Estimating lung cancer mortality from residential radon using data for low exposures of miners.

Some recent estimates of lung cancer risk from exposure to radon progeny in homes have been based on models developed from a pooled analysis of 11 cohorts of underground miners exposed to radon. While some miners were exposed to over 10,000 working level months (WLM), mean exposure among exposed miners was 162 WLM, about 10 times the exposure from lifetime residence in an average house and about three times the exposure from lifetime residence at the "action level" suggested by the U.S. Environmental Protection Agency. The extrapolation of lung cancer risk from the higher exposures in the miners to the generally lower exposures in the home is a substantial source of uncertainty in the assessment of the risk of indoor radon. Using the pooled data for the miners, analyses of lung cancer risk were carried out on data restricted to lower exposures, either <50 WLM or <100 WLM. In the pooled data, there were 115 lung cancer cases among workers with no occupational WLM exposure and 2,674 among exposed miners, with 353 and 562 lung cancer cases in miners with <50 WLM and <100 WLM, respectively. Relative risks (RRs) for categories of WLM based on deciles exhibited a statistically significant increasing trend with exposure in each of the restricted data sets. In the restricted data, there was little evidence of departures from a linear excess relative risk model in cumulative exposure, although power to assess alternative exposure-response trends was limited. The general patterns of declining excess RR per WLM with attained age, time since exposure and exposure rate seen in the unrestricted data were similar to the patterns found in the restricted data. Risk models based on the unrestricted data for miners provided an excellent fit to the restricted data, suggesting substantial internal validity in the projection of risk from miners with high exposures to those with low exposures. Estimates of attributable risk for lung cancer (10-14%) in the U.S. from residential radon based on models from the unrestricted data were similar to estimates based on the data for miners receiving low exposures.