Nuclear Power Plants and Cancer

Published by EMFSA | 26th August 2026

The relationship between nuclear power plants and cancer remains an important question in environmental health research. 

Historical nuclear incidents continue to shape public concern, while research into possible health effects associated with low-level radiation exposure around nuclear facilities faces substantial methodological challenges, including long latency periods for some cancers, exposure misclassification, confounding, ecological-study limitations and limited statistical power to detect associations for less common cancers.  

Continued health surveillance around nuclear facilities remains important, particularly as interest in nuclear energy grows. This post examines three recent peer-reviewed studies investigating cancer outcomes in relation to nuclear power facilities, focusing on their study designs, findings and limitations.

Three Studies at a Glance

Qinshan Nuclear Power Plant, China

Researchers measured external ambient radiation around Qinshan Nuclear Power Plant and examined cancer mortality in Haiyan County. The study found no statistically significant association between nuclear plant operation, environmental radiation levels and cancer mortality within its study framework.

National US Study

Researchers examined cancer mortality across US counties from 2000–2018 using geographic proximity to multiple nuclear power plants as a proxy for exposure. The study reported positive associations between proximity and cancer mortality, particularly among older adults.

Massachusetts Study

Researchers examined cancer incidence across Massachusetts ZIP codes from 2000–2018, again using geographic proximity to nuclear facilities as a proxy for exposure. The study reported positive associations between proximity and cancer incidence, particularly among older adults.

The key distinction: The Qinshan study measured external ambient radiation, whereas the two US studies used geographic proximity as an exposure proxy. 

Living closer to a nuclear power plant does not necessarily mean receiving a higher radiation dose.

Study 1: Qinshan Nuclear Power Plant, China

Study Design

The researchers conducted an ecological and descriptive study of external ambient radiation around Qinshan Nuclear Power Plant (QNPP) Phase I from 2010–2023. They compared the period before the Phase I project’s service extension (2010–2021) with the post-extension period (2022–2023).

They also examined cancer mortality among residents of Haiyan County from 2012–2022.

The study used data from 30 ambient radiation monitoring stations and included age-specific dose and risk calculations based on population-derived occupancy factors.

Authors’ Conclusion

The authors found no statistically significant association between the service-extension period and cancer mortality trends, while reported ambient radiation remained at background levels. 

Importantly, the authors frame this as an absence of a detectable association within the study’s framework, rather than definitive evidence that nuclear plant operation cannot have biological effects. They recommend continued long-term monitoring.

Key Limitations

• Ecological design: The analysis was conducted at the population level and cannot establish a causal relationship between radiation exposure and individual cancer outcomes.
• External exposure only: Internal exposure pathways, including ingestion or inhalation of radionuclides, were not assessed.
• Limited statistical power: The relatively small number of deaths from some specific cancers, particularly leukemia and thyroid cancer, limits the study’s ability to detect modest associations.
• Short post-extension follow-up: The 2022–2023 period is much shorter than the latency period associated with many radiation-related cancers.

What This Study Adds

The study provides data spanning the period before and after the Phase I project’s service extension and adds to the relatively limited epidemiological literature examining low-dose radiation around operating nuclear facilities. It also provides baseline data for future monitoring and comparative studies.

The researchers reported that ambient radiation remained at background levels, while estimated annual effective doses and excess risks remained below applicable standard limits. 

Source: Li Y, Cao Y, Zhang X, et al. External ambient radiation exposure and cancer mortality trends around Qinshan nuclear power plant phase I: long-term study before and after service extension. Frontiers in Public Health. 2026;14:1780605. DOI: 10.3389/fpubh.2026.1780605. Full article

Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

Two US Studies: A Different Approach

The two recent US studies examined cancer outcomes in relation to geographic proximity to nuclear power plants rather than measuring environmental radiation directly.

This difference is important.

The Qinshan researchers used physical monitoring stations to measure external ambient radiation around a specific facility. By contrast, the US studies estimated proximity using the distance between population locations and nuclear facilities.

Because individual radiation doses were not available in these datasets, proximity provided a practical way to investigate whether cancer patterns varied geographically in relation to nuclear facilities.

A proximity measure is therefore an indirect exposure proxy, not a measurement of individual radiation dose. Actual radiation exposure can depend on factors such as radioactive releases, environmental transport, meteorological conditions, exposure pathways and individual behaviour. Because the US studies did not measure radiation releases or individual radiation doses, they cannot establish that differences in proximity corresponded to differences in actual radiation exposure.

The US studies also differed from Qinshan in scale. The national study examined cancer mortality across US counties, while the Massachusetts study examined cancer incidence at the finer ZIP-code level.

These differences mean the studies should not simply be viewed as contradictory. They investigated different populations and outcomes and used fundamentally different approaches to exposure assessment.

Study 2: National US Cancer Mortality Study

Study Design

The researchers conducted an observational ecological study examining the relationship between long-term proximity to nuclear power plants and cancer mortality across US counties from 2000–2018.

Cancer mortality was analyzed across six adult age groups, separately for males and females.

Proximity estimation used a continuous inverse-distance-weighted metric. It incorporated operational nuclear power plants within 200 km of each county centre. 

A 10-year average proximity measure was used to characterize longer-term proximity to nuclear power plants, reflecting the study’s consideration of long-term exposure patterns.

The analysis adjusted for a range of county-level socioeconomic, demographic, behavioral, environmental and healthcare factors.

Authors’ Conclusion

The authors reported that counties closer to operational nuclear power plants had higher cancer mortality rates.

Figure 4 showed the highest model-estimated relative risks at the shortest equivalent plant distances. 

The authors describe these findings as a spatial association rather than proof of causation and call for further research into potential exposure pathways, cancer latency and cancer-specific risks.

Key Limitations

• County-level exposure and residential mobility: Proximity was calculated using county centres and therefore cannot capture where individuals actually lived within a county. The study also did not incorporate individual residential histories, so proximity estimates may not reflect where people lived throughout the study period, potentially leading to exposure misclassification.
• Proximity is only a proxy: The study did not measure individual radiation doses or establish that people living closer to plants received greater radiation exposure.
• Aggregate cancer outcome: The primary outcome was mortality from all malignant cancers combined rather than cancer-specific mortality.
• Residual confounding: Although the analysis adjusted for numerous county-level factors, ecological studies cannot account for every individual-level factor that may influence cancer risk.

What This Study Adds

The study provides national-scale data over 19 years and uses a continuous proximity metric rather than a simple near/far classification. It also incorporates proximity to multiple nuclear power plants and uses a 10-year average measure to examine longer-term proximity patterns.

The observed associations provide a basis for further investigation into potential exposure pathways, cancer-specific risks and latency.

An Author Correction to this article was published on 13 April 2026, and the article has been updated accordingly.

Source: Alwadi Y, Alahmad B, Vieira CLZ, et al. National analysis of cancer mortality and proximity to nuclear power plants in the United States. Nature Communications. 2026;17:1560. DOI: 10.1038/s41467-026-69285-4. Full article

Licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

Study 3: Cancer Incidence in Massachusetts

Study Design

The researchers examined cancer incidence across Massachusetts ZIP codes from 2000–2018 using data from the Massachusetts Cancer Registry.

They used two analytical approaches: longitudinal generalized estimating equation (GEE) Poisson regression for all cancers combined and cross-sectional log-linear Poisson regression for site-specific cancers.

Proximity was estimated using a cumulative inverse-distance-weighted metric incorporating seven nuclear facilities within 120 km.

The analysis incorporated demographic, socioeconomic, environmental and healthcare covariates, including PM2.5, and was stratified by sex and four age groups: 45–54, 55–64, 65–74 and 75+.

Authors’ Conclusion

The authors reported positive associations between residential proximity to nuclear power plants and cancer incidence in Massachusetts, particularly among older adults.

They reported significant positive associations for all cancers combined among people aged 55 and older in both sexes. Relative risks declined with increasing distance and became negligible beyond approximately 25 km.

The authors also reported associations involving several specific cancer types, including lung, prostate, breast, colorectal, bladder, melanoma, leukemia, thyroid, uterine, kidney, laryngeal, pancreatic, oral and esophageal cancers, as well as Hodgkin lymphoma.

The researchers estimated that 20,618 cancer cases could be attributable to proximity under their model assumptions, including 10,815 among females and 9,803 among males. Estimated attributable fractions reached up to 6.3% among females aged 75+.

These attributable estimates require an important qualification: they are model-derived estimates based on the assumption that the observed associations are causal. They should not be interpreted as numbers of cancer cases proven to have been caused by nuclear power plant exposure.

Key Limitations

• Ecological design: ZIP codes provide finer spatial resolution than counties, but the analysis still cannot determine individual exposure. This creates the potential for ecological fallacy: associations observed at the ZIP-code level cannot necessarily be assumed to apply to individuals.
• Proximity as an exposure proxy: The study did not measure individual radiation doses.
• Residential mobility: The study did not incorporate individual residential histories, so ZIP-code proximity may not accurately represent where individuals lived throughout the study period, potentially leading to exposure misclassification.
• Residual confounding: Adjustment for numerous demographic, socioeconomic, environmental and healthcare variables cannot eliminate the possibility that unmeasured factors influenced the results.
• Cancer-specific interpretation: Different cancers have different latency periods and radiation sensitivities, complicating interpretation of both pooled and site-specific associations.
• Childhood cancers: Sparse data prevented meaningful analysis of pediatric outcomes.
• Occupational exposure: The ecological design could not distinguish residents with occupational exposure at nuclear facilities from the broader residential population.

What This Study Adds

The Massachusetts study provides finer geographic resolution than the national US analysis and examines cancer incidence rather than mortality, capturing newly diagnosed cancers regardless of subsequent survival.

It also uses a continuous inverse-distance-weighted proximity metric and includes sensitivity analyses using alternative distance thresholds and temporal averaging windows.

The broadly similar associations reported in this study and the national analysis are noteworthy and merit further investigation.

Source: Alwadi Y, Evans JS, Schwartz J, Vieira CLZ, Christiani DC, Coull BA, Koutrakis P. Residential proximity to nuclear power plants and cancer incidence in Massachusetts, USA (2000–2018). Environmental Health. 2025 Dec 18;24(1):92. DOI: 10.1186/s12940-025-01248-6. PMID: 41408632; PMCID: PMC12713251. PubMed

© The Author(s) 2025. Open Access — licensed under Creative Commons Attribution 4.0 International (CC BY 4.0).

The Two US Studies: Related but Not Independent Replications

The Massachusetts cancer-incidence study and the national cancer-mortality study are closely connected. Both share lead author Yazan Alwadi and colleagues from Harvard T.H. Chan School of Public Health and examine the same broad 2000–2018 period.

They nevertheless examine different populations and outcomes.

The Massachusetts study uses ZIP-code-level cancer incidence and seven nearby nuclear facilities. The national study examines cancer mortality across US counties, providing much broader geographic coverage but at a coarser spatial resolution.

The studies are therefore complementary, but they should not be regarded as independent replications. They share authorship, study period and a similar proximity-based exposure framework.

Their broadly similar associations are noteworthy and merit further investigation, but they do not resolve the central question of causation.

What Can We Conclude From These Studies?

Taken together, these studies illustrate why determining whether nuclear power plants contribute to cancer risk remains scientifically challenging.

The Qinshan study used measured external ambient radiation around a single nuclear facility and found no statistically significant association with cancer mortality within its study framework. However, its relatively small population, limited number of cancer deaths and short post-extension follow-up constrain what can be concluded about rare cancers and long-latency effects.

The two US studies examined much larger populations and reported positive associations between geographic proximity to nuclear facilities and cancer mortality or incidence. Their larger datasets provide greater statistical power to detect associations, but their proximity measures did not establish individual radiation exposure. 

Even within the US studies that reported statistically significant overall associations, some age groups had confidence intervals that included 1.0, indicating that the evidence was not equally strong across all demographic strata. This reinforces why the overall findings should be interpreted cautiously. 

This distinction is critical.

An association between geographic proximity and cancer does not, by itself, demonstrate that radiation from a nuclear power plant caused the cancer.

People living near nuclear facilities may differ from people living farther away in ways that are difficult to fully capture in ecological studies. At the same time, proximity may not accurately represent an individual’s actual radiation exposure.

The evidence therefore does not provide a simple yes-or-no answer.

The Qinshan study does not establish that nuclear power plant operation cannot be associated with cancer risk. Conversely, the US studies do not establish that proximity to nuclear power plants causes cancer.

What they do provide are different pieces of epidemiological evidence: one study based on measured external ambient radiation around a single facility, and two large US studies based on geographic proximity as an exposure proxy.

What Would Strengthen the Evidence?

Research that could substantially strengthen the evidence would involve independent studies with larger populations, individual residential histories, cancer-specific analyses, measured or modelled radiation doses, assessment of relevant exposure pathways and appropriate consideration of cancer latency.

Such research would help determine whether the geographic associations reported in the US studies correspond to differences in actual radiation exposure—and whether those differences could plausibly explain the observed cancer patterns.

For now, the most defensible conclusion is that the studies identify associations that warrant further investigation, but they do not establish that radiation from nuclear power plants caused the observed cancers.

* Illustration by Design Nation on Unsplash