Do Home Radon Levels Affect Risk for Brain Cancer?

A new observational cohort study published in Neuroepidemiology in September raises questions about the likelihood of radon exposure being a risk factor for brain and other central nervous system (CNS) cancers, purporting to show an association between higher home radon levels and incidence of CNS cancers in women. But the heterogeneity of the data makes it difficult to assess how much of a true association there is and what it might mean.The researchers estimated participants’ residential radon exposure concentration by linking their residential addresses from 1993 to 2018 to four data sources of radon measurement: the US Geological Survey (USGS), the US Environmental Protection Agency (EPA) Map of Radon Zones, county-level estimates of radon concentration modeled by the E.O. Lawrence Berkeley National Laboratory (LBNL), and modeled concentration estimates based on zip codes from a 2025 study in PNAS by Li and colleagues.Over 17.7 years of average follow-up, 451 incident cases of CNS cancers occurred, suggesting that residential radon concentration may play a meaningful role in the development of malignant CNS cancers, the authors wrote.But Gideon Meyerowitz-Katz, PhD, MPH, an epidemiologist at the University of Sydney in Sydney, Australia, disagrees with their conclusion.“They’ve basically shown that radon exposure is potentially correlated with cancer, but it almost entirely depends on the statistical model and other analytical choices,” Meyerowitz-Katz said. “If you use the Li et al estimates of radon exposure, the risk is almost double as high in every analysis between the lowest and highest levels. If you use the USGS data, there’s almost no risk difference between low and high exposure.”What We Know About Radon and Cancer Risk?Radon has long been a well-established cause of lung cancer. The US EPA ranks radon as the second leading cause of lung cancer behind smoking and the number-one cause among nonsmokers, responsible for an estimated 21,000 lung cancer deaths annually. The American Cancer Society notes that the gas has also been linked to other cancer types, including leukemia, but the evidence is more mixed than that for lung cancer.No known lower limit of radon has been determined to be safe for human exposure, and the EPA recommends taking mitigation actions for homes with a radon level ≥ 4 picocuries per liter (pCi/L). The agency estimates the average indoor radon concentration in US homes as 1.3 pCi/L.Previous ecological research has suggested a potential link between radon exposure and incidence of CNS cancers, but ecological studies have substantial limitations and tend to be more useful for hypothesis-generating than for clearly linking a risk to an outcome. Their biggest limitation is the assumption that an observation seen in a population necessarily applies to all individuals in that population. Other studies that have attempted to examine radon exposure and CNS cancers have been inconclusive because the low incidence of these cancers requires a large sample size to detect a signal, and it’s challenging to estimate long-term radon exposure.The authors analyzed associations between residential radon concentrations from the four different datasets cited above and the incidence of CNS cancers in participants of the Women’s Health Initiative (WHI).The 159,391 participants with available data were on average 63 years old at baseline and mostly non-Hispanic White (83%). The researchers calculated hazard ratios for risk for incident CNS cancers based on three tertiles of radon exposure, though the cutoffs for each tertile varied by source.For each data source, the authors adjusted the hazard ratio calculations for different sets of confounders, including the design of the WHI study they were enrolled in; socioeconomic and demographic variables; behavioral factors (including smoking history, alcohol consumption, diet quality, and physical activity); and environmental factors, including fine particulate matter (PM2.5), distance to roadway, and solar irradiance.“Across all radon metrics, higher concentration categories accounted for a disproportionately greater share of cases relative to their population distribution, with the greatest magnitudes observed for the LBNL and Li et al tertiles,” the authors reported.The gradual increase in incidence proportions with radon concentrations did not substantially change with the adjustments for potential confounders.What’s the Biological Rationale for How Radon Might Increase Risk for CNS Cancers?Currently, little is known about likely causes of brain and other CNS cancers. Family history is one risk factor, lead author Carson Mowrer, a PhD student in biostatistics at the University of North Carolina at Chapel Hill, told Medscape Medical News. But genetic factors also only account for a small subset of these cancers. Ionizing radiation from medical treatments, such as radiation therapy, is another risk factor but also accounts for only a tiny fraction of cases.There is experimental evidence supporting the biological plausibility of a relationship between radon exposure and brain cancer, although specific mechanisms haven’t been completely characterized, said Mowrer.“Laboratory studies have demonstrated that alpha-particle radiation [which is released by the breakdown of radon and its radioactive decay products] can cause DNA damage and cellular changes associated with carcinogenesis, and there is additional evidence that radiation exposure can contribute to inflammation and oxidative stress, which may also play a role in cancer development,” Mowrer said. Some of these radon decay products can attach to tiny airborne particles that people inhale, and animal research has shown these particles can travel from the nasal cavity to the brain.“Studies examining human brain tissue after death have also detected polonium-210, a radioactive element produced in the radon decay chain,” he said. “These findings suggest a possible pathway through which radioactive materials could reach the [CNS].”Plenty of uncertainties remain, however, including what dose might reach the brain and how CNS tissue responds to alpha-particle radiation, he said.What Did the Results Show?In addition to 451 incident cases of CNS cancers having occurred, the authors found that women living in areas with higher estimated radon concentrations had a consistently greater risk of developing malignant CNS cancer, even at concentrations below the EPA’s recommended action level.In the USGS data, none of the slightly higher hazard ratios for the middle and higher radon concentrations, ranging from 1.04 to 1.16 depending on the adjustment model, reached statistical significance. The EPA data similarly show higher hazard ratios for the middle and higher concentrations, but again, none reach significance.The LBNL data only show a statistically significant association for the highest tertile, > 1.23 pCi/L, relative to 1.43 pCi/L), which is also below the USGS and EPA’s lowest tertile.Mowrer expressed confidence in their conclusion about the associations despite the variations across models.“What’s encouraging is that we did fairly consistently see observed elevated hazard ratios for higher radon concentrations across the exposure metrics, even though the strength and precision of those varied a little bit,” Mowrer said. “It’s important to recognize that if you see a significant association in one model and a nonsignificant association in another, that doesn’t necessarily represent a contradictory finding. The estimates generally point in the same direction.”Countering Mowrer, Meyerowitz-Katz noted that one of the figures graphing the association showed a decline in CNS risk at higher levels of radon exposure, “which contradicts the text that describes these analyses as generally consistent.”If they had relied only on the USGS radon concentration estimates, they would have to conclude there is no relationship, he added.What Are Other Limitations of This Study?Each of the datasets used to estimate radon concentrations also has its own limitations. The USGS and EPA data, the authors noted, was sampled over a couple of days but was used for estimates across many years, and it covered broader areas, such as counties and geographic regions.“One possible explanation for the differences is exposure misclassification in the EPA and the USGS zones.” The categories and the geographic areas are relatively broad. For example, a radon concentration measurement taken in one part of a large county may not apply to an address 10 miles away in the same county where a WHI participant lives, Mowrer said.Another limitation of the data was that it included numbers based on radon surveys from the 1980s — before measurement protocols were published.Mowrer and Meyerowitz-Katz also expressed concerns about the generalizability of the Li and colleagues data.That database relied primarily on property inspections analyzed at two major commercial lab networks, and “homes undergoing property transactions may not be entirely representative of all homes in the US,” Mowrer said. Plus “patterns of repeat testing could introduce additional biases into the underlying data.”Meyerowitz-Katz said their building of “a very complex model that foundationally rests on a single radon measurement per building in a 3-month period as a proxy for the average radon levels in buildings in the area” may not be reasonable.Ultimately, these data are subject to one of the same limitations as ecological studies, he suggested.“We do not know that the people who got CNS cancer were actually exposed to higher rates of radon [or] just that they lived in areas where the modeled concentration may have been higher,” Meyerowitz-Katz said.“This is compounded by the fact that the WHI doesn’t have great national coverage, which could mess up the statistical model further.” He pointed out that several states have only a “handful” of participants (under 100 in 16 states).But the biggest limitation of the study, according to Meyerowitz-Katz, is the fact that four different estimates of radon concentration led to fairly substantial differences in hazard ratios and statistical significance.“I would say that the authors have demonstrated an inconsistent association between radon in the air and that these cancers that would need to be assessed further in future research,” he said. “It’s also important to note that the risk is very small. Even in a worst-case scenario, the rate goes up from about 1 case per 10,000 per year to 2 cases per 10,000 per year between the lowest and highest radon exposure metrics.”The research was funded by the National Institute of Environmental Health Sciences, and the WHI study was funded by the National Institutes of Health and the US Department of Health and Human Services. Neither the study authors nor Meyerowitz-Katz had any disclosures.Tara Haelle has been an independent science/health journalist for over 15 years, with work appearing in over two dozen trade and consumer publications. She is the author of Vaccination Investigation: The History and Science of Vaccines and The Informed Parent: A Science-Based Resource for Your Child’s First Four Years. Tara is the Public Health Beat Leader for the Association of Health Care Journalists, where she develops resources to help other journalists report on medical studies and public health. She is based in Dallas.