Patterns of Mortality Among Plumbers and Pipefitters

Posted on: August 7, 2016 by in Uncategorized
No Comments

Patterns of Mortality Among Plumbers and Pipefitters

Kenneth P. Cantor, PhD, James M. Sontag, PhD, and Marian F. Heid
A proportionate mortality ratio (PMR) study was undertaken of 7,121 members and
retirees of the United Association of Plumbers and Pipefitters in California who died in
1960-79. The PMR for all malignant neoplasms was 1.24, with a major contribution
from lung cancers (PMR = 1.41). Lung cancer PMRs were consistently elevated,
through the 20-year study period, across the pipe trades and within different birth
cohorts. Sixteen mesothelioma deaths occurred, suggesting asbestos as a risk factor.
PMRs for malignancies of the stomach, kidney, brain, and lymphopoietic system were
also elevated, especially among plumbers. Chronic rheumatic heart disease, emphysema,
liver cirrhosis, and all external causes of death were the major non-cancer causes with
significantly elevated PMRs. There were significant deficits in diabetes mellitus, all
pneumonia, chronic nephritis, and vascular lesions of the central nervous system (CNS).
PMRs for successive birth cohorts among all study subjects revealed decreasing emphysema
risk, suggesting previous reduction of a risk factor for this disease. Among
plumbers, PMRs for death due to several non-respiratory malignancies showed an
increasing trend with recency of birth cohort.
Key words: plumbers, pipefitters, asbestos, mesothelioma, lung, stomach, brain and kidney cancer,
Plumbers and pipefitters encounter many hazardous materials, among them
asbestos, metal fumes, and gases from welding, brazing, and soldering and solvents
used to join plastic pipe. Their work includes installation of piping and equipment in
residential, commercial, and industrial construction sites where they may also be
exposed to hazardous materials from the activities of other trade craftspersons.
The United Association of Journeymen and Apprentices of the Plumbing and
Pipe Fitting Industry of the United States and Canada (UA) is the leading labor
representative for plumbers and pipefitters in the United States. In response to
concerns of the UA General Office and the UA’s California Pipe Trades Council
about the health of their members, we embarked on this study to evaluate patterns of
mortality among deceased members and retirees of California locals who died in the
years 1960-79. Much of the data for this study was provided by the individual UA
local unions in the State of California and the Central Data Processing Department of
Environmental Epidemiology Branch, National Cancer Institute, Bethesda, Maryland 20892.
Address reprint requests to: Dr. Kenneth P. Cantor, Enviromental Epidemiology Branch, National
Cancer Institute, Landow Building, Room 3C08, Bethesda, MD 20892.
the UA located in Washington, D.C. Liaison for the study was provided by the
California Pipe Trades Council and the UA’s Department of Safety and Health.
Although today’s plumbers and pipefitters both install and repair pipe and
equipment, there are some differences in the types of systems on which each trade
works. Plumbers install and maintain systems such as potable water, sanitary drainage,
storm drains, and gas supply and distribution. Plumbers also install and maintain
oxygen, nitrous oxide, and vacuum systems in medical facilities. Pipefitters install
and maintain systems such as heating, air conditioning, cooling, and refrigeration;
energy distribution; chemical piping; environmental controls; overland oil and gas
distribution; and power generation.
The trade category “pipefitter” was officially recognized by the UA in 1947.
When the International Union was founded in the late nineteenth century there were
two major trade categories-plumber and steamfitter. Plumbers installed sanitary
plumbing, potable water, and hot water heating. Steamfitters installed steam heating
systems and all other piping. As the industrial and manufacturing capabilities of the
United States became more complex, the work of the steamfitter expanded to include
new systems. This prompted the UA to recognize “pipefitter” as a new classification
of trade category. In current usage, the terms “steamfitter,’ and “pipefitter” are
synonymous. Some UA local unions still use the term steamfitter, but most have
adopted “pipefitter” as the more descriptive term. For the purposes of this study, the
two categories are recognized as the same and are listed in the results as “pipefitter.”
The monthly journal of the UA lists the name and local union affiliation of all
decedents eligible for death benefits. Those eligible include all employed workers and
retirees who remained union members. From lists of deaths that occurred between
January 1, 1960 and December 31, 1979, we identified 7,761 decedents who were
members or retirees of California union locals. Copies of most death certificates were
available from microfilm records at the UA central offices in Washington, D.C.
(deaths occurring in 1968-79) or from California UA locals (deaths in 1960-67). Of
379 certificates not available from the UA, 310 were obtained from state vital statistics
offices. A total of 7,577 (97.6%) death certificates were located. The UA maintains
for most members a record with name, date of birth, dates of union initiation and
separation, local affiliation(s) with dates of changes, and principal trade within the
industry. This record system was automated in 1967-68. Paper records for men who
died in 1960-67 or computer records for those who died in 1968-79 were matched to
7,166 (94.5%) of the death certificates. This study is limited to the 7,121 deaths that
occurred in the U.S. among white male subjects.
All death certificates were coded by an experienced nosologist (MH) according
to the Eighth Revision of the International Classification of Diseases, Adapted for use
in the United States (ICDA-8). Items obtained from death certificates were name,
date of birth, date of death, underlying cause of death, up to three contributing causes
of death, and usual occupation and industry.
The proportionate mortality ratio (PMR) was calculated using an adaptation of
software developed by Monson [1974]. The number of deaths, by cause, was compared
with expected numbers based on cause-specific proportionate mortality by 5-
year age and 5-year calendar period groups for white U.S. males. PMRs were
Mortality Among Plumbers and Pipefitters 75
calculated for the total group and within strata based on principal trade, birth cohort,
year of death, or combinations of these.


Table I shows the number of subjects by year of death and age at death. The
number of decedents increased in each 5-year period, from 1,399 in 1960-64 to 2,026
in 1975-79, for a total of 7,121. This increase may be related to the growth of
California’s population, from 15.7 million (1960) to 23.6 million (1980), and to the
increasing average age of the state’s UA membership. The overall median age at
death was 65, rising from 61 years in 1960-64 to 68 years in 1975-79.
The number of decedents, by last recorded trade was as follows: 3,491 plumbers;
3,225 pipefitters; 163 metal trades apprentices or journeymen; 144 sprinkler fitters;
12 lead burners; and 86 other and trade not recorded. A cross tabulation of last trade
from UA records, with the “usual occupation” from the death certificate, is shown in
Table II. When two occupational groups appeared on the death certificate (eg,
plumber-pipefitter), the decedent was grouped with the first occupation listed. The
agreement between UA records and death certificate occupation was greatest for
plumbers. Of the 3,491 men whose UA records listed “plumber” as last trade, 2,985
(85.5%) also had the trade on their death certificates. Of men with the occupation
“welder” on death certificates, most (89%) were from the pipefitter trade. Shown
also in Table I1 are the median year of birth, age at initiation, and age at death, by
last trade. Differences in these statistics among trades are related primarily to time of
unionization of different segments of the pipe trades.
Table III shows the number of deaths and cause-specific PMRs for all 7,121
decedents. There were significant elevations in deaths from all malignant neoplasms;
stomach cancer; cancer of the lung, bronchus, and trachea; brain cancer; all lymphopoietic
cancer; cancer of “other lymphatic tissue” (25 of the 46 deaths were due to
multiple myeloma); and benign neoplasms. Lung cancer accounted for 185 of the 337
excess deaths from all malignant neoplasms. Mortality from chronic rheumatic heart
disease, emphysema, all diseases of the digestive system, cirrhosis of the liver,
accidents, and suicides was also significantly elevated. The number of deaths from
several causes was significantly lower than expected. These included all infectious



was the underlying cause of death in sixteen cases. Two were expected, based on data from R. Connelly [National Cancer Institute, personal communication]. Twelve of the sixteen mentioned pleura or a site suggesting the pleura, one specified peritoneum, and three occurred at unspecified sites. In each 5- year segment of the study period, the number of reported mesotheliomas increased. There was one in 1960-64, three in 1965-69, four in 1970-74, and eight in 1975-79. None of the mesothelioma deaths was among the 155 decedents whose usual industry cited on the death certificate was a ship, boat, or naval yard. Table IV shows PMRs for deaths from selected causes by major trade. Among plumbers, the PMRs for all malignant neoplasms; stomach cancer; cancer of the lung, bronchus, and trachea; kidney cancer; brain cancer; all lymphopoietic cancer; and lymphosarcoma/reticulosarcoma were significantly elevated. Pipefitters had a significantly elevated PMR for all cancers combined, primarily due to excess lung cancer. In contrast with plumbers, the PMRs among pipefitters for other cancer sites did not show statistically significant excesses. Deaths due to all lymphopoietic malignancies, and especially “other lymphatic cancer,” approached a statistically significant excess. Several non-cancer causes of death had significantly elevated PMRs among both plumbers and pipefitters. These included chronic rheumatic heart disease, emphysema, and cirrhosis of the liver. The number of deaths from all external causes was significantly elevated among pipefitters but not plumbers. Both trades experienced a significantly lower number of deaths than expected due to diabetes mellitus, vascular lesions of the CNS, and pneumonia. Table V shows PMRs for selected causes of death by year-of birth strata for all 7,121 decedents. The PMR for all malignant neoplasms was 1.17 in the earliest birth cohort (< 1900) and 1.38 for those born after 1919. The PMR for all non-respiratory malignant neoplasms was 1.03 in the earliest birth cohort and 1.34 in the most recent birth cohort. The opposite pattern was seen for circulatory diseases. The PMR for lung cancer was elevated in all birth cohorts, with no discernible positive or negative gradient. The PMR for chronic rheumatic heart disease decreased from 2.00 for those born before 1900 to 0.72 for those born after 1919, and the PMR for emphysema decreased from 1.77 to 0.55. Deaths occurring among all trades in each of the four 5-year periods of the study were grouped and cause-specific PMRs within each period were calculated. The PMR for chronic rheumatic heart disease was greater in each successive 5-year time period, ranging from 1.01 (95% CI = 0.63-1.62, 17 observed deaths) in 1960- 64 to 2.22 (95 % CI = 1.55-3.18, 28 observed deaths) for deaths occurring in 1975- 79. When considered with the pattern of PMRs by birth cohort, this suggests a higher risk in older populations due to a long latent period. The opposite pattern was shown for emphysema, with the PMR decreasing from 1.61 (95% CI = 1.13-2.29, 30 observed deaths) in 1960-64 to 0.94 (95% CI = 0.67-1.32, 33 observed deaths) in 1975-79. In combination with the birth cohort trend, the temporal trend for emphysema mortality suggests that an exposure of etiologic importance decreased in intensity or was eliminated in the past.

The PMRs for respiratory cancer were elevated in all birth cohorts in both trades and were statistically significant in all but the most recent cohort of plumbers. No notable pattern of PMRs by birth cohort for other cancers was discernible among pipefitters (Table VII). However, for the combined non-respiratory cancers among plumbers there was a pattern of elevated and increasing PMRs. PMRs for plumbers born after 1919 were elevated and statistically significant for several types of cancer, including stomach, large intestine, skin melanoma, brain, and all lymphopoietic malignancies.


The PMR for lymphosarcomalreticulosarcoma was elevated in the most recent birth cohort but no overall trend was apparent. The PMR for other lymphopoietic malignancies was also elevated in the 1920+ birth cohort and was higher than among decedent plumbers born earlier. Among plumbers, the PMRs for four major non-cancer causes (ie, arteriosclerotic heart disease, vascular lesions of the CNS, non-malignant digestive system disease, and all external causes), varied in a nonspecific manner across birth cohorts and showed no consistent increasing or decreasing patterns. When viewed in conjunction with PMRs by birth cohort, PMRs calculated by calendar year of death (Table VIII) can be helpful in evaluating possible increased risk due to recently introduced materials in the workplace. PMRs for respiratory system malignancies were statistically significantly elevated for each 5-year period but with no apparent trend. The PMR for all non-respiratory cancer was most elevated in the most recent time period, 1975-79. PMRs for several non-respiratory cancers were elevated at statistically significant levels in the 1970-74 or the 1975-79 periods, and included cancers of the stomach, large intestine, testis, kidney, and brain. Of the lymphopoietic cancers, PMRs for lymphoma and reticulosarcoma were elevated in each 5-year period but with no apparent trend. “Other lymphatic cancers,” a grouping that includes multiple myeloma, had slightly elevated PMRs in all periods except in 1960-64 but not at statistically significant levels. There was no consistent pattern of increasing or decreasing PMRs across the 5-year time periods for any of the four major non-cancer causes of death. DISCUSSION We observed elevations in the PMRs for the overall study population for cancers of the lung and brain, all lymphopoietic cancer, and cancer of “other lymphatic tissue.” Sixteen mesothelioma deaths were found. PMRs for lung cancer were consistently elevated when analyses were done by trade, birth cohort, and year of death. Elevations in the PMRs for other cancer sites were primarily restricted to members of the plumbing trade. Plumbers showed excesses in the number of cancers of the stomach, kidney, and brain, as well as all lymphopoietic cancer and lymphosarcoma/reticulosarcoma. Further evaluation of PMRs among plumbers by birth cohort and year of death revealed that the highest PMRs occurred in the most recently born groups and the most recent periods of death. These PMRs were for cancers of the stomach, large intestine, skin, testis, kidney, brain, and all lymphopoietic cancer. Malignancies of the trachea, bronchus, and lung accounted for more than half of the excess deaths included in the “all malignant neoplasms” category. Lung cancer PMRs were significantly elevated in each trade group in all birth cohorts except the earliest (ie, born prior to 1890) and in each 5-year segment of the 1960-79 study period, although no pattern of increasing or decreasing risk by birth cohort or by year of death was apparent. Other studies of plumbers and pipefitters [Englund et al, 1979, Kaminski et al, 19801 and numerous surveys that have evaluated mortality in the pipe trades also have reported elevated risks of respiratory cancer (studies before 1980 reviewed by Kaminski et al, 1980; Dubrow and Wegman, 1984; Peterson and Milham, 1980; Howe and Lindsay, 1983; Walrath et al, 1985). Plumbers and pipefitters have been reported to smoke more than workers in other occupations (Walrath et al, Mortality Among Plumbers and Pipefitters 85 1985; Levin et al, 1985), and some of the excess in respiratory cancer mortality may be tobacco-related. The calculated risk ratio for lung cancer among plumbers and pipefitters due to cigarette smoking is in the range of 1.14-1.29 (smoking rates by occupation provided by L. Levin, Epidemiology and Biostatistics Program, National Cancer Institute). It is thus not likely that tobacco use alone can fully explain the excess in respiratory cancers. Occupational exposure to asbestos has been common among the pipe trades and it is likely an important risk factor contributing to the lung cancer excess. Sixteen deaths were attributed to mesothelioma, twleve of which were of the pleura or “chest. ” Approximately two pleural mesotheliomas were expected, based on national data for the US. [R. Connelly, National Cancer Institute, personal communication]. A relative risk of 3.9 for mesothelioma among plumbers and pipefitters has been reported in a Connecticut case-control study [Teta et al, 19831. A study in Sweden that used tumor registry data linked to census information reported a relative risk of 4.8 [Malker et al, 19851. In another study of the pipe trades, most mesothelioma deaths occurred among steamfitters [Kaminski et al, 19801. In this study, however, mesothelioma deaths were distributed among plumbers and the other trades. This finding suggests that asbestos exposures were not restricted to pipefitterdsteamfitters. An elevated PMR for stomach cancer was observed for plumbers, and a nonsignificant excess was also found among pipefitters. No elevation of gastric cancers among all pipe trades was reported by Kaminski et al [1980], but plumbers were at slightly elevated risk (PMR =1.30, 16 observations). In a Swedish cohort study of 18,521 plumbers, Englund et al[1978] reported a standardized incidence ratio of 1.72 (p < 0.001) for stomach cancer, based on 30 observed cases. A British death certificate review reported a PMR of 1.21 (p < 0.05) for stomach cancer in the pipe trades, with 147 observed cases [Registrar General, 19711. Asbestos exposures may be important in the etiology of stomach malignancies [Selikoff et al, 19791. If asbestos is a major risk factor for this disease, however, it is not apparent why the elevated PMR for stomach cancer was primarily among plumbers, since asbestos exposure appears to be distributed across the trades and lung cancer PMRs are consistently elevated among them. If the elevated PMRs did not arise by chance, other factors unique to plumbers may be important in stomach cancer etiology. The elevated brain cancer mortality among plumbers observed in this study has not previously been reported. However, it has been found in other building trades, including painters, paperhangers, and glaziers; brickmasons, stonemasons, and tile setters; carpenters; and electricians [Thomas and Waxweiler, 19851. Plumbers are more likely than pipefittershteamfitters to work at residential construction sites where other building trades are also employed. Some workplace hazard at residential construction sites may be related to the elevated brain cancer in several of the construction trades. The PMR for kidney cancer was significantly elevated among plumbers, but not among other trades. Similar findings have not been reported elsewhere. A casecontrol study reported elevated renal cell carcinoma among “welders/plumbers” [Asal et al, 19821. Exposure to lead in experimental animals induces renal tumors, but there is no evidence to support the association in humans [McLaughlin and Schuman, 19831. Coke oven workers exposed to polycylic aromatic hydrocarbons (PAH) and other compounds have been reported to have elevated renal cancer risk [Redmond et al, 19721. In the past, plumbers may have been exposed to PAH and 86 Cantor, Sontag, and Heid lead fumes during the sealing of cast iron pipe joints. This process involved pouring molten lead into be1 and spigot joints that were sealed with oakum, a hemp rope impregnated with coal tar pitch. Excess mortality from renal cancer also has been observed in a study of asbestos and insulation workers [Selikoff et al, 19791. Asbestos in drinking water was associated with kidney cancer incidence among women but not among men, in census tracts of the San Francisco Bay area [Kanarek et al, 19801. Our data revealed overall excess in PMR for all lymphopoietic cancer, with variable excesses in lymphosarcoma/reticulosarcoma and “other lymphatic cancer” among trades, birth cohorts, and year-of-death strata. Elevated mortality from lymphosarcoma and reticulosarcoma has also been observed in other studies of plumbers and pipefitters [Guralnick, 1963; Dubrow and Wegman, 1984; Kaminski et al, 19801. A series of seven cases of small cleaved follicular center cell lymphoma, a form of lymphosarcoma, reported among California plumbers and pipefitters by Dolan et a1 [ 19831, raised the possibility of a link with plastic pipe fabrication and installation. Plastic pipe for waste water disposal was introduced in California in the mid-1960s. Our data show elevated PMR for lymphosarcoma/reticulosarcoma both before and after the introduction of plastic pipe, suggesting the involvement of other risk factors. Occupational exposure to asbestos was reported in a case-control study of large-cell lymphomas of the gastrointestinal tract and oral cavity [Ross et al, 19821. Since pathological diagnoses for non-Hodgkin’s lymphomas are unusual on death certificates, we cannot determine whether the lymphomas among plumbers and pipefitters in this study are of the large-cell type or of the small follicular cell variety. Mortality rates for the non-Hodgkin’s lymphomas have been generally elevated in California compared to other areas in the U.S. Thus, some of the excess mortality may be related to geographical influences [Cantor and Fraumeni, 19801 independent of occupation. The pattern of decreasing risk for emphysema with recency of birth is provocative, suggesting earlier removal or reduction in levels of an important occupational risk factor. Other studies of the pipe trades, especially those reporting on deaths before 1970, showed elevated PMRs for emphysema [Milham, 1976; Peterson and Milham, 19801. Excess mortality from emphysema has been observed among coal miners exposed to coal dust [worth, 19841 and burners in shipyards exposed to a variety of metal fumes and particulates [Beaumont and Weiss, 19801. Cadmium has been suggested as a risk factor for emphysema [Hallenbeck, 19841, and the past use of solder or welding rods containing cadmium may be important. Chronic rheumatic heart disease also showed declining mortality in recent birth cohorts. Streptococcal infection is the major risk factor associated with rheumatic heart disease. Other risk factors are low socioeconomic status, overcrowding, ethnic status, and malnutrition [Levy, 19851. The downward pattern in risk with birth cohort may reflect an improved relative socioeconomic status, while the upward trend with year-of-death suggests a higher risk in older populations due to a long latent period. Elevated PMRs among plumbers born after 1919 and/or dying in 1970-74 or 1975-79 were found for the grouped non-respiratory malignancies and, in particular, for cancers of the stomach, large intestine, kidney, and brain. PMRs of two or more were noted for testis cancer and skin melanoma, although the numbers of deaths were small and the PMRs were not statistically significant. These findings may be related to the healthy worker effect, which is primarily operable in populations of working age, since successive birth cohorts had higher proportions of working-age decedents. Mortality Among Plumbers and Pipefitters 87 However, elevated PMRs for these malignancies were not accompanied by concomitant PMR decreases for the four major non-cancer causes of death which would be expected if the healthy worker effect had been an important factor. In addition, the birth cohort analysis of pipefitters did not reveal similar patterns of increasing PMRs. It is unlikely that the analyses by 5-year year-of-death strata were subject to similar differential impacts of the healthy worker effect, because the age at death increased and there were fewer men of working age in successive year-of-death groups. We were not able to determine PMRs by the number of years employed as a plumber or pipefitter, because the available information on the first year employed in the pipe trades was of uneven quality. The UA’s chronological work history started with a member’s initiation date, though many members had already worked in the trade before joining the union (as indicated by a median initiation age of 33 years among plumbers and 38 among pipefitters). At initiation, a prospective member listed previous experience, and this information was available for about half of the study subjects. Since the reported amount of experience may positively affect acceptance as a new member, the self-reported number of years is subject to bias and is thought by UA officials to be of questionable accuracy. The PMR method has widely recognized shortcomings and caution is warranted in interpreting these results. Underrepresentation of some causes of death will result in inflated estimates of risk for other causes. Among employed populations the healthy worker effect may be operative, whereby actively employed persons experience relatively low mortality from chronic disabling diseases of the circulatory and respiratory system [Fox and Collier, 19761. PMR estimates of cancer risk are usually inflated under these conditions [DecouflC et al, 19801. An additional caution is that multiple significance testing may have resulted in associations that arose from chance alone. In principle it is inappropriate to compare indirectly standardized mortality rates or ratios with one another [Monson, 19801, since populations with similar underlying rates may appear to differ, especially if there are differences in their age distributions. For this reason, PMR patterns among successive year-of-birth or year-of-death groups were interpreted qualitatively, and formal tests for trend were not applied. Despite the shortcomings of the PMR methodology, the concordance of our findings with other studies for elevated risk of pulmonary, stomach, and lymphatic cancer provides some assurance of validity. In summary, our results confirm and strengthen earlier findings of increased risk of respiratory cancer and mesothelioma among plumbers and pipefitters, as well as observations made in other settings of excess stomach and lymphopoietic cancer. New findings from this investigation are elevated PMRs for brain and kidney cancers, especially among members of the plumbing trade. The lower PMRs from emphysema and chronic rheumatic heart disease in successive birth cohorts may indicate the earlier removal or reduction of important occupational exposures (emphysema) or very long latency (rheumatic heart disease). The observation of increasing PMRs for several malignancies in successive birth cohorts of plumbers is provocative. These findings raise the possibility that elevated risk may be associated with recently introduced exposures. They deserve further exploration in studies that assess individual exposures and work practices. 88 Cantor, Sontag, and Heid ACKNOWLEDGMENTS We are grateful for the assistance of the General Office, the Data Processing Department, and the Department of Safety and Health of the United Association National Office, as well as individual local unions in California and the California Pipe Trades Council. We also thank A. Blair, S. Hoar, and T. Thomas for helpful suggestions, Stu Kociol for data handling, and Kimberly Young for manuscript preparation.