Mortality among pyrite miners with low-level exposure to radon daughters.

O. Mortality among pyrite miners with low-level exposure to radon daugh ters. Scand J Work Environ Health 14 (1988) 280-285. A cohort mortality study was conducted with regard to a pyrite mine located in central Italy. Exposure to radon ranged from 0.12 to 0.36 worki ng levels (WL) in the work areas; most measurements were around 0.2 WL. The concentration of free silica in the dust was less than 2 %. The cohort was determined from company files and included I 899 sub jects. Mortality was studied for the years 1965-1983. The loss to follow-up was less than 2 010. The stan dardized mortality ratio for all causes and all neoplasms was 97 and 107, respectively. That for lung cancer and for nonmalignant respiratory diseases was 131 (95 % confidence interval 97-175) and 173 (95 % confidence interval 135-231), respectively. It was estimated that the extra cases of lung cancer attributa ble to radon daughters numbered 13 per 10' person-years and working level month in the whole cohort and 21.3 per 10 6 person-years in the subcohort with 10-25 years of exposu re.

K ey terms: cohort study, iron ore mining, lung cance r.
The occurrence of respiratory cancer among miners has been studied in several European countries, the United States, the Soviet Union , Canada, and South Africa , both in uranium and other mine s. Th e inhalation of radon daughters has been con sidered the main risk factor for respiratory cancer, even if th e authors ha ve not always evalu ated other min eral carcinogen s, such as asbe stos, nickel or arsenic, in their studies.
This report considers underground iron ore miners with a low level of expo sure to rado n daughters, and the intention is to pro vide further data for the estimation of risk in the low-dose range. Iron ore miners have been subject to several investigations already, but the importance of radon dau ghter exposure in these mines has not been immed iately clear. Faulds & Stewart (9) noted a high proportion of deaths from lung cancer among hematite miner s in West Cumberland; Boyd et al (4) extended the study and confirmed the finding s. A high proportion of iron ore miners among lung cancer pati ents in Lorraine was reported by Rous sel et al (20), and subseq uently confirmed by Antoine et al (I) and Pham et al (J 8). An excess of lung cancer has been reported among iron min ers from Czechoslovakia (II) . An excess of lung cancer amon g iron ore miners in no rthern Sweden has been detected both in case-referent studies (7, )5) and by a cohort approach (13

280
A series of recent reports ha s shed new light on the issue of lung cancer among iron ore miner s. It ind icates the importanc e o f radon dau ghter exposure in th is context , since lun g cancer seems to be mainl y related to the occurrence of ra do n daughters in the mine atm osphere. Hence, for several iron ore mines, th ere are now estimates available with rega rd to th e excess risk of lung cancer in relation to radon dau ghter exposure. For example, Radford & St Clair Renard (19) found an excess of 19 lung cancer cases per working level month and 10 6 person-years in a miner population without any age restrictions, and Edling & Axelson (8) found 34.8 cases per working level month and 10 6 person-years as an average for persons 50 or more years of age [a working level (WL) is defined as any combinat ion of short-lived radon daughters in I I of air that results in an ultimate emission o f 1.3 . 10 5 MeV of potential alpha energy; a work ing level mon th (WLM ) is equal to an exposure time of 170 h]. On the other hand , a stud y o f hematite miner s in Minnesot a did not pro vide evidence of an y increased risk of lung ca ncer in th e ab sence of sign ificant rad on exposure (16). Nor did a cohort study conducted in Norway (17) among magnetite miners expo sed to a concentration of 0.10-0.15 WL detect any increased risk of lung ca ncer, but the study was limit ed in size, with a relatively low po wer of detecting sma ll risks. It ca n thu s be stated that th e recent evidence shows a con sisten t excess of lung cancer among und erground iron miners in relati on to radon daughter exposure. Cor respondingly, it might also be noted that iron oxide has not been found to be carcinogenic in humans in another occupa tional situation (3). Experiment al studi es reviewed by the International Agency for Research on Cancer (12) suggest tha t ferric oxide lacks any carcinogenic activity, while there is inadequate evidence of carcinogenicity for hematite .
The miners studied in the present investigat ion have extracted pyri te (FeS 2 ) in a mine located in the provi nce of Grosseto (Tuscany, Italy). The radon daughter levels were estimated to have been about 0.2 W L for a long per iod of time, and thereby this study co ntri butes additional information about the risk at low expos ure levels.

Subjects and methods
The mine T he mine in which the study took place was active from 1936 to 1983; the average work force was about 1 200 miners , decreasi ng to about 400 in more recent years . The wo rk to ok place at four different levels, the deepest o ne at about 350 m. Forced venti lat ion provided an airflow rate of 60-80 m v s. Accordin g to measurements performed by the company, the excavated material contained iron (37-44 0/0), sulfur (38-48 %) , and free silica (1-2 % ). T he ore also contained microquantities of magnesium oxide (0.2-1 %), aluminum oxide (0.1-0.4 % ), copper (0.003 %) , zinc and lead (0.010 % ), arsenic (0.0 18-0.80 % ), co ba lt (0.0363 %), and nicke l (0.0029 % ) (5). T he urani um in the surrounding rocks ran ged from 0.0006 % in th e schists to 0.0008°/0 in the qu artzy seams to 0.0013 % in the anh ydrite; these values are somewhat higher than those usually detected in other Ita lian non uranium mine s (6).
About 2 750 t of mineral per da y was delivered to plants which used it in the production of sulfuric acid and "pellets" for the steel industry. Wet drilling was introduced around 1958; the use of diese l engines started around 1974. Smoking was prohibited in the mine in 1952. En viro nmental monitoring of mineral dust was performed from the ea rly 1970s on. Table I summarizes the results of t he air borne dust sampling from 1970 to 1976 (op tica l microscopy method) and from 1977 to 1983 (X-ray diffraction) . It is remarkable that the highest values were reco rded duri ng ore removing in the year s following the entry of diesel engines (1975)(1976). T he lower variability of the data collected after 1977 is presumably due both to the preventive measures and to the new met hod of ana lysis initiated . Levels of airborne dust sometimes exceeded the standard of the American Co nference of Governmental Industrial Hygienists (ACGl H) in the ea rlier peri od , but not in recent years (22). No information on expo sure associated with dry drilli ng was available.
The concentrations of carbon oxide, nitrogen oxide , and sulfur oxide gases in the environment were found to comply with the ACGIH standards. The nitroglycerin and nitroglycollevels, after explosions, showed peak values of up to 0. 20 ppm [mean 0.07 (SD 0.09) ppm, ACGlH time-w eighted average 0.05 ppm] .
Measur ements of exposure to pot ential alpha energy of short -lived 222ra don daughters were carried out in 20 workplaces in 1979 by the Commission of the Institute of Nuclear Protection and Safety of France using individua l active dosimeters (6). The results are summarized in tab le 2.
The radon gas con centration was evaluated in 1981-1984 in various areas of the mine by a team of the National Agency for Energy (ENEA) (23). Individual pa ssive dosimeters were used , and the result s are sho wn in table 3.
The equilibrium factor ranged from 0.50 to 0.80 at the dead end of the drift. The data obtained from passive and act ive do simeters agr eed well with mea surements performed in areas without forced venti lation (23). The authors stated tha t the measurement s carried out for the upcast air may be considered representative of the average radon concentration in several 10cations.
In conclu sion, radon exposure ranged from 0.08 to 0.40 WL; most measurements were about 0.2 WL (unpu blished data from the Ministr y of Industry, 1985), and this level has been considered to be repre sentative of past exposure also .  The information on the vital status and the causes of death of the subjects was obtained from the files of the municipalities where the miners lived and/or died. The loss to follow-up was less than 2 070. Lost subjects were assumed to be alive at the end of the follow-up. The causes of death were independently coded by three physicians, according to the eighth revision of the International Classification of Diseases and Causes of Death (lCD), since the death certificates available in the municipalities do not report the ICD codes (which are available at the Central Institute of Statistics, where data are processed anonymously).
The analysis of the data was performed with the lifetable analysis system of the National Institute for Occupational Safety and Health in the United States. The expected numbers were based on Italian mortality rates, specific for cause, sex, age, and calendar period. A description of the cohort in terms of the distribution of person-years by duration of exposure and calendar time is shown in table 4. Table 4. Distribution of the person-years at risk by calendar time and age.

The cohort
The cohort was determined from the company files and included all the subjects active on 1 January 1950 and still alive on 1 January 1965, along with those subsequently employed until 1 January 1983. The years prior to 1950 were not included because of the low reliability of data of wartime years. The collection of mortality figures was started in 1965 due to considerations of the quality of the certification of causes of death in the area of interest.
Admission to the cohort was restricted to underground workers. Altogether 1 899 male subjects fulfilled the admissibility criteria; their average duration of employment was 13 years. Their mortality experience was studied from 1 January 1965 through 31 December 1983.

Results
The mortality experience of the cohort is shown in table 5. Altogether 463 deaths occurred versus 474.6 expected. The observed deficit in mortality from circulatory diseases (standard mortality ratio 74) is likely to depend on selection to work and other circumstances and is usually referred to as the healthy worker effect. This phenomenon is generally seen in this type of study. No excess of ischemic heart disease or cerebrovascular disease was detected.
The statistically significant excess of nonmalignant respiratory diseases has been presented in greater detail in table 6. Virtually all the extra cases, defined as the difference between the observed and expected figures, occurred after 10 years of latency and 10 years of exposure. Out of 67 cases of nonmalignant respiratory disease, 43 were recorded as silicosis and three as massive silicosis, but no death was referred to as silicotuberculosis.
Among the subjects whose cause of death was silicosis 14 070 had worked up to 10 years, 63 % had worked for 11-20 years, and 23 % had worked for more than 20 years. Sixty-three percent of the deaths occurred in the first 15 years of the study (1965)(1966)(1967)(1968)(1969)(1970)(1971)(1972)(1973)(1974)(1975)(1976)(1977)(1978)(1979), while 37 % took place in the years 1980-1983. Thus the mortality due to silicosis seems to be still increasing. The prevalence of pneumoconiosis among the active miners, in 1983, was 4 %, based on 12 subjects with an average duration of employment of 25 years (22). The mortality figures may overestimate the occurrence of silicosis due to the fact that compensation is awarded to widows and heirs if the death is attributable to silicosis, even in the absence of autopsy.
For most of the sites, cancer mortality in the cohort was in agreement with the national figures (table 5). Excesses of intestinal cancer and lymphohematopoietic  malignancies were detected, but du e to the smallness of the figures and the lack of a priori hypoth eses, no definite conclusions can be drawn . It may be noted, however, that five cases in the latter gro up were lymphosarcoma s and reticulosarcomas (standardized mortalit y ratio 320).
The excess of lung cancer (sta ndardized mortality ratio 131) is presented in more detail in table 7. It may be noted that the extra cases usuall y occurred aft er IO years of latency, especially among subjects with 10-25 years of exposure. The ir mean age at death was 61 years.
Th e concentration of radon daughter s in the mine has been estimated to be about 0 .2 WL, leading to an accumulated exposure of about 2.2 WLM per year. On the basis of the number of cases and person-years in the different expo sure categories, it is possible to obtain a risk estimate of 13.0 cases per 10 6 person-years and working level month. Since most of the extra cases occurred in the exposure category 10-25 years, an estimate restricted to this portion of the cohort would increase to 21.3 cases per 10 6 person-years and working level month .

Discussion
The present study was based on the traditional cohort approach in occupational epidemiology, involving wellknown limitations, especially with regard to phenomena of incomparability between a working population and the general population as providing the rates underlying the expected numbers.
It can be seen that the overall mortality (standardized mortality ratio 92) falls in a range which is relatively typical in occupational mortality studies. The deficit of circulatory deaths is consistent with preemployment selection concerning the fitness of miners with respect to underground work. In this framework the excess of respiratory diseases, both malignant and nonmalignant, appear to be work-related, especially since the extra cases occurred when an adequate latency time had elapsed since first exposure.
A second major limitation of cohort studies of this kind is represented by the absence of the control of smoking, which is somewhat disturbing with regard to lung cancer. In this particular investigation, the smoking habits of all the cohort members were not known, but two comments can be made on the basis of some recent views on this issue (24). First, a crosssectional study carried out with respect to workers active in this mine (22) showed that the proportion of smokers was compatible with that of a random sample of the male population in this area, as shown in table 8. Secondly, several smoking-related causes of death, like larynx and bladder cancer, as well as circulatory diseases, did not occur at an increased rate in this cohort. It may be noted also that cigarette smoking and alcohol consumption tend to be associated, and such causes of death as pharyngeal and esophageal cancer and cirrhosis of the liver did not occur in excess in the cohort.
The current interest in epidemiologic studies on lung cancer incidence among miners is mainly focused on the dose-response relationship to radon daughter exposure. Research conducted in the uranium mines in the United States (2, 21), Canada (10), and Czechoslovakia (14) have considered cumulative exposures ranging from 1 to 1 000 WLM, or even 2000 WLM.
Radford & St Clair Renard (19) estimated that a figure of 18-21 extra cases was derivable from the paper by Kunz et al (14). Howe et al (10), in a recent study, provided an estimate of 21 extra cases, all estimate" without any specification of the age range.
Conventional metal mines generally show cumulative exposure ranging from I to 100-200 WLM.
In the present study, it can be estimated that the highest cumulative exposures were not more than about 80-90 WLM, and the mean cumulative exposure was about 30 WLM. More important however is that, on the basis of these data, it does not seem necessary to assume any substantially stronger effect per dose obtained from low-level exposures than from higher levels. The risk estimates from the present study, though 284 based on estimates of exposure affected by some approximation, appear to be in the range of findings of the aforementioned studies from various countries. Other potential risk factors for lung cancer (silica, arsenic) were present in the mine in low concentrations, whereas diesel equipment was introduced in relatively recent times. Thus the moderate observed excess of lung cancer can, with reasonable confidence, be referred to the exposure of the cohort to radon daughters in the low exposure range.