Portal de Periódicos da CAPES

Long-Term Health Experience of Jet Engine Manufacturing Workers

2013; Lippincott Williams & Wilkins; Volume: 55; Issue: 6 Linguagem: Inglês

10.1097/jom.0b013e3182851605

1536-5948

Gary M. Marsh, Jeanine M. Buchanich, Ada O. Youk, Sarah Downing, Nurtan A. Esmen, Kathleen Kennedy, Steven E. Lacey, Roger P. Hancock, Mary Lou Fleissner, Frank S. Lieberman,

Occupational and environmental lung diseases

In May 2000, the Connecticut Department of Public Health (CTDPH) began an investigation of a perceived increase of brain cancer at the Pratt &Whitney (P&W) jet engine manufacturing facility in North Haven, Connecticut. By August 2001, the CTDPH investigation had identified several cases of primary, malignant brain cancer, all of which were confirmed by the CT Tumor Registry Program. All cases occurred among white male workers, and most of the cases were a common type (glioblastoma [GB]). A preliminary comparative cancer incidence analysis conducted by the CTDPH was inconclusive, and the CTDPH recommended that a more comprehensive and rigorous investigation be undertaken by an independent research group. In August 2001, at the recommendation of the CTDPH, two of the authors (G.M. and N.E.) were asked by P&W to evaluate the feasibility of conducting a formal epidemiological investigation of the suspected brain cancer excess. The feasibility study, which included an evaluation of information assembled by P&W for North Haven and five other central CT plant sites (East Hartford, Middletown, Rocky Hill, Southington-Aircraft Road, and Southington-Newell Street), concluded that sufficient data were available for these plants to warrant a formal epidemiological investigation and recommended that this include two other P&W central CT plants (Cheshire and Manchester Foundry). Only two of these plants (East Hartford and Middletown) are currently operating. The other plants were closed and the operations contained therein were transferred during the years 1988 to 2011. In 2002, we began work on a large, multipart investigation. Our primary objectives were to determine whether mortality or incidence rates from central nervous system (CNS) neoplasms, including GB, were elevated among workers at the North Haven site or seven other P&W facilities serving as comparison sites, and whether these rates were associated with specific workplace experiences or exposures. The epidemiological and biostatistical component of the study, conducted by the Department of Biostatistics, Center for Occupational Biostatistics and Epidemiology, University of Pittsburgh, included historical cohort mortality and cancer incidence studies, a nested case-control study of malignant and benign neoplasms of the brain and other CNS cancer sites, and a genetics study based on tissue specimens obtained from cases of malignant brain neoplasms (the genetics study was not completed because of lack of relevant results from the main studies). A companion exposure assessment project, conducted independently by the University of Illinois at Chicago, Division of Environmental and Occupational Health Sciences, characterized the historical work practices and exposures that occurred within the P&W study plants. Because the only well-established association between an environmental or occupational exposure and the occurrence of CNS neoplasms is with exposure to high doses of ionizing radiation, no specific occupational factors at the North Haven or other P&W study plants (ie, exposures to specific agents or particular job assignments) were hypothesized as risk factors for the perceived brain cancer excess. Thus, we considered our epidemiological investigation as exploratory in nature and did not attempt to test any specific a priori etiologic hypotheses. Comprising close to a quarter million subjects and a 53-year observation period (1952 to 2004), our study is one of the largest and most comprehensive cohort studies ever undertaken in the occupational setting and the first and only large-scale study of workers in this industry. For example, our mortality study cohort of 222,123 subjects contributed more than 7.6 million person-years of observation, of which about 1.4 million or 18% were among workers employed more than 5 years and followed for 20 or more years from first employment. We reported results of our investigation in nine papers,1–9 six of which appeared in this journal.1,2,6–9 Two papers in the series discussed methodological issues with the accurate assessment of incidence CNS neoplasm cases4 and the analysis of nonmalignant CNS neoplasms.5 Our exposure assessment considered 11 chemical or physical agents on the basis of known or suspected carcinogenic potential in the CNS and/or other organ systems. We generated quantitative estimates for soluble and mineral oil metalworking fluids, nickel, cobalt, chromium, solvents, and the incomplete combustion aerosol from metalworking fluids generated during high-speed and high-temperature grinding (“blue haze”) that was unique to North Haven. Qualitative estimates (exposed/unexposed) were assigned for ionizing radiation, electromagnetic fields, polychlorinated biphenyls, and lead-cadmium. Our estimates showed decreasing trends for all exposures during the study period and the quantitative exposure levels were similar to or less than published data from other industries.7 Because the 11 agents selected for focused examination may not have included all possible neurocarcinogens, the assessment also involved estimation of worker time spent in 20 categories of “parts produced” and 16 categories of “processes performed.” The examination of parts and processes functioned as a screening tool to detect possible excess GB risk in relation to some 3000 exposure agents that were present at P&W over the study period but could not be characterized in the exposure assessment.8 The historical cohort study also provided the basis for a nested case-control study of 723 incident cases of CNS neoplasms (malignant, benign, and unspecified), including 277 GB cases, aimed at obtaining detailed information on personal lifestyle, behavior, and medical histories unavailable at the cohort level. While poor subject participation rates precluded an analysis of the survey data, the case-control study provided the foundation for a more refined exposure assessment. Rule-based job class assignments and imputations for missing work history information were used for the cohort and yielded discrete exposure classes. In the case-control study, line-by-line examination of individual work histories not feasible at the cohort level allowed for job class and imputation adjustments, where necessary, and resulted in continuous exposure estimates.6,7 For CNS neoplasms, the results of our initial mortality and incidence studies were generally consistent, revealing no statistically significant elevations in overall rates among the P&W workforce compared with the corresponding rates in the general populations of the United States and Connecticut. In both studies, internal comparisons within the P&W workforce revealed elevated rates of CNS neoplasms (and GB in the incidence study) among workers in the North Haven plant compared with other plant groups and in certain subgroups of workers from North Haven, but no evidence of an association with general workplace factors at P&W such as duration of employment.1,3 In general, the elevated rates in North Haven were small to moderate and not statistically significant. For example, in our updated cancer incidence study, overall GB incidence among North Haven workers was only 7% higher than that in the general population of Connecticut, and not statistically significant.6 On the contrary, internal cohort comparisons of North Haven workers with workers from the largest study plant (East Hartford [EH]) revealed a moderate 44% elevation in GB incidence that was statistically significant. The statistical significance of this moderate elevation was driven, however, by a statistically significant 31% deficit in GB incidence in the EH plant used as the baseline for comparison.6 Our subsequent evaluations of CNS neoplasm incidence in relation to estimated workplace exposures and our evaluation of GB incidence in relation to workplace experiences with parts and processes also revealed no workplace associations overall and did not explain the elevated rates of GB in North Haven.6,8 Because the core of our investigation was a large, long-term historical cohort study, we also evaluated whether mortality rates from all causes of death combined or from any malignant or nonmalignant cause of death category (excluding CNS neoplasms) were elevated or related to workplace experiences or exposures. With the exception of elevated rates in two study plant groups for certain chronic obstructive pulmonary disease (COPD)–related categories, our evaluation of total and cause-specific mortality rates (excluding CNS neoplasms) found no evidence of elevated rates. The COPD-related mortality excesses were not related to the P&W workplace exposure factors studied. We could not rule out occupational exposures received outside of work at P&W or uncontrolled positive confounding by smoking as reasons for these excesses.2,9 As we end our exploratory investigation prompted in 2000 by the perceived unusual occurrence of GB at the P&W North Haven plant, we draw several conclusions about occupational exposure conditions and the long-term mortality and CNS neoplasm incidence experience of the P&W workforce overall and among workers at the North Haven plant: Occupational exposures to chemical or physical agents with known or suspected carcinogenic potential in the CNS and/or other organ systems have decreased over the time frame of the study and quantitatively estimated levels were similar to or less than published data from other industries. With the exception of elevated COPD-related mortality in two of five study plant groups, our evaluation of total and cause-specific mortality rates (excluding CNS neoplasms) found no evidence of elevated rates. The elevated COPD-related mortality rates in two study plant groups were not related to the occupational factors studied. Occupational exposures received outside of work at P&W or uncontrolled positive confounding by smoking cannot be ruled out as reasons for these excesses. For CNS neoplasms, including GB, the results of our incidence study revealed no statistically significant elevations in overall rates among the P&W workforce compared with rates in the general populations of the United States and Connecticut. Some small to moderate elevations in GB incidence were observed in the North Haven plant, where the index GB cases arose, but most of these were not statistically significant. Statistically significant elevations occurred only when GB incidence in North Haven was compared with GB incidence in the EH baseline plant, where a statistically significant deficit in GB incidence was observed. If not due to chance alone, the small to moderately elevated GB rates in North Haven may reflect unmeasured external occupational factors or nonoccupational factors unique to North Haven or to the baseline EH plant used in the internal comparisons.