EPIDEMIOLOGY OF THERADON RISK IN FRANCE

The radon risk constitutes a recent concern which shows a correlation with the appearance of certain forms of cancer. We wish to thank Ms Margot TIRMARCHE, a specialist on this question, for having discussed this topical issue on behalf of I.E.C.

Introduction

Radon, a radioactive gas of natural origin, is produced by the disintegration of the uranium and radium located in the earth's crust. It constitutes the chief source of natural exposure to ionising radiation. Radon exposure is omnipresent for the general public, but with variable concentrations, because radon mainly comes from granitic and volcanic subsoils as well as from certain construction materials.

In 1987, the International Centre for Cancer Research (ICCR) declared radon to be carcinogenic for humans (lung cancer) based on the results of experimental animal studies and epidemiological studies among uranium miners. Results available at the time were related to very high exposure levels, and the extrapolation of this risk for the general public exposed to far lower levels in a domestic environment raised numerous debates.

Today, the data for the annual exposure of uranium miners are closer to the concentration levels inhaled in certain houses, and current discussions bear more particularly on the interaction of radon with other components present in the mine and absent from housing, which might interact when a cancer develops, or on the possibility of a multiplicative action between radon and tobacco.

Inside the most exposed housing (more than 1000 Bq/m 3 ) 1 , annual exposure levels are close to or higher than the exposure limits accepted in professional environments, and the results of the study of cohorts of uranium miners show an excess of lung cancer for these concentrations cumulated over 15-20 years of exposure. Below 1000 Bq/m3 the results currently available relating to environmental exposures are not always consistent and some uncertainties remain.

The respective influences of the duration and the intensity of exposure are as yet not well known. It is not certain that an exposure to 400 Bq/m 3 for 10 years presents the same risk as an exposure to 4000 Bq/m 3 during one year. Yet the importance and the urgency of the prevention message to be passed on to the public depend on the scale of the risk for a given level of exposure, taking into account age, gender, tobacco habits and other exposure factors.

Human exposure to radon and its descendants

The concentration of radon in the air is higher in underground galleries than in houses and depends on ventilation and airing conditions as well as on the life-style of the occupants. The respiratory tract is radon's chief way of penetration into the human organism. Following inhalation, radon is rapidly exhaled because it has little affinity with the biological milieu, particularly the lungs. On the contrary, its radioactive descendants are deposited all along the pulmonary aerial tracts according to a distribution that is related to their granulometry. The chief descendants have a radioactive period of the order of a few minutes, which limits their  action to the lung tissues close to the deposit site, particularly the cells of the bronchial epithelium. Irradiation by these descendants is of the alpha type, that is to say with high energy deposits.

Analysing the radon-related health risk is a complex business because the cumulated exposure is the result of a variable concentration of the air in a given place, weighted by the length of stay in that place.

Study of the effects of radon exposure among miners

Because of their professional exposure, pit workers and above all uranium miners were among the first population groups to be subjected to epidemiological monitoring in order to evaluate the risk of cancer related to radon inhalation. The studies effectively carried out are cohort studies, that is to say studies that follow miner populations from the day they are hired by the mine, with a systematic and individual recording of their exposure and a health supervision even after they leave the company. All the studies presented reveal higher cancer-induced mortality when compared with a control-population not subjected to professional exposure. They have all concluded that there exists indeed an excess mortality due to lung cancer and have shown conclusively that the excess risk relative to lung cancer increases in linear fashion according to the cumulated exposure to radon and its descendants. The results of these studies are remarkably concordant.

Some studies suggest an inverse effect to the rate of exposure: with an identical cumulated exposure, a low rate seeming to induce a higher risk than the same cumulated exposure over a shorter period. These results have not been confirmed by other human or animal studies.

In the first publication relative to French uranium miners, in 1993, a majority of the exposures to radon and its descendants were at low annual levels, and the duration of exposure was on average 14 years. An excess mortality due to lung cancer was shown and the increase in the risk of this cancer with cumulated exposure follows a linear relationship making it possible to estimate the gradient, called the coefficient of excess relative risk per unit of exposure. An extended study of this cohort and an additional study of a second population of French miners has made it possible to make a new estimate of this coefficient which stands at between 0.35 and 0.8% per unit of exposure, depending on the study under consideration. The results are perfectly concordant with and close to the coefficient estimated from the international joint study. This study groups 11 cohorts of miners and was coordinated by the National Cancer Institute (U.S.A.). It made possible the calculation of the risk based on the 2620 cases of death by lung cancer observed in all these studies. The relative excess risk so obtained is 0.49% per WLM 2 .

This analysis also confirms the plausibility of the linear relationship and shows that the excess risk per unit of exposure decreases according to the age reached and to the time lapse since exposure or since the end of exposure. In addition, in those studies where data were available for tobacco consumption, the radon-tobacco interaction was also studied. Results show that this interaction is more than additive; it is in fact multiplicative or sub-multiplicative. A joint international analysis studying the risk of cancers other than lung cancer has not shown, for any other localisation, an increase in the risk as a function of cumulated radon exposure.

Studies of the cancer risk relative to domestic exposure to radon

The risk of lung cancer relative to domestic exposure to radon was initially estimated by extrapolating the results observed in the professional environment to the general population.

These estimates were done by committees of international experts. However this approach has its limitations:

Ecological studies

Ecological-type studies have been carried out to verify the existence of a link between domestic exposure to radon and lung cancer at a geographical level, by comparing the rates of bronchial-pulmonary cancer in regions with a high and a low exposure to radon. The data used are, at the level of a same region, the average of the measurements of domestic exposure to radon and the average rate of death by cancer. Such studies have been strongly criticised. Persons dying in a given region, and therefore forming the basis of the calculation of the annual death rate may not necessarily have lived for a very long time in that region, whereas the latency time for bronchial cancer after a given exposure is between 10 and 20 years. In addition, the average of the radon measurements at the level of a region (or of a département) is not an exact exposure indicator because intra-regional variability is often higher than inter-regional variability. The chief criticism levelled at these studies is that of a poor taking into account of one important risk factor, tobacco, as compared to a relatively low risk, radon. Indeed, tobacco consumption multiplies the risk of lung cancer by a factor of 5 to 10 among smokers as compared to non-smokers, whereas in the regions where the domestic concentration of radon is higher than the current recommendations given by the European Community (200 to 400 Bq/m 3 ) the expected risk increase would be of the order of 1.2 to 1.5. The major role played by tobacco may explain the inverse relationship between radon and the bronchial cancer rate observed in certain ecological studies, in particular the Cohen study in the United States.

Ecological studies therefore do not provide, in cases of domestic exposure to radon, precise information sufficient to establish an estimate of the risk based on cumulated exposure.

Control-case studies

Studies of control-cases among the general population have been set up, particularly during the last 10 years, with the aim of verifying in the field the results obtained in the context of professional exposure. Moreover, they aim to take into account co-factors such as active or passive tobacco consumption, something which has only rarely been done at the level of miners' cohorts, and they will also make it possible to study the risk for the female population.

The principle of these studies is to question persons suffering from lung cancer and "control" persons, in order to reconstitute a history of the houses occupied over a period of at least 30 years before inclusion and to measure the concentration of radon in these dwellings. The minimum duration of 30 years was retained based on the studies of miners' cohorts showing that the risk diminishes over time since exposure. Distant exposures intervene less in the risk of cancer than more recent exposures.

Certain studies, in particular the one done by Pershagen et al in Sweden, show a significant increase of the relative risk according to the level of domestic exposure. In this study, the relative risks are the following: 1.0; 1.1; 1.3; 1.8 (for, respectively: 3/4 50 Bq/m 3 , > 50-80 Bq/m 3 , >80-140 Bq/m 3 , > 140-400 Bq/m 3 , > 400 Bq/m3). With tobacco consumption that remains constant, this trend persists but in a non-significant way. It is worth noting that the excess is very high (relative risk = 32.5) for smokers of more than 10 cigarettes a day, living in houses where radon concentrations are higher than 400 Bq/m 3 . Other studies, most notably the Canadian study in Winnipeg and also a Finnish study, do not highlight any positive trend of the relative risk with cumulated exposure to radon. A recent study, undertaken in Missouri, where the cases are limited to women who were former smokers or non-smokers shows an absence of risk relative to radon after adjustment for former or passive tobacco consumption. A joint analysis, taking up the basic data from the 3 studies and limited to women (n = 1000), equally does not show a positive trend relative to cumulated exposure to radon. Other studies, such as the recent study by Darby et al show a positive trend, close to that of the Pershagen study, but which is not statistically significant.

Several factors may explain the non-concordant results of these studies of control-cases:

A meta-analysis has been carried out of the complete set of control-case studies currently published. It covers a total of 4263 lung cancers and 6612 control-subjects and shows a positive trend of the risks relative to exposure levels. For an annual exposure of 150 Bq per m 3 over 25 years, the relative risk of lung cancer is estimated at 1.14 (IC95% = 1.01-1.3). This estimate is very close to that obtained from the analysis of miners subjected to low exposure if we admit that a miner's cumulated exposure of 25 WLM is comparable to 25 years of domestic exposure at 231 Bq/m 3 . However there is a great variability in the risk estimates from one study to another. Moreover, this estimate is strongly dependent on a single study, that of Pershagen et al.

The setting-up a vast study of control-cases bringing together a large population has been decided at the European level in order to refine the estimate of the risk of lung cancer related to domestic radon inhalation. It is being carried out in parallel in Germany, France, Belgium, Luxemburg and the United Kingdom. Its aim is to evaluate the risk of lung cancer based on exposure to radon during the 30 years preceding the disease. The radon measurement is carried out in each of the houses occupied during this period, for 6 months with 2 dosimeters per house. The possible exposure to other factors known to be carcinogenic for lungs is also taken into account: active smoking, passive smoking for non-smokers, professional exposure known to be carcinogenic for the lungs.

In France, the study will be conducted in regions susceptible of having housing with high concentrations of radon: Auvergne, Ardennes, Brittany, Languedoc-Roussillon, Limousin. The French study plans to include 600 cases and 1200 control-individuals. The results of the French and European studies will make it possible to provide a better estimate of the risk of lung cancer due to domestic radon inhalation and to better define the interaction between radon and tobacco. Such findings will be essential for guiding the policy for radon-related risks in France and in Europe.

Conclusion

To estimate the risk of lung cancer after inhalation of radon and its descendants, we dispose today of a large number of studies, both epidemiological studies of uranium miner cohorts and control-case studies targeting domestic exposure. Very few environmental factors can avail themselves of such a vast tool, to which should be added the results of tests carried out on animals. As early as 1987, the ICCR considered that studies on animals and on uranium miners were providing sufficient evidence to conclude that radon is indeed carcinogenic for man, but these results came from high annual exposures. The question now is to assess the risk of lung cancer relative to lower exposures, be it at a professional or domestic level. The studies of control-cases currently under way in many countries should make it possible to provide a more precise estimate of the risk among the general population, particularly calculating the risk for smokers and non-smokers. These studies also show that for environmental exposures, which are low and variable and dependent on external conditions that are hard to control, an accurate measurement of individual annual exposure is an essential criterion for assessing the quality of a study and consequently for a correct evaluation of the risk of cancer.

 

Margot TIRMARCHE

Epidemiologist

Institut de Protection et de Sûreté Nucléaire (France).

1. A bécquerel means an event by second. Itvaries a lot so with the nature of this event.

2. WLM : Working Level Month : 1 WLM corresponds to an exhibition during 170 hours to 1 Working Level ( WL). 1 WL is the level of concentration of an atmosphere the descendants of the radon of which deliver an energy alpha of 1,3 105 Mev by

Translation by Andrew WILES

     

© Institut Européen de Cindyniques - Lettre 28 - Nov 99