Lung cancer risk from residential radon: meta-analysis of eight epidemiologic studies.

BACKGROUND Studies of underground miners exposed to radioactive radon and its decay products have found that exposure increases risk of lung cancer. Consequently, when radon was found to accumulate in houses, there was concern about the public health impact from exposure to a known carcinogen. Estimates on the basis of studies of underground miners suggest that indoor radon may account for 6000-36,000 lung cancer deaths each year in the United States. Because of differences between working in underground mines and living in houses, estimates are subject to major uncertainties. Numerous case-control studies were launched to assess directly the lung cancer risk from indoor radon. Some studies report positive or weakly positive findings, while others report no increased risk. Thus, the potential hazard from indoor radon remains answered only indirectly through miner studies, experimental animal studies, and cellular studies. PURPOSE To provide more information on the risk of lung cancer from indoor radon, we conducted a meta-analysis of all case-control studies that included at least 200 case subjects each and that used long-term indoor radon measurements. METHODS Eight studies were available and included a total of 4263 lung cancer case subjects and 6612 control subjects. From the published results of each study, confounder-adjusted relative risk (RR) estimates and 95% confidence intervals (CIs) for categories of radon concentration were obtained, and weighted linear regression analyses were performed. RESULTS The combined trend in the RR was significantly different from zero (two-sided P = .03), and an estimated RR of 1.14 (95% CI = 1.0-1.3) at 150 Bq/m3 was found. An influence analysis indicated that no single study dominated the combined results. The exposure-response trend was similar to model-based extrapolations from miners and to RRs computed directly from miners with low cumulative exposures. However, there were significant differences in the study-specific estimates of the exposure response (two-sided P < .001), which were not explained by study differences in percent of the defined exposure interval covered by radon measurements, mean number of residences per subject, and other factors. CONCLUSIONS Meta-analyses are valuable for identifying differences among studies and for summarizing results, but they should be interpreted cautiously when expected RRs are low as with indoor radon exposure, when there is study heterogeneity and where there is the potential for confounding and exposure misclassification. Nonetheless, the results of this meta-analysis suggest that the risk from indoor radon is not likely to be markedly greater than that predicted from miners and indicate that the negative exposure response reported in some ecologic studies is likely due to model misspecification or uncontrolled confounding and can be rejected. IMPLICATIONS Until ongoing case-control studies of indoor radon are completed and the data are pooled and analyzed, the studies of underground miners remain the best source of data to use to assess risk from indoor radon. This meta-analysis provides support for their general validity.