|Comments | Additions/Updates|
RADON UPDATE (Continued)
The EPA has issued a series of documents on radon that provide useful guidance to the home owner. A report which provides detailed and practical information on mitigation strategies for existing homes is found in the Citizens Guides (1,3). Practical information concerning methods for reducing radon levels in new construction are given by the EPA as are specifically outlined abatement methods (2,4).
Water purification systems and aeration techniques can be useful in areas with high levels of radon in the home water supply. Typically these are charcoal filter systems, but the filter itself presents potential difficulty in disposal and is a potential source of elevated external radiation dose. Air cleaning systems are not recommended because they have not been found to be effective.
Major attention is given to methods in which natural or forced ventilation is increased to diminish indoors levels of radon gas. These range from simply opening windows, for example, to increasing ventilation in homes with low but elevated levels, to forced ventilation systems when higher levels need to be abated. In winter, two-fold reductions can be obtained by the use of simple rotating household fans, such as commonly used for summer ventilation during the winter months. (38).
Covering exposed earth reduces ingress of radon, as does sealing cracks and openings in ground level walls and floors. Drain-tiles can be placed surrounding the foundation and vented away from the house (drain-tile suction method). This method is designed to pull radon from the soil surrounding the house and vent it away from the house. Sub-slab suction is more difficult to accomplish as it involves placing pipes under the house (laterally through side walls or by drilling holes in the concrete slab. A fan is used to vent these pipes away from the house. The walls of concrete block houses can be vented by sucking air from the hollow spaces in the wall and venting away from the house to prevent radon from entering from this route. Lastly, methods are described for decreasing negative pressures within the house by bringing air into the house in proportion to losses from chimneys, dryers, etc., or by positive pressure including basement pressurization by blowing air from upper floors into the sealed basement.
A comparison of the features of different systems is given in Table 8
*the costs provided in the exhibit represent the range of typical costs for reducing radon levels in homes above 4 pCi/liter down to radon levels below 4 pCi/liter. In most cases, homes are reduced to an average of about 2 pCi/liter. Adapted from reference 13.
The major health risk from exposure to radon progeny is bronchogenic carcinoma. There are two major sources of human data. These are:
Almost all large epidemiologic studies of lung cancer in miners indicate an excess mortality in groups receiving cumulative exposures of >120 WLM. However, dosimetry measurements made in working mines in different countries many years ago, especially prior to l950, are subject to considerable uncertainty (8). Moreover, interpretation of dose response curves for alpha particles is complicated. Evidence derived from radiobiology indicates that densely ionizing radiations, such as those from a-particles, show dose response curves which increase linearly from low doses to a maximum value, above which cancer induction rates fall due to wasted radiation, i.e. an additional dose to transformed cells is less efficient since the affected cells are already damaged; hence further dose either has no additional effect, or results in cell killing (14).
Figure 9 shows representative data from studies of underground miners. A positive linear high slope region is seen following exposures below 200-300 WLM, which falls off and becomes negative at higher doses presumably due to cell killing.
The risk is expressed as the attributable annual risk per WLM per million persons. (11)
It is assumed that the increased lung cancer risk to miners is due to Rn-222 and its daughters, but the cofactor role of the other dusts they breath in the mine has long been debated (39).
A retrospective cohort study conducted in Southern China in collaboration with the U.S. National Cancer Institute involved 175,143 person-years of observation of workers in a tin mine (40). Eighty-percent of the workers were employed underground, and were exposed to radon, along with arsenic-containing dusts. Death was attributed to lung cancer in 981 individuals. This is the largest study reported to date, and is the first in which these detailed relationships could be tested with a reasonable statistical power. In addition to lung cancer, statistically significant increases in mortality were also observed for leukemia (12 deaths), lymphoma (5 deaths), pneumoconiosis (32 deaths), other respiratory diseases (63 deaths), coronary heart disease (47 deaths), cerebral vascular disease (302 deaths), and accidents (81 deaths).
Table 9 shows age-adjusted relative risk in relation to exposure. Level 0 is non exposed, and increasing levels are graded by quartiles. The excess relative risk (ER) of lung cancer per WLM (ER/WLM) from radon fell from 0.6% to 0.2% when adjustment was made for arsenic exposures. The increase in relative risk with increasing levels of arsenic exposure is much stronger than the increase with level of radon exposure. Radon exposures ranged to greater than 800 WLM with the majority of exposures exceeding 400 WLM, and arsenic exposures ranged to greater than 10 mg-years m-3 with the average exposure in the 3-5 interval. The study is the largest of its kind and permits analysis of several other important factors. The ER/WLM declined significantly with: 1) increasing exposure rate (cumulative WLM/duration of exposure), 2) with years since last exposure, and 3) with increasing attained age. These effects only became significant after adjustment for the exposure effect from arsenic. In this cohort, 41% of the underground workers were <15 years old when they started mining, however, lung cancer risk did not vary consistently with age at first radon exposure.
* From reference 40
Figure 10 shows the relative risk estimates for different groups of miners, which indicate a wide spread in cancer induction rates observed (41).Great variation in the ER/WLM from lung cancer has been seen in the different miner studies with the lowest risk observed in the Port Radium and American uranium miners, and the highest risk observed in the Swedish and Beaverlodge miners. Whether the differences are due to errors in dose estimation, failure to correct for smoking and other life style cofactors, or to other exposures received in the mines can not be assessed at this time. Studies are going on in these mines to establish and corroborate dosimetry estimates and to measure other materials, such as from arsenic, to which the miners may have been exposed. Continuing follow-up is proceeding and more definitive information may be forthcoming from these studies, although the uncertainties in doses received by miners many years ago will be very hard to overcome.
The effect of smoking as a cofactor in these studies is well accepted. Tumors appear earlier in smoking miners, and smoking is a significant cofactor NAS-BEIR IV (8) estimated that smokers have a 10 times higher risk per unit absorbed dose than do non smokers). It is also true that exposure to passive smoking has not been controlled in any of the miner studies, and this may be as important as the radon exposures themselves.
The major uncertainties in the miner studies arise from: uncertainties in dosimetry, uncertainties in exposures to other carcinogens and/or promoters in the mine, and difficulty in controlling for smoking. The studies being conducted in China (40,28) point out the importance of controlling for other exposures in the mines (arsenic in this case), and their use of Pb-210 skull measurements provides a potential means of improving the dosimetry, which has been a problem in all radon epidemiology studies.
A second source of data on radon risk comes from epidemiologic studies of persons living in homes with increased radon levels. A large Canadian study was conducted in 18 cities involving 14,000 homes (42). A statistically significant correlation was found for smoking and lung cancer mortality in males, but the correlation was negative for mortality on measures of radon daughter concentration for males and positive for females neither of which were statistically significant. Multiple linear regression analysis revealed that radon daughter concentrations did not add significantly to the effect of smoking on lung cancer rates. The authors concluded that any effect of radon, if present, was so small in comparison to the effect due to smoking, that it could not be detected in this type or size of study.
Case-control studies are ongoing in 10 countries attempting to relate radon exposure and lung cancer risk. Features of the different studies including the prevalence of homes with exposures greater than 4 pCi/liter are enumerated by Neuberger (43) . The studies range in size from 32 to 3200 lung cancer cases with equal or greater numbers of controls in each study. The total number of subjects under study include 12,273 lung cancer cases, and 19,082 controls. Sample sizes needed to reach statistical significance at different exposure levels are calculated and suggest that many of the studies have adequate statistical power to reject the null hypothesis at high doses. But Neuberger believes that radon health effect studies at low doses could provide an opportunity to test the linear hypothesis and assist in deciding whether and in what circumstances the costs of radon remediation could be justified. He notes that through 1990 only about 25% of radon studies found statistically significant associations (44). A large number of the studies found lower than expected hazard from low doses, but these effects are almost never statistically significant. The conclusions from these studies is that deleterious effects of low doses, if present, are too low to detect in human population studies.
Letourneau has recently completed a large case-control study in Winnipeg involving 750 histologically-confirmed lung cancer cases, age and sex matched against 750 controls. Winnipeg was studied because it has the highest radon levels in urban Canada. Over 80% of the residences were measured with alpha-track detectors. They adjusted for occupational factors, active smoking and ethnicity and found no evidence of a correlation between lung cancer and residential radon levels. (Letourneau, E. personal communication).
An NIH-sponsored case-control study in Sweden investigated the correlation between radon exposure and lung cancer in 210 women with lung cancer and 400 control subjects (45). Smoking and residential history were obtained by interviews, and radon measurements were made in a small fraction of the homes lived in by the subjects over their lifetime. Time weighted radon measurements were made using either alpha track detectors (one year average level per household measured), or thermoluminescent dosimeters which recorded radon levels during a 2-week period in the heating season.
The authors indicate that lung cancer risk tended to increase with estimated radon exposure, reaching a relative risk of 1.7 (1.0-2.9 CI) in women exposed to average radon level above 4 pCi/liter. They note that these risk estimates are within the range reported for radon exposed miners. The risk was 14 times higher in smokers than non smokers in the lowest exposure group (<2 pci/liter), while in the higher exposure groups it was 6 times higher. in none of the smoking groups was there a significant trend relating level of radon exposure to cancer risk. they found an increased trend with exposure in young women based on 5 cases in the low dose group, and 11 in the high dose group, and the under age 55 group was the only one in which a significant trend was noted. they report a stepwise increase in relative risk for lung cancer in non smokers with a p value of 0.04. this correlation was strongly dependent on the dose intervals chosen, for when dose was treated as a continuous variable, no significant correlation was obtained (p=0.5) (lubin, jh, personal communication ). a problem common to all residential radon studies is the difficulty of locating homes in which cases and controls lived, especially in the remote past (46). three methods of assigning dose to missing time periods were used, and in only one did they find a significant correlation. in neither of the other two methods of adjustment was a significant trend noted. this study does not provide strong support for a positive association between residential radon exposure and an increased risk of lung cancer in sweden with its high residential levels.
A more recent report was released by the Swedish group at a press conference in Feb, 1993. (47). They conducted a larger case control study based on 1360 lung cancer cases diagnosed between 1980 and 1985 and 2,857 matched controls. Track etch dosimetry was obtained during the winter season in aprox. 70% of their residences . Regression analyses included smoking as a variable along with radon exposure, age, degree of urbanisation and occupation. They found a relative risk of 1.3 ( 1.1-1.6) at 4-11 pCi/liter and 1.8 (1.1-2.9) at exposures >11 pCi/liter, and attributed 15% of the lung cancer cases to radon. They also found a greater than multiplicative role for smoking. (47)
A large study is being conducted in the high background region of China in the Guangdong Province region, and an adjacent control region. (48) The study involves 2 million person years of observation equally divided between the two regions. The Rn-222 levels differ by a factor of 3 in the two areas, but the rates of lung cancer mortality were reversed in relation to radon dose There were 25 lung cancer deaths in the high background area (Av. lung and tracheobronchial dose= 300-400 mrem), and 35 in the control region (Ave. lung and tracheobronchial dose = 100 mrem),i.e., a 25% higher lung cancer mortality rate in the low background region.
Levels of radon in the Reading Prong region in the United States are very high in certain areas of Pennsylvania and New Jersey. A case:control study was carried out in New Jersey in 433 women with lung cancer and 402 controls (49). They reported a statistically significant positive trend, compatible with increasing risk of elevated radon residential exposures. Only a small fraction of the cases and their residences could be located and radon levels measured. The study revealed a high relative risk associated with the highest exposed individuals. The authors urged caution in interpretation because of selection biases and the small numbers of subjects in the high exposure group.
Umhausen, Austria, is a small village (2600 inhabitants) in the West Tyrol in which very high radon concentrations (median=50 pCi/liter) are found in an area between two rivers. In the rest of the town, radon levels are lower (median=5 pCi/liter). The median lifetime radon exposure in these two areas is 242 and 23 WLM, with relative risks of 6.17 (4.4-8.4) and 1.43 (0.7-2.7) respectively. The rates in the very high exposure group are comparable to those observed in uranium miners, whereas the rate in lower exposed group is not significantly elevated. (50)
In the United States, Cohen (51) has studied lung cancer rates in 965 counties, and in all States. He found a strong negative slope, which is highly significantly different from the slope predicted using linear/non-threshold models and BEIR IV data. (Fig 11).
An extension of that study includes data from 1600 U.S. counties and compares mortality rates for the various types of cancer to average radon levels. The strongest correlations are found with lung cancer, and the sign of the correlation is negative. Cohen BL. Relationship between exposure to radon and various types of cancer. (52)
All of the environmental radon epidemiology studies have serious methodological problems. One problem involves uncertainties in dosimetry. This arises from difficulty in locating former residences, in measuring the cumulative dose to assign to each individual in case:control studies, as well as the very important question as to what dose to assign to the TBE cell from which radon-induced lung cancer is thought to arise. The new methods being used for radon measurements should provide some help on the data collection side. Remaining major problems common to all epidemiology studies are the difficulty in identifying and controlling for the presence of confounding variables, such as smoking (active and passive), along with the problems in identifying and correcting for various selection and ascertainment biases.
Because of the uncertainties, the size of the study group needed to establish statistical confidence is so large that the power of the statistical tests is often too weak to establish a significant difference between no risk from residential radon and increased risk at the level found in miner studies. A reasonable conclusion from these studies is that deleterious effects of naturally-occurring background levels, if present, are too small to detect in most residential radon epidemiology studies. Evidence derived from ecologic studies has been critically reviewed recently with special relevance to radon. (53). The authors conclude that the 15 largest ecologic studies they reviewed did not contribute to better understanding of the quantitative risks of indoor radon.
Fig. 11. Age-adjusted lung cancer rates are plotted vs. radon levels in counties in the study data base. The abscissa is divided into ranges shown at the top which also shows the number of counties in each range. Lung cancer rates for each set of counties are plotted along with mean, standard deviation, and first and third quartiles. Least squares best fit lines are plotted for the mean values.(42) [80k]
The American Cancer Society estimated that there were 136,000 deaths from lung cancer in 1987, and that about 113,000 of these were the direct result from cigarette smoking. This assumption would leave 23,000 lung cancer deaths that may arise from all other causes. Using the average continuous radon exposure of 0.75 pCi/liter (0.19 WLM/year) and the NAS-BEIR IV risk estimates, the number of radon induced lung cancer deaths expected annually can be computed. Assuming a population of 240,000,000 in the USA, between 4500 and 23,000 lung cancer deaths could be attributed to radon exposure annually (54). The average risk from NAS-BEIR IV (3.5 x l0-4/WLM) would predict 16,000 deaths. Since there must be other causes of lung cancer besides cigarette smoking and radon progeny, many scientists involved in radiation protection matters believe that the hazards of radon exposure are significantly overestimated. In any event, the cheapest and most effective way of diminishing the lung cancer risk is to decrease or eliminate cigarette smoking.
Based upon the results of studies in miners, the estimated risk of lung cancer from exposure to radon progeny from ICRP, NCRP, and BEIR IV, are shown in Table 10. The estimates average 3.5 x l0-4 /WLM.
The Second International Workshop on Residential Radon (46) discussed the various ongoing case:control studies of residential radon exposure and lung cancer risk. Over 10,000 lung cancer cases are included in these investigations. A tabular summary of these studies is given by the EPA (13). Given the large number of studies now being conducted, and the difficulties in establishing meaningful dosimetry, correcting for confounders, and finally in pooling data, the DOE report concluded that it was unlikely that meaningful low dose risk estimates could be derived from additional radon epidemiology studies.
Currently there are no U.S. statutory limits covering naturally occurring radioactive materials such as radon and its progeny. However, both the NCRP and EPA have published guidelines for acceptable levels of radon in the home (1,16). The NCRP recommends that in single family homes remedial action be taken to reduce radon levels if the average annual exposure exceeds 2 WLM/year (equal to 8 pCi/liter assuming radon daughters are in 50% equilibrium with Rn-222).
EPA recommendations are based on average airborne radon levels in the home, and they recommend a graded scale of actions. These are presented in Table 11. Their recommendations suggest action at a lower dose (factor of 2) than NCRP, but otherwise there is no major difference. The recently passed radon act 51(55) poses a long term goal of remediation to outdoor levels of 0.2 - 0.7 pCi/liter, which would require many billions of dollars to accomplish. The urgency of recommended actions depends on the average radon levels in the living areas of individual homes, and not simply on the highest level in an uninhabited portion of the house. The amount of time spent in the home and where one spends most of that time needs to be considered when making decisions on corrective actions. If high levels are found in high occupancy areas, remedial action should be considered and advice obtained from experts. Radiation control officials at the state or local level can suggest additional kinds of measurements, as well as recommend remedial actions, if indicated.
The EPA estimates approx. 22,000 lung cancer deaths per year may be related to radon exposure in the USA. (56). Over a period of 70 years, with 75% of a persons day spent in the home, they calculate that an indoor level of 4 pCi/L, with a 50% equilibrium between radon and its daughters would result in 54 WLM cumulative exposure. Assuming 0.25 WLM/yr and 240 million persons results in 60 million person WLM. They then assume 360 deaths per million WLM (an average between the lower BEIR IV, and higher EPA estimates), from lung cancer (age-averaged rate for the US population), and compute 21,600 deaths due to lung cancer due to radon per year. The ICRP gives a range of 8,600-25,900 to these estimates.
Much controversy surrounds the true magnitude of health risks from radon, and the appropriate actions to be taken at different measured levels in the home or workplace. The issue boils down to understanding the magnitude of the health and economic risks, the costs and benefits of different responses. The ICRP (7) recommends that "proposed interventions should ...be sufficiently (beneficial) to justify the harm and the costs, including social costs, of the intervention. The form, scale and duration of the intervention should be chosen so that the net benefit of the reduction of dose, i.e. the benefit of the reduction in radiation detriment, less the detriment associated with the intervention, should be maximized." (57)
The issue comes down to cost and benefit. The EPA has estimated the cost per life saved (by averting a predicted lung cancer from radon) for various action levels that might be chosen. The numbers range from 1.1 million dollars at 2.0 pCi/L to 0.7 million at 4 pCi/L and 0.4 million at the NCR level of 8 pCi/L.(13). The cost per life saved from other non radiological risks can reach the 0.4 million figure (57).
*From reference 1.
Radon is a naturally-occurring element which in high doses has been shown to cause lung cancer. Miners exposed to high doses have an increased lung cancer risk which is significantly enhanced by smoking. Radiobiology data reveal a linear dose response following exposures to alpha particle emitters in the low dose region with saturation at high exposure levels (>200-400 WLM). A resident of a 4 pCi/liter house (0.02 WL) could be exposed at a rate of 0.52 WLM /year. A small fraction of homes have much higher radon concentrations, in some cases exceeding levels in mines. It is clear that these homes need to be identified and their levels reduced. The cost of remediation of an individual dwelling is reasonably cheap at low radon levels, and there are many such houses, while very high radon level houses are more difficult and expensive to mitigate, but they are relatively rare. The estimated costs are not very high at the 4 pCi/liter level, and these costs are consistent with the costs society has and does spend for various health and safety problems.
To date the EPA has had little success in stimulating home owners to measure levels in their homes which would be the first step in the process of deciding on a course of action if a high radon level is found. This is partly because it is difficult to get people concerned that their home, a place that one looks to for security, is a potential source of hidden danger. Also, it has not yet been possible to generate convincing data on increased risk at or below 4-8 pCi/liter. However, it is prudent for people living in areas in which high levels do exist to test their home. Then, based on individual attitudes toward acceptable risks, appropriate action depending upon their resources and competing needs. This is also the position that society, faced with a multitude of costly options must take. The cost/benefit ratio for radon abatement needs to be carefully considered in that context. Based on currently available data, the Committee concludes that the costs of remediation exceed the anticipated potential benefits for radon levels less than 8 pCi/liter.
From reference 1.
For more information on Radon, try out Information Sources page