POSTED 11 AUGUST 2005
How much is too much?
How dangerous is ionizing radiation? You'd think, after 110 years of experimenting with radiation, that scientists could give us a straight answer. Nobody doubts the danger of huge doses, which kill blood stem cells in bone marrow. This "acute radiation sickness" killed some people at Hiroshima and Chernobyl. Nobody doubts that moderate doses can damage dividing cells, cause cancer, or harm genes, causing birth defects.
But should we worry about even lower doses -- the ones we can get from nuclear power plants, nuclear accidents, or advanced medical imaging techniques?
We really don't know. Why?
Radiation comes in many flavors and doses.
The effects of radiation are governed by chance.
People get plenty of cancer from other causes, which can hide additional cases of cancer.
As the effect gets smaller, you need a larger the study to find it.
You can't experiment on people.
Low-level exposure, here, is defined as radiation doses that are slightly above background, such as what you get from a few X-rays, a mammogram, or living near a leaking nuclear waste site. Before you insist these small doses will kill you, remember that something can be dangerous in large doses but innocuous or even helpful at low doses (think oxygen, vitamin A or video games).
Still, the most widespread assumption about radiation is called the linear, no-threshold model. "Linear" because dose is related one-to-one with effect: if dose "X" causes cancer in 1 person in 10, dose X/10 would cause one cancer in 100 people. And "no-threshold," because there is no safe level of radiation. Even small increases over background are dangerous.
The Hiroshima data support linear, no-threshold, says Dale Preston, who spent many years interpreting statistics at the Radiation Effects Research Foundation. "In the A-bomb survivor data, there seems to be some increase [in solid-cancer risk] at any dose, so the linear, no-threshold seems to hold." Much of the evidence for the dose-response curve comes from moderately high doses, he admits, but "at low doses, there is some suggestion of increase; I don't think there is any reason to think there is really a threshold."
The linear model is controversial, even though it does seem logical: If one particle of radiation could hit a cell and cause cancer, wouldn't 1,000 particles be likely to hit 1,000 cells? Yet some scientists question the linear relationship. Animals, plants and people, we know, evolved amid natural radiation exposures, which now average about 0.29 millisievert in the United States, far more than the average person gets from artificial sources. (Total average radiation exposure, adding in artificial sources, comes to 0.36 millisievert.) Background may have been even higher in the distant past, due to the nuclear fissioning of uranium in the early Earth.
That evolutionary history apparently gave us the ability to repair our own DNA, the info-chemical that carries the digital genetic code. But since DNA has been under assault by natural radiation since life started, living things must have evolved a mechanism for repairing damage to DNA due to radiation and other causes.
It's the standard evolutionary situation: you sink, or you swim.
There are hints that a previous exposure to radiation could increase that ability to repair DNA. According to a 2002 article by physicist Bernard Cohen, low-level radiation, "can be shown to stimulate production" of self-repair enzymes. Radiation may also stimulate programmed cell death and immune responses that kill mutant cells. In Iran, Cohen notes, when blood cells from people who lived in intense natural radiation were irradiated in a lab, they had fewer chromosomal breaks (see "Very High Background..." in the bibliography).
The existence of "all of these biologic defense mechanisms" means that "the basis for the linear no-threshold theory is far too simple," Cohen maintained. Or as Arizona State University professor of Medicine Kenneth Mossman told us in 1996, "a small radiation dose has a minimal impact on the total amount of DNA damage that's done spontaneously." In other words, adding a straw to a camel that's already bearing a heavy burden is unlikely to break its back, since the original load has already made it stronger.
Responding to the dose
The existence of a threshold would make life easier for regulators and the nuclear industry: The United States spends billions on radiation protection, and some of it may be wasted, says Mossman. It's hard to study the dangers of low-level radiation, he says, because the "signal" of disease is so low, while the "noise" is so high. About 20 percent of Americans die of cancer, Mossman says, and "With such a large noise factor, it's very difficult to detect a small increase in cancer as a result of radiation exposure from mammograms, nuclear power plant operations, or other routine use of radiation in society. ... Epidemiology is not good enough to detect small risks at low doses, because the spontaneous incidence of cancer is very high in the population."
BTW, we didn't speak to anyone who thought you should skip a medical test out of concern about radiation exposure.
The whole effort to look at the health effects of low-level radiation is lagging, says Armin Weinberg, professor of medicine, and director of the Chronic Disease Prevention & Control Research Center at Baylor College of Medicine. "Given the state of interest in potential nuclear terrorism, it's very important that these questions resurface and be addressed straight up or down," he says. "I think the time for not being willing to openly discuss radiation, both the positive and negative aspects, has long since passed."
Are people irrationally fearful of radiation? Hard to tell, says Weinberg, but he thinks the failure to build a new nuclear plant in the United States since the 1979 accident at Three Mile Island does indicate intense concern about radiation. Compared to Chernobyl, Weinberg says, TMI was probably quite small, "yet it was serious in its consequences, in terms of the societal response to the fear of radiation exposure." Fear may be justified, he says, "But if we don't have a complete science program to answer the question, we are operating with less than a full deck."
Topsy-Turvy, very curvy
We've tried to shield you from graphs, but the argument over low-level radiation comes down to this: What is the shape of the dose-response curve, the one we use to extrapolate from the high exposures that we can study to the low ones that we can't? Is the curve linear, meaning a doubling of the dose would double the toll of disease? Does it bulge at the bottom, so low doses would be more or less dangerous than predicted by the one-to-one model? Or do different cancers have different shapes?
That last idea is a real possibility, argues Mossman, who says bone and skin cancers show "convincing evidence" for a threshold, and "There is good evidence of linearity for breast and thyroid cancer." For leukemia, he sees "excellent evidence" that low doses have less effect than the linear model predicts.
That is three dose-response curves, he maintains, not one, and if you try to compress them into one curve, you could make an expensive mistake.
Mossman says linear, no-threshold is unsupportable. "I have a great deal of difficulty with the notion that the linear, no-threshold can be selected on scientific grounds... It cannot be supported at small doses of radiation, because of the paucity of direct observations at small doses."
Basing health policy on linear, no-threshold might minimize the risk of disease, but it would also raise the price of radiation protection. "What we have done, for purposes of policymaking, is to adopt linear, no-threshold as a general theory that will describe all cancers. If we want to make that decision on a policy basis, that's fine," so long as it's clear that the science does not support it.
Caution: Low ceiling
Instead of today's radiation protection schemes, which set exposure ceilings, Mossman would prefer to make exposures "as low as reasonably achievable.... We want to reduce exposures as far as we can, given our societal, technological and economic capacity."
Since can't know the danger of any particular low-dose exposure, "Risk-based decision-making is difficult," Mossman says. Dose, on the other hand, "can be measured at levels that are orders of magnitude smaller than risks."
But if we don't know the risk of any particular dose, why reduce it? Because we know higher doses bring higher risks, he says. "We say the risk is being reduced, but we don't know how much. We are not making any claim as to what the risk is at a particular dose, because we don't have that information."
Yet in questioning linear, no-threshold, Mossman seems to be in the minority. In the convoluted science of radiation protection, 2005 was a good year for the linear linkage of dose and disease. For example, in July, a study of cancer in 407,000 nuclear workers in 15 countries (see "Risk of Cancer after ..." in the bibliography) associated a slight increase in cancer risk with low-level exposure:
"1-2 percent of deaths from cancer among workers in this cohort may be attributable to radiation. These estimates, from the largest study of nuclear workers ever conducted, are higher than, but statistically compatible with, the risk estimates used for current radiation protection standards. The results suggest that there is a small excess risk of cancer, even at the low doses and dose rates typically received by nuclear workers in this study."
And at least two expert groups agree. Behind a haze of jargon, the International Commission on Radiological Protection seems solid on linear, no-threshold: "For cancer and hereditary disease at low doses/dose rates the use of a simple proportionate relationship between increments of dose and increased risk is a scientifically plausible assumption."
In other words, even at low doses, increasing the dose increases the risk of cancer and hereditary disease.
We read the same story in the 2005 Biological Effects of Ionizing Radiation VII report, which concluded, "A comprehensive review of available biological and biophysical data supports a 'linear-no-threshold' risk model-that the risk of cancer proceeds in a linear fashion at lower doses without a threshold and that the smallest dose has the potential to cause a small increase in risk to humans."
Although linear, no-threshold seems to be having a good year, we'd like to throw a monkey-wrench into these proceedings by quoting one prominent nay-sayer. John Cameron, late professor of medical physics at University of Wisconsin-Madison, insisted that low-level radiation, far from being harmful, was actually beneficial. In 2003, for example, he pointed to studies that showed that people with industrial exposures had actually lived longer than matched controls (see "Longevity Is..." in the bibliography.
British radiologists from the early 20th-century got high radiation exposures, and they developed more cancer than other docs -- but they also lived longer, on average. The same thing arose from a study of 28,000 U.S. nuclear shipyard workers, Cameron noted: "... workers with the largest cumulative doses had a death rate from all causes 24 percent lower than did the 32,000 age-matched and job-matched unexposed shipyard workers."
Where did this whole radiation-health debate start?
Megan Anderson, project assistant; Terry Devitt, editor; S.V. Medaris, designer/illustrator; David Tenenbaum, feature writer; Amy Toburen, content development executive