Living a few dozen kilometers from a nuclear power plant is part of the daily life of millions of people, including in France. Is this proximity dangerous for health? A new American study provides a new element of response by looking not at an isolated site, but at the entire nuclear fleet of an entire country.
Researchers from the Harvard TH Chan School of Public Health analyzed cancer mortality in every county in the United States between 2000 and 2018. And their main result is clear.
A first national study on American nuclear power plants
Published on February 23, 2026 in the journal Nature Communicationsthe study covers 93 commercial reactors spread across 54 sites in 28 states. The researchers used cancer mortality data from the Centers for Disease Control and Prevention and plant information from the U.S. Energy Information Administration. They adjusted their models for many factors: education level, median income, racial composition, smoking, body mass index, temperature, humidity and distance to the nearest hospital.
To measure exposure, the team did not draw a simple circle of 10 or 20 km around each site. It calculated “continuous proximity”: researchers used sophisticated statistical modeling to assess the overall impact of nuclear power plants located within a 200 km radius, rather than focusing on a single facility. This approach was applied over a 10-year period to take into account the time to onset of cancers. Information regarding the location and commissioning dates of nuclear power plants in the United States, as well as some nearby Canadian facilities, was provided by the Agency American Energy Information System (EIA) County-level cancer mortality data were obtained from the Centers for Disease Control and Prevention (CDC).
This choice also makes it possible to capture the cumulative effect of several nearby power plants. Sensitivity tests with other radii (100 to 200 km) or other time windows (2 to 20 years) produce comparable results.
Excess mortality from cancer, especially among the elderly
The most exposed counties have higher cancer mortality rates than remote counties, all else equal. In short, the more proximity to nuclear power plants is stronger, the more cancer deaths there are, even in areas that appear socially or medically similar. And the risk decreases with distance.
The authors estimate that approximately 115,000 deaths from cancer over the period, or nearly 6,400 per year, are associated with the proximity of the power plants. The signal is particularly marked among women aged 55 to 64 and men aged 65 to 74, with a relative risk which reaches around 1.19 to 1.20, or almost 20% additional mortality in the closest areas.
Among people aged 65 and over, cancer mortality linked to this proximity represents on average 4,266 deaths per year. To situate the magnitude, the authors recall that another study associated coal-fired power plants with 20,909 deaths from all causes per year. Cancers potentially linked to nuclear power plants would therefore represent approximately 20% of this burden, even if the comparison remains imperfect. The biological link remains plausible: ionizing radiation is a well-established carcinogen.
A correlation without direct proof of cause
However, the researchers highlight important limitations. The analysis is called “ecological”: it is done at the county level, without tracking individuals or their exact addresses over time. No direct measurement of radioactivity was included and all plants were assumed to have the same potential impact. The types of cancer are not distinguished and childhood cancers are not studied. The authors note that their results are in the same direction as a previous study carried out in Massachusetts, which had already observed an increase in the incidence of cancers near nuclear installations.
“We recommend that additional studies be carried out on the impacts of nuclear power plants on health, particularly at a time when nuclear energy is presented as a clean solution to climate change.“, declared Petros Koutrakis, lead author. For highly nuclearized countries like France, this type of statistical signal fuels the call for finer health monitoring around sites, detailed dose measurements and robust cancer registries, in order to better document the possible long-term effects of this energy source.