Cancer incidence among workers exposed to radon and thoron daughters at a niobium mine

S. Cancer incidence among workers exposed to radon and thoron daughters at a niobium mine. Scand J Work Environ Health II (1985) 7-13. The aim of this study was to investigate the incidence of cancer among 318 male employees of a niobium mining company which was only operated between 1951 and 1965. Many of the workers, especially underground miners, were exposedto the daughters of radon and thoron and also to thorium. The accu mulated doses to the workers from short-lived radon and thoron daughters in the mine atmosphere were assessed to be relatively low; up to 300working-level months. During the follow-up period 1953-1981, 24 new cases of cancer were observed compared to an expected number of 22.8. Twelve cases of lung cancer had occurred versus 3.0 expected. Among the 77 miners, 9 cases of lung cancer were observed against0.8expected. Associationsbetweenthe occurrenceof lungcancerand exposureto alpha radiation and smoking werefound. For the radon and thoron daughter exposure, about 50excess casesper million person-years at risk per working-level month were observed.

underground mine was exca vated to a depth of 135 m belo w ground level. This mine became the main provider of raw material until the works were closed down in 1965.
There were five main act ivities in the mining works . They were mining; washing, crushing and sorting; ferroniobium production; maintenance work, both on the surface and underground; and office work.
In addition to 0.3-0.5 % niobium ox ide, the mined rocks contained 238U (uranium) and 232T h (thorium). The concentration of uranium was 0.3-2 ppm and that of thorium was 50-300 ppm (21). The average concentration of 232Th in the rocks was 2-10 ppm. In the radioactive decay series of 238U and 232Th, radon and thoron , respectively, are found .

Study design
The names of the employees were obtained from three different sources, ie, the company's payroll, two reco rds at the local office of the National Insurance Administration , and a letter of congratulation sent by the miners to one of their colleagues. From the se three sources it was possible to reconstruct a complete list of employees in every year with the exception of the construction years 1951 and 1952 and the production years [1955][1956]. Three hundred and twenty-two males were identified. From an analysis of the employment numbers in the company's payroll, it can be assumed that about 40 men have not been traced for these four year s. It can also be inferred tha t all these men had been employed for a rather short time , ie, from a few day s to less than two years. The list of employees ha ving worked for two 7 years or more is complete . Four more workers had to be excluded because of incomplete data . The study group, therefore, consisted of 318 men, born between 1889 and 1947.
Information on the complete occupational history, work at the mine, the type of work before and after employment at the mine, and asbestos exposure was obtained for each individual. Information on tobacco usage was also recorded. The information was obtained in the following way: 26 men were interviewed personally; 195 employees filled out a questionnaire, and about 60 070 of the 195 were afterwards contacted by telephone for supplementary information; 16 were interviewed by telephone alone . Three former workers could not be traced. No one refused to give information.
For all the deceased members of the cohort except one , information on exposure was obtained from close relatives and/or colleagues.
The Norwegian Cancer Registry has acquired information on all new cases of cancer in the country since 1953. The registration is based on compulsory reporting from all the hospitals and pathological laboratories. The material from the mine was matched against that of the Cancer Registry with the aid of the nat ional birth numbers.
All members of the study group were observed from the beginning of 1953 or, in the case of later first employment. from the time they were appointed at the mine. They were observed until the end of 1981 or to the time of death. The study was based on a comparison of obser ved and expected total mortality and incidence of cancer, the method described by Pedersen et al (18) and Langard & Nor seth (13) being used. This method includes calculation of the risk of getting cancer for each person separately; it involves the use of age-specific incidence rate s of cancer in the Norwegian male population (6,7).
It was assumed that the observed numbers of cancer cases followed the Poisson distribution with the expected numbers as the modes. Poisson distributions were used for significance testing. The test used has been described by Axelson (3).
After the identity of the cancer cases was disclosed, the medical files of the lung cancer cases were obtained , with the exception of one file which could not be found .

Exposure history
The cohort members were mainly recruited from industry and agriculture. Some were school dropouts . The 77 underground miners in the cohort were mainly recruited from the building and construction industry, but about one-fourth of them also had previous experience in mining and tunnel work.
The employment time and job description at the mine could be found in the company records for almost all employees in the study group. The information was confirmed and supplemented by the 8 questionnaires and the interviews. For the workers who had different jobs at the mine, the occupation in which the worker spent the longest time was used to describe his job.
The work time in months at the different sites in the mining works was assessed from a combination of the company records and the recollection of the workers themselves or their colleagues.

Assessment of exposure
Asbestos was rarely used at the mine, and only in the repa ir of machinery . An assessment of exposure to tobacco and radon and thoron daughters was attempted.
Tobacco. Information about the degree and type of tobacco usage was obtained from all but 5 of the 318 employees. For the deceased members of the cohort, the information was not so exact, but an estimate was made of how many years they had smoked, the type of smoking (ie, pipe, cigarettes or both), and quantity (either grams /week or cigarettes /day). Based on these data, a lifetime dose in grams was calculated for each of them up to the end of 1981.
The cohort was grouped into the following three categories: never or light smokers (up to a lifetime dose of 20 000 g), medium smokers (from 20 000 g up to 200 000 g), and heavy smokers (more than a 2000oo-g lifetime dose).
Radon and thoron daughters. The National Institute of Radiation Hygiene performed measurements of the alpha activity in the air in the underground mine in 1959 (unpublished). From the basic data of these measurements, the short-lived alpha activity concentrations in the mine air were assessed in terms of work ing level (WL)4 (table I).
The alpha activity was also measured in 1959 in the ferroniobium factory and in the washing plant. The activity of the radon daughters was found to be less than 0.02 WL in both places. Thoron daughters were not measured.
The accumulated dose of alpha radiation to the individual employee was calculated separately on the basis of information about the underground worktime in month s and the mean levels shown in table I. For all other sites in the works the radiation level was considered to be zero.
The International Commi ssion on Radiological Protection (ICRP) recommends a conversion factor from working-level months (WLM)4 to an effective dose equivalent of 10 mSv/WLM for radon daughters and 3.4 mSv/WLM for thoron daughters (10). For this reason one cannot simply add WLM from thoron daughters to WLM from radon daughters to get a total exposure level. The exposure must be weighed according to the aforementioned conversion factors . This procedure was carried out when the total dose of inhaled alpha activity was calculated for the workers. These dose assessments made it possible, with reasonable confidence, to subdivide the cohort into five exposure categories.
The employees' place of residence at the time they were employed at the mine was recorded . No measurements or assessments of radiation in the houses of the employees were made.
All employment classifications, exposure characterizations, and subdivision into categories were done before the cancer cases were identified .

Results
The numbers of deaths and cancer cases in the population exposed to radiation and other hazardous factors are presented as ratios between expected and observed numbers of deaths and cancer cases. The population was followed from 1953 to the end of 1981. Table 2 presents the overall mortality and incidence of cancer in the whole cohort and in different occupati ons. During the follow-up period 78 deaths had occurred versus 67.86 expected. The total num-ber of observed cancer cases was 24 against 22.8 expected.
The 24 cancer cases were distributed as follows: 12 in the lung, 2 in the prostate, 2 in the large intestine, and 2 were leukemias . Furthermore, there was one in each of the following locations: the hypopharynx, the esophagus, the stomach, the pancreas, the kidney, and the central nervous system.
There were two cases of leukemia versus 0.7 expected (not presented in the tables). Both of these workers had only been employed in the washing plant and were not characterized as radiation exposed. One of them had worked mainly at the stone mill for nine years and had been exposed to dust the whole time. The other had been employed for only eight months.
The distribution of the lung cancer cases among the different occupations is presented in table 3. Nine cases of lung cancer were observed in the 77 millers, against 0.81 expected.
The expected number of lung cancer cases in the whole study group , based on the risk in the average  " p < 0.05, "" P < 0.01, """ P < 0.001.  (22). The smoking habits of the study group at the end of 1964 can therefore be compared with the results of this survey.
The survey showed that the percentage of current smokers and nonsmokers was 60 and 40, respectively, for the male population of Southeast Norway. In the present study group the percentage was 75 and 25, and as high as 85 and 15, respectively, for the underground miners. The relative distribution of heavy, medium, and light smokers in the study group was approximately the same as in the survey. Table 5 shows the distribution of lung cancers in five categories of assessed cumulated alpha activity exposure in working-level months. The table shows a dose-related increase in the lung cancer rate. Table 6 shows the 12 lung cancer cases with relevant exposure factors. All of these workers started work at the mine in middle age; only two were younger than 40 years. Three of them lived in the Fen area during the work period. The mean age of diagnosis among the 12 cases was 67.3 years.
Radiology reports in the 11 hospital records showed that seven of the cancers were centrally situated, and four were in the periphery of the lung. The pathology reports showed that eight of the cancers were probably small squamous cell carcinoma, and three were probably small cell carcinoma.

Discussion
The present study shows an association between lung cancer and cumulative alpha radiation exposure and also between lung cancer and smoking. However, some confounding factors have to be considered.
About 40 short-time workers employed before 1957 could not be identified. They had all been employed less than two years. Because of the limited exposures which would have been possible during a maximum of two years, all these workers would have belonged to the lower two categories for alpha radiation exposure. The missing persons could not, however, affect the results in the higher three exposure categories for radiation.
The reliability of the exposure data can be questioned. The assessment of the alpha radiation exposures were based on insufficient measurements, both in number and quality. Radon concentrations in mine air are known to show large spatial and temporal variations. The measurements were only per-formed during a 2-d period at a few sites during the winter. During winter, the ventilation system was only used periodically. The temporal variations in the radon concentrations thus were probably large.
The measurement technique was also rather crude, and the uncertainties of the calibration procedure are not easy to assess. The dust concentration in the air was also very high at the time of the measurements, and the self-absorption in the filters could have been significant.
After the physicist who performed these measurements was consulted, the data on the alpha radiation exposure were reviewed. It then ensued that the minimum measurements in table 1 were made after a trial period of exceptionally good ventilation, which was not typical for the general work conditions in the mine.
It is therefore assumed that a mean level of 2 WL radon daughters in the mine, in table 1, is a more correct assessment than 1 WL. The calculated doses in tables 5 and 6 should therefore also be increased roughly by a factor of 2 to give a more correct appraisal.
Despite these limitations, we consider that major errors in the exposure characterization are unlikely to have occurred.
Other exposure factors may also have been of importance. The employees' place of residence at the time they worked at the mine was recorded, but no measurements of radiation in their homes were carried out. Measurements in some houses (21) in the area where the mine was situated have shown an exposure of up to 1.3 WLM/year. Only 3 out of the 12 workers with lung cancer lived in the area of the mine at that time. The radiation in the homes, therefore, seems to be of limited importance.
The thorium content of the rocks might also have contributed to the radiation dose to the lungs. The critical tissues, however, are bone surfaces (9). As no bone cancers have appeared, there are so far no indications that this exposure has been of major importance.
It is also possible that other carcinogens may exist in the rocks mined. The mineral fiber content in one rock specimen from the site of the mine has been analyzed. Fibers compatible with actinolite were found, but these have not previously been associated with an increased risk of lung cancer. As it has not been possible to make further geochemical analyses of the rocks to search for possible carcinogenic elements, previous studies have been consulted (5,14,15), but with negative results. A few of the workers were also exposed to diesel combustion exhaust, but, as the main mining activity was in the deep mine where the equipment was electrically driven, this factor seems to be of very little importance. The miners may occasionally have been heavily exposed to blasting gases, but it is difficult to assess the importance of this factor further.
There were more smokers in the study group than in the reference population, and especially in the subcohort of miners, even though the distribution of light, moderate, and heavy smokers was identical. In accordance with estimations performed by Axelson (2), it is not likely that the smoking habits of the study cohort have contributed to a great extent to the increase in lung cancer cases.
It might have been more appropriate to use a local reference cohort rather than the national rates. The lung cancer incidence, in the county of Telemark where the mine was located, is about 90 0J0 of the national rates (6); therefore the national rates were used.
The longest follow-up period in this group was 28 years, and the shortest was 16 years. As cancers may occur after a considerably longer latency period, the numbers of the cancer cases are only preliminary. The assumed latency period for cancer development has not been subtracted from the calculation of the expected figures in this study; therefore the expected figures may have been overestimated (12).
When the factors to which the workers with lung cancer may have been exposed are considered (table  6), it can be assumed that for some cases the occupational exposure during previous employment may have been of importance. Three workers had been exposed to asbestos. Some of the cases in table 6 had previous exposures to radon and thoron daughters in underground work. Case 3 worked as a blacksmith in an iron mine for nine years (1917)(1918)(1919)(1920)(1921)(1922)(1923)(1924)(1925)(1926). Recent measurements in this mine suggest that the thoron daughter concentration was about 3 WL and the radon daughter concentration about 0.1 WL (21). It is assumed that the blacksmith spent about 30 lifo of his workhours in this mine, and this figure indicates an exposure of about 30 WLM. This is significant exposure when compared to his exposure in the niobium mine.
Case 12 had worked 20 years in a molybdenum mine. The radon daughter concentration in that mine has been assessed to be about 0.2 WL (17). Therefore the case involved a previous exposure of nearly 50 WLM.
The other cases with earlier underground exposures (cases 4, 6, and 8) were exposed in tunnels for limited periods. The radon daughter concentrations in tunnels are normally lower than in mines, and, in view of the relatively short exposure times, these exposures are hardly significant. Table I indicates that the thoron daughter exposure to the workers at the mine was relatively low. After inhalation of thoron daughters, tissues at bone surfaces and in the kidneys are considered at greatest risk, together with lung tissues. This situation is in contrast to the inhalation of radon daughters, where only the lung is considered at greatest risk (11). In the present study group, no bone cancer and only one cancer of the kidney was found. This finding suggests that the thoron daughters seem to have contributed relatively little to the radiation hazard among the workers.
A commonly applied method for calculating risk among miners is to calculate the excess cases per million person-years (pY) at risk per WLM. After correction for the assumed underestimation of the alpha radiation level in the mine, the excess lung cancers for the exposed workers in the study group appears to be 50 cases' 10-6 Py-I • WLM-l. ICCR publication 32 (10) regards 5-15 cases' 10-6 py-l. WLM-1 as the most probable range averaged over all age periods during occupational work. The risk, however, seems to be rather age-dependent, and the BEIR III report (8) concludes that for persons with age of diagnosis over 65 years a likely risk factor is about 50 cases' 10-6 py-l . WLM-1 • In the present material the mean age at diagnosis was 67.3 years.
The study has demonstrated an increase in the lung cancer incidence related to both the intensity of smoking and the alpha radiation dose.