Introduction

The mining of radioactive ores in the Erz Mountains in eastern Europe was the first occupation associated with an increased risk of lung cancer. Metal ores were mined in Schneeberg, on the German side of the mountains, beginning in the fifteenth century, and in Joachimsthal, on what is now the Czechoslovakian side, beginning in the sixteenth century.^26,30 Both areas were later mined for radioactive ores. As early as the sixteenth century, Agricola¹ described exceptionally high mortality from respiratory diseases in miners in this region. The lung-cancer hazard was first recognized by Harting and Hesse¹⁹ and was reported in 1879. Their report provided clinical and autopsy descriptions of intrathoracic neoplasms in miners, which they classified as lymphosarcoma. In a work force of about 650 men, Harting and Hesse counted 150 deaths from "miner's disease" between 1869 and 1877; in retrospect, most of these deaths were probably from lung cancer. During the early twentieth century, histopathological review of a series of cases established that the malignancy prevalent among miners in the Erz Mountains was primary cancer of the lung.^5,49

The problem was not recognized in the miners on the Czechoslovakian side of the Erz Mountains until 1929, when two cases of lung cancer were reported in Joachimsthal miners. In 1932, Pirchan and Sikl⁴⁶ described the autopsy findings in nine miners with lung cancer. These 9 miners were among 19 miners in Joachimsthal who died during 1929–1930. Formal epidemiological studies of the Schneeberg and Joachimsthal miners were not carried out, but published reports documented that about 50% of the miners eventually died from lung cancer.⁵³ Peller⁴⁴ calculated lung-cancer mortality rates for the Schneeberg miners during 1875–1912 and found that they were about 50 times those in Vienna males during 1932–1936.

Many authors offered explanations of the excess cancer in the Schneeberg and Joachimsthal miners (see references 26, 30, and 63 for reviews). Early theories emphasized dust exposure, metals in the ore (particularly arsenic), and increased susceptibility as a result of inbreeding in small mining communities. In 1924, Ludwig and Lorenser³¹ reported that radioactivity could be measured in the air and water in the mines of Schneeberg and might contribute to the development of lung cancer. Pirchan and Sikl⁴⁶ suggested in 1932 that radioactivity was the most probable cause of the Joachimsthal cancers, on the basis of the finding of radioactivity in both Schneeberg and Joachimsthal mines, the occurrence of lung-cancer in both locations, and the long exposure of underground miners to radioactivity. Teleky's opinion in 1937 was similar.⁶³ He could find no other satisfactory explanation and concluded that the high level of radioactivity, thought not to be present in other mines, led to the apparently unique lung-cancer problem of Schneeberg and Joachimsthal miners. In 1944, Lorenz³⁰ argued that radon alone could not be the cause of lung cancer and proposed that genetic susceptibility to lung cancer might be unusually high in the miners. However, during the 1950s and 1960s, as the biological basis of respiratory carcinogenesis became better understood and additional mining groups were studied, it came to be accepted that inhaled radon progeny were the cause of lung cancer in the Schneeberg and Joachimsthal miners and other exposed miners.^26,33,53

After World War II, several new epidemiological studies were initiated to determine the safety of exposure to radon progeny in mines. Unlike early studies, the newer surveys addressed such important biological questions such as the shape of the dose-response curve, the influence on risk of age at exposure, the effect of dose rate, the temporal expression of risk after exposure, and the interaction of radon daughters with other substances associated with lung cancer. This appendix reviews the epidemiological literature that is now available for addressing these issues.

Colorado Plateau Study

Beginning in the late 1940s, the American uranium industry grew rapidly in the Colorado Plateau, a mountainous region of southwestern Colorado and southeastern Utah. In 1949, in response to concerns about the health hazards to workers in this industry and with awareness of the high lung-cancer incidence in European miners in Joachimsthal and Schneeberg,²¹ the U.S. Public Health Service (PHS) began to investigate the uranium mines and mills in the Colorado Plateau region. The investigation combined an industrial-hygiene survey with a medical study of the workers. A prospective cohort study of miners and millers was carried out later, first by the PHS and then by the National Institute for Occupational Safety and Health (NIOSH). Until recently, this study offered one of the few epidemiological data bases for estimating the lung-cancer risk associated with exposure to radon progeny.

Field teams from PHS periodically conducted medical surveys of miners and millers between 1950 and 1960.³³ Before 1954, the teams did not attempt to examine all workers, but during 1954–1960, they tried to attain complete coverage. From among the examined miners, a group was assembled for follow-up that included miners who had worked at least a month underground in a uranium mine by January 1, 1964.³³ The number of subjects varied in the reports of this investigation (Table IV-1); in 1971, Lundin et al.³³ provided data on 3,366 white and 780 nonwhite subjects.

TABLE IV-1

Results of Colorado Plateau Study (Summarized from Principal Reports) of Male Uranium Miners.

An exposure data base was developed from diverse sources: PHS, state agencies, and the mining companies. Holaday (quoted by Lundin et al.³³ ) has provided a chronology (Table IV-2). During the period 1951–1968, for which cumulative exposure in working-level months (WLM) was initially calculated, nearly 43,000 measurements of radon-daughter concentrations were made in the approximately 2,500 mines that were worked (Table IV-3).³³ In discussing sources of potential inaccuracy in the working-level (WL) data, Holaday (quoted by Lundin et al.³³ ) pointed out that the measurements taken after 1960 were primarily for control purposes and might have led to overestimates of the exposures to miners.

TABLE IV-2

Chronology of Radon and Radon-Daughter Measurements in Colorado Plateau Study.

TABLE IV-3

Number of Mines Visited and Number of Measurements Made in Colorado Plateau Study.

Because coverage was not comprehensive for all mines in all years, several different estimation procedures were used to fill the gaps in the exposure data. These estimation procedures were more important in the earlier years, when exposures to radon daughters were higher and fewer measurements were available.

To make estimates for missing data in the temporal series of WL measurements for a particular mine, the investigators interpolated and extrapolated earlier and later concentrations of radon daughters. When gaps in the data were too wide, area averages by locality, district, and state were used. For 1950 and earlier years, WL values were estimated on the basis of the few available radon measurements and the investigators' knowledge of the mining conditions. Many of the miners worked in other types of hard-rock mines before becoming uranium miners. For exposures to radon daughters in the hard-rock mines, WL values were based on calendar year: 1.0 WL for years Before 1935, 0.5 WL for 1935–1939, and 0.3 WL for years 1940 and later.³³

The arithmetic average of the individual WL measurements made within a mine in a given calendar year was assigned to the mine for that year. Use of the arithmetic average implicitly weighted all measurements equally; error would have been introduced if the numbers of workers exposed at the concentrations indicated by the measurement were not uniform. The arithmetic average could also be strongly influenced by outlying high values.

To calculate WLM, the WL estimates were combined with work-history information obtained from annual censuses of active miners and from questionnaires. Apparently, a 170-h work month was assumed; and time for vacations, sick leave, or other absences from work was not subtracted from the number of underground hours estimated from the work history.⁵² However, cumulative exposures were also not adjusted for time worked beyond 170 h/month, a common practice in the early years of the industry.⁵² The investigators did not have enough information to consider work location within a specific mine or job classification, which might have influenced ventilatory demands.

Because WL measurements were sparse in relation to the numbers of mines that were worked, the WLMs accumulated by most miners were based on both measurements and estimates. In fact, WLM totals were calculated solely from measurements on only 10.3% of the white miners. For 36.1% of the white miners, some type of estimation was involved in the calculation of all WLM values; for the remainder, some WLM estimates were based on WL values derived by one of the estimation procedures.³³ In reports published to date, the WLM estimates have extended through September 1969. Information on cigarette smoking was obtained during the survey examinations, at the annual censuses of miners, and from mailed questionnaires.^33,69 As described by Whittemore and McMillan,⁶⁹ information on smoking was obtained on one to four occasions between 1950 and 1960, when the surveys were conducted, and at other times between 1963 and 1969.

Mortality in the cohort was determined with follow-up techniques that included records of the Social Security Administration and the Internal Revenue Service, direct contact, and other approaches.^33,67 Only a few subjects could not be traced, and nearly all death certificates were obtained. Most published reports are based on analysis with a modified life-table approach, which is a conventional method for longitudinal studies that compares observed with expected numbers of deaths by cause. More recently, several investigators have applied modeling techniques to the data.^23,24,34,69 In cohort analyses based on an external referent population, expected numbers of deaths were calculated with mortality rates for the western states where the mines were or with the rates for all U.S. white males.

Table IV-1 summarizes the principal reports for the white male miners. At all follow-up intervals, statistically significant excesses of lung-cancer deaths were reported; the standardized mortality ratios (SMRs), which are age- and calendar-year-adjusted ratios of observed to expected deaths, ranged from approximately 4 to 6, without an obvious temporal trend. In several reports, the investigators used stratified analysis to examine the exposure-response relationship of lung-cancer mortality with cumulative WLM by calculating standardized mortality ratios within strata of increasing WLM.^3,4,32,33,66 In one report,³² the mortality rates were standardized for cigarette smoking; in another,⁴ they were stratified by cumulative WLM and smoking. Lundin et al.³³ adjusted the expected numbers of lung-cancer deaths for cigarette smoking. The investigators usually provided tables stratified by the interval after the start of employment in uranium mining.

Lundin et al.³³ compared observed with expected numbers of lung-cancer deaths in six strata of lifetime cumulative WLM (Table IV-4). A statistically significant excess was present in all categories of exposure, except in the category of less than 120 WLM. Archer et al.⁴ provided mortality rates by exposure and cigarette smoking but did not include expected numbers of deaths.

TABLE IV-4

Lung-Cancer Deaths by Cumulative WLM in White Underground Miners in Colorado Plateau Study.

Mortality from causes other than lung cancer was also examined. Significant excesses were not observed for cancers at sites other than the respiratory system.^4,33,67 Greater than expected numbers of deaths occurred from tuberculosis, nonmalignant respiratory diseases, accidents, and suicides. The 1981 report by Waxweiler et al.⁶⁷ showed a statistically significant excess of deaths (SMR, 202) attributable to the grouping of chronic and unspecified nephritis and renal sclerosis.

The data on the white underground miners have also been analyzed with other statistical approaches. Lundin and coworkers^33,34 developed a descriptive model for the development of lung cancer after radon-daughter exposure; the model was based on the assumption of a time-latency distribution with the same shape and dispersion as that of leukemia incidence after a single radiation exposure. They used the model to examine the effects of latent period, age at exposure, dose rate, and cigarette smoking and to compare absolute- and relative-risk models for the effect of radon-daughter exposure. They found that the relative-risk model was preferable to the absolute-risk model and that a 10-yr latent period gave the best fit. Effects of age at first exposure and of exposure rate on lung-cancer risk were not demonstrated. With regard to cigarette smoking, Lundin et al.^33,34 concluded that nonsmokers had much less radiation-induced lung cancer and that the excess radiation-induced lung cancer in smokers was not heavily influenced by the extent of smoking.

Assuming an exponential form for the relative hazard, Hornung and Samuels²⁴ used the Cox proportional-hazards model on data accumulated through the 1977 follow-up date. They found that a lag period of 6–11 yr for exposure was most compatible with the data. The modeling also showed that the exposure-response curve was downward at higher doses; that is, lower exposure rates led to greater effects. On a multiplicative scale for assessing the effects of exposures on lung-cancer risk, smoking and radon-daughter exposure had statistically significant effects, but a cross-product term of the two exposures was not statistically significant. These analyses were limited, however, to examination of only the exponential form of the relative risk.

More recently, Hornung and Meinhardt²³ reported on a proportional-hazards analysis of data based on follow-up of the cohort through December 31, 1982. A total of 255 deaths from lung cancer was identified by that date. Hornung and Meinhardt considered exponential, linear, and power-function models of risk and chose the power-function model, because it provided the Best fit to the data. The model was developed with a stepwise approach; the data were best fitted by variables for cumulative WLM, cumulative smoking (in packs), and age at initial exposure. In the power-function model, the coefficient for the interaction of radon-daughter exposure and cigarette smoking was negative, although it was of borderline statistical significance (P = 0.058). This finding implies a submultiplicative, rather than purely multiplicative, interaction between cigarette smoking and radon-daughter exposure.

Hornung and Meinhardt²³ assessed the effects of several temporal factors: exposure rate, calendar year, age at exposure, and cessation of exposure. They found increasing risk with decreasing exposure rate, greater risk for more recent birth, greater risk for those first exposed at a greater age, and decreasing risk with cessation. The last two effects were thought to suggest a late-stage action of radon daughters, in the context of a multistage model.

Hornung and Meinhardt²³ used their power-function model to develop risk estimates for occupational exposures. Quantitative relative-risk estimates were made for occupational exposure beyond an assumed background exposure rate of 0.4 WLM/yr. For a 30-yr working lifetime, risk estimates were made for exposures of 30–120 WLM (1–4 WLM/yr). The relative risks ranged from 1.42 at 30 WLM to 2.07 at 120 WLM.

Whittemore and McMillan⁶⁹ used a case-control approach to examine additive and multiplicative models for the relationship of lung-cancer mortality to radon-daughter exposure and cigarette smoking. The results of their analyses are discussed Briefly here and more fully in Appendix VII. A multiplicative linear model, with excess relative risk given by the product of the risk associated with radon-daughter exposure and that associated with cigarette smoking, fitted the data better than an excess-relative-risk model in which excess risks associated with radon and smoking were added. A series of multiplicative relative-risk models was evaluated By the investigators. They found a Better fit for a model that incorporated the effects of smoking and WLM on relative risk as simple linear variables than for one that included exponential representations of these factors. Cumulative exposure variables fitted the data better than measures of exposure rate. Risk was not affected by age at the start of underground mining.

The PHS study cohort also includes nonwhite male miners, primarily American Indians. These subjects are of particular interest because of the low incidence of lung cancer in American Indians of the Southwest—a pattern probably attributable to a low prevalence of cigarette smoking.^4,50 Less information has been reported on the nonwhite subjects (Table IV-5). No cases of lung cancer among American Indians were observed initially, But a statistically significant excess was present in the 1974 follow-up.⁴ In fact, the expected numbers of cases were probably overestimated because of the use of mortality rates for all nonwhites rather than for American Indians alone. In New Mexico during 1969–1977, for example, the average annual lung-cancer mortality rate in American Indian males was 8.6/100,000, whereas the rate for non-Hispanic white males was 60.8/100,000.⁵⁰ Lung-cancer mortality rates for black males have generally been equal to or higher than rates for white males.

TABLE IV-5

Data on Nonwhite Male Underground Uranium Miners in Colorado Plateau Study.

Two other reports have addressed lung-cancer risks in American Indians employed in the Colorado Plateau mines. Gottlieb and Husen¹⁸ reported a case series of 17 Navajo males diagnosed as having lung cancer at the Shiprock Indian Health Service Hospital. Ail but one had worked as a uranium miner, and only two had smoked cigarettes; cumulative WLM ranged from 59 to 2,125. Samet et al.⁵¹ conducted a population-based case-control study to assess the association between uranium mining and lung cancer in Navajo males. Of 32 lung-cancer cases diagnosed between 1969 and 1982, 23 had a documented history of uranium mining. None of the 64 matched controls had been uranium miners. The results imply an extremely high relative risk in this nonsmoking population, but individual WLM estimates were not available for all miners, and the data cannot be used for quantitative risk estimation.

The Colorado Plateau study was designed and implemented 35 yr ago. Its strengths include the size of the cohort, the long duration of followup, the estimation of WLM for individual subjects, and the availability of cigarette-smoking histories. Application of new techniques to the data set has helped to explore the interaction between cigarette smoking and radon daughters and the effects of time-dependent factors such as dose rate and lag times. Even though investigators have dealt pragmatically with the severely limited number of WL measurements in calculating WLM estimates, the quality of the exposure information must be considered in interpreting the results of the study. Both random error and systematic bias might affect WLM estimates. Much of the exposure occurred before extensive measurement procedures were in place. For example, 36.1% of the total WLM ultimately accumulated by the cohort of white miners occurred before 1956 (Richard W. Hornung, NIOSH, personal communication, 1986). Few measurements were taken during the early years, when exposure rates were highest, so the higher exposures were probably estimated less accurately than later ones. If higher exposures were subject to a greater misclassification, the risk coefficients that have been calculated for the higher WLM values might be artificially low. Bias could also have been introduced by the investigators' decision to rely on measurements taken for control purposes after 1960, in that such measurements can over-represent higher exposures. Finally, the cohort had relatively high exposures and thus provides little information on the results of cumulative exposures of less than 100 WLM.

Czechoslovakian Uranium Miners

The retrospective cohort mortality study of the Czechoslovakian uranium miners was initially reported in 1971,⁵⁵ and periodic updates have been published.^{22,28,29,47,54,56,58} The cohort consisted of miners who began mining uranium ore in 1948–1957. However, the results in the more recent reports are limited to 2,433 miners⁵⁵ who began in 1948–1952. The selection criteria for the cohort have not been specifically described. The investigators have not reported whether the study cohort included all eligible miners in a particular geographic area or only a sample, what procedure was used and what records were reviewed to identify the cohort, the total number of miners who died from any causes other than lung cancer, and the distribution of the cohort members by birth year, age, or age when first exposed.

Individual work histories were abstracted from payroll cards for all miners (Langon Swent, personal communication, 1984) from 1948. For each miner, WLM was estimated from radon gas measurements and the number of months of employment at each mine in each calendar year. Since 1948, more than 120,000 radon gas measurements were made by measuring ionization current in an ionization chamber by electrometer. Yearly numbers of radon measurements were not given, but the lowest reported mean number of measurements for a year was 101 ± 8/mine. The range of coefficients of variation of average yearly radon concentrations in mines was 3.5–20.0%. Radon gas concentrations were converted to WL on the basis of ventilation conditions and practices, emanation rates from different types of ores, and after 1959, radon-daughter measurements. Since 1968, each miner's WLM has been determined from individual personal dosimetry cards. Assessment of dosimetry errors was based on the magnitudes of coefficients of variation, which do not provide information on the validity of the dosimetry data.

The cohort was followed with lung-cancer registrations administered by the authors in health facilities, the records of the hygiene service in the uranium industry, and oncology notification cards from throughout the country. The latter two served as independent follow-up sources after 1960. Until 1960, only 12 deaths due to lung cancer occurred. The success of this approach for identifying lung-cancer cases is not established, and the number of persons lost to follow-up is not given in the 1976 report by Sevc et al.⁵⁶ Except for a paper on skin cancer, health effects other than lung cancer have not been reported.

In analyses of this cohort, observed lung-cancer mortality was compared with that expected on the basis of age- and calendar-period-specific rates of the male population in Czechoslovakia. In the 1976 report by Sevc et al.,⁵⁶ person-years at risk for each subject were classified by the final cumulative WLM category, rather than being distributed across the appropriate WLM categories u they accumulated. This error was corrected in later analyses,^28,29 and only the later analyses are considered here. According to Swent (personal communication, 1984), a miner must have worked at least 4 yr underground to be eligible for inclusion in the cohort. However, person-years at risk were counted from the first date underground, rather than from the date of eligibility, so expected deaths were slightly overestimated.

Cigarette smoking was not assessed for all cohort members individually, but results of studies on a random group of 700 miners indicated that about 70% of the uranium miners were smokers. Data were not given on the amount smoked or the age when smoking started. According to Sevc et al.,⁵⁶ the prevalence of smoking in the general male population of Czechoslovakia was comparable with that in the sample of miners. Radon-daughter and other exposures from prior hard-rock mining were not evaluated, because less than 2% of the cohort miners had previously mined nonuranium ores. Other characteristics of this sample have not been reported.

The most recent and thorough analyses were based on follow-up through 1975 of miners who began exposure in 1948–1952. Follow-up averaged 26 yr.²⁹ In these modified life-table analyses, observed minus expected (based on the male population in Czechoslovakia) lung-cancer deaths were calculated for five categories of cumulative WLM (less than 100, 100–199, 200–399, 400–599, and 600 and over) and, with further stratification, for three categories of duration of exposure (0–7.9 yr; mean, 5.6 yr; 8–11.9 yr; mean, 9.5 yr; and 12 yr or longer; mean, 14 yr) or for three temporal exposure patterns. A temporal pattern of exposure was modeled for each miner individually by the regression, cumulative WLM = a ^b, where cumulative WLM was calculated for each year of work, t. The cohort was then divided into three groups based on individual members' value of b: group A, b significantly less than 1, implying a high rate followed by a low rate of exposure; group B, b not significantly different from 1, implying a fairly constant rate of exposure; and group C, b significantly greater than 1, implying a low rate followed by a high rate of exposure.

The authors reported two general findings. First, the analyses indicated significant effects of cumulative WLM, duration of exposure, and their interaction. In this study, excess risk is expressed as the excess number of lung cancers per 1,000 miners (Table IV-6) and not per person-year, as is reported in most other cohort studies reviewed here. For those exposed 12 yr or longer, this risk was linearly related to cumulative WLM for miners overall (Figure IV-I), but not for the two shorter exposure periods. Second, excess risk was linearly related (Figure IV-2) to cumulative WLM for miners in groups A and B, but not group C (low followed by high exposure rate). Other analyses of the data²⁹ indicated a significant effect of cumulative WLM on excess risk, but not of exposure pattern or their interaction. From Figures IV-I and IV-2, it appears that the 95% Poison-based confidence intervals are wide enough to allow nonlinear interpretations of the relationship between excess risk and Cumulative WLM within separate groups of exposure duration or temporal pattern.

TABLE IV-6

Lung-Cancer Mortality among Czechoslovakian Uranium Miners.

Figure IV-I

Relation between additional lung-cancer frequency and cumulative radiation exposure in three groups of Czechoslovakian uranium miners by mean duration of exposure. Source: Kunz et al.

Figure IV-2

Relation between additional lung-cancer frequency and cumulative radiation exposure in three groups of Czechoslovakian uranium miners by time course of exposure accumulation (see text). Source: Kunz et al.

An earlier report²⁸ of follow-up through 1973 is the only report on the cohort of Czechoslovakian uranium miners that provided observed and expected mortality rates per 10,000. person-year and observed to expected lung-cancer mortality ratios, in addition to excess lung-cancer deaths. However, only one independent variable, cumulative WLM (<100, 100–199, 200–399, and >400), was reported (Table IV-7).

TABLE IV-7

Lung Cancer Among Czechoslovakian Uranium Miners in Relation to Cumulative Radon-Daughter Exposure Based on Modified Life-Table Method.

Ontario Uranium Miners

A retrospective cohort study of Ontario miners^37–39 engaged in various types of mining included a subcohort of uranium miners who met the following criteria:

• received a miner's physical examination required annually by the company any time in 1955–1977 (uranium mining began in 1955 in Ontario);
• worked at least 1 month as an underground uranium miner; and
• had not worked in a job with any known asbestos exposure, in uranium processing (except in mills), or in any uranium mining in another province as an employee of Eldorado Nuclear.

Radon-daughter exposure was estimated by different methods for 1967 and earlier and for 1968 and later. For 1968 and later, exposure records of WLM maintained by the mining companies were used. For 1957–1967, the investigators calculated WLM by combining WL information with work histories.³⁸ Because of the variability of radon-daughter concentrations, the investigators developed two separate sets of WL values for this earlier period. The standard (or lower) WL values were the averages of the four quarterly averages or three 4-month averages for a particular year. To calculate the special (or upper) WL values, the investigators weighted the average of the four highest quarterly measurements or the three highest 4-month measurements in headings, stoops, and raises (a total of 12 or 9 measurements, respectively) by 0.8 and the average of the four highest quarterly or three highest 4-month measurements in travel ways by 0.2. The difference between the standard and special WL values varied with mine and year;³⁸ for some mines in some years, the special and standard values were equivalent, but the special values were up to 4 times the standard WL estimates in the years and mines for which both were available.³⁸ The investigators considered that the true exposure of each man lies within this range. During 1958–1967, 13,081 measurements were taken (Table IV-8). For one large mine, WL data for the 4 yr from 1957, when the mine started operating, through 1960 had to be rejected, because they were shown to be unreliable. The values for the missing years were estimated by taking into account tonnage mined, ventilation, and dust concentrations at various times.

TABLE IV-8

Numbers of Mines and Measurements in Study of Ontario Uranium Miners.

Work-history information was obtained primarily from records of pre-employment and yearly examinations carried out by Ontario government agencies,³⁸ Additional information related to the first 5 yr of employment in the mining industry was collected from work-history cards.

The WLM values for 1955–1967 were calculated by combining the work-history information with a matrix of annual WL values for each mine in each year. Adjustment was made for deviations from normal working hours in a mine, considered to be 2,000 h/yr. No estimates of WL were made for prior gold-mining experience, but persons with such experience were analyzed separately, because Ontario gold miners were at increased risk of lung cancer.⁴⁰ It should be noted that the committee's analysis of the Ontario miners, described in Annex 2A, excluded miners with previous gold-mining experience.

Because the WL measurements did not cover the complete working experience of the cohort, some estimation of exposures before 1954 was necessary. These years included the period of highest exposures and, as Muller et al.³⁸ reported, 22% of the total WLM accumulated by the cohort is based on extrapolation from measured values, with account taken of, for example, ventilation. For one large mine, this percentage includes extrapolation up to 1960. The period of extrapolation weighted by WLM is, however, less than 2 yr.

Follow-up through 1981 was carried out by computer linkage with national mortality data bases combined with manual cross-checking to resolve problems. The investigators did not report on the percentage lost to follow-up or on the percentage of death certificates not obtained. However, on the basis of a sample of known deaths, 6.3% were not identifiable as deceased with the same follow-up method. Death certificates were the only source of information on cause of death.

Using the modified life-table technique, Muller et al.³⁸ compared observed with expected mortality (based on the Ontario general male population rates with adjustment for age and calendar period). Results for causes of death other than lung cancer were available only for 1955–1977. The authors did not have information on cigarette-smoking habits of the miners.

The mean cumulative WLM of miners with no previous gold-mining experience was 40 (lower estimate) to 90 (upper estimate). All other descriptions of the cohort included those who had previously mined gold. The median year of birth of the cohort was 1932, and the median year first employed in a mine in Ontario was 1957; thus, the median age at first employment in a mine in Ontario was probably about 25 yr. The median duration of work in a mine was 1.5 yr.

Among uranium miners without any gold-mining experience, Muller et al.³⁸ found that observed to expected ratios for lung-cancer deaths increased across the six categories of cumulative WLM (Table IV-9). When the upper estimated exposures were used, the first definite excess occurred at a cumulative WLM of 100–170 (mean, 130), with 14 observed and 6.9 expected lung cancers. When the lower estimated exposures were used, there was a definite excess at a cumulative WLM of 40–70 (mean, 53), with 13 observed and 7.0 expected lung cancers. Muller et al.³⁹ reported that linear regression of the dose-response relationship, weighted by number of person-years at risk (PYAR), showed similar fits for the excess- and relative-risk models.

TABLE IV-9

Observed and Expected Lung-Cancer Deaths by Cumulative WLM among Ontario Uranium Miners with No Gold-Mining Experience.

A 5- and 10-yr exposure lag did not change the slope of the relative-risk model (0.5% excess relative risk per WLM for the upper exposure estimates and 1.3% for the lower ones), but slightly increased the slope of the excess-risk model (from 4.8 to 7.2/million WLM for the lower exposure estimates and from 2.0 to 2.8 for the upper ones). With either model, the use of the upper exposure estimates decreased the dose-response slope by more than 50%. However, dose-response analysis for two age groups of PYAR indicated that the slopes for the relative-risk model were age-independent, whereas the slopes for the excess risk model were not.

Eldorado Uranium Miners

Howe et al.²⁵ conducted a retrospective cohort mortality study of all 10,945 male employees who had worked at the Eldorado Uranium Mine in Beaverlodge, Saskatchewan, anytime between 1948 (when the mine opened) and December 31, 1980. The cohort was identified from company employment and payroll records. The final study group included 8,487 subjects; 1,782 (16%) persons were excluded because of missing or incorrect information, and another 676 (6%) were excluded because they had worked at other company sites. The authors were unable to detect any bias due to these exclusions. Follow-up from 1950 through 1980 was carried out by linkage with a national mortality data base. Only one person was lost to follow-up.

The WLM values for Beaverlodge uranium miners were calculated by Eldorado Resources Ltd., which operated the mine. Different approaches were used for 1966 and earlier years and for 1967 and later years. For 1966 and earlier, the WL estimates were based on all available measurements of radon and radon daughters (Table IV-10). Equilibrium between radon and its daughters was estimated by comparing paired measurements of radon and radon-daughter concentrations. When paired measurements were unavailable for a particular year, the average of the equilibrium factors for adjoining years was interpolated. Because the distribution of measurements was strongly skewed toward higher values, the annual median, rather than the mean, was used to calculate exposure for each year.^* For 1967 and later, radon-daughter measurements were generally available. Geometric means or averages of geometric means were used for the calculations. For some locations, adjustments were made on the basis of working conditions.

TABLE IV-10

Numbers of Radon-Daughter and Radon Measurements in Eldorado Beaverlodge Uranium Miner Study, 1954-1966.

In calculating the WLM for the work force, the WL values for each year were adjusted for the extent of underground exposure sustained by workers in eight occupational categories. Dates of employment were used to determine the number of weeks worked in each year. Four weeks of holiday time each year were assumed, and adjustments were made for the changing duration of working hours over the study.

Silica exposures to this cohort were always very low, and diesel machinery was never used underground. Potential confounding from other mining exposures was addressed in one analysis by excluding the 540 men who were included in the Ontario miner study³⁸ and by excluding miners who had reported previous mining experience elsewhere. No measures of cigarette smoking were reported for cohort members individually.

The final cohort consisted of three groups: surface workers only (48%), underground workers only (45%), and both surface and underground workers (7%). The mean years of first exposure for these three groups were 1966, 1966, and 1963, respectively. The mean ages at first employment were 27.7, 28.8, and unreported, respectively. The mean periods of follow-up were 13.9, 13.5, and 17.3 yr, respectively. The mean durations employed were 22.2, 15.0, and 43.9 months, respectively. The means of cumulative WLM were 2.8, 16.6, and 28.9, respectively.

A modified life-table analysis was carried out. Comparisons were made with 5-yr, age- and calendar-period-specific mortality rates for the general male population of Canada.

The finding of no lung-cancer excess among those with less than 5 WLM (19 observed versus 18.36 expected) was interpreted as evidence against strong confounding by cigarette smoking in the entire cohort. Furthermore, among those with greater than 5 WLM, no excess of lung cancer was found within the first 5 yr after exposure began. The authors excluded the first 10 yr of follow-up from further analyses, to be consistent with procedures in other studies, although an excess risk of lung cancer was found at higher doses within 5–9 yr after first exposure (6 observed versus 1.54 expected).

The SMRs for lung cancer increased monotonically (Table IV-11) from the lowest to the highest category of cumulative WLM (0–4, 5–24, 25–49, 50–99, 100–149, 150–249, and 250+). The authors used weighted least-squares regression to describe the exposure-response relationship. Exposure within each category was represented by the mean cumulative WLM, and PYAR was used for weighting. The addition of a quadratic term did not significantly improve the fit of the linear model to the data. When the authors multiplied simple linear functions by exponential terms to represent a cell-killing parameter, they were unable to fit a biologically appropriate model to the data. Furthermore, a 5-yr lag of exposures changed the linear-regression coefficients by less than 10%, compared with no lag. Howe et al.²⁵ also investigated the effects of age at first exposure and age at observation. In both cases, the attributable risk was found to be much more dependent on age than was the relative risk.

TABLE IV-11

Observed and Expected Lung-Cancer Deaths by Cumulative WLM, 1950-1980 (First 10 Years of Followup Excluded) Among Eldorado Beaverlodge Uranium Miners.

The Beaverlodge miners have also been included in a larger study of Eldorado Resources Ltd. employees. Nair et al.⁴¹ conducted a retrospective cohort mortality study of all males employed before 1981 at four major operations: a pitchblende mine at Port Radium from 1932 to 1940 (time period excluded from study) that was later a uranium mine during 1942–1960; a refinery at Port Hope, Ontario, which opened in 1932, refined radium until 1954, and refined uranium and converted it to uranium dioxide and uranium hexaflouride until the present; a uranium mine at Beaverlodge from 1953 until the present; and other sites.

The cohort was assembled from a company employee roll that included full name, sex, place and date of birth, and last year known alive. The Port Radium cohort was divided into those who ever and those who never worked underground. Follow-up was limited to computer linkage with a national mortality data base for 1950–1980. Bias might have been introduced by the rejection of a large percentage of each cohort (Port Hope, 38%; Port Radium, 44%; and Beaverlodge, 13%) because of inadequacy in personal data or loss to follow-up before 1950.

Preliminary findings on lung-cancer mortality were reported only as observed deaths due to lung cancer for each work group versus those expected based on national rates (Table IV-12).

TABLE IV-12

Standard MoRality Ratios for Lung Cancer among Eldorado Employees.

Later examination determined that the Port Radium surface cohort included a number of underground miners. These results are not useful for assessing dose-response relationships, because data on WLM were not available.

French Uranium Miners

Tirmarche et al.⁶⁴ carried out a retrospective cohort study of all men who began underground uranium mining during 1947–1972 in any of 12 French mines and worked a minimum of 3 months. For 1947–1955, WLM values were based on a few radon measurements, ventilation conditions, ore characteristics, and working methods. Extensive radon measurements were taken later; there were an average of 20–30 values taken per mine/year during 1957–1970 and twice that during 1971–1980. WL was estimated retrospectively using the current equilibrium factor of 0.22. The only epidemiological results were for lung-cancer mortality: 36 observed versus 18.77 expected among the entire cohort without any lag in exposure or latency considerations. No dose-response results were reported. It appears that PYAR for each miner began inappropriately at the date of first employment and not after 3 months of mining. National mortality rates were used for comparison; however, the mines are all in agricultural areas. The WLM data in this study are potentially limited by the lack of measurements for 1947–1955 and the retrospective estimation of the equilibrium factor. Furthermore, cause of death was not known for 25% of the deceased subjects.

Cornish Tin Miners

High concentrations of radon and its daughters have been measured in tin mines in Cornwall, England. Fox et al.¹⁷ conducted a retrospective cohort study of mortality in 1,333 men employed in two tin mines in Cornwall during 1939. In comparison with mortality rates for England and Wales, lung-cancer mortality was increased in the underground miners (SMR, 211), but not in surface workers (SMR, 74) or in workers who were not classifiable into either of these two categories (SMR, 94). WLM were not estimated for the subjects. The authors reported government estimates of 25 WLM and 15 WLM, respectively, annually for the two mines.

Chinese Tin Miners

Tin has been mined in the Yunnan region of China for centuries,⁶² and the miners in this region are known to have arsenic and radon-daughter exposures. Wang et al.⁶⁸ identified a cohort of 12,243 underground miners and followed them from 1975 to 1981 for lung-cancer incidence and mortality. Information has not been reported on the selection of study subjects, their duration of work, latency distribution, smoking distribution, followup methods, or losses to follow-up. The age distribution of the cohort was not given, but it has been reported that many persons began underground mine work between the ages of 8 and 14 yr;⁶² this practice was phased out around 1949. WLM were calculated from detailed individual work histories and systematic radon-daughter measurements at underground work boxes during 1972–1980. Only natural ventilation was used in the mines in 1953–1972, so exposures were assumed to be constant during this interval. Before 1953, some of the mines were smaller, no wet-mining methods were used, and proportionate adjustments were made. Another adjustment was made for exposures before 1949, when more primitive mining methods, including back-carrying of ore through narrow tunnels, were used.

During the follow-up period, lung-cancer incidence was 515/100,000 (499 cases) among underground miners, 41.3/100,000 (59 cases) among surface workers. The investigators did not report the incidence data by dose, years worked, or latency. Lung-cancer mortality for the underground miners was also compared with that of the Shanghai population (Table IV-13), but apparently without adjustment for age, sex, calendar period, or smoking.

TABLE IV-13

Lung-Cancer Mortality by Cumulative Radiation Exposure among Chinese Tin Miners.

In addition to radon daughters, exposure to arsenic was considered to play an etiological role in the lung-cancer excess. Ore samples contained 1.5–3.5% arsenic; the investigators estimated that a miner's respiratory tract was exposed to 1.97–7.43 mg of arsenic/yr during the years immediately after 1949. Total dust concentrations were estimated at 30 mg/m³, with peaks during dry drilling of 344 mg/m³ in the earlier years.

Sun et al.⁶² described 929 lung-cancer cases (755 deaths) ascertained during 1954–1978 among workers at three mines of the Geiju Tin Mine Company in Yunnan. Death certificates, histological-cytological reports, and chest x rays were cross-checked to confirm the cases. The little available information on relative risks was based on a crude cohort study that calculated expected deaths using the age distribution of the workers in one of the mines in 1975. The cohort analyses that controlled for duration of mining indicated significant differences in SMRs by age at which the miners began mining. However, for those who began mining before age 14, risk did not increase with duration of mining. The smoking habits of 17,287 miners were recorded. The authors reported that the relative risk for lung cancer in smoking miners was about 20 times higher than that in nonsmokers. There was no significant relationship between latent period and degree of smoking.

The age distribution of the work force in the three mines in 1973 was trimodal, with peaks at 20, 25, and 40 yr, which is a reflection of temporal changes in hiring practices. This unusual age distribution and the emphasis on case ascertainment (i.e., follow-up) during 1971–1978 obscures the relationships among age at which miners started mining, latency period, length of follow-up, and risk. However, of the large number of persons who began work underground before the age of 14, few developed lung cancer before the age of 35.

Canadian Fluorspar Miners

The open-pit mining of fluorspar (calcium fluoride) in St. Lawrence, Newfoundland, began in 1933. Underground mining began in 1936 and has been carried out in 12 mines. After the discovery in the 1950s of an unusual excess of lung-cancer deaths among the miners, a retrospective-prospective cohort study was undertaken.^13,14,35,36 The ore itself is not radioactive,¹⁴ but the substantial amounts of water seeping through the mines contain radon gas.

Exposures have been estimated on the basis of occupational histories that include type and place of work and hours of work by year. For years before and to 1960, work hours were converted to working months (167 or 170 h) and used to calculate WLM. WL values were estimated retrospectively from measurements made in only one mine in one year, 1959. Before 1960, the mines were ventilated primarily by natural draft, occasionally aided by small blowers, and the ventilation varied greatly in each mine, as did the amount of water seepage. Radiation measurements were infrequent during 1960–1968, but were taken daily after 1968.

Recent follow-up of the cohort has been primarily through linkage with the nationwide mortality data base³⁶ but has also included a small number of deaths certified by local clergy, parish records, hospitals, and relatives, rather than medically. All persons not definitely identified as deceased were assumed to be alive. Initial analyses of lung-cancer mortality in 1952–1960 showed 21 deaths among the miners, compared with 0.7 expected from age-adjusted rates for the remainder of Newfoundland. Recent analyses^35,36 used standard modified life tables of PYAR and age-specific lung-cancer rates among the surface workers for comparison. This analysis is limited by the small number of lung-cancer deaths (seven) in the comparison group and possibly by migration between surface and underground work. Follow-up was accomplished with the use of the company, union, and medical files. Analysis based on follow-up through 1978 (ignoring the first 10 yr of risk after hiring) showed a strong dose-response relationship between lung-cancer risk and cumulative WLM (Table IV-14).³⁶ Smoking-specific findings are not reviewed because of the lack of adequate ascertainment of smoking status. The authors found that latency periods decreased for men first exposed when older and for men exposed during the earlier years, when exposures were presumably higher.

TABLE IV-14

Lung-Cancer Mortality by Cumulative Radiation Exposure among Canadian Fluorspar Miners.

Swedish Iron Miners: Malmberget

A retrospective cohort mortality study by Radford and Renard⁴⁸ included miners from two iron mines (in Malmberget and Koskoskulle) owned by one company (LKAB). Selected for study were the 1,415 men born in 1880–1919 who were alive in 1930 and who worked underground for more than i calendar year during 1897–1976. The cohort was identified principally from company and union records of active and pensioned miners that dated back to 1900. Additional men were identified from medical surveys and a few were identified from parish records. Time worked underground was determined from company and union records and medical files. Work histories appear to contain data only by year; July 2 was assumed as a starting and stopping date for underground work. For those who stopped and restarted in 1 yr or started and died in 1 yr, April 1 and October 1 were assumed, respectively. The extent to which the cohort covered all employed miners was evaluated for the years 1942–1946 by comparing person-years underground from two sources: the work histories of the cohort and company records.

The WLM values for this analysis were those calculated by Radford and Renard.⁴⁸ As described in their 1984 report, radon dissolved in water was assumed to be a major source of radon daughters in the mines. Comparison of radon measurements in water taken in 1915 with data from 1972 and 1975 indicated constant radon concentrations in groundwater. The first measurements of radon in mine air were obtained in 1968. Radon and radon daughters were later measured by the mining company and by the National Radiological Institute. Past concentrations were then reconstructed on the basis of these measurements in combination with information on ventilation conditions. Radford and Renard assumed that ventilation conditions in 1968–1972, when the measurements were made, were not greatly different from those in the past. In support of this assumption, they cited a pattern of natural convection and data on quarts dust concentrations that extended to the 1930s. Radon daughters were found to be at about 70% equilibrium with radon.^*

Time worked underground was determined from company and union records and from medical files. Adjustments were made for variation in the average number of hours worked underground in a month. Average yearly WLM were calculated for each 10-yr calendar period from the average number of hours per month underground and from radon-daughter concentration in each area, with weighting by the company data on the numbers of person-hours worked underground in each section of the mines.

Follow-up of the cohort depended on parish records and was thorough (only one person was untraceable through 1976). Of the lung-cancer deaths, 70% had been confirmed by autopsy or thoracotomy, but only death-certificate information was used for comparisons. The expected number of cases was based on age- and calendar-year-specific national mortality rates for males since 1951. Accordingly, PYAR and expected deaths were calculated from the later of two dates: January 1, 1951, and January 1 of the year after a miner began work underground. Induction latency periods were considered in two ways: by excluding PYAR for each miner for 10.5 yr after mining was begun and by lagging the cumulative WLM by 5 yr.

Information on cigarette smoking was not reported for all cohort members individually, but only for a sample of the responses to a 1972–1973 survey of active miners and surface workers and from a 1977 survey of pensioners in the study cohort. In addition, smoking histories were obtained for each lung-cancer death. The authors estimated smoking-specific lung-cancer SMRs for two categories: smokers combined with recent ex-smokers and all others. They based these SMRs on the ratios between a sample of the responses from the surveyed miners and a national population study of the age-specific proportion of smokers and the amount smoked. Interpretation of the SMRs must be constrained by the lack of information for all cohort members on smoking as presently reported for the surveyed miners (556 of 1,294, or 43%), by differences in the periods associated with the questionnaire data from the miners (1972 and 1977) and from the national population sample (1963–1972), and by the use of information provided by the next of kin for deceased lung-cancer cases.

Other potentially confounding variables for lung cancer were considered. Silicosis, examined in a case-control study nested within the cohort, was found to be equally severe and prevalent in lung-cancer victims (14/50) and in age- and work-period-matched controls (26/100). Diesel equipment, with its exhaust, was not introduced into the mines until the 1960s, by which time 70% of the persons who later developed lung cancer had terminated work. Arsenic, chromium, and nickel—known respiratory carcinogens—were virtually absent in analyses of bedrock. X-ray diffraction of airborne dust samples from the mine showed no identifiable asbestos fibers. Indoor radon concentrations in miners' homes ranged from 0.002 to 0.03 WL, but had been measured in a sample of homes selected because of potentially high concentrations. The lung-cancer rates among nonminers in this region are lower than Swedish national rates.

Of the mining groups exposed to radon daughters, this cohort offers one of the longer follow-up experiences. Over 41% of the cohort (532/1,294) were deceased. The average year first employed underground was 1932, the average age at first employment was 27.8, and the average duration underground was 19.5 yr. The average exposure rate was 4.8 WLM/yr, resulting in an average cumulative WLM of 93.7 (range, 2–300 WLM). Cause-specific and total mortality were assessed with a modified life-table analysis. Excesses of observed deaths were found for total mortality, lung cancer (50 observed versus 12.8 expected), stomach cancer (28 observed versus 15.1 expected), and all causes except cancer combined (393 observed versus 312.6 expected). The latter excess was due to silicosis, occupational accidents, and cardiovascular disease, according to Radford and Renard.⁴⁸

Lung-cancer mortality was studied in detail. Excess risk was not evident until at least 20 yr after the start of underground work. Significantly increased risks were found for both smokers (32 observed versus 11.0 expected; SMR, 291) and nonsmokers (18 observed versus 1.8 expected; SMR, 1,000). The excess-risk coefficient for smokers was 21.8/million person-yr WLM, and for nonsmokers, 16.3. The combined effect of smoking and radon-daughter exposure was reported as nearly additive,⁴⁸ although formal statistical testing, as described in Appendix VII, was not carried out. (The rate ratio for smokers versus nonsmokers based on the Swedish population study was estimated by the authors to be 7.4.)

Dose-response relationships were evaluated for five categories of lagged cumulative WLM (0–49, 50–99, 100–149, 150–199, and over 199). An excess of lung cancer was found even in the lowest dose category (8 observed versus 3.4 expected), and the dose-response data were equally consistent with absolute- and relative-risk models, as measured by weighted correlation coefficients. Assessments of effects of age at beginning of work, year of beginning work, and age at risk were undertaken separately and not by multivariate modeling.

Swedish Iron Miners: Kiruna

A proportionate-mortality-ratio study was carried out in Kiruna, Sweden, to compare cause-specific mortality distributions among underground iron miners from two companies (LKAB and TGA), surface miners and workers (LKAB), and all other male deaths in Kiruna.²⁷ Selected for study were all deaths registered in Kiruna that occurred in 1950–1970 among men aged 30–74. Because rates of emigration from Kiruna were very small, the authors considered that nearly all deaths among the miners would have been registered there. Lung-cancer deaths were verified from hospital records for 41 of 42 cases, and autopsies were performed in all 13 cases among miners. An additional analysis attempted to calculate SMRs among active employees of the mines (surface and underground combined), but was limited by the absence of data.

After age adjustment, underground miners experienced 13 lung-cancer deaths (12 were after 1957) versus 4.5 expected based on the cause-specific distribution of deaths among all other residents of Kiruna and versus 4.2 based on the cause-specific distribution of deaths among the entire Swedish male population in 1951–1966. Analyses were not presented on dose-response relationships, latency, or interaction of cigarette smoking and exposure to radon daughters.

The iron mine was an open-pit mine until the 1950s, when underground mining began. Diesels were introduced in the late 1950s. Quartz concentrations were around 7% of the particles smaller than 5 µ. The concentration of radon daughters, measured only since 1970, was 10–30 pCi/liter in most places and much higher in some unventilated areas. In 1966, a survey of all employees showed that about two-thirds of both underground and surface workers were smokers. Information from coworkers and next of kin indicated that 12 of the 13 lung-cancer cases among underground miners were smokers (four of these smoked only a pipe).

Swedish Iron Miners: Kiruna and Gallivare

A case-control study¹¹ was carried out on 604 lung-cancer victims who died during 1972–1977 in three counties in northern Sweden. These counties contained two major iron mines, which were in two separate municipalities, Kiruna (containing the Kirunavaara mine) and Gallivare (containing the Malmberget mine). The investigators used next-of-kin interviews to determine underground mining and smoking histories. No estimates were made of WLM or duration of mining. The investigators concluded that their data showed that relative risks for smoking and underground iron mining were between additive and multiplicative in their combined effect. However, the wide confidence intervals in their data are consistent with an additive, a multiplicative, or a more extreme interpretation.

A recent extension of this case-control study¹² included 69 deaths during 1972–1982, but was limited to Kiruna and Gallivare. The median age of the subjects was 66. WLM were not estimated, but lung-cancer risks by duration of underground iron mining and lifetime number of cigarettes smoked were found to fit a multiplicative-risk model based on linear logistic regression. Unfortunately, statistical testing of the model was not reported, and the data were limited by the small numbers of nonsmoking miners (four) and nonsmoking nonminers (two) among the cases.

Swedish Iron Miners: Grangesberg

Edling¹⁵ carried out a case-control study of all male residents known to have died of lung cancer during 1957–1977 in the iron-mining town of Grangesberg, Sweden. The unmatched controls (897), who all died of other causes, and cases were submitted to the local iron-mining company for identification of history of underground mining. The author found an age-adjusted rate ratio for lung-cancer deaths associated with employment at a mine (16.6) that was significant (95% confidence interval, 7.7–35.3). Strikingly, 42 of the 47 lung-cancer cases had mined underground.

A separate analysis in the same report of the effect of cigarette smoking¹⁵ added cases through 1980, but included only persons who had been underground miners. A new set of controls (individually matched for age, sex, and year of death) who died from causes other than malignancy and had been underground miners was selected (44 pairs). Smoking histories were obtained from next of kin by telephone. A risk ratio of 2.0 (95% confidence interval, 0.7–5.7) was found for smoking and lung cancer; the author interpreted that as not fully consistent with the general experience of at least a 5-fold to 10-fold risk ratio.

A second case-control analysis on the same population¹⁶ used only controls who died during 1966–1977 at ages over 50. The authors found a lower age-standardized lung-cancer death rate ratio than in the previous analysis (relative risk, 11.7; 95% confidence interval, 5.3–26.0). A separate analysis of smoking similar to the one described above resulted in a risk ratio of 1.5 (95% confidence interval, 0.4–5.3) for smoking and lung cancer, on the basis of 28 matched pairs.

Edling and Axelson¹⁶ also estimated a lung-cancer excess risk per million person-years WLM for miners aged 50–64 (26 excess cases) and aged 65 or greater (54 excess cases). These estimates were made by multiplying the number of miner person-years at risk in Grangesberg during 1966–1977, as estimated from town censuses, by the proportion of controls from the case-control study who had previously been underground miners. Cumulative WLM was estimated by multiplying the average number of years worked underground by the product of number of cases and 0.5 WL—an exposure assumed to apply to the entire period, according to 1969 mine measurements. Although these risk estimates were based on extensive assumptions, the authors noted that they were in agreement with estimates in the report by the Committee on the Biological Effects of Ionizing Radiations (BEIR III).⁴²

General Swedish Miners

Snihs⁵⁹ presented an epidemiological study of miners in all districts of Sweden. Sweden had 60 underground mines; all mined ferrous and sulfide ores, and none mined uranium. Radon measurements were made in all mines in 1969–1970 with 4.80-liter propane containers and ionization chambers. Radon daughters were sampled with conventional glass-fiber filters and analyzed by the Kusnetz method. In March 1972, exposures were limited by regulation to an annual average of 30 pCi/liter. Snihs⁵⁹ reported that air Brought into the miners for ventilation, rather than water or rocks in the working areas, was the predominant source of exposure in 17 of 22 mines.

A nationwide cohort of all miners aged 20–64 and employed during 1961–1968 was followed during 1961–1971. The follow-up and analytical methods were not described. It is unclear whether follow-up extended beyond 5 yr after employment ended.

Observed and expected lung-cancer deaths during 1961–1968 were compared among underground miners, aboveground miners, and nonminers in the mining districts. The methods were not fully described, but it appears that the estimated annual number of active miners aged 20–64 during this period was multiplied by district age-specific lung-cancer rates to estimate the expected number of deaths. Observed deaths were included if they occurred within 5 yr of cessation of mining. Annual WL was estimated from measurements only after 1969. Cumulative WLM was estimated from the WL estimates and duration of exposure estimates for all workers based on data on those dying of lung cancer. Limitations of the study noted above make it difficult to judge the validity of the dose-response relationship results.

Swedish Lead-Zinc Miners: Hammar

Axelson and Sundell⁷ studied Kiruna (Hammar parish) lead-zinc miners with a case-control study embedded in a crude cohort study. In the case-control study, the 29 cases included all men who died from lung cancer during 1956–1976 in the parish surrounding two physically connected lead-zinc mines. Controls (174) consisted of the first three deaths other than from lung cancer listed before and after the case in the chronologically ordered parish registry, but matching was dropped in the analysis. The authors believed that the registry included fairly complete diagnoses from the death certificates. The local mining company assessed the underground work experience of all subjects.

Smoking habits of the miners were learned from medical files and interviews with two retired foremen. For 2 of 10 subjects on whom smoking information was independently obtained from more than one source, the information was conflicting.

The age-standardized rate ratio for lung cancer among lead-zinc miners was 16.3 (90% confidence interval, 7.8–35.3). Among underground miners, those who had never smoked (nine) appeared to have longer work-related induction latent periods than smokers (nine) (respective medians of 49 versus 37 yr) and to have a greater risk of developing lung cancer.

Norwegian Niobium Miners

Solli et al.⁶⁰ followed a cohort of employees at a niobium-mining company that operated from 1951 to 1965. Niobium itself is not considered to be carcinogenic, but the ore also contained ²³⁸U (0.3–2 ppm) and ²³²Th (50–300 ppm). Exposure estimates for the cohort were of questionable quality. The WLM from both radon and thoronium progeny was calculated for the employees on the basis of measurements of alpha activity during 2 days in 1959. Among the employees, a strong dose-response relationship was found between lung-cancer risk and cumulative WLM (Table IV-15). Poor dosimetry probably resulted in underestimation of exposures by about a factor of 2, according to the authors. Lifetime occupational histories indicated that three of the subjects had been previously exposed to asbestos and one had mined iron. In addition, 75% of the employees were smokers, compared with 60% of the Norwegian population.

TABLE IV-15

Lung-Cancer Mortality among Norwegian Niobium Mine Workers.

Florida Phosphate Workers

Some U.S. phosphate ore contains uranium and radium. Workers involved in the mining and the processing of the ore might be exposed to radon and radon daughters. Two retrospective cohort studies of mortality in Florida phosphate workers have been conducted recently; each was performed because of concern raised by apparent clusters of lung cancer.

Stayner et al.⁶¹ conducted a study of 3,199 workers employed at a phosphate fertilizer plant. Seven samples were taken for radon progeny; the range was 0.00–0.02 WL. Overall respiratory-cancer mortality was not significantly increased (SMR, 113). Further analysis did not show trends of respiratory-cancer mortality with duration of employment or length of follow-up in white men. In black men, respiratory-cancer mortality was significantly increased in those with more than 20 yr of employment. However, only five cases were identified in black men, and two were in the index cluster.

In a larger study, Checkoway et al.¹⁰ examined mortality in 17,601 white and 4,722 nonwhite male employees of the Florida phosphate industry. Lung-cancer mortality was not significantly increased in either group, in comparison with rates for Florida. When mortality from lung cancer was examined in the workers considered to have potential exposure to alpha radiation, a significant excess was apparent (SMR, 1.08).

These studies do not have sufficiently detailed information on exposure for risk estimation. Individual exposures to radon progeny were not estimated, and information on cigarette smoking was not collected. Furthermore, the limited measurements that have been made indicate radon-daughter concentrations only slightly above background concentrations.

Residential Exposure

Within a building, radon-progeny concentrations are determined by the strength of the source and the rate of air exchange with the outside. Most of the radon in buildings enters from the underlying soil and building materials, although water and utility gas can also contribute radon progeny to indoor air.⁴³ A wide range of radon-daughter concentrations in dwellings has been demonstrated, with different radium concentrations in soil and building materials and different air-exchange rates largely explaining the size of the range.

Epidemiological investigations of domestic radon progeny as a risk factor for lung cancer are still preliminary. Both descriptive and analytical approaches have been used to examine the association between radon-daughter exposure in the home and lung cancer. Techniques for estimating lifetime exposure of people to radon daughters from indoor air are not yet available, and surrogates based on residence type or a few limited measurements have been used in the analytical studies. The available studies are insufficient for the development of quantitative risk estimates for associating exposure to radon progeny in the home and lung cancer.

In the descriptive studies, incidence or mortality rates for lung cancer within geographic units have been correlated with measures of exposure for inhabitants of the units. Edling ^{15 a} compared mortality rates for different Swedish counties with background gamma radiation, described as being correlated with indoor exposure to radon and its daughters. For lung-cancer mortality, the correlation coefficients were 0.46 for males and 0.55 for females. Hess and colleagues²⁰ performed a similar analysis for lung-cancer mortality during 1950–1969 in the 16 counties of Maine. Using average radon concentrations in water as the measure of exposure, they calculated correlation coefficients of 0.46 for males and 0.65 for females. In a study of 28 Iowa towns served by deep wells, lung-cancer incidence increased with the concentration of ²²⁶Ra, a possible surrogate for the radon concentration in the water.⁹ These descriptive studies, which did not consider the exposures of people to radon daughters and other agents, provided only suggestive evidence that radon progeny exposure in the home increases lung-cancer risk.

The association has been more directly tested in case-control and cohort studies. Axelson et al.⁸ conducted a case-control study in a rural area of Sweden. The investigation included 37 cases and 178 controls. Exposure to radon progeny was inferred from the characteristics of the subjects' residences at the time of death. Those who lived in stone houses were assumed to be most heavily exposed to radon daughters, and those who lived in wooden houses were assumed to be least exposed; other types of dwellings were considered to be sources of intermediate exposure. In spite of the crudeness of this exposure classification, residence in stone houses was associated with a significantly increased odds ratio, in comparison with the reference category of wooden houses (by Mantel-Haenszel method; odds ratio, 5.4; 90% confidence interval, 1.5–19). Data concerning cigarette smoking and residence history were not obtained.

Edling and Axelson ^{16 a} conducted a similar case-control study in a rural area of Sweden. The study subjects were residents of the island of Oeland who died during 1960–1978. The geological characteristics of this island were thought to result in strong differences in background radon concentrations within a small area. Inclusion in the study population required at least 30 yr of residence at the same address before death; 23 lung-cancer cases and 202 controls who died from causes other than lung cancer met this criterion. Most of the dwellings were monitored for radon daughters during 3 months of summer and I month of winter. The dwellings were also classified on the basis of structural characteristics, as in the earlier study by Axelson et al.,⁸ and cigarette-smoking information was obtained from next of kin. Lung-cancer risk was significantly associated with radon-daughter exposure, as assessed by either the measured concentration or the characteristics of the dwelling, and both crude and smoking-adjusted risk estimates were significantly increased. Logistic analysis yielded smoking-adjusted odds ratio, comparing most with least exposed, of 3.9, and the 90% confidence interval was 1.5–10.0.

Pershagen et al.⁴⁵ reported the findings of two small case-control studies in Sweden on domestic radon-daughter exposure, one drawn from a larger study in northern Sweden and the other from a twin registry. The investigators assembled each series with 30 case-control pairs, divided equally between smokers and nonsmokers. Exposure to radon was estimated from information on dwelling type; the investigators attempted to consider all residences lived in by the subjects. In the study group from northern Sweden, imputed radon exposures were significantly higher in smokers than in their smoking controls. Estimated exposures to radon progeny were similar in the nonsmoking cases and controls in the series from northern Sweden and in the smoking and nonsmoking cases and controls in the second series (selected from the twin registry).

In the United States, Simpson and Comstock⁵⁷ examined the relationship between lung-cancer incidence and housing characteristics. During a 12-yr period in Washington County, Maryland, lung-cancer incidence was not significantly affected by the type of basement construction or building materials. No measurements of radon or its daughters were made. Rather, dwelling-related variables were assumed to be surrogates for radon daughter exposure.

Summary

Cause-specific mortality risks for a number of the miner groups discussed above are listed in Table IV-16. Without exception, these studies indicate an excess probability of death due to lung cancer and, in many cases, other causes of death as well. Continued follow-up of these miner groups will provide additional information on the association of radon-daughter exposure to lung cancer and perhaps other diseases. As discussed in Chapter 2 and Appendix VII, epidemiological information that includes the smoking status of each participant is of paramount value. The committee suggests that every effort be made to collect and report such information for the studies described in this appendix.

TABLE IV-16

Cause-Specific Risks of Mortality among Miners Exposed to Radon Daughters.

References

1.

Agricola, G. De Re Metallica. Basel, 1556. New York: Dover Publications. English reprint (Hoover translation). 1950, p. 214.

2.

Archer, V. E., H. J. Magnuson, D. A. Holaday, and P. A. Lawrence. 1962. Hazards to health in uranium mining and milling. J. Occup. Med. 4:55–60. [PubMed: 13862081]

3.

Archer, V. E., J. K. Wagoner, and F. E. Lundin, Jr. 1975. Lung cancer among uranium miners in the United States. Health Phys. 25:351–371. [PubMed: 4361970]

4.

Archer, V. E., J. D. Gillam, and J. L. Wagoner. 1976. Respiratory disease mortality among uranium miners. Ann. N.Y. Acad. Sci. 271:280–293. [PubMed: 1069515]

5.

Arnstein, A. 1913. Sozialhygienische untersuchungen über die Bergleute in den Schneeberger Kobaltgruben. Wein. Arbeit. Geb. Sos. Med. 5:64–83.

6.

Axelson, O., and C. Edling. 1980. Health hazards from radon daughters in dwellings in Sweden, Pp. 79–87 in Health Implications of New Energy Technologies, R. Rom, editor; and V. Archer, editor. eds. Ann Arbor, Mich.: Ann Arbor Science Publishers, Inc.

7.

Axelson, O., and L. Sundell. 1978. Mining, lung cancer and smoking. Scan. J. Work Environ. Health 4:46–52. [PubMed: 644266]

8.

Axelson, O., C. Edling, and H. Kling. 1979. Lung cancer and residency: A case-referent study on the possible impact of exposure to radon and its daughters in dwellings. Scand. J. Work Environ. Health 5:10–15. [PubMed: 441704]

9.

Bean, J. A., P. Isaacson, R. M. A. Hahne, and J. Kohler. 1982. Drinking water and cancer incidence in Iowa. Am. J. Epidemiol. 116(6):924–932. [PubMed: 7148818]

10.

Checkoway, J., R. M. Mathew, J. L. S. Hickey, C. M. Shy, R. L. Harris, E. W. Hunt, and G. T. Waldman. 1985. Mortality among workers in the Florida phosphate industry. I. Industry-wide cause-specific mortality patterns. J. Occup. Med. 27:885–892. [PubMed: 4087054]

11.

Damber, L., and L. G. Larsson. 1982. Combined effects of mining and smoking in the causation of lung carcinoma-case-control study in northern Sweden. Acta Rad. Oncol. 21:305–313. [PubMed: 6297249]

12.

Damber, L., and L. G. Larsson. 1985. Underground mining, smoking and lung cancer: A case-control study in the iron ore municipalities in northern Sweden. J. Natl. Cancer Inst. 74(6):1207–1213. [PubMed: 3858594]

13.

deVilliers, A. J., and J. P. Windish. 1964. Lung cancer in a fluorspan mining community. Br. J. Ind. Med. 21:94–109. [PMC free article: PMC1038330] [PubMed: 14142524]

14.

deVilliers, A. J., J. P. Windish, F. de N. Brent, B. Holloywood, C. Walsh, J. W. Fisher, and W. D. Parsons. 1971. Mortality experience of the community and of the fluorspan mining employees at St. Lawrence, Newfoundland. Occup. Health Rev. 22:1–15. [PubMed: 5158833]

15.

Edling, C. 1982. Lung cancer and smoking in a group of iron ore miners. Am. J. Ind. Med. 3:191–199. [PubMed: 6291384]

15a. Edling, C., P. Combs, O. Axelson, and V. Flodin. 1982. Effects of low-dose radiation—a correlation study. Scand. J. Work Environ. Health 8(Suppl. 1):59–64. [PubMed: 7100858]

16.

Edling, C., and O. Axelson. 1983. Quantitative aspects of radon daughter exposure and lung cancer in underground miners. Br. J. Ind. Med. 40:182–187. [PMC free article: PMC1009169] [PubMed: 6830715]

16a. Edling, C., H. Kling, and O. Axelson. 1984. Radon in houses—a possible cause of lung cancer. Scand. J. Work Environ. Health 10:25–34. [PubMed: 6740274]

17.

Fox, A. J., P. Goldblatt, and L. J. Kinlen. 1981. A study of the mortality of Cornish tin miners. Br. J. Ind. Med. 38:378. [PMC free article: PMC1069292] [PubMed: 7317301]

18.

Gottlieb, L. S., and L. A. Husen. 1982. Lung cancer among Navajo uranium miners. Chest 81:449–452. [PubMed: 6279361]

19.

Harting, F. H., and W. Hesse. 1979. Der lungenkrebs, die Bergkrankheit in den Schneeberger gruben. Vjschr. Gerichtl. Med. Offentl. Gesundheitswesen 31:102–132,313–337.

20.

Hess, C. T., C. V. Weiffenbach, and S. A. Norton. 1983. Environmental radon and cancer correlations in Maine. Health Phys. 45:339–348. [PubMed: 6885433]

21.

Holaday, D. A., W. D. David, and H. N. Doyle. 1952. An Interim Report of a Health Study of the Uranium Mines and Mills by the Federal Security Agency, Public Health Service, Division of Occupational Health, and the Colorado State Department of Public Health. Washington D.C.: U.S. Public Health Service.

22.

Horaceck J., V. Placek, and J. Sevc. 1977. Histologic types of bronchogenic cancer in relation to different conditions of radiation exposure. Cancer 40:832–835. [PubMed: 196745]

23.

Hornung, R. W., and T. J. Meinhardt. 1987. Quantitative risk amassment of lung cancer in U.S. uranium miners. Health Phys. 52:417–430. [PubMed: 3032855]

24.

Hornung, R. W., and S. Samuels. 1981. Survivorship models for lung cancer mortality in uranium miners—is cumulative dose an appropriate measure of exposure? Pp. 363–368 in International Conference, Radiation Hazards in Mining: Control, Measurement, and Medical Aspects, M. Gomez, editor. , ed. New York: Society of Mining Engineers of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc.

25.

Howe, G. R., R. C. Nair, H. B. Newcombe, A. B. Miller, S. E. Frost, and J. D. Abbatt. 1986. Lung cancer mortality (1950–1980) in relation to radon daughter exposure in a cohort of workers at the Eldorado Beaverlodge uranium mine. J. Natl. Cancer 77(2):357–362. [PubMed: 3461198]

26.

Hueper, W. C. 1966. Occupational and Environmental Cancers of the Respiratory System. New York: Springer-Verlag.

27.

Jorgensen, H. S. 1973. A study of mortality from lung cancer among miners in Kiruna 1950–1970. Work Environ. Health 10:126–133.

28.

Kuns, E., J. Sevc, and V. Placek. 1978. Lung cancer mortality in uranium miners. Health Phys. 35:579–580. [PubMed: 748269]

29.

Kunz, E., J. Sevc, V. Placek, and J. Horacek. 1979. Lung cancer in man in relation to different time distributions of radiation exposure. Health Phys. 36:699. [PubMed: 224006]

30.

Lorenz, E. 1944. Radioactivity and lung cancer; a critical review of lung cancer in the miners of Schneeberg and Joachimsthal. J. Natl. Cancer Inst. 5:1–13.

31.

Ludwig, P., and E. Lorenser. 1924. Untersuchungen der Grubenluft in den Schneeberger gruben auf den Gehalt and Radiumemanation. Strahlentherapie 19:428–435.

32.

Lundin, F. D., Jr., J. W. Lloyd, E. A. Smith, V. E. Archer, and D. A. Holaday. 1969. Mortality of uranium miners in relation to radiation exposure, hardrock mining and cigarette smoking—1950 through September 1967. Health Phys. 16:571–578. [PubMed: 5795875]

33.

Lundin, F. D., Jr., J. K. Wagoner, and V. E. Archer. 1971. Radon Daughter Exposure and Respiratory Cancer, Quantitative and Temporal Aspects. Joint Monograph No. 1. Washington, D.C.: U.S. Public Health Service.

34.

Lundin, F. D., Jr., V. E. Archer, and J. K. Wagoner. 1979. An exposure-time-response model for lung cancer mortality in uranium miners—effects of radiation exposure, age and cigarette smoking. Pp. 243–264 in Proceedings of the Work Group at the Second Conference of the Society for Industrial and Applied Mathematics, N. E. Breslow, editor; and A. S. Whittemore, editor. , eds. Philadelphia: Society for Industrial and Applied Mathematics.

35.

Morrison, H. I., D. T. Wigle, H. Stocker, and A. J. deVilliers. 1981. Lung cancer mortality and radiation exposure among the Newfoundland fluorspar miners. Pp. 372–376 in International Conference on Radiation Hazards in Mining: Control, Measurements and Medical Aspects, M. Gomez, editor. , ed. New York: Society of Mining Engineers of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc.

36.

Morrison, H. I., R. M. Semenciw, Y. Mao, D. A. Corkill, A. B. Dory, A. J. deVilliers, H. Stocker, and D. T. Wigle. 1985. Lung cancer mortality and radiation exposure among the Newfoundland fluorspar miners. Pp. 354–364 in Occupational Radiation Safety in Mining, H. Stocker, editor. , ed. Proceedings of the International Conference. Toronto: Canadian Nuclear Association.

37.

Muller, J., W. C. Wheeler, J. F. Gentleman, G. Suranyi, and R. Kusiak. Pp. 359–362 in The Ontario Miners Mortality Study, General Outline and Progress Report: Radiation Hazards in Mining, M. Gomez, editor. , ed. New York: Society of Mining Engineers of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc.

38.

Muller, J., W. C. Wheeler, J. F. Gentleman, G. Suranyi, and R. A. Kusiak. 1983. Study of Mortality of Ontario Miners, 1955–1977, Part I. Toronto: Ontario Ministry of Labour, Ontario Workers Compensation Board.

39.

Muller, J., W. C. Wheeler, J. F. Gentleman, G. Suranyi, and R. Kusiak. 1985. Study of mortality of Ontario miners. Pp. 335–343 in Occupational Radiation Safety in Mining, Proceedings of the International Conference, E. Stocker, editor. , ed. Toronto: Canadian Nuclear Association.

40.

Muller, J., R. A. Kusiak, G. Suranyi, and A. C. Richie. 1986. Study of Mortality of Ontario Gold Miners, 1955–1957, Part II. Ottawa, Canada: Ontario Ministry of Labour.

41.

Nair, R. C., J. D. Abbatt, G. R. Howe, H. B. Newcombe, and S. E. Frost. 1985. Mortality experience among workers in the uranium industry. Pp. 154–164 in Occupational Radiation Safety in Mining, H. Stocker, editor. , ed. Proceedings of the International Conference. Toronto: Canadian Nuclear Association.

42.

National Research Council, Committee on the Biological Effects of Ionizing Radiations (BEIR). 1980. The Effects on Populations of Exposure to Low Levels of Ionizing Radiation. Washington, D.C.: National Academy Press. 524 pp.

43.

Nero, A. V. 1983. Airborne radionuclides and radiation in buildings: A review. Health Phys. 45:303–322. [PubMed: 6350229]

44.

Peller, S. 1989. Lung cancer among mine workers in Joachimsthal. Human Biol. 11:130–143.

45.

Pershagen, L. D., and R. Falk. 1984. Exposure to radon in dwellings and lung cancer: A pilot study. Pp. 73–78 in Indoor Air, Vol. 2. Radon, Passive Smoking, Particulates and Housing Epidemiology, B. Berglung, editor; , T. Lindvall, editor; , and J. Sundell, editor. , ed,. Stockholm: Swedish Council for Building Research.

46.

Pirchan, A., and H. Sikl. 1980. Cancer of the lung in the miners or Jachymov (Joachimsthal). Am. J. Cancer 4:681–722, 1932.

47.

Placek, V., A. Smid, J. Sevc, L. Tomasek, and P. Vernerova. 1983. Late effects at high and very low exposure levels of the radon daughters. In Proceedings of the 7th International Congress of Radiation Research. Amsterdam: Martinus Nijhoff Publishers.

48.

Radford, E. P., and K. G. St. Clair Renard. 1984. Lung cancer in Swedish iron miners exposed to low doses of radon daughters. N. Engl. J. Med. 310(23):1485–1494. [PubMed: 6325913]

49.

Rostoski, O., E. Saupe, and G. Schmorl. 1926. Die bergkrankheit der Erzbergleute in Schneeberg in Sachsen (''Schneeberger Lungenkrebs"). Z. Krebforsch. 23:360–384.

50.

Samet, J. M., C. R. Key, D. M. Kutvirt, and C. L. Wiggins. 1980. Respiratory disease mortality in New Mexico's, American Indians and Hispanics. Am. J. Public Health 70:492–497. [PMC free article: PMC1619430] [PubMed: 7377419]

51.

Samet, J. M., D. M. Kutvirt, R. J. Waxweiler, and C. R. Key. 1984. Uranium mining and lung cancer in Navajo men. N. Engl. J. Med. 310:1481–1484. [PubMed: 6717538]

52.

Schiager, K. J., and L. W. Hersloff. 1984. Review of radon daughter exposure measurements in U.S. uranium mines—past and present. In Assessment of the Scientific Basis for Existing Federal Limitations on Radiation Exposure to Underground Uranium Miners, Appendix F. Toronto: SENES Consultants Ltd.

53.

Seltser, R. 1965. Lung cancer and uranium mining: A critique. Arch. Environ. Health 10:923–935. [PubMed: 14308832]

54.

Sevc, J., and V. Placek. 1973. Radiation induced lung cancer: Relation between lung cancer and long-term exposure to radon daughters. Pp. 305–310 in Proceedings of the 6th Conference on Radiation Hygiene. Jasna pod Chopkom, CSSR.

55.

Sevc, J., V. Placek, and J. Jerabek. 1971. Lung cancer risk in relation to long-term radiation exposure in uranium miners. Pp. 315–326 in Proceedings of the 4th Conference on Radiation Hygiene. Parezska, Lhota, CSSR.

56.

Sevc, J., E. Kunz, and V. Placek. 1976. Lung cancer mortality in uranium miners and long-term exposure to radon daughter products. Health Phys. 30:433–437. [PubMed: 955899]

57.

Simpson, S. G., and G. W. Comstock. 1983. Lung cancer and housing characteristics. Arch. Environ. Health 38:248–251. [PubMed: 6615007]

58.

Staid, A., J. Sevc, V. Placek, and E. Kunz 1983. Lung Cancer in Exposed Human Populations and Dose/Effect Relationship. Presented at the 7th International Congress on Radiation Research. Amsterdam: Martinus Nijhoff Publishers.

59.

Snihs, J. O. 1974. The approach to radon problems in non-radium uranium miners in Sweden. Proceedings of the 3rd International Congress of the International Radiation Protection Association. CONF-730907. Oak Ridge, Tenn.: U.S. Atomic Energy Commission.

60.

Solli, M., A. Andersen, E. Straden, and S Langand. 1985. Cancer incidence among workers exposed to radon and thoron daughters at a niobium mine. Scan. J. Work Environ. Health 11:7–13. [PubMed: 2986282]

61.

Stayner, L. T., T. Meinhardt, R. Lemen, D. Bayliss, R. Herrick, G. R. Reeve, A. B. Smith, and W. Halperin. 1985. A restrospective cohort mortality study of a phosphate fertilizer production facility. Arch. Environ. Health 40:133–138. [PubMed: 2992396]

62.

Sun, S., X. Yang, Y. Lan, M. Xionyu, L. Shengen, and Y. Zhanyun. 1984. Latent period and temporal aspects of lung cancer among miners. Radiat. Prot. 4(5) (English translation).

63.

Teleky, L. 1937. Occupational cancer of the lung. J. Ind. Hyg. Toxicol. 2:73–85.

64.

Tirmarche, M., J. Brenot, J. Piechowski, J. Chameaud, and J. Pradel. 1985. The present state of an epidemiological study of uranium miners in France. Pp. 344–349 in Proceedings of the International Conference, Occupational Radiation Safety in Mining, Vol. 1. Canadian Nuclear Association, Toronto, Ontario, Canada.

65.

Wagoner, J. K., V. E. Archer, B. E. Carroll. 1964. Cancer mortality patterns among U.S. uranium miners and millers, 1950 through 1962. J. Natl. Cancer Inst. 32:787–801.

66.

Wagoner, J. K., V. E. Archer, F. E. Lundin, D. A. Holaday, and J. W. Lloyd. 1965. Radiation as the cause of lung cancer among uranium miners. N. Engl. J. Med. 273:181–188. [PubMed: 14306333]

67.

Waxweiler, R. J., R. J. Roscoe, V. E. Archer, M. J. Thun, J. K. Wagoner, and F. E. Lundin, Jr. 1981. Mortality follow-up through 1977 of the white underground uranium miners cohort examined by the United States Public Health Service. Pp. 823–830 in International Conference, Radiation Hazards in Mining: Control, Measurement and Medical Aspects, M. Gomez, editor. , ed. New York: Society of Mining Engineers of the American Institute of Mining, Metallurgical, and Petroleum Engineers, Inc.

68.

Wang, X., X. Huang, et al. 1984. Radon and miners lung cancer, Zhonghua Fangshe Yixue yu Fanghu Zazhi 4:10–14 (English translation).

69.

Whittemore, A. S., and A. McMillan. 1983. Lung cancer mortality among U.S. uranium miners: A reappraisal. J. Natl. Cancer Inst. 71(3):489–499. [PubMed: 6577225]

Footnotes

*

A recent review of these calculations submitted to the committee, ''Beaverlodge Working Level Month Calculations," Draft 4 by S. E. Frost, has suggested possible underestimation of exposures. New calculations for some years indicate that the choice of the median WL value and the method used to determine equilibrium factors might have resulted in bias toward low WLM estimates. For the years reviewed, use of the arithmetic mean, rather than the median, increases the annual WL value.

*

A report submitted to the committee, "Comments to the U.S. Mine Safety Health Administration for the American Mining Congress," by Swent and Chambers, questioned some assumptions underlying the historical reconstruction of the exposures for the years before measurements were taken. Because Radford and Renard⁴⁸ assumed that water was the major source of radon in the mines and its strength was constant, Swent and Chambers argued that changing mining practices might alter radon influx into a mine, even in the face of a constant concentration in water. In addition, changing ventilation practices over the years could have influenced exposures. In discussing potential bias in the exposure estimates used by Radford and Renard,⁴⁸ Swent and Chambers suggested that the direction of changes in exposures would have been downward. If the exposures were, in fact, underestimated, the estimated risk coefficients would exaggerate the actual risk.