Many publications have claimed that increased residential radon levels increase lung cancer risk. However, there is a major problem with such publications.
Whereas high-dose radiation, received in a short time, has been observed to increase cancer risk as observed in the atomic bomb survivors, low radiation doses have been observed to reduce the cancer risk in many studies (Berrington et al, 2001; Doss, 2018; Linet et al, 2017; Pollycove & Feinendegen, 2008; Sponsler & Cameron, 2005). In view of such data, it would not be appropriate to use the linear no-threshold (LNT) model for analyzing the radon-lung cancer data in order to determine the dose-response shape or to estimate the lung cancer risk due to low levels of residential radon. Therefore, a vast majority of the publications on radon and lung cancer risk, which utilize the LNT model for analyzing the data, including meta-analyses of such publications, do not have any validity and so will not be discussed here.
When publications have compared the lung cancer rates between residences with different residential radon levels, a reduction of lung cancer with increasing radon level has been reported in case-control studies (Thompson, 2011) and ecological studies (Cohen, 1995; Denton & Namazi, 2013).
A study of lung cancer in women who predominantly never smoked has shown evidence for a U-shaped dose-response shape that is consistent with radiation hormesis and inconsistent with the LNT model (Bogen & Cullen, 2002).
A recent ecological study (Simeonov & Himmelstein, 2015), which examined lung cancer incidence in the USA counties, and considered a large number of demographic variables, concluded that lung cancer risk increases with smoking prevalence. It also concluded that lung cancer incidence decreases with elevation, and that this correlation can explain the observed negative correlation between radon levels and lung cancer incidence in the USA counties. Using the data in this publication (available at: https://dfzljdn9uc3pi.cloudfront.net/2015/705/1/county-data.txt ), and considering counties within a narrow range of values for elevation and smoking prevalence, a scatterplot has been generated between residential radon levels and lung cancer rates, and it shows a negative slope indicating reduction of lung cancers with increasing residential radon levels (unpublished graph).
Confounding by smoking prevalence or by elevation wound not be able to explain this negative slope, as the data include only a narrow range of smoking prevalence and elevation.
All these data are consistent with radiation hormesis and inconsistent with the LNT model.
Berrington, A., Darby, S. C., Weiss, H. A. & Doll, R. (2001) 100 years of observation on British radiologists: mortality from cancer and other causes 1897-1997. Br J Radiol, 74(882), 507-19. http://www.ncbi.nlm.nih.gov/pubmed/11459730
Bogen, K. T. & Cullen, J. (2002) Residential Radon in U.S. Counties V Lung Cancer in Women Who Predominantly Never Smoked. Environmental Geochemistry and Health, 24(3), 229-247. http://dx.doi.org/10.1023/A:1016051322603
Linet, M. S., Kitahara, C. M., Ntowe, E., Kleinerman, R. A., Gilbert, E. S., Naito, N., Lipner, R. S., Miller, D. L., Berrington de Gonzalez, A. & Multi-Specialty Occupational Health, G. (2017) Mortality in U.S. Physicians Likely to Perform Fluoroscopy-guided Interventional Procedures Compared with Psychiatrists, 1979 to 2008. Radiology, 284(2), 482-494. https://www.ncbi.nlm.nih.gov/pubmed/28234559
Thompson, R. E. (2011) Epidemiological Evidence for Possible Radiation Hormesis from Radon Exposure: A Case-Control Study Conducted in Worcester, MA. Dose Response, 9(1), 59-75. http://www.ncbi.nlm.nih.gov/pubmed/21431078