Secondary Logo

Journal Logo

ORIGINAL ARTICLES

Lung Cancer Among Rock and Slag Wool Production Workers

Kjærheim, Kristina1 2; Boffetta, Paolo1; Hansen, Johnni3; Cherrie, John4 5; Chang-Claude, Jenny6; Eilber, Ursula6; Ferro, Gilles1; Guldner, Karlheinz7; Olsen, Jørgen H.3; Plato, Nils8; Proud, Louise4; Saracci, Rodolfo1 9; Westerholm, Peter10; Andersen, Aage2

Author Information
  • Free

Abstract

Man-made vitreous fibers (MMVF) are used for thermal and acoustic insulation purposes. These comprise four major groups of products: rock and slag wool (RSW), glass wool, continuous glass filament, and refractory ceramic fibers. Production in Europe started in the 1930s, and use has increased greatly during the last decades. In analogy with asbestos, the fibrous properties of MMVF have raised the question of their possible carcinogenicity, especially on the lung. Animal models with intracavitary instillation of MMVF have shown increased tumor incidence; however, inhalation studies have not shown associations. 1 Studies indicate that biopersistence in the lung is lower for MMVF than for asbestos. 2–5

Two large cohort studies have been conducted on MMVF production workers, one in Europe and one in the United States. The European study was established in 1977, and includes approximately 22,000 production workers from 13 plants in seven countries. Several analyses with increasing periods of follow-up for cancer mortality or incidence have been published, 6–9 showing mostly small elevations of the standardized mortality and incidence ratios of lung cancer. Among the RSW workers, an elevated risk of lung cancer by indirect indicators of exposure (such as time since first employment and technological phase at first employment) has been suggested. Elevation in risk has been found 20 or more years after first employment among workers with long employment periods. In a study based on mathematical modeling of past plant-specific fiber exposure, no increased risk was found among RSW workers employed more than 1 year. 10 No individual or job-specific exposure estimates have been available for these cohort analyses. Also, because no information on smoking habits or exposure to other possible carcinogenic substances was available, it has not been possible to adjust for possible confounding from these factors.

The U.S. cohort has likewise been updated several times, 11–13 and overall increases in standardized mortality ratios for respiratory cancers have been found among RSW workers. The latest cohort update and nested case-control study 14 revealed no consistent evidence for an association with estimated fiber exposure, regardless of adjustment for other occupational exposures and tobacco smoking.

Neither of these two international cohorts has indicated elevations of lung cancer risk similar to those found for glass wool production workers. 8,12,13 Analyses of lung cancer risk among continuous glass filament production workers have also not suggested increased mortality (overall or by time since first employment). 8,13,15–17

The aim of the present study was to analyze the association between exposure to RSW and risk of lung cancer while controlling for potential confounders such as tobacco smoking and occupational exposures other than RSW. A secondary aim was to investigate whether exposure to other known or suspected carcinogens was directly associated with the lung cancer risk in this group of workers. To do this, we conducted a nested case-control study in the RSW part of the cohort of European MMVF production workers. The RSW workers were chosen because the most consistent elevations of risk, especially during the early production period, had been found in this part of the cohort. 18

Methods

Male blue-collar workers from seven RSW plants in Denmark, Norway, Sweden, and Germany were included in the present case-control study. Workers employed from the start of production (between 1937 and 1950) until the end of 1976 were included in the cohort from which the cases and controls were selected. During this period, the cohort comprised 9,174 workers. The period for identification of cases and controls was from 1971 until 1996 in Denmark, until 1995 in Norway and Sweden, and until 1991 (deaths only) in Germany. A total of 196 lung cancer cases were identified: 134 in the mortality study (160 minus the 26 that occurred before 1971), 8 45 additional lung cancer cases in the incidence study, 9 and 17 cases from the extended follow-up in Denmark. Of these cases, 101 had been employed in the RSW industry for more than 1 year.

Because collection of information from different sources for deceased cases and living controls might introduce information bias into the study, we selected two control series. Control series I comprised controls who were alive at the time of diagnosis or death of the corresponding case but who died before being interviewed. Workers who died from non-neoplastic respiratory diseases (International Classification of Diseases, 9th revision [ICD-9], codes 460–519) were excluded as possible controls because of the strong association with tobacco smoking. To minimize selection bias, control series II comprised incidence density-sampled controls. For the living cases (N = 2), only incidence density-sampled controls were selected. We excluded workers who died from respiratory cancers (ICD-9 codes 160–165) and ill-defined cancers (ICD-9 codes 195–199) from both control series, because of possible misclassification of lung cancer. All controls were individually matched to cases on plant and date of birth (±3 years). Because preliminary analyses showed that the two control series were very similar with respect to the exposures under study, we combined the two groups. All results presented here are based on this combined control group.

The study proposal was approved by relevant ethical committees; subjects gave informed consent to participate in the study.

Tracing of each subject and identification of one or more next of kin for deceased workers was done differently in each country, according to differences in national legislation and registration practices. The “next of kin” to be interviewed were, in order of preference, the spouse, descendants, siblings, other relatives, neighbors, and friends. Table 1 presents the number of cases and controls identified, contacted, interviewed, and included in the analysis. The present study includes 133 lung cancer cases and 513 controls. This gives a response rate among cases of 68%. For the controls, 59% of interviews were performed out of the number of possible respondents contacted (608 of 1,027). Two of the cases (2%) and 188 (37%) of the controls were alive at the time of interview. On average, there were 3.9 controls per case, with a range of 1–6 (only one set had just one control).

Table 1
Table 1:
Numbers of Cases and Controls Identified, Contacted, and Interviewed, by Country

Trained interviewers interviewed all respondents. The interviews focused on history of tobacco smoking, residence, and occupation outside the MMVF industry. The data on occupational history outside the RSW industry were classified independently by two industrial hygienists according to a standard protocol. Table 2 shows the substances assessed. Target values were defined as the current exposure limit values, against which assessments were made. The probability of exposure for each job and period was coded on a three-level scale (no or unlikely, possible, and probable). Intensity of exposure was coded on a four-level scale (no, low, medium, and high). A four-level summary exposure index was composed out of the combination of probability and intensity of exposure (no, low, medium, and high). High exposure was defined as the combination of probable exposure (>70% chance of exposure) and high or medium intensity (at or above the target value).

Table 2
Table 2:
List of Substances Assessed

Assessment of exposures within the RSW industry was based on information obtained from expert panels of experienced foremen and workers that were set up at each plant. The expert panels first reconstructed each individual’s work history, with information on job titles, dates of employment, and work areas and tasks. The expert panels then characterized the various job titles and work environments using a standard form to identify such details as how tasks were carried out, which materials were used, how they were handled, the local environment, and other pollutants. For each plant, task, and period, this information was coded independently by three specifically trained industrial hygienists and combined to form overall job exposure assessments. The assessments were made using the methodology described by Cherrie and coworkers 19 and later validated. 20,21 The model for historical exposure used in the assessment of individual exposure to MMVF has been developed on the basis of measurements done in the years 1977–1980, 22 a historical environmental investigation, 23 a simulation experiment of past production conditions, 24 and dust emission measurements on old and new products. 25 The final outcome was a fraction of the defined target value (Table 2). High exposure inside the RSW industry was defined as average exposure exceeding 70% of the target value.

Exposure variables were constructed according to the source of exposure, ie, exposure within the RSW industry (seven agents), exposure outside the RSW industry (12 agents), and the combination of exposure within and outside the industry (six agents). Variables measuring ever-exposure, ever-high exposure, and duration of exposure were developed for all three groups. For exposures within the RSW industry, additional variables of cumulative exposure were calculated. For all variables of duration of exposure and cumulative exposure, we constructed 15-year lagged variables; for the RSW variables a 30-year lag was also applied. Lag periods were chosen a priori on the basis of previous research on asbestos and lung cancer. 26 To ensure adequate numbers of cases in each group, the continuous variables were categorized according to the distribution of values among the cases into approximate quartiles (when less than 36% of cases were unexposed) or tertiles (when 36–89% of cases were unexposed), keeping the unexposed group as reference whenever possible. When more than 90% of cases were unexposed, only the dichotomous variables were used.

Smoking histories were collected in detail, and several smoking variables (smoking status, cumulative tobacco consumption, and combinations of smoking status and cumulative tobacco consumption) were computed. These different constructs produced very similar results. In the final analyses, we used a variable combining information on smoking status and cumulative tobacco (in kilograms from cigarettes, cigars, cigarillos, and pipe) to adjust for the possible effect of confounding by smoking, as this variable gave the most stable risk estimates for tobacco smoking. Educational level was measured as number of years at school, and urban residence was operationalized as the proportion of life spent in urban areas. Stratified analyses were done by selected indicators of quality of information, such as completeness of interview, quality of interview, histologically verified cases and whether the next of kin knew the index person at the time of his death, while he was working in the RSW industry, or before he was 40 years old. In addition, we computed variables similar to those used in the cohort study (ie, time since first employment, duration of employment, and technological phase at first employment). As in the previous cohort studies, technological phase was classified as early, intermediate, or late, as defined by the periods of introduction of dust-suppressing agents and the end of the operation of a batch process involving labor-intensive and hand-operated production methods. 27

All analyses were based on conditional logistic regression models. We calculated odds ratios (ORs) of lung cancer and 95% confidence intervals (CIs), adjusted for potential confounders such as tobacco smoking and for occupational exposures other than RSW to which a sufficient number of workers were exposed. Linear trends were tested by assigning equidistance scores to the values on the categorized exposure variables and using them as continuous variables. All main analyses were performed on all subjects and after excluding workers employed in the RSW industry for 1 year or less. The rationale behind the exclusion of short-term workers was that the work history within the RSW industry is likely to be less accurate for these transient workers. Subjects with missing values on specific variables were excluded from the models in which these variables were used. Hence, the total numbers in the different analyses are not always the same. For many of the non-RSW exposure variables, the proportion of missing values was somewhat higher for cases than for controls, but this was so only for exposures outside the industry.

We assessed to what extent the subjects included in the case-control study were representative of the cohort that provided the basis of the present study. We did so by multivariate logistic regression analysis in the cohort, evaluating how the exposure indicators previously used (ie, time since first employment, duration of employment, and technological phase at first employment) predicted inclusion in the case-control study. We computed this for all subjects combined and then computed it for cases and controls separately. To aid the comparison between previous results from the cohort studies and the present case-control study, the number of current smokers in 1985 among the controls included in the present study was compared with the age-adjusted expected number derived from national surveys in 1985/1986. 28 Because of the age restriction on the population data and the requirement to be alive in 1985, only 58% of controls contributed to this comparison.

Results

Table 3 shows the risk of lung cancer according to selected RSW exposure variables for all workers and for workers employed more than 1 year. Because possible confounding from smoking was an important research question, results are given both unadjusted and adjusted for smoking in the analyses including all workers. The ORs or trends in ORs were essentially unchanged for all exposure variables after adjustment for smoking. Results were similar for all workers and for those employed more than 1 year. Ever-exposure to RSW was associated with a reduced risk of lung cancer, whereas ever-high exposure to RSW was associated with an elevated risk (Table 3); neither, however, was substantial.

Table 3
Table 3:
Lung Cancer According to Selected Rock and Slag Wool (RSW) Exposure Variables, for All Workers and for Workers Employed More than 1 Year

The analysis of duration of exposure to RSW within the industry showed a slightly decreasing trend in the ORs with a longer duration, but when a 15-year lag was introduced the trend was relatively flat. The P-value for trend was 0.96 for the smoking-adjusted analysis among workers employed more than 1 year.

A similar but more marked pattern was seen when analyzing the combined duration of exposure to RSW within and outside the industry. In this case, increasing duration gave a decreasing trend in the ORs, whereas ORs above 1 were seen when a 15-year lag was applied to the same variable. In the highest quartile of cumulative exposure to MMVF, mean exposure was 5.1 and 5.9 fibers per milliliter for cases and controls, respectively. With increasing cumulative exposure to RSW, a decreasing trend in lung cancer risk was found. When a 15-year lag was applied, the trend was no longer monotonic. For none of the variables did the application of a 30-year lag or the splitting of the highest-exposure group into two equally sized groups give indications of elevated risks in the highest-exposed groups (data not shown).

No major changes in the trends were observed when control for polycyclic aromatic hydrocarbons, asbestos, or silica (in addition to adjustment for smoking) was performed (data not shown). Adjustment for exposure to formaldehyde, welding fumes, and diesel exhaust also did not change the risk estimates. Stratified analyses by the previously mentioned indicators of quality of information did not change the results.

In country-specific analyses, the ORs associated with exposure to RSW were above 1 more often in the Danish workers, whereas in German workers significantly negative trends were seen when exposure outside the industry was taken into account. However, when testing for heterogeneity among countries, differences were seen only for ever-exposure and lagged cumulative exposure, and only when all workers were included.

Table 4 provides ORs for factors other than RSW with known or possible etiologic relevance to lung cancer. A cumulative amount of tobacco of more than 400 kg among current smokers resulted in an OR of 8.8 (CI = 3.9–19.8) when all workers were included in the analysis. Neither number of years at school or urban residence was associated with risk. Ever-exposure to asbestos, polycyclic aromatic hydrocarbons, or silica did not give elevated ORs. Ever-exposure to welding fumes, diesel exhaust, or paints also did not show elevation of risk. The ORs associated with 15-year lagged cumulative exposure for the substances assessed inside the RSW industry showed no increasing trends by increasing exposure (Table 4).

Table 4
Table 4:
Lung Cancer According to Selected Variables Other than Rock and Slag Wool (RSW)

In the analysis of lung cancer risk according to time since first employment (Table 5), no monotonic trend was seen. Increasing duration of work in the RSW industry was associated with decreasing ORs for all workers combined, but ORs were around 1 when short-time workers were excluded. No trend with technological phase at first employment was observed.

Table 5
Table 5:
Lung Cancer According to Exposure Indicator Variables Used in Previous Follow-Up Studies, Adjusted for Smoking

The observed smoking prevalence in 1985 was 20% higher among the controls than expected according to national rates. Among workers employed more than 1 year, those first employed in the late technological phase had a slightly higher probability of being included than workers from the early phase (OR = 1.4; CI = 0.9–2.4; not shown in ). This tendency was more pronounced among the cases (OR = 5.0; CI = 0.5–48.7) than among the controls (OR = 1.4; CI = 0.8–2.3). Also in the group with more than 1 year of employment, cases with a long period since first employment had a lower probability of being included in the study. The opposite was found for controls.

Discussion

In this study from four northern European countries, no increased lung cancer risk was found among men who worked in the production of rock wool and slag wool. In some of the analyses, a significantly negative trend, implying decreased risk, was seen when unlagged exposure variables were included in the models. However, when a 15-year lag was introduced, no association was seen with duration of RSW exposure. A nonsignificant decrease in risk was seen with cumulative exposure to RSW. Given the long latency shown by other lung carcinogens such as asbestos, it seems reasonable to regard the results of the lagged analyses as the most informative.

Tobacco smoking was, as expected, an important predictor of risk, with ORs for the highest-exposed group around 10. There were no indications of a confounding effect from tobacco on the association between RSW and lung cancer. There were also no indications of a confounding effect from other occupational exposures such as polycyclic aromatic hydrocarbons and silica, whereas for asbestos there was moderate negative confounding.

We assessed workers’ exposure to 13 occupational substances known or suspected to be lung carcinogens, within or outside the RSW industry. Most of these exposures showed only very modest, if any, association with lung cancer. Asbestos exposure was low among both cases and controls, and mostly stemmed from work outside the RSW industry. Arsenic and chromium exposure both gave elevated ORs, as did paints (to some extent), but these exposures were too rare to have much explanatory power.

Our results are in agreement with the results from case-control studies among RSW workers in the U.S. cohort, 14,29 which have given little support to the hypothesis of an association between exposure to RSW and lung cancer risk. In addition, case-control studies among glass wool production workers have given little support to such an association, 30–32 a finding in agreement with the results from a recent case-control study of glass wool workers nested in the U.S. cohort. 33

The previous follow-up studies in the European cohort had data on employment periods within the RSW industry only until 1977. In contrast, the present case-control study included complete work histories for jobs inside the RSW industry as well as work outside the industry until the year of diagnosis/death of case. This has made it possible to assess not only individual RSW exposure, but also individual exposure to other known or suspected occupational carcinogens. In addition, detailed tobacco smoking histories were collected.

A pilot study was initially undertaken to ensure that valid information could be obtained from RSW workers and their next of kin and to evaluate the feasibility of obtaining interviews with relatives of deceased workers. 34 The quality of information from relatives of living workers in general was good except in the case of information on specific occupational exposures, for which other sources of information should be sought.

The study was designed to minimize biases. The identification of cases occurred before the exposure assessment, which was done blindly as to the case-control status of the workers. The method of exposure assessment has been validated, with correlation coefficients between the estimated and the measured values ranging from 0.5 to 0.9 20 and from 0.7 to 0.9. 21 Analyses by the separate series of controls gave similar results, and analyses stratified by a series of indicators of quality of information did not change the results.

Response rates among both cases and controls were higher than in the pilot study. 34 However, selection bias is still a concern. The analysis of how the technological phase of first employment predicted inclusion in the present case-control study showed that the probability of being included was lower for subjects first employed in the early phase, when fiber exposure supposedly was at its highest, than for subjects employed later (OR of inclusion for workers first employed in the late vs the early phase = 1.4; CI = 0.9–2.4), and even lower for cases than for controls. This finding indicates that because of some degree of nondifferential selection we are not able to study those in the early phase, and it additionally suggests some differential bias in the selection of cases and controls. Therefore, if all excess risk was associated with the early production period, and the study does not cover appropriately the whole exposure experience, the study would fail to detect most or all of the association between RSW exposure and lung cancer. From the cohort studies, there were several indications that the elevated risk was present particularly among those with a long period since first employment. For instance, a relative risk of 4.6 (CI = 1.2–17.6) was found among workers with 30 years or more since first employment and a duration of employment of 20 years or more (based on 15 lung cancer cases). 8

In the latest follow-up of incidence in the cohort, 9 the overall standardized incidence ratio among workers employed 1 year or more was lower than previously (1.1; CI = 0.9–1.4), whereas there was still an increasing trend by time since first employment. This is consistent with a cohort that in the past experienced exposure leading to an increased cancer risk, whereas workers hired in later years have been less exposed. Some indication of this was given by the present study: cases and controls who were first employed during the early technological phase had an average RSW exposure of 1.27 and 0.53 ml × fibers per year, respectively. In the later periods, there was no difference in fiber exposure between cases and controls, with yearly averages of 0.48 and 0.47 ml × fibers in the intermediate phase and 0.23 and 0.26 ml × fibers in the late phase. Only four cases from the early phase were included in the present study, precluding a separate case-control evaluation. An elevated risk among workers first employed in the early phase may have been related to fiber concentration, to the chemical or physical properties of the fibers, or to other exposures within or outside the RSW industry.

We designed and conducted a nested case-control study with power sufficient to detect ORs in the range 1.7–2.0 for binary exposure variables (ie, cumulative exposure divided at the median). The only strong and consistent risk factor for lung cancer was tobacco smoking. It must be considered whether the elevations of overall standardized mortality or incidence ratios in the follow-up studies may actually have been caused by smoking. The group of controls, for whom we were able to compare smoking habits with the general population, had a 20% higher prevalence of ever-smokers. Although these matched controls are not in a strict statistical sense representative for the cohort, the comparison is based on age-adjusted smoking rates and may thus be viewed as being a conservative estimate of the smoking prevalence of the age groups in the cohort giving rise to the lung cancer cases. This difference in smoking prevalence would be expected to result in a moderate confounding effect in the calculation of mortality ratios based on external rates. According to the model proposed by Axelson, 35 a 30% increase in lung cancer risk could be approximately explained by a 20% higher smoking prevalence, assuming a relative risk of 10 for ever-smoking. New data from the U.S. cohort also suggest that the elevated standardized mortality ratios for respiratory cancer were at least partly caused by uncontrolled confounding by smoking. 36,37

In the European cohort, measured mean levels in most plants were between 0.01 and 0.1 fibers per milliliter around 1980, 38 increasing to 0.1–1.0 when extrapolated back to the 1950s. 39 Exposure levels in some of the user industries apparently have been higher, with means from 0.5 to 1.0 fibers per milliliter. 40,41 Previous cohort studies of workers in the user industries have not demonstrated any increased lung cancer risk, 42,43 but a recent large case-control study including workers in industries handling MMVF products has found positive associations based on self-reported exposure. 44 Exposure assessments based on expert rating in a subset of the study subjects, however, revealed no risk associated with cumulative fiber exposure. Exposure assessment in these industries is a very difficult task, and these results may be distorted by residual confounding, particularly from asbestos. The findings leave open the possibility of an association with high exposure levels.

In conclusion, the present study offers no support to the hypothesis of an association between exposure to RSW and lung cancer risk as experienced by RSW production workers. Although the cases and controls may not cover the earliest periods when exposures were highest, we do not believe that the exposure estimates, and thereby the risk estimates, are biased in any important way. We therefore consider the results valid for exposure experiences in the industry during the last four to five decades. This interpretation does not exclude a possible carcinogenic effect from RSW in the earliest periods. Previous indications of elevated risk in the European cohort study may reflect this early effect and the contribution of smoking and social class.

References

1. Hesterberg TW, Miller WC, Thevenaz P, Anderson R. Chronic inhalation studies of man-made vitreous fibres: characterization of fibres in the exposure aerosol and lungs. Ann Occup Hyg 1995; 39: 637–653.
2. Bignon J, Saracci R, Touray JC, eds. Biopersistence of respirable synthetic fibers and minerals. Environ Health Perspect 1994;102(suppl 5):3–283.
3. Hesterberg TW, Miller WC, Musselman RP, Kamstrup O, Hamilton RD, Thevenaz P. Biopersistence of man-made vitreous fibers and crocidolite asbestos in the rat lung following inhalation. Fundam Appl Toxicol 1996; 29: 267–279.
4. Hesterberg TW, Chase G, Axten C, et al. Biopersistence of synthetic vitreous fibers and amosite asbestos in the rat lung following inhalation. Toxicol Appl Pharmacol 1998; 151: 262–275.
5. Rogers RA, Antonini JM, Brismar H, et al. In situ microscopic analysis of asbestos and synthetic vitreous fibers retained in hamster lungs following inhalation. Environ Health Perspect 1999; 107: 367–375.
6. Saracci R, Simonato L, Acheson ED, et al. Mortality and incidence of cancer of workers in the man made vitreous fibres producing industry: an international investigation at 13 European plants. Br J Ind Med 1984; 41: 425–436.
7. Boffetta P, Saracci R, Andersen A, et al. Lung cancer mortality among workers in the European production of man-made mineral fibers: a Poisson regression analysis. Scand J Work Environ Health 1992; 18: 279–286.
8. Boffetta P, Saracci R, Andersen A, et al. Cancer mortality among man-made vitreous fiber production workers. Epidemiology 1997; 8: 259–268.
9. Boffetta P, Andersen A, Hansen J, et al. Cancer incidence among European man-made vitreous fiber production workers. Scand J Work Environ Health 1999; 25: 222–226.
10. Consonni D, Boffetta P, Andersen A, et al. Lung cancer mortality among European rock/slag wool workers: exposure-response analysis. Cancer Causes Control 1998; 9: 411–416.
11. Enterline PE, Marsh GM, Esmen NA. Respiratory disease among workers exposed to man-made mineral fibers. Am Rev Respir Dis 1983; 128: 1–7.
12. Enterline PE, Marsh GM, Henderson V, Callahan C. Mortality update of a cohort of U.S. man-made mineral fibre workers. Ann Occup Hyg 1987; 31: 625–656.
13. Marsh GM, Enterline PE, Stone RA, Henderson VL. Mortality among a cohort of U.S. man-made mineral fiber workers: follow-up. J Occup Med 1990; 32: 594–604:1985.
14. Marsh GM, Stone RA, Youk AO, et al. Mortality among United States rock wool and slag wool workers: 1989 update. J Occup Health Safety 1996; 12: 297–312.
15. Shannon HS, Jamieson E, Julian JA, Muir DC. Mortality of glass filament (textile) workers. Br J Ind Med 1990; 47: 533–536.
16. Chiazze L, Watkins DK, Fryar C. Historical cohort mortality study of a continuous filament fiberglass manufacturing plant. I. White men. J Occup Environ Med 1997; 39: 432–441.
17. Watkins DK, Chiazze L, Fryar C. Historical cohort mortality study of a continuous filament fiberglass manufacturing plant. II. Women and minorities. J Occup Environ Med 2000; 39: 548–555.
18. Boffetta P, Kjaerheim K, Cherrie J, et al. A Case-Control Study of Lung Cancer Among European Rock and Slag Wool Production Workers. IARC Internal Report No. 00/004. Lyon, France: International Agency for Research on Cancer, 2000.
19. Cherrie JW, Schneider T, Spankie S, Quinn M. A new method for structured, subjective assessments of past concentrations. Occup Hyg 1996; 3: 75–83.
20. Cherrie J, Schneider T. Validation of a new method for structured subjective assessment of past concentrations. Ann Occup Hyg 1999; 43: 235–245.
21. Semple SE, Proud LA, Tannahill SN, Tindall ME, Cherrie JW. A training exercise in subjectively estimating inhalation exposures. Scand J Work Environ Health 2001; 27: 395–401.
22. Cherrie J, Dodgson J, Groat S, Maclaren W. Environmental surveys in the European man-made mineral fiber production industry. Scand J Work Environ Health 1986; 12 (suppl 1): 18–25.
23. Cherrie J, Dodgson J. Past exposure to airborne fibers and other potential risk factors in the European man-made mineral fiber production industry. Scand J Work Environ Health 1986; 12 (suppl 1): 23–33.
24. Cherrie J, Krantz S, Schneider T, Ohberg I, Kamstrup O, Linander W. An experimental simulation of an early rock wool/slag wool production process. Ann Occup Hyg 1987; 31: 583–593.
25. Dodgson J, Cherrie J. Estimates of past exposure to respirable man-made mineral fibres in the European insulation wool industry. Ann Occup Hyg 1987; 31: 567–582.
26. Blot WJ, Fraumeni JF. Cancers of the lung and pleura. In: Schottenfeld D, Fraumeni JF, eds. Cancer Epidemiology and Prevention. 2nd ed. New York: Oxford University Press, 1996; 637–665.
27. Simonato L, Fletcher AC, Cherrie J, et al. The man-made mineral fiber European historical cohort study: extension of the follow-up. Scand J Work Environ Health 1986; 12 (suppl 1): 34–47.
28. Nicolaides-Bouman A, Wald N, Forey B, Lee P. International Smoking Statistics: A Collection of Historical Data from 22 Economically Developed Countries. Oxford: Oxford University Press, 1993.
29. Wong O, Foliart D, Trent LS. A case-control study of lung cancer in a cohort of workers potentially exposed to slag wool fibres. Br J Ind Med 1991; 48: 818–824.
30. Gardner MJ, Magnani C, Pannett B, Fletcher AC, Winter PD. Lung cancer among glass fibre production workers: a case-control study. Br J Ind Med 1988; 45: 613–618.
31. Chiazze L, Watkins DK, Fryar C. A case-control study of malignant and non-malignant respiratory disease among employees of a fibreglass manufacturing facility. Br J Ind Med 1992; 49: 326–331.
32. Chiazze L, Watkins DK, Fryar C, Kozono J. A case-control study of malignant and non-malignant respiratory disease among employees of a fiberglass manufacturing facility. II. Exposure assessment. Br J Ind Med 1993; 50: 717–725.
33. Marsh GM, Youk AO, Stone RA, et al. Historical cohort study of U.S. fiber glass production workers. I. Initial findings of the follow-up. J Occup Environ Med 2001;1992:43:741–756.
34. Hansen J, Boffetta P, Andersen A, et al. Comparison of information on occupation and lifestyle habits obtained from European man-made vitreous fibre production workers and their relatives. Int J Epidemiol 1997; 26: 1009–1016.
35. Axelson O. Aspects of confounding in occupational health epidemiology. Scand J Work Environ Health 1978; 4: 98–102.
36. Buchanich JM, Marsh GM, Youk AO. Historical cohort study of U.S. man-made vitreous fiber production workers. V. Tobacco-smoking habits. J Occup Environ Med 2001; 43: 793–802.
37. Marsh GM, Buchanich JM, Youk AO. Historical cohort study of U.S. man-made vitreous fiber production workers: VI. Respiratory system cancer standardized mortality ratios adjusted for the confounding effect of cigarette smoking. J Occup Environ Med 2001; 43: 803–808.
38. Boffetta P. Exposure to man-made vitreous fibres and cancer risk: a review of epidemiological studies. In: Peters GA, Peters BJ, eds. Current Asbestos Issues. Sourcebook on Asbestos Diseases. vol. 18. Charlottesville, VA: Lexis Publishing, 1998;191–218.
39. Krantz S, Cherrie JW, Schneider T, Ohberg I, Kamstrup O. Modelling of past exposure to MMMF in the European rock/slag wool industry. Arbete och Hälsa (Solna National Institute of Occupational Health) 1991;1:1–42.
40. Hallin N. Mineral Wool Dust in Construction Sites. Report 1981-09-01. Stockholm: Bygghalsan (The Construction Industry’s Organization for Working Environment, Safety, and Health), 1981.
41. Esmen NA, Sheehan MJ, Corn M, Engel M, Kotsko N. Exposure of employees to manmade vitreous fibers: installation of insulation materials. Environ Res 1982; 28: 386–398.
42. Engholm G, Englund A, Fletcher AC, Hallin N. Respiratory cancer incidence in Swedish construction workers exposed to man-made mineral fibres and asbestos. Ann Occup Hyg 1987; 31: 663–675.
43. Gustavsson P, Plato N, Axelson O, et al. Lung cancer risk among workers exposed to man-made mineral fibers (MMMF) in the Swedish prefabricated house industry. Am J Ind Med 1992; 21: 825–834.
44. Pohlabeln H, Jockel KH, Bruske-Hohlfeld I, et al. Lung cancer and exposure to man-made vitreous fibers: results from a pooled case-control study in Germany. Am J Ind Med 2000; 37: 469–477.
Keywords:

lung neoplasms; man-made mineral fibers; rock wool; slag wool

© 2002 Lippincott Williams & Wilkins, Inc.