For decades, studies have shown that even low doses of radiation are harmful to humans.

This week, the Associated Press reported that the Trump administration may be reconsidering that. The Environmental Protection Agency seemed to be looking at raising the levels of radiation considered dangerous to humans based on a controversial theory rejected by mainstream scientists. The theory suggests that a little radiation might actually be good for our bodies. In April, an EPA press release announced the proposal and included supporting comments from a vocal proponent of the hypothesis, known as hormesis. It prompted critical opinion pieces and sparked worry among radiation safety advocates.

Those comments back in April were made by Edward Calabrese, a toxicologist at the University of Massachusetts, Amherst, who also testified before Congress on the issue this week. And in the initial release, Calabrese hailed the EPA’s decision to move away from the radiation dose model widely accepted by the scientific mainstream. But by Friday, the EPA backed away from Calabrese’s stance in comments to Discover.

The debate cuts to the heart of the debate over the effects of low doses of radiation and reveals how difficult it is to craft clear guidelines in an area where scientific evidence is not clear cut.

Radiation Debate

When radiation damages our DNA, the body steps in to make repairs. Hormesis suggests that hitting the body with a little more radiation should kick our defensive mechanisms into overdrive. According to proponents of the theory, this results in the production of anti-oxidants and anti-inflammatory compounds that reduce our risk for cancer and heart disease, among other things. That’s why hormesis backers want the EPA to raise the level of acceptable radiation, pointing out that it would also save millions in safety costs.

It sounds convincing, and proponents have dozens of studies to point to that they say back up their claims. But, there’s never been a large-scale human study of hormesis. And while studies of low-dose radiation are very hard to do, so far, most suggest that radiation is indeed bad for us, at any dose.

“Large, epidemiological studies provide substantial scientific evidence that even low doses of radiation exposure increase cancer risk,” says Diana Miglioretti, a professor in biostatistics at the University of California, Davis in an email. “Risks associated with low-doses of radiation are small; however, if large populations are exposed, the evidence suggests it will lead to measurable numbers of radiation-induced cancers.” 

Long-term studies of Hiroshima and Nagasaki bombing survivors show higher cancer risks. Marshall Islanders exposed to radiation from atomic bomb tests suffered a higher risk of thyroid disease. And patients who get CT scans, which deliver a dose of radiation equal to thousands of X-rays, saw cancer risks go up afterward. Researchers also found that radiation from childhood CT scans can triple the risk of leukemia and, at higher doses, triple the risk of brain cancers as well. Another found that low-dose radiation increased the risk of breast cancer among some some women.

And large-scale reviews of the evidence for hormesis find that it is decidedly lacking. Two studies, one in 2006 by the National Research Council, and another in 2018 by the National Council and Radiation Protection and Measurements looking at 29 studies of radiation exposure find no evidence for hormesis, and reiterate that the evidence points toward radiation being bad for us even at low doses.

Scientific Uncertainty

It’s difficult to study low doses of radiation, though, and that’s where much of the controversy comes from. At doses below a few hundred millisieverts (mSv), a radiation unit that accounts for its effects on the body, it becomes extraordinarily hard to separate out the effects of radiation from other things like lifestyle or genetics. Research on the effects of these small radiation doses often use data sets involving thousands of people to compensate for the minimal effect sizes, but even then it’s often not enough to be certain what’s happening. 

“Data collected at low doses (defined by the scientific community [as] exposures less than 100 mSv) suffers from a ‘signal to noise’ problem which limits our ability to conclusively state effects one way or another,” says Kathryn Higley, head of the school of nuclear science and engineering at the University of Oregon in an email.

A single CT scan delivers anywhere from 1 to 15 mSv, but some patients need many scans during the course of their treatment, increasing the total dose. Workers cleaning up after the Fukushima meltdown received radiation doses above 100 mSv in some cases. And current U.S. standards limit radiation workers to no more than 50 mSv of exposure per year.

Many studies indicate that there are dangers at that level, but it’s often an assumption. Those studies base their suppositions on what’s called the linear no-threshold model, which extrapolates more reliable data from studies of higher doses of radiation to lower doses. Though it may be an educated guess, for decades large-scale studies have indicated this is true.

Muddying the Waters

But hormesis researchers say that those studies don’t add up. Mohan Doss, a researcher at the Fox Chase Cancer Center, believes that studies of CT scan patients are biased toward populations that might be predisposed to cancer. Similarly, Doss has a different interpretation of studies looking at Hiroshima and Nagasaki survivors. A 2017 study of bomb survivors actually shows that cancer risk does not increase at low levels of radiation for men, according to Doss. He says it’s evidence for hormesis.

The study’s authors are more measured. They found that women were likely at risk even at low levels of radiation, and they call for more study of what they say is an unresolved question.

Doss says that radiation levels of up to 300 or 400 mSv would likely be safe, and he supports increasing the maximum allowed dose in the U.S.

“The current limits for the public are insane,” he says. “Everyone is going to be safe at 100 mSv a year.”

Doing so would not only save money, he says, but also make it easier to study cancer and Alzheimer’s and encourage more people to undergo chemotherapy, potentially saving lives.

Doss is convinced that studies of low dose radiation do not provide evidence of harms. He points to several studies that indicate the opposite. For example, back in 1957, some 10,000 residents were evacuated from villages near the Mayak Nuclear Weapons Facility in Russia after a nuclear contamination incident. When scientists studied those same residents in 1994, they actually found a dip in cancer rates among people exposed to low levels of radiation.

It lines up with what researchers would expect if hormesis was true, he says, though the study authors do note that more refinement of their methods would be necessary to prove the absence of radiation effects.

Other research into the effects of radiation exposure at Mayak has found evidence of an increase in cancer rates following the event.

The EPA in recent days appeared to back away from the suggestion that it supported hormesis. The agency released a statement in response to the AP story affirming that it intends to continue using the linear no-threshold model when constructing radiation guidelines, something that contradicts Calabrese’s comments in the April press release.

“The proposed regulation doesn’t talk about radiation or any particular chemicals. EPA’s policy is to continue to use the linear-no-threshold model for population-level radiation protection purposes which would not – under the proposed regulation that has not been finalized – trigger any change in that policy,” said an EPA spokesman in response to a request for comment.

Playing It Safe

But radiologist Rebecca Smith-Bindman says the vast bulk of the evidence suggests even small amounts of radiation are harmful. We shouldn’t base our policies on an unproven theory, she adds. 

“There is extensive evidence that ionizing radiation will cause cancer,” says Smith-Bindman, a professor of radiology at the University of California, San Francisco in an email exchange. “These data come from a range of different sources, including epidemiological data (such as studies of patients who have received diagnostic and therapeutic radiation and from environmental exposures and accidents), from animal studies and from basic science studies. While it is more difficult to precisely quantify the exposures — which will vary by many factors, such as age at exposure, and source of radiation, etc. — there is no uncertainty among the scientific community that radiation will cause cancer.”

She says that pointing to issues with the linear no-threshold model misses the point. Though it may not be totally accurate at very low doses, she says it’s unfair to use that uncertainty to cast doubt on data about radiation where there’s solid evidence.

It’s still not clear, for example, how normal background radiation affects humans. That dose averages 3.1 mSv a year in the U.S., though it can vary considerably depending on where you live. Workers cleaning up after the Fukushima power plant spill were exposed to levels far beyond that: Over 150 people received doses in excess of 100 mSv, and six were exposed to more than 250 mSv. That’s two and five times, respectively, the current maximum dose recommended by the U.S. for radiation workers, though still within the range of acceptable short-term doses for emergency workers prescribed by several countries. It’s estimated that the 240,000 workers who helped to clean up the Chernobyl spill were likely exposed to radiation doses about 100 mSv, as well.

When it comes to actual health effects, though, the data is mixed. A WHO report found that risks of cancer from exposure at Fukushima were negligible. Yet cancers often don’t appear until years later, making it difficult to estimate the risks right now. After Chernobyl, workers cleaning up the spill and those from the area saw rates of thyroid cancer spike in the years afterward, though Doss also disagrees with how that data is interpreted. CT scan studies also fall in this range, and multiple groups of researchers have published data showing that the scans increase cancer risk.

Because data from studies isn’t totally conclusive, there’s some disagreement in the field over what constitutes an acceptable level of radiation. Some scientists do think the maximum acceptable doses could be higher.

“In terms of public health, even if we raise the standards to where they were back in the 1960s, we’re not likely to observe an adverse impact with the current scientific tools available to us,” Higley says. “There are populations around the world exposed to natural radiation at levels up to 100 mSv per year, and we haven’t seen measurable health effects.”

Those populations are small, though, so the uncertainties in that data are large, she adds.

Miglioretti disagrees: “Based on the large body of evidence to date, I believe that revising the regulations to increase allowable radiation exposure limits will lead to an increase in the number of radiation-induced cancers in this country.”

That’s in line with what multiple experts Discover contacted believe — that radiation can harm even at low doses and raising limits would endanger the public, though the increase in risk would likely be small.

It’s not clear at the moment whether the EPA proposal to raise limits will pass, though it does follow in the footsteps of other Trump administration proposals to weaken safety standards. At the moment, it’s unclear what the effects on the public if the EPA raises radiation limits.

“Perhaps it might make nuclear power plants less expensive to build. It might lower the cost of cleanup of radioactively polluted sites,” says David Brenner, director of the Center for Radiological Research at Columbia University in an email. “But [it] begs the question of whether cleanup to a less rigorous standard is desirable.”