Conversion factors and units

The SI unit for the activity of a radioactive substance is the becquerel (Bq), which is one radioactive decay per second. In indoor air, the degree of radioactive equilibrium between its airborne short-lived progeny and radon gas depends on several factors, principally on the aerosol concentration and its size distribution, the surface-to-volume ratio of the room and the air exchange rate. The degree of equilibrium is usually expressed in terms of the equilibrium factor (F factor), whereby an F factor of 1 means full radioactive equilibrium between radon and its airborne short-lived progeny. The F factor is important for determining the dose to the lungs from radon progeny. Measurements in several countries have shown F factors in dwellings to generally lie between 0.2 and 0.8. The United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR) and the International Commission on Radiological Protection (ICRP) have adopted a typical worldwide F factor of 0.4 for indoor air (1,3). The F factor in outdoor air is usually somewhat higher, in the range 0.6–0.8.

Owing to these considerations, there are several measures used to describe airborne radon decay products. A commonly used measure is the equilibrium equivalent radon concentration, which is the activity concentration of radon (given in Bq/m³) in equilibrium with its short-lived decay products that would have the same potential alpha energy concentration as the actual non-equilibrium mixture present in the air being measured. The potential alpha energy concentration is the sum of the potential alpha energy per unit volume of the short-lived radon progeny in the decay chain down to ²¹⁰Pb (half-life 22.3 years); its SI unit is J/m³.

The potential alpha energy exposure of workers was historically expressed in terms of the working level month (WLM). Here one working level (WL) is any combination of short-lived radon progeny that will result in the ultimate emission of 1.3 × 10⁵ MeV of potential alpha energy. This is approximately the alpha energy released by the decay of radon progeny in equilibrium with 3.7 Bq of radon. The WLM exposure unit was introduced and is still used to specify occupational exposure to 1 WL for a working month of 170 hours.

Analytical methods

Most of the radiation dose and hence the risk from radon is due to its short-lived alpha-particle-emitting polonium decay products (polonium-218 and polonium-214). Radon gas itself contributes a much smaller dose than these decay products. As an equilibrium factor of 0.4 is taken as being representative of indoor air in most homes, a measurement of radon gas is considered in general to be a good surrogate for estimating the concentrations of these decay products. Radon gas measurement is also technically much simpler and cheaper than measuring the decay products. Owing to both the effects of building usage practices (i.e. ventilation and heating) and meteorological variables, indoor radon concentrations may exhibit quite large diurnal and seasonal variations. In the case of meteorological variables, their effects on radon exhalation from the ground and entry to dwellings is quite complex, with changes in atmospheric pressure, wind speed and precipitation being most important in this regard. As a consequence of these weather effects, in addition to diurnal and seasonal variations, indoor radon concentrations show substantial year-to-year variability.

To make a reliable estimate of the radon risk, it is thus necessary to make long-term (three months to a year) measurements of radon (4). For measurements of a few months, a seasonal adjustment factor, if available, may be applied to obtain an estimate of the annual average value. Measurements made during the course of a single year also need correction for the aforementioned year-to-year variation (5). From a health assessment perspective, short-term measurements of radon (duration of some days) are of limited use but may be of use in radon screening surveys to identify locations with a potential for high radon concentrations or when remediating a dwelling with a radon problem.

Owing to their low unit cost and reliability, the most popular devices used for making long-term radon measurements in European countries are small, passive devices using alpha-particle-sensitive material. These solid state nuclear track materials record the damage in the form of sub-microscopic latent tracks caused by alpha particles from radon and its decay products striking their surface. The latent tracks caused by the alpha particles striking the detector material are enlarged and made visible for optical microscopy by chemical or electrochemical etching (6). These radon detectors are very simple and rugged in construction. They consist of a small piece of the alpha track material mounted inside a pillbox-sized chamber into which radon gas may diffuse. These detectors are passive as they do not require electrical power. The most commonly used alpha-particle-sensitive materials used in these detectors are polyallyl diglycol carbonate (CR-39), cellulose nitrate (LR-115) and polycarbonate (Makrofol). After exposure to radon and subsequent processing, the measured alpha track density is directly proportional to the integrated radon exposure in Bq/hour per m³. The conversion from track density to the mean radon concentration over the exposure period is achieved by controlled exposure of the detector to a calibrated concentration of radon in a sealed exposure chamber. Comparisons between laboratories in Europe measuring radon take place regularly at the Radiation Division of the Health Protection Agency in the United Kingdom (formerly the National Radiological Protection Board), having originally been organized jointly with the European Commission (7).

In addition to the alpha track passive detectors described above, passive charged electret radon detectors are also available, as well as a range of electronic continuous monitors of both radon gas and its decay products (4,8,9).

These techniques are useful in determining contemporary indoor radon levels. In residential radon epidemiological studies, contemporary radon levels in the present and in previous dwellings of a person are generally used as surrogates for the unknown radon levels in these dwellings in the past. Indoor radon levels can, however, be quite variable on diurnal, seasonal and annual timescales and can also be affected by changes in indoor living habits over time. Because of this, using contemporary indoor radon measurements as surrogates for indoor radon levels in past decades poses challenges to the assessment of long-term past exposure to radon.

Retrospective techniques based on the measurement of the build-up in a dwelling over many years of long-lived radon progeny such as polonium-210 in glass (the surface trap technique) or in porous media (the volume trap technique) can give some insight into the indoor radon concentration in a dwelling in past decades (10,11). Direct in vivo measurements of lead-210 in the human skeleton have also been used to assess exposures in the past (12).

Indoor air

All rocks contain some uranium, typically at concentrations of 1–3 ppm. The uranium content of a soil will be about the same as the uranium content of the rock from which the soil was derived. As radium-226, the immediate parent of radon, is a decay product of uranium, the higher the uranium content of the soil the greater the radium concentration and the higher the chance that houses built on such soil will have high levels of indoor radon. The main source of indoor radon is the radon produced by the decay of radium in the soil subjacent to a house. Soil gas containing radon enters a house through cracks and fractures in the foundations by pressure-driven flow, as the air in a house is generally warmer and therefore at a lower pressure than the subjacent soil gas (13). Radon concentrations in soil air/gas typically range from less than 10 000 Bq/m³ up to 100 000 Bq/m³. Most houses draw less than 1% of their indoor air from the soil; the remainder comes from outdoor air, which is generally quite low in radon. Houses with poorly sealed foundations, built on high-permeability ground and with several entry points for soil gas may draw more than 10% of their indoor air from the soil. Even if the soil air has only moderate levels of radon, levels inside such houses may be very high.

In comparison to soil gas, the radon exhaling from building materials in most cases does not significantly contribute to indoor radon levels. The uranium and radium content of building materials will be similar to the rock or clay from which they are made. While this is generally low, there are some building materials that may have high concentrations of radium. Examples of these are alum shale concrete and building materials made of volcanic tuff, by-product phosphogypsum, and some industrial waste materials (14).

Water supplies can also contribute to indoor radon levels. River and surface reservoir water supplies usually contain very little radon but groundwater may contain high concentrations, depending on the uranium/radium content of the aquifer formation. Public waterworks using groundwater and private domestic wells often have closed systems and short transit times that do not remove radon from the water or permit it to decay. This radon is out-gassed from the water to the indoor air when the water is used for washing, cooking and other purposes in a house. The areas most likely to have problems with radon in groundwater are those that have high levels of uranium in the underlying rocks. Radon concentrations can reach several thousand Bq/l in water from drilled wells in regions with granite rock or other uraniferous rocks and soils (15). This contributes to indoor radon and to exposure via ingestion but the dose to the lung per unit exposure arising from inhalation is much higher than that owing to ingestion (16). A very rough rule of thumb for estimating the contribution of radon in domestic water supplies is that house water with 10 000 Bq/m³ radon contributes about 1 Bq/m³ to the level of radon in the indoor air.

The range and distribution of indoor radon levels in many countries have been determined both by national surveys and in other investigations. Table 7.1 gives a summary of indoor radon surveys that have been carried out in a number of European countries (17). It should be noted, however, that the survey design was not the same for each country. In some countries, dwellings were selected on the basis of population density. In this approach, more measurements are made in large centres of population than in sparsely populated rural areas. This enables estimates to be made of the collective exposure and health risk of the general population in a country. Such information is useful for developing national radon control strategies by the relevant authorities. Some national surveys were made on a geographical basis, where the strategy was to achieve the same density of dwelling sampling per unit area irrespective of the national population density distribution. Notwithstanding these differences, the data presented here give a reasonably accurate overview of average radon concentrations in contemporary European dwellings. It should also be noted that the maximum radon concentration values quoted in Table 7.1 are the maximum values found in national survey data. In many countries, much higher indoor radon concentrations have been found in targeted surveys carried out in areas where high radon levels were expected to be present on the basis of geological characteristics. The results of such surveys yield an erroneously high average radon concentration when extrapolated to the whole country. Table 7.2 gives a summary of indoor radon data for a number of large non-European countries (1). It should be noted that representative national surveys of indoor radon have not yet taken place in countries with the largest populations, such as China and India.

Table 7.1

Radon surveys in dwellings in some European countries.

Table 7.2

Radon concentrations in dwellings from surveys in some non-European countries.

The only other radon isotope that can occur indoors in significant amounts is radon-220 (half-life 55.6 seconds). Radon-220 is referred to as thoron and is a member of the thorium-232 (half-life 1.4 × 10¹⁰ years) decay series. Its immediate parent is radium-224 (half-life 4.6 days). It should be noted that there has been an increasing interest in indoor thoron in recent years. Owing to its short half-life, thoron in soil gas beneath a building, in most situations, cannot survive long enough to enter a house and thereby contribute to the level of thoron in indoor air. Indoor thoron is due to the exhalation of thoron from thorium that may be present in the materials forming the internal surfaces of the building. Some building materials, such as volcanic tuff in Italy, have been found to have a high thoron exhalation rate. While in general indoor thoron levels are low, research in recent years has identified uncommon situations, such as cave dwellings, where the doses from airborne thoron decay products can be significant and can even exceed those from the radon decay products in the same location (18). In this context, it should be noted that for the same exposure (i.e. concentration by time) the dose from thoron decay products is estimated to be about four times that of radon decay products (1,19). From the perspective of radiation dose to lung tissue due to inhalation, the most important airborne thoron decay product is lead-210 (half-life 10.64 hours). While lead-210 itself is a beta particle emitter when it decays in the lung, it gives rise to the alpha-emitting decay products bismuth-212 (half-life 60.5 minutes, 36% alpha particle energy Eα = 5.5–6.1 MeV) and polonium-212 (half-life 3 ×10⁻⁷ seconds, alpha particle energy Eα = 8.68 MeV).

Outdoor air

Land masses are the sources of outdoor radon while sea waters, having minimal radium concentrations, act as sinks. As a consequence, outdoor air radon levels are much lower (circa 0.1 Bq/m³) over oceans and seas than over a continental land mass such as mainland Europe (20). Outdoor radon levels are determined mainly by the soil characteristics (uranium/radium content, porosity and the consequent radon exhalation rate), local topology and meteorological conditions. In some situations, such as atmospheric temperature inversions in valleys with high radon fluxes from the soil, short-term elevated outdoor radon levels have been observed. High outdoor radon levels are rare but could be of local health significance in areas such as former uranium mining districts, where elevated radon exhalation from tailing ponds combined with meteorological and topological conditions could give rise to high outdoor radon levels of seasonal duration. A direct proportionality in risk between indoor and outdoor radon exposures based simply on radon concentration and duration of exposure cannot, however, be assumed. This is because factors that influence the lung dose, such as the equilibrium factor between radon and its decay products (which are generally higher outdoors than indoors) and also aerosol characteristics, will be different indoors than outdoors.

National data on average outdoor radon levels are quite limited. It seems that they lie between 5 and 20 Bq/m³ (21). The ratio of the radon concentration in outdoor air to the mean indoor radon concentration in European countries (see Table 7.1) would appear to be in the range of about 7% (Czech Republic) to 20% (United Kingdom).

Routes of exposure

The most important route of exposure to radon and its decay products is inhalation. It is also possible to be exposed to radon by ingestion of water containing high radon concentrations but the dose and risk in this case are very small in comparison to those due to inhalation. In indoor air, radon produces a series of short-lived decay products that may attach to aerosol particles present in the air or deposit on room surfaces (22). It is the inhalation and deposition in the airways of short-lived airborne radon decay products that give rise to irradiation by alpha particles of sensitive cells in lung tissue, such as the basal cells of the bronchial epithelium (15). From considerations of their respective radioactive half-lives as well as their physical and chemical properties, the radiation dose delivered to the lung from inhaled radon decay products is dominated by the alpha particles emitted by the short-lived radon decay products polonium-218 (half-life 3.05 minutes, alpha particle energy E_α = 6.00 MeV) and polonium-214 (half-life 1.64 × 10⁻⁴ seconds, alpha particle energy E_α = 7.68 MeV). Because these alpha particles have respective ranges of only 48 μm and 71 μm in tissue, they deliver a high density of DNA damage to cells in these short distances.

Absorption and doses

Dosimetry of inhalation of radon and its decay products is important in understanding the biological mechanisms and in estimating the effects of different factors that contribute to carcinogenesis (21). Estimates of the absorbed dose per unit of radon exposure to various organs and tissues can be derived from information on, for example, the unattached fraction, the activity/size distribution of the radon progeny aerosol, breathing rate, fractional deposition in the airways, mucus clearance rate, location of the target cells in the airways, and lung-to-blood absorption parameters. The doses to various lung tissues may be calculated using the ICRP human respiratory tract model (23) and other models. Dose estimates strongly depend on the choice of input parameters and other model assumptions, thus leading to some uncertainty in estimated absorbed doses (24). The various dose calculation procedures and assumptions have been reviewed in detail in several reports (1,21).

Kendall & Smith (16) estimated the doses to various organs and tissues from radon and its decay products, either by inhalation or ingestion or by deposition on the skin. With respect to inhalation, about 99% of the lung dose arises from radon progeny and not from the gas itself, as almost all of the gas that is inhaled is subsequently exhaled (25). Radon decay products are largely deposited on the surface of the respiratory tract. Because of their relatively short half-lives, they decay mainly in the lung before being cleared either by absorption into the blood or by particle transport to the gastrointestinal tract. Thus, in the case of inhalation, the highest doses are to the lung and to the extra-thoracic airways (i.e. the nose, pharynx and larynx), while dose estimates to other organs and tissues were at least one order of magnitude lower. Outside the respiratory tract, the kidney is the organ most exposed to radon decay products. In general, the doses from radon gas are much lower than those from radon decay products. However, radon is more soluble in tissues with a higher fat content. As fat receives the highest dose of all tissues outside the lung, the doses to the red bone marrow and the female breast are somewhat higher. Kendall & Smith (16) also investigated the dose to the fetus. As the fat content of the fetus is low, its dose is assumed to be similar to that of the maternal muscle, which is estimated to be about 3–4 orders of magnitude smaller than the dose to the lung.

Kendall & Smith (26) also considered doses from radon and decay products when inhaled by one-year-old infants and ten-year-old children. They found that the general pattern of doses to organs is broadly similar to that of adults. The largest dose is received by the respiratory tract. Even though dose coefficients for children are generally larger than those for adults, the total annual doses are more similar because of the smaller intake of air by children. Radon decay products deposited on skin may be able to induce skin cancer. However, the location of the sensitive cells is not known with certainty and they may lie too deep to receive significant doses. On irradiation, it is likely that doses to children would be larger than those to adults.

Marsh et al. (25) recently provided a mathematical model to calculate the individual annual absorbed doses arising from radon and its decay products to regions of the lung, red bone marrow, liver and kidney among uranium miners of the Czech, French and German cohort studies. Several exposure scenarios (wet/dry drilling, good/medium/poor ventilation, diesel engines, underground/surface, etc.) and levels of physical activity had been evaluated. For example, the scenario of underground work with wet drilling, medium ventilation and medium physical activity was estimated with the following annual absorbed doses in mGy/WLM: bronchial region 7.3, bronchiolar region 7.3, alveolar-interstitial region 0.45; red bone marrow 0.031, kidney 0.02 and liver 0.0065. As expected, the dose to the lung is the highest, as most of the short-lived radon progeny decay before they leave the lung. For the red bone marrow, the dose arising from the radon gas is greater than that from the radon progeny. Overall, the doses to red bone marrow, liver and kidney were appreciably lower than those to the lung.

Experimental animal studies

Animal studies have been conducted for several decades to evaluate the biological effects of inhaled radon and its decay products, mainly in rats but also in mice and beagle dogs. These studies systematically examined the pathogenic role of radon and its decay products, either alone or in various combinations with uranium ore dust, diesel-engine exhaust and cigarette smoke. In the late 1960s and early 1970s, it was proved that radon and its decay products, either alone or in combination, produce lung tumours (21). Only a few laboratory animal studies investigated the risk of non-respiratory neoplasms, producing inconsistent results (1,21,27,28).

A number of experimental animal studies examined the effects of exposure rate on induction of lung cancer, particularly at low cumulative exposures comparable to current underground mining exposures or to lifetime exposure in houses with high radon levels (29–33). The results indicate that at low cumulative exposures, the risk of lung cancer increased with increasing exposure rate, while at high cumulative exposures (> 100 WLM), the reverse was observed (decreasing risk with increasing exposure rate). These data are consistent with that of underground miners showing an inverse exposure-rate effect at high cumulative exposures but a reduction of this effect at cumulative exposures lower than 50–100 WLM (28,34–36). When biologically based models were applied to the various animal experimental data, the obtained set of significant model parameters appeared to compare reasonably well with that from similar models derived from studies on uranium miners (37–40).

Molecular and cellular studies

Molecular and cellular radiobiology studies are important in understanding the mechanisms involved in carcinogenesis caused by ionizing radiation. In 1996, Jostes (41) provided an overview of the genetic, cytogenetic and carcinogenic effects of radon. He reported that radon and radon progeny cause cell transformation, changes in chromosome structure and gene mutations containing a wide range of deletions, as well as base-pair changes. It is thus possible that even exposure to low radon concentrations such as in homes adds to the genetic load for cancer risk. Since then, a comprehensive review on cellular and molecular responses to various forms of radiation has been given by the Committee on the Biological Effects of Ionizing Radiation (BEIR VI) (28) and UNSCEAR (1). The UNSCEAR report (1) includes specific annexes on DNA repair and mutagenesis, biological effects of low doses of ionizing radiation and the combined effects of exposure to radiation and other agents. An extensive update of this report, with specific focus on radon, is given in UNSCEAR's 2008 report (21).

A number of in vitro studies of cells exposed to alpha-particle radiation demonstrated not only direct effects in irradiated cells but also non-targeted effects such as the bystander effect (21). Bystander effects occur when irradiated cells emit signals that result in damage to nearby non-irradiated bystander cells. Brenner et al. (42) suggested that bystander effects can result in non-linear dose–response relations and inverse dose-rate effects, and thus make it difficult to extrapolate risks based on linear models of miner studies to the risk from residential radon (43–45).

Chromosomal aberrations are among the most useful biomarkers of effects and doses from radon exposure (1,21). Associations between chromosomal aberrations and cancer incidence have been observed in radon-exposed miners, while correlations between radon exposure and chromosomal aberrations have been found in radon-exposed miners and to some extent also in the general population through residential radon exposure (21).

BEIR VI (28) and WHO (4) argued that it is possible that radon-related DNA damage can occur at any level of exposure to radon, since even a single alpha particle can cause major genetic damage to a cell. Therefore, it is unlikely that there is a threshold concentration below which radon does not have the potential to cause lung cancer.

Identification of studies

Health effects of radon were identified by hand searching references in former reviews by UNSCEAR (1,21), the National Research Council (15,28), IARC (46) and WHO (4). All these reports were published between 1988 and 2009. The detailed UNSCEAR report Sources-to-effects assessment for radon in homes and workplaces from 2008 (21) and the WHO radon handbook from 2009 (4) formed the major basis for the text. Next to that, electronic searches were made in PubMed in January 2009, with an update in December 2009 in order to identify newly published papers. The keywords were: “radon” and “cancer” or “health effects” or “mortality”. Moreover, recent papers known to the experts were included. For the present review, next to the above-mentioned summary reports, approximately 50 publications on health effects in relation to radon exposure were selected. About 70% of them concerned studies on miners with occupational underground radon exposure and 30% concerned indoor radon studies in the general population.

Effects on humans: lung cancer

Residential radon studies

There is substantial uncertainty in the extrapolation of the risk of lung cancer from the miners studies to the risk of lung cancer from radon exposure in the home (28). For this reason, a series of epidemiological studies directly investigated the association between indoor radon and risk of lung cancer since the 1980s (1,4,21,28). The first generation of these studies were ecological studies, in which average radon concentrations were correlated with average lung cancer rates at an aggregated geographical level. This type of study is known to be prone to bias because of several methodological problems (1,21,67). Later, a number of case-control studies were carried out that gathered detailed information on smoking history and other risk factors for lung cancer and assessed the radon exposure retrospectively by measuring radon in the current and previously occupied homes of the study participants.

A detailed review of the results of these individual case-control studies is given in the 2008 report by UNSCEAR (21). The majority of the studies showed a positive association between radon exposure and risk of lung cancer; however, the estimated risk coefficients often did not reach statistical significance in the individual studies. Moreover, there was a substantial variation in the estimated radon-related risk as published in the individual studies. Several meta-analyses had been undertaken to summarize the findings (68-70). Differences in the methodology used to analyse the different studies, such as adjustment for smoking and exposure quantification, however, limit the interpretation of these meta-analyses. For this reason, the original data of the individual studies were brought together and collaborative analyses were performed on the individual data of 13 European studies (71,72), 7 North American studies (73,74) and 2 Chinese studies (75).

The largest of these pooled studies is the European pooling study published by Darby et al. (71,72). It includes 7148 cases and 14 208 controls from 13 European indoor radon case-control studies on lung cancer, all with detailed information on smoking histories and radon measurements in homes that the individual had occupied for the past 15 years or more. The available radon measurements covered a mean of 23 years in the relevant radon exposure period 5–4 years prior to interview. Individual exposure to radon (called “measured” radon concentration) was calculated as the time-weighted average of the radon concentrations in all the homes occupied over the past 5–34 years, with missing radon values substituted by the mean concentration of the controls in that region. A statistical model was fitted in which the additional risk of lung cancer was proportional to measured radon concentration. In addition, radon exposure was subdivided into categories and the relative risk across categories of measured radon concentrations was plotted against the mean level in these categories. In both models, detailed stratification was performed for study, age, sex, region of residence and 25 smoking categories. Since no statistically significant heterogeneity in the radon-associated risk between the studies was present, the data were pooled.

In the pooled analysis, the excess relative risk of lung cancer per 100 Bq/m³ “measured” radon concentration was 8% (95% CI 3–16). This proportionate increase did not differ significantly by study, age, sex or smoking history. The corresponding risk estimates for lifelong non-smokers, ex-smokers and current cigarette smokers were 11% (95% CI 3–28), 8% (95% CI 3–21) and 7% (95% CI –1 to –22), respectively. The exposure–response relationship appeared to be approximately linear, with no evidence for a threshold below which there was no risk. In particular, the results were incompatible with a threshold exposure below which there is no risk above 150 Bq/m³. Even when the analysis was restricted to individuals with measured radon concentrations below 200 Bq/m³, the exposure-response relationship remained statistically significant. The risk of lung cancer was 1.2-fold (95% CI 1.03–1.30) higher among individuals with measured radon concentrations of 100–199 Bq/m³ than in those with concentrations < 100 Bq/m³ and the increase was statistically significant.

Analysis based on the so-called “long-term average radon concentration”, which takes into account the random year-to-year variability in measured radon concentration in the homes, led to a doubling of the excess relative risk of 16% (95% CI 5–31) per 100 Bq/m³. Again, the risk did not differ significantly by study, age, sex or smoking status, and the exposure–response relationship was approximately linear (Fig. 7.2).

Fig. 7.2

Relative risk of lung cancer vs long-term average residential radon concentration. Source: Darby et al. (72).

Darby et al. (72) also reported in detail on the combined effects of smoking and radon within the European pooled study. Table 7.3 gives information on the cumulative risk of death from lung cancer by the age of 75 years for lifelong non-smokers and continuing smokers of 15–24 cigarettes a day (“current smokers”). For these calculations, the estimated excess relative risk of 16% per 100 Bq/m³ of long-term average radon concentration, which was independent of smoking status, was used. The relative risk for current smokers of 15–24 cigarettes per day compared to lifelong non-smokers was estimated as 25.8-fold. For lifelong non-smokers, it was estimated that living in a home with a long-term average radon concentration of 0, 400 or 800 Bq/m³ was associated with a cumulative risk of death from lung cancer of 41, 67 or 93 per 1000. For current smokers, the corresponding values would be 101, 160 or 216 per 1000, respectively. For those having stopped smoking, the radon-related risks are substantially lower than for those who continue to smoke, but they remain considerably higher than the risks for lifelong non-smokers.

Table 7.3

Cumulative risk of death from lung cancer by the age of 75 years for lifelong non-smokers and continuing smokers of 15–24 cigarettes per day at various radon concentrations .

Krewski and co-workers (73,74) reported on the results of the pooled analysis of seven indoor radon case-control studies in Canada and the United States, which included a total of 3662 cases and 4966 controls. Residential radon levels were measured for one year by alpha-track detectors. For each individual, the time-weighted average of the radon concentrations in the homes was calculated, with a focus on the period 5–30 years prior to the date of interview. Because no statistically significant heterogeneity of radon-related risk was found between the studies, a combined analysis was performed. Based on this joint analysis, the risk of lung cancer increased by 11% (95% CI 0–28) per 100 Bq/m³ increase in measured radon concentration. The trend was consistent with a linear exposure–response relationship. There was no apparent difference in the proportionate increase in risk by sex or smoking history, although there was some evidence of decreasing radon-associated lung cancer risk with age. Analyses restricted to individuals with presumed “more accurate dosimetry” resulted in increased risk estimates. For example, for individuals who lived in only one or two homes in the 5–30-year period and for which alpha-track measurements covered at least 20 years of this period, the proportionate increase in lung cancer risk was 18% (95% CI 2–43).

The Chinese pooled study published by Lubin et al. (75) included 1050 cases and 1996 controls from two studies in two areas, Gansu and Shenyang. As in the North American pooled study, the time-weighted average of the radon concentration in homes was calculated within the exposure period 5–30 years. No significant heterogeneity in the associated risk was present between the two studies. For the pooled data, the increase in risk per 100 Bq/m³ increase in measured radon concentration was 13% (95% CI 1–36) and the results were consistent with a linear exposure–response relationship with no threshold. When analyses were restricted to individuals resident in only one home and with complete measurement coverage in the relevant period, the proportionate risk per 100 Bq/m³ increased to 33% (95 CI 8–96).

In the WHO handbook on indoor radon (4), a review and comparison of the risks are provided from all three pooled residential radon studies and the miner studies (Table 7.4). The radon-related risk estimates in the three pooled indoor radon studies were very similar. In each study, the exposure–response relationship appeared linear, without evidence of a threshold, and there was no statistically significant evidence that the radon-related risk varied by age, sex or smoking status. A weighted average of the three pooled risk estimates was provided, with weights proportional to their variances, resulting in a joint estimate of 10% proportionate increase in lung cancer risk per 100 Bq/m³ measured radon concentration. WHO (4) estimated that, based on long-term average radon concentration instead of measured radon concentration, this 10% estimate could even increase to 20% per 100 Bq/m³ if it is assumed that the effect of adjusting for year-to-year random variation in the three pooled studies combined is the same as in the European study.

Table 7.4

Summary of risks of lung cancer from indoor radon based on international pooling studies that have combined individual data from a number of case-control studies and on studies of radon-exposed miners.

In general, a direct comparison of the risks of lung cancer between residential and miner studies is difficult. This is due to the higher average radon exposures among miners and the time-dependent modifying factors, but primarily the inverse exposure rate effect. Thus, the summary estimates of the joint 11 miner studies and the German miner study are somewhat lower, with 5% and 3% per 100 Bq/m³, respectively, than in the residential studies (4). When analyses in the BEIR VI study were restricted to cumulative exposures below 50 WLM, which is comparable to living in a house with a radon concentration of around 400 Bq/m³ for 30 years, the estimated risk coefficient increased to 14% per 100 Bq/m³ (76) and even to 30% per 100 Bq/m³ after additional restriction to miners with radon concentrations lower than 0.5 WL (i.e. < ∼ 2000 Bq/m³). No such risk analyses had been performed within the German study; assuming that the same restriction as in the BEIR VI study has the same effect, however, then the corresponding risk would be around 18% per 100 Bq/m³. Based on the joint analysis of the Czech and French cohorts, which is characterized by low levels of cumulative exposures, an increase of about 29% per 100 Bq/m³ would be expected in the exposure window 5–34 years.

Based on these comparisons, WHO (4) concluded that, in summary, the radon-related risk estimates for lung cancer from residential radon studies and studies of underground miners with relatively low cumulative exposures accumulated at low concentrations are in good agreement.

Effects on humans: diseases other than lung cancer

A number of studies focused on the relationship between radon and leukaemia in children and adults. Laurier et al. (77) reviewed 19 ecological studies, 8 residential case-control studies and 6 miner cohort studies published between 1997 and 2001. While the ecological studies suggested a positive correlation between residential radon exposure and leukaemia at a geographical level, the case-control studies and cohort studies did not. Overall, the authors concluded that the available data did not provide evidence of an association between radon and leukaemia (77). Since then, a positive association between leukaemia incidence and exposure to radon has been reported in a case-cohort study among Czech uranium miners (78), in a Danish case-control study on residential radon and childhood leukaemia (79) and in a French ecological study on childhood leukaemia (80), while no evidence for an increased risk of leukaemia by exposure to radon was reported in two independent studies among German uranium miners (81,82).

Overall, individual miner cohort studies have provided little evidence for an increased risk of cancers due to radon other than lung cancer (28,58,83,84). However, most of these studies were limited owing to small numbers of cases. The two largest and most informative studies are the pooled analysis of 11 miner cohorts (85) and the German Wismut cohort study (82). In the pooled study, the observed mortality from extrapulmonary cancers combined (O) was close to that expected from national rates (E) (n = 1179; O/E = 1.01; 95% CI 0.95–1.07). The trend with cumulative exposure was statistically significant only in the first decade since start of employment.

Among individual sites examined, a statistically significant excess in mortality was found for leukaemia and cancers of the stomach and liver (in the period less than 10 years since starting work). For none of the examined cancer sites was mortality significantly related to cumulative radon exposure, except for pancreatic cancer, which might be a chance finding.

In common with the pooled analysis of 11 miner cohorts, no excess in the overall mortality from extrapulmonary cancers was observed in the German Wismut cohort when compared to the general population (n = 3340; O/E = 1.02, 95% CI 0.98–1.05) (82), while statistically significant excesses in mortality for cancers of the stomach and liver were present. When the relationship with cumulative radon exposure was considered, a statistically significant relationship was found for all extrapulmonary cancers combined (ERR/WLM = 0.014%; 95% CI 0.006–0.023) and cancers of the extra-thoracic airways and trachea (ERR/WLM = 0.062%; 95% CI 0.002–0.121) (64). The majority of non-respiratory cancer sites investigated revealed positive exposure–response relationships, which were non-significant however. The authors concluded that the study provides some evidence for an increased radon-related risk of cancers of the extra-thoracic airways and some other non-respiratory cancer sites; this is in line with calculations of organ doses, though chance and confounding cannot be ruled out. Based on a large case-control study on laryngeal cancer among German uranium miners, Mohner et al. (86) reported no relationship with cumulative radon exposure.

Epidemiological studies on diseases other than cancer mainly focused on the relationship between radon exposure and cardiovascular disease among miners. None of these studies found any evidence that radon causes cardiovascular diseases (64,87–91). A Norwegian study demonstrated an association between multiple sclerosis and indoor radon, but this study was prone to bias owing to the ecological study design (92).

Overall, the currently available epidemiological evidence indicates that there is only suggestive evidence that radon causes a material risk for diseases other than lung cancer.

Definition of health outcomes

Health effects of radon, most notably lung cancer, have been investigated for several decades. An increase in the risk of lung cancer with increasing exposure to radon was first consistently demonstrated in studies on underground miners (15,28). Based on these results, radon was classified by IARC in 1988 as a Group 1 human pulmonary carcinogen (46). In addition, there is direct evidence from epidemiological indoor radon studies that radon in homes increases the risk of lung cancer in the general population (4,21).

The pooled analysis of data from the European (71,72), North American (73,74) and Chinese (75) residential radon studies consistently demonstrated that the risk of lung cancer increases approximately linearly with increasing long-term radon exposure. There is no known threshold below which radon exposure presents no risk. The increase is statistically significant even below 200 Bq/m³. Risk estimates from epidemiological studies of miners and residential case-control studies are remarkably coherent. There is limited, though inconsistent, evidence of other cancer risks due to radon.

When radon gas is inhaled, densely ionizing alpha particles emitted by deposited short-lived decay products of radon can interact with biological tissue in the lungs, leading to DNA damage. Molecular and cellular studies demonstrated that it is possible that radon-related DNA damage can occur at any level of exposure to radon, since even a single alpha particle can cause major genetic damage to a cell (1,4,21).

Relevance for health of current indoor exposures in various regions of the world

Radon is a major contributor to the ionizing radiation dose received by the general population. Outdoor radon levels are usually very low, with average values in the range of 5–20 Bq/m³. National indoor radon surveys show that the distribution of radon concentrations in dwellings is approximately log-normal, with average values ranging between 20 and 150 Bq/m³ (4,93). Published estimates of the proportion of lung cancers attributable to residential radon exposure range from 3% to 14%, depending on the average radon concentration in the country concerned and the calculation methods (4). As most people are exposed to low or moderate radon concentrations, the majority of lung cancers related to radon are caused by these exposure levels rather than by higher concentrations. In many countries, radon is the second cause of lung cancer after smoking. Most of the radon-induced lung cancer cases occur among smokers and ex-smokers owing to a strong combined effect of smoking and radon exposure. Nevertheless, radon exposure is the primary cause of lung cancer among people who have never smoked.

In summary, there is sufficient evidence to conclude that radon causes lung cancer, even at concentrations typically found in indoor air. There is suggestive evidence of an association with other cancers, in particular leukaemia and cancers of the extra-thoracic airways.