is evident for those under 40 years of age. In other words, fire fighters under 40 die less frequently than comparable members of the general population. However, by the time a fire fighter reaches 40, he is no longer at a reduced risk of death when compared to the general population. Furthermore, all-cause mortality shows an increasing trend in SMRs as years since first exposure (YSFE) increase. Analysis by 10 year periods indicates that a strong healthy worker effect is evident for those with less than 10 YSFE, but not evident for those workers with more than 10 YSFE. In other words, although the fire fighters experience a lower mortality rate than comparable members of the general population, this is only so for approximately the first ten years since start of employment. After this ten year period, for many causes of death their mortality experience is similar to or higher than that of the general population.

In anticipation of an overall healthy worker effect in these results, the protocol for this study provided for a comparison of the fire fighters to another working population, the Northern Ontario Miners' Cohort. The comparison to miners would have been subject to some severe limitations for several reasons. In any event, because no strong healthy worker effect overall is evident in this study, this comparison is not necessary. Because of the pattern of increasing SMRs with increasing YSFE, there is some difficulty in interpreting the SMRs. For some causes an SMR of 100 may be an excess relative to an SMR of about 85 for all-cause mortality that has been shown for several working populations (Meijers et al., 1989). Therefore, depending on what one judges as the most appropriate comparison group for the fire fighters, the real magnitude of some of the risks seen here may be greater than the SMRs indicate.

In the other cohort study of Canadian fire fighters, with the Albertan population as a reference group, a healthy worker effect was also not seen (Guidotti, 1992). It has been reported, however, in several studies of U.S. fire fighters (Beaumont et al., 1991; Demers et al., 1992, for example). Dr. Guidotti explains that a healthy worker effect is not seen in his study because few subjects are lost to follow-up (as in this study), there is a long follow-up period (as in this study), and the use of a provincial instead of national comparison may reduce the effect since Alberta males are generally healthier than the national average.

The Survivor Effect

Another effect that could mask the results is the survivor effect. Because adequate health is required for continuing employment and unwell workers may selectively leave the fire department before 30 years of employment, it may be that only the healthiest workers continue to be employed long-term. If duration is the only proxy for exposure and there is a survivor effect, a true exposure-response relationship over the range of employment duration may be masked (Robins, 1987). This may explain in part, for example, the increases in SMRs seen until 30+ years' duration, and the decline seen after this point for some causes of death. The healthy worker effect is generally more pronounced for cardiovascular disease than for cancer causes of death.

Circulatory System

This overall group of causes of death ("Major cardiovascular diseases") is comprised of four sub-groups, with the first accounting for the majority of deaths: ischaemic heart disease, cerebrovascular disease, arteriosclerosis, and diseases of the veins, lymphatics and other circulatory system. Within ischaemic heart disease, the largest sub-group is acute myocardial infarction, and within arteriosclerosis, the largest sub-group is aortic aneurysm.

Major cardiovascular diseases (SMR=99)

The group of causes labelled "major cardiovascular diseases" was responsible for almost half of the deaths in this study, as it is for the general population. The SMR hovers around 100 across all calendar periods and for all categories of YSFE and age at risk. There is no obvious pattern according to duration of employment. While risk is low among current workers (SMR=62), it is higher among terminated workers (SMR=117). Within this overall cause of death category, ischaemic heart disease accounted for a majority of the deaths, again with no apparent excess, and also displayed no pattern of risk with the temporal or exposure variables.

This differs from the findings of Mastromatteo (1959) who reported an excess in City of Toronto fire fighters for deaths occurring between 1921 and 1953 from cardiovascular-renal diseases. This overlaps with the present study by four years of follow-up, a period in which there is no apparent excess in our study.

Coronary heart disease (CHD) was also elevated in the study by Bates (1987) of fire fighters in the City of Toronto who worked six years or more. He compared mortality rates to those for males in the City only. Follow-up was from 1949 through 1984, and of 52 deaths known through pension records, 47 death certificates were obtained and examined for any circulatory system cause of death. For the remaining deaths, cause was abstracted from newspaper or coroner reports. The resulting SMR for CHD was 173 based on 21 deaths. A possible explanation for these discrepant results is the different methodologies (different inclusion criteria, method of death enumeration, method of cause of death coding, and comparison populations) used by Bates and in the present study. It is important to note that in an unpublished survey of smoking among fire fighters in 1980 in the City of Toronto, about one-third smoked. This is approximately the same as the proportion of males in the general population who smoked, and smoking can therefore be discounted as an explanation of the increased risk seen for CHD. Dr. Bates is continuing work on this issue and is looking at the possible contribution of cholesterol levels and hypertension. Two studies have shown increasing SMRs for circulatory system diseases with increasing years of work (Vena and Fiedler, 1987; Heyer et al., 1990, and updated in Demers et al., 1992). Guidotti found no association with duration of employment for this cause (Guidotti, 1992).

A recent review discusses the possible role of occupational factors in the development of cardiovascular diseases (Kristensen 1989a and b). Some non-chemical occupational factors which could contribute to cardiovascular disease risk in fire fighters are sudden exertion during waiting or sleeping periods, shift work, environmental stress (heat, cold, noise), physical stress (from heavy equipment, prolonged work), and emotional stress (critical incidents such as witnessing fire victims, response to emergencies, responsibility for the lives of others). Substances which may potentiate a cardiovascular reaction in fire fighters include carbon monoxide and dioxide, acrolein, benzene, cyanide, hydrogen chloride, parti-culates, sulphur oxides, and organic solvent.

Myocardial infarction (SMR=107)

Although the overall SMRs for acute myocardial infarction are not statistically remarkable, they are higher than would be expected from a population whose work requires a degree of physical fitness higher than that of the general population.

The SMRs do not vary substantially from 100 over the range of YSFE, while analysis by age indicates that the risk is slightly higher for those under 60 years. There is no apparent trend with duration of employment. Again, risk is low among current workers (SMR=71) and high among terminated workers (SMR=129). Although the overall SMRs for acute myocardial infarction are not statistically remarkable, they are higher than would be expected from a population whose work requires a degree of physical fitness higher than that of the general population.

For ischaemic heart disease as a whole, Hansen (1990) reports a slight excess, while most cohort studies of fire fighters, which use general population comparisons, do not show any evidence of an increased mortality risk for ischaemic heart disease or myocardial infarction.

Arteriosclerosis (SMR=141)

The excess seen for arteriosclerosis does not exhibit any pattern according to YSFE. Risk is slightly elevated in younger men and even higher in older men. Risk is relatively high and constant across all duration categories.

Aortic aneurysm (SMR=226)

The majority of deaths attributed to the larger category of arteriosclerosis are due to aortic aneurysm. Of the 19 deaths, no death occurred before 20 YSFE, but at 20-30 YSFE the SMR rises to 303 and at 40-49 YSFE, there is a statistically significant SMR. Risk is higher among those over age 60 than in those under age 60. Risk also increases from the first to second category of duration of employment, then remains stable. Although one death occurred in a fire fighter with 1-4 years of employment, all others occurred in those with 15 or more years of employment and risk is elevated only for terminated workers.

Smoking, hypertension, diabetes, atherosclerosis, trauma and stress have been identified as possible risk factors (Lindsay and Hurst, 1979).

Diseases of the veins, lymphatics and other circulatory system (SMR=167)

No death occurred before 10 YSFE or after 50 YSFE, or in those with less than 15 years employment. Risk is elevated only for those with 15-29 years duration of employment, but is identical for current and terminated workers.

Cancer

The results for this section are organized by cancer site. For five cancer sites, a review has combined results from six cohort studies of fire fighters (Howe and Burch, 1990). Since that publication, four additional cohort studies of fire fighters are available which present results on the sites of cancer covered in the review (Tornling et al., 1990; Beaumont et al., 1991; Demers et al., 1992; Guidotti, 1992). For three of these cancer types (colon, malignant melanoma, and multiple myeloma), results are unremarkable in the present study, and are not covered in this summary report. For the other two sites (brain and other central nervous system tumours, and lung cancer), we attempt here to combine our study results with all results available to us.

Cancer of the brain and other nervous system (SMR=200)

With 13 deaths from this site, there is statistically significant excess mortality.

With 13 deaths from this site, there is statistically significant excess mortality (SMR=200, 95% CI: 109-336). Excess risk is seen in all calendar periods examined except 1975 through 1984. Large excesses are also seen for 0-19 YSFE and 40-59 YSFE, but not for 20-39 YSFE. The level of risk is identical for relatively young and older fire fighters. Breakdown by five year groupings of duration of employment revealed that no deaths occurred in those with less than 5 years of employment, but that the risk is statistically significant at 5 to 9 years duration. According to broad categories of employment duration, risk exhibits a bimodal pattern, showing an excess for employees working less than 15 years and more than thirty years, but not differing from that in the general population in the intervening period.

According to information received from the OCTRF, one subject whose primary cause of death was certified as a different cancer site death was recorded as an incident case of brain cancer in the Ontario Cancer Registry. The certification of the death can be thought of as an error, in that the other site was most likely secondary to the primary brain cancer (Dr. Holowaty, OCTRF, personal communication). The histological type was glioma. The remaining nine subjects who died from brain cancer after 1963 were registered as incident brain cancer cases at the OCTRF and six of these were histologically confirmed, all as gliomas.

[Within the total cohort enumerated, there was one additional death certified as cancer of the brain and other nervous system, but the subject was excluded from the analysis based on the criterion that the death occurred before 1950.]

When the results of the present study are combined with the previous studies' results, the overall SMR is 164 (95% CI: 126-210). In addition, two recent studies, not included in the combined results because they are not cohort studies, have reported a possible increased risk of brain cancer in fire fighters.

When the results of the present study are combined with the previous studies' results, the overall SMR is 164 (95% CI: 126-210). In addition, two recent studies, not included in the combined results because they are not cohort studies, have reported a possible increased risk of brain cancer in fire fighters. A study of fire fighters in Hawaii produced a proportionate mortality ratio of 378 based on 3 deaths (Grimes, 1990), and a case-control study in Massachusetts gave a standardized morbidity odds ratio of 86 when the state pop-ulation was used as a reference, and 152 when police were used as a reference (Sama et al., 1990).

A comprehensive review of brain cancer and occupational risk factors found numerous studies showing increased risk of brain cancer in a wide variety of occupational groups (Thomas and Wax- weiler, 1986). The lack of exposure information in most of these studies makes it difficult to determine exactly what any of them might have in common with fire fighting. However, the authors mention, among other substances, vinyl chloride, formaldehyde, acrylonitrile, polycyclic aromatic hydrocarbons, diesel and other motor fuel exhausts, trichloroethylene, asbestos and solvents as exposures that are prevalent among many of the occupational groups with increased risk of brain cancer. A recent study has demonstrated a brain cancer excess in workers exposed to poly-chlorinated biphenyl (Sinks et al., 1991).

Other malignant neoplasms (SMR=282)

With 20 deaths from this cause, statistically significant excess mortality is evident. There is some excess risk apparent in all calendar periods. No death occurred before 10 YSFE, and after this time the SMR increases as YSFE increases, until 50+ YSFE.

According to age at risk, there are two periods of statistically significant excess: at ages 50-59 and 70-85. According to five year periods of employment, there are two statistically significant excesses, at 10-14 and 35+ years. Although one death occurs in a worker with 1-4 years of employment, all others occur in those with 10 or more years of employment. This cause of death is elevated among current and terminated workers alike.

Of the 20 "other malignant neoplasm" deaths, 17 were diagnosed with cancer after 1963. The OCTRF certified 11 of these subjects as "ill-defined site" incident cases. Of the remaining six subjects, two had cancer incidence of the same site and the other four had cancer at different sites.

Cancer of the pharynx (SMR=139)

There is a suggestive increase in mortality from cancer of the pharynx. All deaths attributed to cancer at this site occurred after 20 YSFE and all in terminated fire fighters. Three of the 4 cases occurred in subjects with 35+ years of employment and 2 of these 3 cases were cases of cancer of the oropharynx, resulting in a statistically significant excess (SMR=1176). All four deaths were histologically confirmed cases of pharyngeal cancer (two oropharynx and two hypopharynx), according to the OCTRF.

Cancer of the rectum and rectosigmoid junction (SMR=171)

Also suggestive of excess risk is death certified to this site. No death occurred before 10 YSFE, and there is little evidence of a pattern according to YSFE. Cancer of the rectum increases with increasing age. All deaths occurred in subjects with 15 or more years of employment, and the SMRs are elevated for the 15-29 and 30+ duration categories. Of the 13 deaths, according to the OCTRF, 10 were incident cases of cancer of the rectum, while 3 were colon cancer cases. All were histologically confirmed, excluding two rectum cancer cases.

Cancer of the pancreas (SMR=140)

Of the 14 deaths, none occurred before 10 YSFE, and there is no apparent trend by calendar period or YSFE, although the overall SMR is somewhat elevated. Risk increases with increasing age. Deaths occur only in those with 10 years or more of employment.

Cancer of the bronchus and lung (SMR=95)

Although there is no evidence of increased risk for lung cancer in this study, this site is covered in the Howe and Burch review (1990). There is little evidence of consistently increasing SMRs as YSFE increases. There was no death before 10 YSFE or after 60 YSFE. Likewise, no pattern is apparent according to calendar period. Whereas about 7 deaths are expected before age 50, deaths from this cause occur only after age 50, but there is no pattern according to age among those over 50. Risk is elevated for those in the shortest duration category and statistically significantly decreased among current workers (SMR=34). All 51 lung cancer deaths were verified as lung cancer incident cases, excepting two certified to other sites. Of the lung cancer cases, all but seven were histologically confirmed.

Eight studies combine to give an SMR of 93 (95% CI: 84-103). Not included in this calculation is a study by Hansen (1990) comparing the mortality of male fire fighters to civil servants and salaried employees classified as such from census information. This study found a statistically significant increase in lung cancer for fire fighters (SMR=317). Hansen also reports an excess for "non-pulmonary" cancer, but provides no detail of specific cause.

Cancer of the prostate (SMR=132)

None of the 16 deaths occurs before 20 YSFE, and excluding this period the SMR is 136. While there is no pattern according to age at risk or calendar period, there is a statistically significant SMR for 1980-1984 (SMR=303). According to the five year groupings of duration of employment, there is one statistically significant excess at 25-29 years (SMR=329). While one death occurs in a subject with 5-9 years of employment, all others occur in those with 25 years or more duration.

Cancer of the testis (SMR=246)

Although there are only three deaths from cancer of the testis and they occurred in three different calendar periods, it is important to note that all three occurred among current workers within 10 years of first employment, giving a statistically significant SMR of 500. Two of the three deaths occurred after 1963, and both were histologically confirmed incident cases at the same site.

Lymphosarcoma and reticulum cell sarcoma (SMR=150)

There is a suggestion of increased mortality risk for this cancer site. Three of the four deaths are lymphosarcomas (SMR=203). There is some evidence of increasing risk as YSFE increases and an excess is seen at 10-14 years of employment. The risk is approximately equal for current and terminated workers. Of the four deaths, three had cancer incidence records after 1963 and two of these were histologically confirmed.

Lymphatic leukaemia (SMR=190)

Despite only four deaths for this cause, an elevated SMR is apparent. All four deaths occurred since 1970 in subjects 60 years of age and older, with at least 30 YSFE, and at least 30 years of fire department service. Records were available from the OCTRF for three of the four deaths, and all were incident cases of lymphatic leukaemia; however, only one was histologically confirmed.

Respiratory System Diseases

Chronic bronchitis, emphysema and asthma (SMR=130)

These diseases are usually lumped together under the term, Chronic Obstructive Pulmonary Disease or COPD. This group includes any disorder which is marked by persistent slowing of bronchial air flow (Higgins and Thom, 1988). There is some disagreement over whether or not asthma should be included in the definition, and in this study the one death from asthma is included.

After more than 20 YSFE, SMRs increase consistently with increasing 10 year periods of YSFE: 124, 137, 161 and finally 250 (not shown), indicating a diminishing healthy worker effect.

After more than 20 YSFE, SMRs increase consistently with increasing 10 year periods of YSFE: 124, 137, 161 and finally 250 (not shown), indicating a diminishing healthy worker effect. In the most recent calendar period, there is almost a statistically significant increase (SMR=206) for COPD. While there is low risk at less than 60 years of age, the SMR for those 60 years and older is significantly increased. SMRs vary substantially over the range of employment duration. The only death attributed to asthma was that of a man who had worked 35 years or more.

Non-occupational risk factors include smoking and air pollution. Some researchers contend that occupational factors include toluene diisocyanate, grain, wood and hardrock dust, and several irritant gases (Becklake et al., 1988). While chronic changes in pulmonary function have been documented in fire fighters (Sparrow et al., 1982; Peters et al., 1974), as has acute airway responsiveness (Sherman et al., 1989), many investigators contend that this is not reflected in cohort studies of mortality because risk is masked by the healthy worker effect (Becklake et al., 1988; Rosenstock et al., 1990). In comparison to police officers, one study has shown that fire fighters experience an excess risk of non-malignant respiratory diseases (SMR=141, Rosenstock et al., 1990). In the study of fire fighters in Edmonton and Calgary, COPD was moderately increased (SMR=157), but showed no relation to duration of employment (Guidotti, 1992).

Digestive System Diseases

Cirrhosis of the liver (SMR=133)

Although cirrhosis of the liver exhibits no excess overall, there is a statistically significant SMR during 1980-1984. Of the 26 cirrhosis deaths, no death occurred before 10 YSFE, and risk is highest after 20 YSFE. The SMRs are similar for the two broad age categories, but there is a statistically significant excess at ages 40-49. Excess risk is seen only for those working for 15-29 years. There is elevated risk only among terminated workers (SMR=207).

Digestive system diseases as a whole were elevated in San Francisco fire fighters due to excesses of cirrhosis and diseases of the stomach and duodenum, and the authors did not know whether to attribute this excess to fire fighting exposures, alcohol or their interaction (Beaumont, 1991).

Gallbladder diseases (SMR=296)

Gallbladder diseases account for only 5 deaths, but the overall SMR is elevated at almost 300. There is one statistically significant excess for the calendar period 1955-1959. All five deaths occurred at 30+ YSFE, at 60 years or older, and at 30+ years of employment.

Gallbladder diseases include cholelithiasis (gallstones), cholecystitis (inflammation of the gallbladder), and haematobilia (bleeding into the biliary passages). Known risk factors include obesity, trauma and a diet high in cholesterol (Schoenfield, 1977). We are not aware of any occupational risk factors that have been identified for mortality from these diseases, and results for this cause of death have not been reported in other studies of fire fighters.

Summary of Findings

Although there was no apparent excess risk due to mortality from cardiovascular diseases as a group in this cohort of fire fighters, for one specific cause of death in this group, aortic aneurysm, a statistically significant excess was demonstrated. Further, risk of death from acute myocardial infarction and diseases of the veins, lymphatics and other circulatory system diseases was somewhat elevated, particularly when one considers that these fire fighters were much healthier than the general population when they entered their profession.

In terms of cancer, statistically significant excesses were seen for cancer of the brain and other nervous system tumours and for "other" malignant neoplasms. Other cancer sites exhibiting elevated, although not statistically significant, risks were: pharynx, rectum and rectosigmoid junction, pancreas, prostate, lymphosarcoma and reticulum cell sarcoma, and lymphatic leukaemia. Some degree of elevated mortality risk was also seen for respiratory and digestive system diseases as follows: chronic bronchitis, emphysema, and asthma as a group, cirrhosis of the liver, and gallbladder diseases.

Discussion

This epidemiologic study suggests that fire fighting is a hazardous occupation in terms of increasing risk of death due to several causes, such as cancer of the brain and other tumours of the nervous system, "other" malignant neoplasms, and aortic aneurysm. There are also suggestive increases in risk for several other causes of death, including various cancers, and diseases of the respiratory, circulatory and digestive system.

Study limitations

As in any epidemiologic study, the associations reported here may be due to chance, bias, or the effects of factors not measured, such as smoking, alcohol or diet. In terms of smoking, since there is no elevation in risk for lung cancer, excess smoking by fire fighters compared to the general population is most likely not an explanation for any of the risks seen in this study.

Causal relationships are rarely established in one study. In addition to the strength of an association, it is biological plausibility and consistency across several studies that will establish a causal link between an occupation and a disease. It is also important to note that an epidemiologic study cannot definitively establish the absence of a risk but can often place an upper bound on overall risk. Furthermore, this study examined mortality. The extent to which this is an indicator of morbidity depends on the specific cause of death.

This study shares a common feature of occupational cohort studies in that no information is available on an individual's exposure. In the absence of information on individual fire fighter activity and fire hall assignment and in light of the small variation in job assignment of fire department employees, the only feasible index of exposure for this cohort study is number of years of employment. In this respect, the results are directly comparable to those from other studies.

Potential exposures

There is very little documentation of the prevalence of actual exposures which fire fighters receive. In an effort to obtain information specific to urban Canadian fire fighters, a group of researchers was commissioned to provide exposure data on fire fighters as part of their on-going study of cancer in Metropolitan Montreal (Siemiatycki, 1991). Exposure assessment is done by a team of industrial hygienists and chemists based on a detailed personal interview. Forty-seven fire fighters were interviewed, 40 of whom were from urban areas. Several substances were identified as exposures for the majority of fire fighters (85%-98% prevalence of exposure): polycyclic aromatic hydrocarbons (PAHs), pyrolysis fumes, aliphatic aldehydes, carbon monoxide, formaldehyde, particulates from wood and soot, nitrogen oxides, sulphuric acid, chlorinated alkanes, benzopyrene, chrysotile asbestos, sulphur dioxide, monocyclic aromatic hydrocarbons, chlorine, phosgene, engine emissions, carbon tetrachloride, solvents, alkanes, hydrogen cyanide, hydrogen chloride, plastics pyrolysis products, benzene, cyanide, and lubricating oils. It is known, however, that very different exposures can occur for two fire fighters at the same fire even if they are side by side (Brandt-Rauf et al., 1988). In experimental studies, carbon monoxide, formaldehyde and acrolein are among chemicals identified as pyrolysis products which occur even at relatively low temperatures (i.e., in the overhaul stage of a fire) (Edgerley, 1980).

In a recent study which measured substances in 22 fires (15 residential, 6 training, and 1 car fire) which occurred in various urban fire departments across the U.S., carbon monoxide was the most common contaminant (Jankovic et al., 1991). Distinctions were made between measurements taken in the knockdown (bringing the main body of fire under control) and overhaul (searching for and extinguishing hidden fires), and those taken inside and outside self-contained breathing apparatus. Ranked in order of decreasing occurrence in the knockdown phase, the substances found were: carbon dioxide, carbon monoxide, ethylene, methanol, methane, acetaldehyde, acrolein, benzene, hydrogen cyanide, and hydrochloric acid. Several of these chemicals were also measured, although at lower concentrations, in the overhaul phase, which is necessary mainly for larger fires. These investigators found that masks are usually not worn during the overhaul phase and that the low concentrations found which were attributed to "inside mask" were due either to early removal or non-use of masks and not to leakage of the masks themselves. The increased use of synthetic materials since the 1950's has increased concern about fire fighters' exposure to potentially toxic chemicals.

Conclusion

This epidemiologic study suggests that fire fighting is a hazardous occupation at least in terms of increasing risk of death due to several causes, such as cancer of the brain and other tumours of the central nervous system, "other" malignant neoplasms, and aortic aneurysm. There are also suggestive increases in risk for several other causes of death, including various cancers, and diseases of the respiratory, circulatory and digestive systems. For several causes of death we have argued that, because fire fighters are selected for their excellent health, where moderate increases are apparent compared to the general population, these are probably underestimates of the magnitude of the true hazards of fire fighting.

TABLE 4 Standardized Mortality Ratios by Years Since First Employment for Selected Causes
	Years since first exposure
	< 20 Years			20-29 Years			30 Years
Cause	Obs.	SMR	95% Cl	Obs.	SMR	95% Cl	Obs.	SMR	95% Cl
All Causes	120	79	65-94	143	95	80-112	514	99	91-108
All Cancers	28	120	77-170	36	94	66-131	135	106	89-125
pharynx	0	0	--	1	122	3-679	3	188	39-549
rectum	1	125	3-696	2	146	18-527	10	179	86-329
pancreas	1	100	3-557	2	95	12-344	11	157	78-281
lung	1	23	1-128	13	103	55-176	40	100	71-136
melanoma	1	91	2-507	1	130	3-724	0	0	--
prostate	0	0	--	2	244	30-881	14	126	69-211
testis	3	333	69-973	0	0	--	0	0	--
brain & other nervous system	6	286	105-623	2	99	12-356	6	222	81-483
lymphatic leukaemia	0	0	--	0	0	--	4	267	73-684
other malignant neoplasms	1	143	4-797	4	288	78-737	15	306	171-505
Circulatory System	37	93	65-128	77	118	93-147	267	96	85-108
acute myocardial infarction	22	102	64-154	48	126	93-167	135	102	86-121
arteriosclerosis	1	100	3-557	3	165	34-482	20	141	86-218
aortic aneurysm	0	0	--	3	303	62-886	16	232	133-377
diseases of veins,     lymphatics, etc.	1	143	4-797	2	211	25-760	5	161	52-376
Chronic bronchitis, asthma   & emphysema	0	0	--	2	80	10-290	26	146	95-214
Cirrhosis of the liver	4	87	24-223	10	152	73-280	12	141	73-246
Gallbladder diseases	0	0	--	0	0	--	5	385	125-898

TABLE 5 Standardized Mortality Ratios by Age for Selected Causes
	Age
	< 60 Years			60 Years
Cause	Obs.	SMR	95% Cl	Obs	SMR	95% Cl
All Causes	337	91	81-101	440	97	88-107
All Cancers	87	107	86-132	112	104	86-125
pharynx	1	61	2-340	3	240	49-701
rectum	4	138	38-353	9	191	87-363
pancreas	4	97	26-248	10	170	81-313
lung	22	91	57-138	32	97	66-137
melanoma	2	93	11-336	0	0	--
prostate	2	153	19-553	14	130	71-218
testis	3	268	55-783	0	0	--
brain & other nervous system	10	198	95-364	4	204	56-522
lymphatic leukaemia	0	0	--	4	336	92-860
other malignant	9	304	139-577	11	266	133-476
Circulatory System	150	110	93-129	231	93	81-106
acute myocardial infarction	93	117	94-146	112	100	82-120
arteriosclerosis	4	113	31-289	20	149	91-230
aortic aneurysm	3	163	33-473	16	245	140-398
veins, lymphatics & other	3	148	31-433	5	181	59-422
Respiratory System	7	57	23-117	38	112	79-154
chronic bronchitis, asthma & emphysema	2	43	5-155	26	155	101-227
Digestive System	20	96	58-148	24	156	100-232
cirrhosis of the liver	18	130	77-205	8	138	60-272
gallbladder diseases	0	0	--	5	420	136-980

References

Bates, J.T. Coronary artery disease in the Toronto fire department. Journal of occupational medicine. Vol.29,no.2 (Feb., 1987). p.132-135.

Beaumont, J.J.; Chu, G.S.T.; Jones, J.R.; Schenker, M.B.; Singleton, J.A.; Piantanida, L.G.; Reiterman, M. An epidemiologic study of cancer and other causes of mortality in San Francisco firefighters. American journal of industrial medicine. Vol. 19 (1991). p.357-372.

Becklake, M.R.; Bourbeau, J.; Menzies, R.; Ernst, P. The relationship between acute and chronic airway responses to occupational exposures. Current pulmonology. Vol.9 (1988). p.25-66.

Breslow, N.E.; Day, N.E. Statistical Analysis in Cancer Research, V. II: The Design and Analysis of Cohort Studies. Lyon: IARC Scientific Publications No. 82, 1987.

Demers, P.A.; Heyer, N.J.; Rosenstock L. Mortality among firefighters from three northwestern United States cities. British journal of industrial medicine. Vol.49 (1992) (in press).

Edgerley, J. Oxidative pyrolysis of plastics. Fire and materials. Vol.4,no.2(1980). p.77-82.

Eliopulos, E.; Armstrong, B.K.; Spickett, J.T.; Heyworth F. Mortality of fire fighters in western Australia. British journal of industrial medicine. Vol.41,no.2(May, 1984). p.183-187.

Grimes, G.; Hirsch, D.; Borgeson, D. Risk of death among Honolulu fire fighters. Hawaii medical journal. Vol. 50, no. 3 (1991). p.82-85.

Guidotti, T.L. Mortality of urban firefighters in Alberta, 1927-1987. American journal of industrial medicine (1992) (in press).

Hansen, E.S. A cohort study on the mortality of firefighters. British journal of industrial medicine. Vol. 47. (1990). p.805-809.

Heyer, N.; Weiss N.S.; Demers, P.; Rosenstock L. Cohort mortality of Seattle fire fighters: 1945-1983. American journal of industrial medicine. Vol.17 (1990). p.493-504.

Higgins, M.; Thom, T. Epidemiology and natural history of COPD. Current pulmonology. Vol.9 (1988). p.1-24.

Howe, G.R.; Burch, J.D. Fire fighters and risk of cancer: an assessment and overview of the epidemiologic evidence. American journal of epidemiology. Vol. 132, no. 6 (1990). p.1039-1050.

Jankovic, J.; Jones, W.; Burkhart, J.; Noonan, G. Environmental study of firefighters. Annals of occupational hygiene. Vol.35 (1991). p.581-602.

Kristensen, T.S. Cardiovascular diseases and the work environment. A critical review of the epidemiologic literature on chemical factors. Scandinavian journal of work, environment and health. Vol. 15 (1989a). p.245-264.

Kristensen, T.S. Cardiovascular diseases and the work environment. A critical review of the epidemiologic literature on nonchemical factors. Scandinavian journal of work, environment and health. Vol. 15 (1989b). p.245-264.

Lindsay, J.; Hurst, J.W. The Aorta. New York: Grune and Stratton, Inc., (1979).

Mastromatteo, E. Mortality in city firemen--II. A study of mortality in firemen of a city fire department. A.M.A. archives of industrial health Vol.20 (Sept., 1959). p.227-233.

Meijers, J.M.M.; Swaen, G.M.H.; Volovics, A.; Lucas, L.J.; Van Vliet, K. Occupational cohort studies: The influence of design characteristics on the healthy worker effect. International journal of epidemiology. Vol.18,no.4 (1989). p.970-975.

Musk, A.W.; Monson, R.R.; Peters, J.M.; Peters, R.K. Mortality among Boston fire fighters, 1915-1975. British journal of industrial medicine. Vol.35,no.2 (May, 1978). p.104-108.

Peters, J.; Theriault, G.; Fine, L.; Wegman, D. Chronic effects of fire fighting on pulmonary function. New england journal of medicine. Vol.291 (1974). p.1320-1322.

Robins, J.M. A graphical approach to the identification and estimation of causal parameters in mortality studies with sustained exposure periods. Journal of chronic diseases. Vol.40(Suppl.2) (1987). p.139S-161S.

Rosenstock, L.; Demers, P.; Heyer, N.J.; Barnhart, S. Respiratory mortality among fire fighters. British journal of industrial medicine. Vol.47 (1990). p.462-465.

Sama, S.R.; Martin, T.R.; Davis, L.K.; Kriebel, D. Cancer incidence among Massachusetts firefighters, 1982-1986. American journal of industrial medicine. Vol. 18 (1990). p.47-54.

Shannon, H.; Jamieson, E.; Walsh, C.; Julian, J.; Fair, M.; Buffet, A. Comparison of individual follow-up and computerized record linkage using the Canadian National Mortality Data Base. Canadian journal of public health. Vol.80 (1989). p.54-57.

Sherman, C.B.; Barnhart, S.; Miller, M.F.; Segal, M.R.; Aitken, M.; Schoene, R.; Daniell, W.; Rosenstock, L. Firefighting acutely increases airway responsiveness. American review of respiratory disease. Vol.140 (1989). p.185-190.

Shoenfield, L.J. Diseases of the gallbladder and biliary system. New York John Wiley and Sons (1977).

Siemiatycki J. Risk factors for cancer in the workplace. Boca Raton: CRC Press (1991).

Sinks, T.; Smith, A.B.; Rinsky, R.; Watkins, K.; Shults, R.; Steele, G. Health Hazard Evaluation Report HETA-89-116-2094 Westing-house Electric Corporation, Bloomington, Indiana. Cincinnati, Ohio: NIOSH (Jan., 1991).

Sparrow, D.; Bosse, R.; Rosner, B.; Weils, S. The effect of occupational exposures on pulmonary function--a longitudinal evaluation of fire fighters and non-firefighters. American review of respiratory diseases. Vol.123 (1982). p.319-322.

Thomas, T.L.; Waxweiler, R.J. Brain tumours and occupational risk factors: A review. Scandinavian journal of work, environment, and health. Vol.12 (1986). p.1-15.

Tornling, G.; Gustavsson, A.; Gustavsson, P.; Hogstedt, C.; Soder-holm, M. Dodsorsaker bland brandman i Stockholm. Omradesvisa undersokningar av kemiska miljoer (25: Dec., 1990).

Vena, J.E.; Fiedler, R.C. Mortality of a municipal-worker cohort; IV. fire fighters. American journal of industrial medicine. Vol.11, no.6 (1987). p.671-684.