6

EPIDEMIOLOGICAL STUDIES

THE Navy Environmental Health Center prepared a comprehensive review of the epidemiological literature in support of its proposed occupational standard for manufactured vitreous fibers (MVF) (NEHC 1997b). There have been well over two dozen publications of epidemiological studies of MVF production facilities in North America and Europe that have examined large cohorts of workers. Those studies have used both historical cohort mortality and case-control study designs. Interest has centered primarily on lung cancer, although other cancer sites and causes of death have been considered. Lee et al. (1995) have reviewed the published work with respect to respiratory system cancers associated with MVF.

The Navy's documentation in Man-Made Vitreous Fibers evaluated data primarily from the three large cohort studies conducted in Europe, the United States, and Canada. Since publication of the Navy's report, several new epidemiological studies (Chiazze et al. 1997; Watkins et al. 1997; Hansen et al. 1999) and updated analyses (Boffetta et al. 1997; Boffetta et al. 1999; Chiazze et al. 1999; Marsh et al. 1996) of earlier studies of MVF have been published. The subcommittee suggests that the Navy update its report to discuss the strengths and limitations of the recent cohort and case-control studies so that their results can be adequately assessed for the Navy's needs. More recent studies that examined the health effects of refractory-ceramic-fiber (RCF) production workers might be able to account for potential biases and confounders such as smoking that limit the older epidemiological studies of MVF. In analyzing and interpreting results from these studies, it is imperative that the Navy assess their underlying limitations.



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers 6 EPIDEMIOLOGICAL STUDIES THE Navy Environmental Health Center prepared a comprehensive review of the epidemiological literature in support of its proposed occupational standard for manufactured vitreous fibers (MVF) (NEHC 1997b). There have been well over two dozen publications of epidemiological studies of MVF production facilities in North America and Europe that have examined large cohorts of workers. Those studies have used both historical cohort mortality and case-control study designs. Interest has centered primarily on lung cancer, although other cancer sites and causes of death have been considered. Lee et al. (1995) have reviewed the published work with respect to respiratory system cancers associated with MVF. The Navy's documentation in Man-Made Vitreous Fibers evaluated data primarily from the three large cohort studies conducted in Europe, the United States, and Canada. Since publication of the Navy's report, several new epidemiological studies (Chiazze et al. 1997; Watkins et al. 1997; Hansen et al. 1999) and updated analyses (Boffetta et al. 1997; Boffetta et al. 1999; Chiazze et al. 1999; Marsh et al. 1996) of earlier studies of MVF have been published. The subcommittee suggests that the Navy update its report to discuss the strengths and limitations of the recent cohort and case-control studies so that their results can be adequately assessed for the Navy's needs. More recent studies that examined the health effects of refractory-ceramic-fiber (RCF) production workers might be able to account for potential biases and confounders such as smoking that limit the older epidemiological studies of MVF. In analyzing and interpreting results from these studies, it is imperative that the Navy assess their underlying limitations.

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers MANUFACTURED VITREOUS FIBERS Chiazze et al. (1997) conducted a historical cohort mortality study of a continuous-filament glass-fiber manufacturing plant in Anderson, South Carolina. The investigation was undertaken in light of evidence of a statistically significant increase in the proportionate mortality ratio for lung cancer among white men at the plant. A nested case-control study of lung-cancer deaths among white men was incorporated into the study design. Information on sociodemographic factors—including smoking, alcohol consumption, and medical history—was obtained through an interview survey; exposure assessments for respirable glass fibers, total particles, asbestos, RCF, respirable silica, formaldehyde, total chrome, and arsenic were gathered through reconstruction of historical data. Respirable glass fibers were produced at the plant only in 1963-1968. Lung-cancer odds ratios (ORs) among workers exposed to respirable glass fibers were below unity, as were ORs for exposures to asbestos, RCF, respirable silica (except for the lowest exposure level), total chrome, and arsenic. Those results suggest that none of the plant exposures resulted in an increase in lung-cancer risk for this population, although the lung cancer standardized mortality ratio (SMR) is slightly increased (observed, 47; SMR, 126). The authors conclude that neither glass fibers nor any of the substances investigated as part of the plant environment were associated with an increase in lung-cancer risk in the study population. In a separate cohort study conducted at the same continuous-filament glass-fiber plant in Anderson, South Carolina, Watkins et al. (1997) examined mortality in women and minority groups. The study population consisted of 1074 white women, 130 black women, and 494 black men who worked at least one year from the opening of the plant in 1951 through 1991. The women made up the largest cohort of white women assembled to date in a wool or continuous-filament glass-fiber manufacturing facility, and this was the first study of black men and women in the MVF industry. Over 95% of the women and minority-group members included in the study held production positions in the plant. Results of the analyses demonstrated that there were no significant excesses or deficits in mortality among white women by cause of death—including cancer—compared with national mortality, except for motor vehicle accidents. Among black men, SMRs for all cancers combined were below unity on the basis of local and national standards; and lung-cancer SMRs

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers were below unity among white women and black men. However, the authors mentioned that the statistical power of the study was low and possibly not adequate to detect risks of the magnitude typically of interest in studies of men. Hansen et al. (1999) conducted a cross-sectional survey of 235 Danish rock-wool production workers with at least 5 years of exposure in 1955 and 1987. The reference group consisted of 243 nonexposed subjects randomly sampled from the general population in the same area. Respiratory health was assessed by questionnaires, spirometry, and measurement of diffusion capacity of the lungs. The authors determined, after adjusting for age and smoking, that there was no association between MVF and respiratory symptoms as reported in the questionnaires. Self-reports of emphysema were more common in exposed workers (3.8%) than in nonexposed workers (0.9%) (RR, 4.5), but only nine exposed workers and two nonexposed workers reported emphysema. Of the lung-function parameters measured, only the lower FEV1/FVC ratio (ratio of forced expiratory volume in 1 second to forced vital capacity) was significantly associated with MVF exposure after adjusting for age, height, and smoking. The prevalence of airflow obstruction was greatest among exposed workers with more than 40 pack-years of smoking; this suggested a synergistic effect between smoking and MVF. A similar interaction between smoking and MVF was reported by Trethowan et al. (1995) in ceramic-fiber workers. Several updated analyses of data sets have been published recently. Marsh et al. (1996) published an update of the 1946-1989 mortality experience of the rock- and slag-wool subcohort of the North American Insulation Manufacturers Association (NAIMA) cohort. (The NAIMA cohort comprises production and maintenance workers in 17 of the oldest and largest MVF plants in the United States.) The study authors reported a statistically significant increase in lung cancer in the subcohort (observed, 70; SMR, 129). However, as in all the historical cohort mortality studies of MVF, SMRs were not adjusted for the possible effect of confounding by cigarette smoking. There were no statistically significant increases in the SMRs for any of the remaining cause-of-death categories except nephritis and nephrosis (observed, 12; SMR, 204). Exposures were estimated quantitatively for respirable fibers, asbestos, formaldehyde, and silica and qualitatively for arsenic, asbestos, asphalt, polycyclic aromatic hydrocarbons, phenolics, radiation, and urea. The exposure estimates were used in relative-risk regression modeling, which

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers yielded no consistent evidence of an association with any of the respirable-fiber measures considered. The findings were corroborated in a nested case-control study that adjusted for smoking. Boffeta et al. (1997) published results of an updated mortality followup of a cohort of MVF production workers in Denmark, Finland, Norway, Sweden, the United Kingdom, Germany, and Italy (Simonato et al. 1987). The population studied consisted of people employed in 13 factories in 1933-1950 (when production was started), and followed through 1990, 1991, or 1992, depending on the country. Historical exposure information on MVF was based on production periods when workers were employed in plants (Simonato et al. 1987); no quantitative estimate of MVF exposure was available. Among specific cancer causes of death, SMRs were statistically significantly increased only for lung cancer in the rock- and slag-wool subcohort (observed, 97; SMR, 1.34), with 42 of the deaths in workers in one factory in Germany, and the glass-wool subcohort (observed, 140; SMR, 1.27), with 109 of the deaths in workers in one plant in the United Kingdom. The SMR was increased for workers with less than 1 year of MVF employment (SMR, 1.48) and with more than 1 year (SMR, 1.29). There were no statistically significant increases for cancer in the continuous-filament subcohort (Boffeta et al. 1997). The authors conclude that the results were not sufficient to find that the increased lung-cancer risk is specifically related to exposure to MVF. However, they note that insofar as respirable fibers were an important component of ambient concentrations in the workplace, these fibers might have contributed to the increased risk. In a further followup analysis of the rock- and slag-wool and glass-wool cohorts in Denmark, Norway, Sweden, and Finland, Boffeta et al. (1999) found an increased standardized incidence ratio of 1.15 for lung cancer in the combined cohorts. However, it could not be correlated with fiber exposure. Time since first employment and duration of employment were used as surrogate measures when no other exposure measures were available. It is important to note that none of the SMRs in the European study was adjusted for the possible confounding effect of cigarette smoking; exposures to other carcinogens, polycyclic aromatic hydrocarbons, and asbestos might also have occurred. In another European study, Sali et al. (1999) investigated the nonneoplastic mortality of rock- and slag-wool, glass-wool, and continuous-filament workers in MVF factories in Denmark, Finland, Norway, Sweden, the United Kingdom, Germany, and Italy. Exposure to MVF was based on various measures, including time since first employ

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers ment, duration of employment, and technological phase at first employment. The authors determined that mortality from nonneoplastic diseases did not appear to be related to employment in the MVF industry. However, they believe that an increased risk of death from nonmalignant renal diseases among rock- and slag-wool workers with increasing duration of employment and an increased risk of ischemic heart disease in rock- and slag-wool and continuous-filament workers with more than 30 years since first employment warrant further investigation. Recently, Chiazze et al. (1999) conducted a case-control study based on the Owens Corning Surveillance System to investigate the question of whether there is an association between silica or respirable glass fiber exposure and mortality from nephritis and nephrosis among workers in fibrous glass wool manufacturing facilities. Quantitative estimates of exposure to silica and respirable fibers were obtained through historical environment reconstruction. Information on sociodemographic factors—including education, marital status, income, drinking, and smoking—was gathered through interview surveys. The study authors found no consistent relationship for silica or respirable fibers when the analysis was based only on underlying cause of death. When the analysis was based on both underlying and contributing causes of death, the odds ratios for respirable fibers were below unity for both the lowest exposure (OR, 0.81; 95% CI, 0.30-2.13) and the highest exposures (OR, 0.53; 95% CI, 0.22-1.31). The results were similar for silica: ORs were below unity except for the highest exposure, for which the OR was 1.04 (95% CI, 0.24-4.46). The authors state that although the results do not prove the lack of association between nephritis and nephrosis and glass-fiber or silica exposure, they do not support such an association. However, the findings do suggest that all information on the death certificate, not just the underlying cause, should be used to capture the most accurate picture of renal disease. Those cohort and case-control studies have contributed to expanding the existing database on effects of exposures to MVF. However, the confounding factors and limitations in the exposure information that affect the interpretation of the results must be considered. Typically, exposures in the plants have been relatively low—average plant concentrations were reported to be 0.005-0.292 f/cm3 for fibrous glass and 0.194-0.426 f/cm3 for mineral wool (Marsh et al. 1990)—and workers were exposed to multiple types of fibers, and to other potential carcinogens at the plants, including asbestos. Because of the retrospective nature of the studies, direct exposure measurements typically do not

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers exist, and exposure information must be reconstructed from employment histories and plant records. Inability to control for potential confounders, in particular cigarette smoking, is especially important because smoking is such a strong predictor of respiratory system cancer. For instance, Chiazze and Watkins (1994), in their reanalysis of data from the Owens Corning plant in Newark, Ohio, demonstrated how important it is for lung-cancer SMRs based on national data to take the confounding effect of smoking into account. They derived an overall estimate of the average prevalence of ever having smoked cigarettes in 1940-1980 (the followup period of the Newark cohort) for white men in the United States at least 20 years old from the National Health Interview Survey special smoking supplements for 1978-1980. The estimate was based on survey participants who were self responders with known age, race, sex, smoking status, and age when they began smoking. After adjustment for the confounding factor of cigarette smoking, the adjusted SMR was found to be below the level of significance and quite similar to the SMR obtained by using local rates. Several review articles (McClellan et al. 1992; Lee et al. 1995) have recommended that future studies provide information on cigarette smoking and other workplace carcinogens because future studies would not provide useful information without such data. The issue of statistical significance is moot unless potential extraneous factors are taken into account. A statistical-significance test (or any mathematical model) cannot compensate for lack of appropriate data on exposures and confounders (Chiazze and Watkins 1994). As Lee et al. (1995) state in their review article, “available data indicate that among those occupationally exposed, glass fibers do not appear to increase the risk of respiratory system cancer. Exposure to rock and slag wool may increase the risk of such cancers; however, the data do not convincingly prove that this association is causal. In addition to inconsistent patterns from cumulative exposure, latency, and duration of employment, the potential for confounding by cigarette smoking and other workplace carcinogens exists in these studies of rock and slag wool fibers.” REFRACTORY CERAMIC FIBERS In the last several years, there has been increased use of RCF. There is also a growing database of studies examining the health effects on workers employed in the manufacture of RCF. These studies, of which

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers several are cross-sectional or prospective analyses (Lemasters et al. 1998; Lockey et al. 1998), might be better able to account for confounding factors, including cigarette smoking, in their analyses. Lockey et al. (1996) reported the results of a retrospective cohort study and a nested case-control study designed to evaluate chest radiographs of 652 workers involved in the manufacture of RCF at two sites. A latency-validation review was also conducted in which historical chest xrays were examined to determine whether there was a biologically plausible latency between initial RCF exposure and first appearance of pleural plaques. The study included men and women employed at the time of interview—October 1987-December 1991—and former employees who had been employed for at least 1 year. The cohort study demonstrated an association between exposures to RCF and the occurrence of pleural plaques as measured by three exposure indexes: time from beginning RCF production job, years of RCF-production employment, and cumulative exposure in fiber-months per milliliter. The highest prevalence of pleural plaques was observed in the highest-exposure groups for each exposure index. The nested case-control study, which controlled for the possibility of past asbestos exposure as a potential confounding factor, corroborated the findings of the cohort study. The latency-validation review of historical xrays confirmed that a biologically plausible period had elapsed between initial exposure and development of pleural plaques. The authors concluded that “the cohort and nested case-control studies and the latency validation review support a causal hypothesis for RCF exposure and the development of pleural plaques.” In another study, Lockey et al. (1998) conducted a prospective analysis to evaluate the relationship between RCF exposure and pulmonary-function changes in 361 men employed at five RCF manufacturing plants in 1987-1984. Workers included in the analysis provided at least five pulmonary-function tests. The exposure-response relationship was modeled with two exposure variables: years in a production job and cumulative fiber exposure (fiber-months per milliliter). Analyses by Rice et al. (1997) reconstructed exposure estimates for workers in two plant locations, calculating cumulative RCF exposures from pre-1987 exposure data and current exposure data. However, cumulative exposures could not be calculated for the other three plants, because pre-1987 exposure data were not available. The authors determined that employment in RCF production and cumulative RCF exposure were not associated with a decrement in longitudinal FVC or FEV1 values from 1987 through 1994. However, a cross-sectional analysis of the initial pulmonary-function test

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers results found that production workers exposed for more than 7 years had lower FVC and FEV1 than baseline nonproduction workers; the decrease might have been caused by earlier exposure to RCF concentrations that were greater than those in 1987-1994. Those results correspond with the historically greater exposure levels in the 1950s—an estimated maximal exposure of 10 fibers/mL, compared with recent exposures that ranged from below the detection limit to 0.66 fiber/mL (Rice et al. 1997). The authors state that “the decrease in RCF exposure levels over the last 10 years through engineering and work practice changes has reduced any detectable continued effect of RCF exposure on FVC and FEV1.” They plan to study workers with fewer than five spirometry tests to further assess participation bias because it was noted that smokers with reduced lung function were less likely to have produced at least five spirometry tests which were required for inclusion in the longitudinal analysis. Lemasters et al. (1998) also conducted an industrywide pulmonary-morbidity study to evaluate the respiratory health of employees engaged in the manufacture of RCFs at five U.S. sites in 1987-1989. Of the 753 eligible employees, 742 provided occupational histories and completed the American Thoracic Society respiratory-symptom questionnaire; and 736 of the 742 also underwent an initial pulmonary-function test. Exposure to RCF was assessed by classifying workers as production or nonproduction employees and computing the duration of time the former spent in production. The results of the cross-sectional study demonstrated an increase in respiratory symptoms in production versus nonproduction workers and declines in FVC and FEV1 for each 10 years of employment in excess of the declines attributed to smoking, age, and other factors. Among current and past male smokers, a significant decline in FVC was associated with employment in the production of RCF. A significant decline in FEV1 in men who were current smokers was associated with work in RCF production A significant decline in FVC was seen in production women but not production men who had never smoked. The results suggest that there might be important sex differences in responses to occupational or environmental exposures. CONCLUSIONS In a review of the published epidemiologic literature with respect to respiratory system cancer, Lee et al. (1995) concluded that “ available data indicate that among those occupationally exposed, glass fibers do

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers not appear to increase the risk of respiratory system cancer. Exposure to rock or slag wool may increase the risk of such cancers; however, the data do not convincingly prove that this association is causal. ” Recent studies, including case-control studies, make it clear that any lung-cancer SMRs based on national data must take into account the potential confounding effect of smoking. Evidence from the case-control studies demonstrates that there is no significant association between fiber exposure and lung cancer or nonmalignant respiratory disease in the MVF manufacturing environment. It is clear, for example, that of the Newark, Ohio, plant workers (who made up some 35% of the U.S. cohort) exposure to MVF, including respirable glass fibers, was not responsible for any increase in lung cancer risk (Chiazze et al. 1993). A recent agreement between the Occupational Safety and Health Administration and manufacturers and users of MVF for monitoring of the exposure of production workers and users might provide important exposure data for future studies (NAIMA 1999). That would be particularly valuable in light of the sparse epidemiological data on special-purpose fibers and RCF. More data are also needed on exposure of women and members of racial and other ethnic minorities to MVF. Nested case-control studies that can provide information on potential confounders might be the best means of addressing continued questions surrounding the effects of exposures to MVF. Both cohort and case-control methods have been used to describe the distribution of disease in a population and to derive statistical associations in the epidemiology of MVF. Regardless of the approach, to be useful, epidemiological investigations must demonstrate internal and external validity and freedom from bias. In the Navy review of epidemiological studies, insufficient consideration was given to either internal or external validity. Internal validity refers to the design and conduct of a study that is free of systematic errors or biases, which are always of concern in epidemiological investigations. A number of biases, such as exposure and diagnostic misclassification, and confounding by lack of data on cigarette smoking are limitations in many of the available epidemiological investigations, but little consideration is given to such potential problems in the Navy review of the epidemiological studies of MVF. For the Navy, external validity refers to the generalizability or applicability of results from the study populations to Navy personnel working with MVF. However, because Navy personnel are largely involved in the removal of MVF, their exposure could differ—in physicochemical properties of the fibers and in the health status of workers—from exposures

OCR for page 41
Review of the U.S. Navy's Exposure Standard for Manufactured Vitreous Fibers experienced by manufacturing workers, who have been the focus of epidemiological investigations. Therefore, the difference in exposures between the two workforces raises concerns about inferring health risks to Navy personnel on the basis of data derived from the MVF manufacturing industry.