Ann Coulter says: Radiation is good for you!

Here at Science-Based Medicine, we try to be relatively apolitical. We might not always succeed, but in general our main concern is not so much with right-wing or left-wing politics, but rather with how prevailing government policies and regulation impact the delivery of medical care, in particular whether they tend to prevent, do nothing about, or promote the proliferation of non-science-based medical care. Consequently, when Kimball or I call for the disbanding of the National Center for Complementary and Alternative Medicine (NCCAM), it does not matter one whit to us who is President or which party controls Congress. All that matters is that we see NCCAM as a government entity that, through credulously studying many “alternative medicine” modalities, ends up inadvertently promoting them and providing them with the imprimatur of government approval. The same concept applies to state medical licensing boards licensing pseudoscientific modalities, such as naturopathy, acupuncture, and homeopathy. By regulating these “disciplines,” states also provide them with an unmerited mantle of respectability through their imprimatur of regulating them as professions, just like medicine and nursing.

As far as political views, although all of us have them and they occasionally even come to the fore in disagreements (remember when Wally Sampson occasionally clashed with others with differing political viewpoints?), we generally subsume them for purposes of the SBM blog experience into our advocacy for basing medicine on the best science available. Sometimes, however, when a pundit or politician makes claims that are either contrary to or distort science for ideological or political advantage, I feel the need to discuss those claims, sometimes even sarcastically. Such was the case last week, when Ann Coulter wrote a blisteringly ignorant column, entitled A Glowing Report on Radiation. She wrote this article in the wake of the fears arising in Japan and around the world of nuclear catastrophe due to the damage to the Fukushima nuclear power plant caused by the earthquake and tsunami that hit northern Japan on March 11. Coulter was subsequently interviewed by Fox News pundit Bill O’Reilly on The O’Reilly Factor on Thursday evening:

Yes, according to Coulter, radiation is good for you, just like toxic sludge! Even more amazing, in this video Bill O’Reilly actually comes across as the voice of reason, at least in comparison to Ann Coulter. He’s very skeptical of Coulter’s claims and even challenges her by saying, “So by your account we should all be heading towards the nuclear reactor.”

So, fellow SBM aficionados, is Coulter right? Are all those scientists warning about the dangers of even low-level radiation all wrong? Should we start hanging out in radioactive mine shafts, as Coulter mentions in her column (seriously) in order to boost our health and decrease our risk of cancer?

Not so fast, there, Ann. Here’s a hint: If Bill O’Reilly can lecture you on science and look more reasonable than you, you’re off the rails.

Coulter, hormesis, and “Don’t worry, be happy!”

Actually, the scientific assessment of what levels of exposure to ionizing radiation are dangerous is, as you might imagine, a wee bit more complicated than my little sarcastic rejoinder makes it, but you’d never know that from Ann Coulter’s article and her interview with Bill O’Reilly. The reason for my sarcastic characterization of Coulter’s scientific nonsense is because her article uses many of the same tactics as any denialist. Chief among these is that Coulter takes the germ of a scientific controversy and then uses it to try to imply that the scientific consensus is fatally flawed. In this case, the scientific controversy is over how dangerous low level exposure to radiation is used to imply that the radiation from a nuclear disaster is not potentially harmful. All you former residents of Chernobyl, take note! It’s fine to move back to your homes that you were forced to abandon 25 years ago!

Here is what Coulter claims in her article:

With the terrible earthquake and resulting tsunami that have devastated Japan, the only good news is that anyone exposed to excess radiation from the nuclear power plants is now probably much less likely to get cancer.

This only seems counterintuitive because of media hysteria for the past 20 years trying to convince Americans that radiation at any dose is bad. There is, however, burgeoning evidence that excess radiation operates as a sort of cancer vaccine.

As The New York Times science section reported in 2001, an increasing number of scientists believe that at some level — much higher than the minimums set by the U.S. government — radiation is good for you. “They theorize,” the Times said, that “these doses protect against cancer by activating cells’ natural defense mechanisms.”

What Coulter is referring to is the phenomenon of radiation hormesis. This is nothing more than a biphasic dose-response curve to radiation in which the curve initially goes down with increasing dose (less risk of disease with increasing radiation exposure) and then curves upward and at some point crosses a threshold where radiation exposure is no longer beneficial but harmful with further dose increases. Basically, it’s a scientific model wherein low level exposure to radiation is not only not harmful but in fact beneficial. The reason for this effect, if it exists in humans, is hypothesized to be that low level radiation activates DNA damage repair and other protective mechanisms that are not activated in the absence of radiation; moreover, it is further hypothesized that these mechanisms are activated more than they need to be, so that low level radiation is actually protective against radiation-induced diseases such as cancer.

The radiation hormesis model is markedly different from the currently prevailing model that is used for regulatory purposes by most governments, the linear no-threshold (LNT) model, which states that there is no such thing as a “safe” dose of radiation and that radiation dammage accumulates in a linear fashion with dose. For completeness sake, I will note that there is also at least one other model for the biological effects of radiation, specifically a model in which there is a threshold dose under which radiation is not harmful. In practice, distinguishing between a threshold model and a hormesis model can be very difficult.

In order to give you an idea of what hormesis would look like in a radiation dose-response curve, I stole this graph from Wikipedia. Actually, I didn’t steal it; it’s public domain because it’s a product of a U.S. government agency. However, it illustrates the concept of hormesis quite well:

Curve A demonstrates supralinearity, in which toxic effects are actually more intense per unit of radiation at lower doses; there is no evidence that this is indeed the case. Curve B is linear, and Curve C is linear-quadratic, in which low doses of radiation are less harmful per unit of radiation than higher doses. Curve D represents hormesis, where low doses of radiation are actually protective up to a certain threshold, where the curve shifts from a protective effect to a harmful effect with increasing radiation. The main contenders for the model that best describes radiation effects are either curve B, C, or D.

The key aspect of Coulter’s article that makes it so irresponsible is what she leaves out. What she neglects to mention is that, even if hormesis is an accurate model for radiation effects in humans, it only applies for very low dose exposures. (More on how low in the next section.) True, Coulter does at one point concede that it is “hardly a settled scientific fact that excess radiation is a health benefit,” throughout the rest of her article she presents the idea of hormesis as though it were–you guessed it!–a settled scientific fact. Indeed, Coulter’s earlier assertion that “excess radiation acts as sort of a cancer vaccine” is the sheerest exaggeration, even if hormesis is an accurate model of radiation exposure. Aside from this major exaggeration, how do Coulter’s assertions, which appear to be based largely on studies cited in a single NYT article that is nearly a decade old, stack up against science?

Not very well. Surprise! Surprise! As is the case with many denialists, Coulter takes a germ of actual science and then twists and exaggerates it beyond all recognition in order to support a preconceived notion, namely that those pointy-headed (and, of course, liberal) environmentalists are hiding the evidence that radiation at low doses is good for you. To accomplish this, Coulter cherry picks studies, failing to put them into their proper context with existing research, all for the purpose of advancing her ideological viewpoint.

Radiation hormesis: Ann Coulter’s claims versus reality

Before I discuss what the data regarding radiation hormesis actually show, it’s essential to discuss briefly why it is that the LNT model predominates when it comes to policy-making and setting limits on what is considered “safe” radiation exposure. The reason is not that biased scientists are “hiding” the evidence that radiation is good for you. Rather, it boils down to a few reasons. The first is probably that an LNT model is the simplest, most conservative model that can be fit to currently existing evidence. The problem with the LNT model is the same as the problem with the hormesis model. While at higher radiation doses, effects due to radiation are, like effects due to pretty much any other high-level environmental exposure, much more robust and reproducible, at lower radiation doses, the effects are weaker, and the scatter in the data is much greater. In other words, at low doses the signal-to-noise ratio is much lower due to a lot more “noise” and a lot less signal in the data. Moreover, the data are more difficult to collect, and variability from system to system, organism to organism, and cancer to cancer is likely to be much greater.

As imperfect as it is, the LNT model is a reasonable approximation for purposes of policy-making because it is conservative and safe. Admittedly, there are problems applying such a model when the doses get really low, as in lower than the normal background radiation that we all live in, but it’s a useful approximation. When it is very hard to distinguish between an LNT model and a hormesis model at very low radiation exposures, until better data can be gathered that clearly demonstrate the superiority of one model over another, the responsible and safe model to choose is the most conservative one that fits reasonably well. Basing public policy on a model that, if incorrect, has the potential to result in considerable harm in the form of increased radiation-induced disease prevalence is not wise policy at all, at least when the alternate model is not demonstrably wrong.

As far as Coulter’s reliance on an old NYT article, I thought I’d take a look at the article itself. As an aside, I can’t help but note that I really hate it when the online version of an article doesn’t include links to cited articles, and Coulter is no different in this regard. However, I do believe I managed to find this 2001 NYT article anyway from November 27, 2001, entitled For Radiation, How Much Is Too Much? It’s by Gina Kolata and discusses the controversy that had begun to bubble up about what doses of ionizing radiation might be considered safe. If you read it, you’ll see that it’s much more balanced than how it is portrayed by Coulter. For example, here is what Coulter writes about two studies cited by Kolata:

Among the studies mentioned by the Times was one in Canada finding that tuberculosis patients subjected to multiple chest X-rays had much lower rates of breast cancer than the general population.

Here is what Kolata actually wrote about these studies:

Now, some scientists even say low radiation doses may be beneficial. They theorize that these doses protect against cancer by activating cells’ natural defense mechanisms. As evidence, they cite studies, like one in Canada of tuberculosis patients who had multiple chest X-rays and one of nuclear workers in the United States. The tuberculosis patients, some analyses said, had fewer cases of breast cancer than would be expected and the nuclear workers had a lower mortality rate than would be expected.

Dr. Boice said these studies were flawed by statistical pitfalls, and when a committee of the National Council on Radiation Protection and Measurement evaluated this and other studies on beneficial effects, it was not convinced. The group, headed by Dr. Upton of New Jersey, wrote that the data “do not exclude” the hypothesis. But, it added, “the prevailing evidence has generally been interpreted as insufficient to support this view.”

Notice how the finding in “some analyses” that there were fewer cases of breast cancer than might be expected has magically morphed into “tuberculosis patients subjected to multiple chest X-rays had much lower rates of breast cancer than the general population” in Coulter’s words. Also note that this appears to be the NCRPM report that analyzed the data. Unfortunately, it would have cost me $40 to download the PDF; so I didn’t. But what about these studies?

The first study to which Coulter refers appears to be a study from Canada that was reported in the New England Journal of Medicine in 1989. This study examined the mortality from breast cancer in a cohort of 31,710 women who had been treated for tuberculosis at Canadian sanatoriums between 1930 and 1952. A significant proportion (26.4%) of these women had received radiation doses to the breast of 10 cGy or more from repeated fluoroscopic examinations during therapeutic pneumothoraxes. It should be noted that these sorts of doses of radiation are far in excess of anything likely to be received using modern radiological equipment, in particular given that we no longer perform fluoroscopy and therapeutic pneumothorax to treat tuberculosis. Interestingly, this is how the abstract summarizes the results of this study:

Women exposed to ≥ 10 cGy of radiation had a relative risk of death from breast cancer of 1.36, as compared with those exposed to less than 10 cGy (95 percent confidence interval, 1.11 to 1.67; P = 0.001). The data were most consistent with a linear dose–response relation. The risk was greatest among women who had been exposed to radiation when they were between 10 and 14 years of age; they had a relative risk of 4.5 per gray, and an additive risk of 6.1 per 104 person-years per gray. With increasing age at first exposure, there was substantially less excess risk, and the radiation effect appeared to peak approximately 25 to 34 years after the first exposure. Our additive model for lifetime risk predicts that exposure to 1 cGy at the age of 40 increases the number of deaths from breast cancer by 42 per million women.

Oops! Maybe I found the wrong study! On the other hand, this is a Canadian study that looked at women with tuberculosis who received numerous chest X-rays (fluoroscopy, actually), and I can’t find another one like it. I also couldn’t find other publications with other analyses. The analysis that exists in the published literature, for better or for worse, concludes that the risk of breast cancer is elevated with exposures to radiation greater than 10 cGy. So, what are these other “analyses” that purport to claim that these patients actually had a lower risk of mortality from breast cancer? I smelled a rat.

My first hint came from an article published in the Journal of the Association of American Physicians and Surgeons (JPANDS) by Bernard Cohen entitled The Cancer Risk From Low Level Radiation: A Review of Recent Evidence. I’ve discussed JPANDS and how it plays fast and loose with science for ideological reasons before, in particular its antivaccine views and its publishing studies so bad that laughter is the only appropriate response. In his article, Cohen claims that hormesis “found for breast cancer among Canadian women exposed over longer periods of time to X-ray fluoroscopic examinations for tuberculosis (13); when appropriately evaluated, this evidence shows a decrease in risk with increasing radiation dose at least up to 20 cSv (20 rem).” Unfortunately, no evaluation of this evidence is included; Cohen simply asserts that this is so.

Fortunately, it didn’t take long for me to find other JPANDS articles making the same argument. For example, this one by Joel M. Kauffman. In it, Kaufmann divides up the subjects into several radiation dose ranges, while rejecting data from Nova Scotia because “too few” low radiation points were included. Conveniently he fails to define what “too few” is. However, if one looks at Table I in the NEJM paper, it’s obvious that in the dose range between 10 and 99 cSv, the death rate in Nova Scotia was much higher than the other provinces. One wonders if that had anything to do with leaving out the data, rather than writing the authors for a more detailed breakdown of the data between those dose levels, one does. In any case, what Kaufmann appears to have done is what JPANDS writers frequently do: Cherry pick the data. He took the lower end of the dose ranges, used “eyeball” fitting instead of statistical fitting to models, and left out any hint of a statistical analysis. The authors of the NEJM article went to great lengths to demonstrate that a LNT model was the best fit to their data; Kaufmann expects you to “eyeball” his graph and accept his claim of hormesis. Similarly, Jerry Cuttler and Myron Pollycove, in another JPANDS article, plotted the Canadian data on a semilog scale to make a hormesis effect look far more convincing than the actual data support, all the while simply claiming that a hormesis model fit the data better than an LNT model. Unfortunately, they didn’t “show their work,” so to speak. No discussion of how they modeled the data is included. No wonder the NCRPM found these “other” analyses unconvincing. Also, while it’s not surprising that Coulter would have gotten her data on this from JPANDS, it’s rather disappointing that Kolata didn’t look deeper back in 2001.

The second study cited by Kolata and exaggerated by Coulter was a study of U.S. nuclear industry workers. Regarding this sort of data, the scientists at the Lawrence Berkley National Laboratory have included on their website this analysis:

The results of individual studies have been inconclusive, and to investigate the matter further a combined analysis has been carried out of seven studies–three for sites in the United States (Hanford, Oak Ridge, and Rocky Flats), three for sites in the United Kingdom, and one for Canada. A total of 95,673 workers was included, of whom 60% received effective doses above 10 mSv (1 rem). In the entire population, there were 15,825 deaths, of which 3,976 were from cancer. The comprehensive results for all cancers taken together showed a very slight decrease in cancer rate with increasing dose. However, this result had no statistical significance. Of possible greater statistical significance is a slight increase with radiation dose for some types of leukemia. Overall, the statistical uncertainties were large enough that the analysis did not rule out linearity or any of the other alternative dose-response curves indicated in Figure 15-1–although it does set an upper limit on the possible magnitude of a hypothesized supra-linearity effect.

The study being discussed it this one, which, by the way, concludes:

These estimates, which did not differ significantly across cohorts or between men and women, are the most comprehensive and precise direct estimates of cancer risk associated with low-dose protracted exposures obtained to date. Although they are lower than the linear estimates obtained from studies of atomic bomb survivors, they are compatible with a range of possibilities, from a reduction of risk at low doses, to risks twice those on which current radiation protection recommendations are based. Overall, the results of this study do not suggest that current radiation risk estimates for cancer at low levels of exposure are appreciably in error.

Coulter also makes much of a study of shipyard workers from 1991:

A $10 million Department of Energy study from 1991 examined 10 years of epidemiological research by the Johns Hopkins School of Public Health on 700,000 shipyard workers, some of whom had been exposed to 10 times more radiation than the others from their work on the ships’ nuclear reactors. The workers exposed to excess radiation had a 24 percent lower death rate and a 25 percent lower cancer mortality than the non-irradiated workers.

The reference for this is:

Matanoski, G. M. (1991) Health Effects of Low-Level radiation in Shipyard Workers, Final Report, DOE/EV/10095-T2, National Technical Information Service, Springfield, Virginia, USA.

Unfortunately, I couldn’t get a hold of this report online over the weekend. I did, however, find the more recent reanalysis of the data from 2008 by Matanoski et al published in the Journal of Radiation Research. What Matanoski found wa that most of the differences in mortality and cancer rates found between shipyard workers who serviced nuclear ships and shipyard workers who did not were not significant, although there did appear to be trends towards increased risk of leukemias and other cancers with increasing dose. Overall, as far as saying anything about the association between radiation exposure and cancer, at best this study could be described as inconclusive. Certainly it’s exceedingly thin gruel to make such definitive statements about hormesis. As for the lower all-cause mortality among the nuclear workers, that is almost certainly due to phenomenon known as the “healthy worker effect“; i.e., the selective recruiting of healthier than average persons into the industry who have continued access to better than average health care.

Similarly thin gruel is this claim by Coulter:

In 1983, a series of apartment buildings in Taiwan were accidentally constructed with massive amounts of cobalt 60, a radioactive substance. After 16 years, the buildings’ 10,000 occupants developed only five cases of cancer. The cancer rate for the same age group in the general Taiwanese population over that time period predicted 170 cancers.

The people in those buildings had been exposed to radiation nearly five times the maximum “safe” level according to the U.S. government. But they ended up with a cancer rate 96 percent lower than the general population.

Not exactly. Actually, not at all. It’s not even thin gruel; it’s misrepresentation, either intentional or through Coulter’s laziness in researching the article. Coulter, as usual, is exhibiting willful ignorance by citing old data. In fact, more recent analyses of the Taiwanese population that lived in these buildings do not support her claims at all. The most recent followup study I could find was published in 2006 in the International Journal of Radiation Biology by Hwang et al. The results were:

A total of 7271 people were registered as the exposed population, with 101,560 person-years at risk. The average excess cumulative exposure was approximately 47.8 mSv (range 5 1 – 2,363 mSv). A total of 141 exposed subjects with various cancers were observed, while 95 developed leukemia or solid cancers after more than 2 or 10 years initial residence in contaminated buildings respectively. The SIR were significantly higher for all leukemia except chronic lymphocytic leukemia (n1⁄46, SIR1⁄43.6, 95% confidence interval [CI] 1.2–7.4) in men, and marginally significant for thyroid cancers (n1⁄46, SIR 1⁄4 2.6, 95% CI 1.0 – 5.7) in women. On the other hand, all cancers combined, all solid cancers combined were shown to exhibit significant exposure-dependent increased risks in individuals with the initial exposure before the age of 30, but not beyond this age.

Hwang et al concluded:

The results suggest that prolonged low dose-rate radiation exposure appeared to increase risks of developing certain cancers in specific subgroups of this population in Taiwan.

So, basically, Coulter is completely wrong about the Taiwan incident. There is an increased incidence of cancer in young people, at least, who lived in those apartment buildings. Science is hard, isn’t it? Coulter’s also on seriously dubious footing when she cites Professor Bernard L. Cohen, whose various studies of the relationship between radon and lung cancer buck the established consensus that radon is a risk factor for lung cancer. (Yes, this is the very same Bernard Cohen who wrote the JPANDS article I mentioned earlier in this post; to me his having published in JPANDS is to me a huge hit on any credibility he might have had.) It turns out that Cohen probably didn’t control adequately for smoking in his studies because a reanalysis of his reported data demonstrated similar, strongly negative correlations between radon exposure and cancers strongly linked to cigarette smoking and weaker negative correlations between radon and cancers moderately associated with smoking. No such correlation was found for cancers not linked to smoking. These results strongly suggest that Cohen didn’t adequately control for smoking in his analysis. Another criticism points out that Cohen fell prey to the ecological fallacy and suggested that county-level data probably do not represent the best units to detect a correlation between radon and lung cancer.

Coulter’s final claims center on the Chernobyl disaster and victims of the atomic bombings of Hiroshima and Nagasaki. In particular, she claims that only 30 people died in the plant as a direct result of the disaster and further downplays the risk of cancer in the survivors, stating:

Even the thyroid cancers in people who lived near the reactor were attributed to low iodine in the Russian diet — and consequently had no effect on the cancer rate.

As is usually the case for any scientific claims made by Coulter, this is utter rubbish. Unfortunately for Coulter, her timing in publishing her article was exquisitely bad. On the very next day after her article was published, the National Cancer Institute released the most comprehensive study yet of thyroid cancer in Chernobyl survivors. The findings indicated that radioactive iodine (131I) from the fallout from the reactor was likely responsible for thyroid cancers that are still occurring among people who lived near the reactor and that the risk of this cancer is not declining. In other words, no, Ann, the hugely elevated levels of thyroid cancer among people who live near Chernobyl when the reactor disaster occurred are not due to iodine deficiency in the Russian diet. There is some evidence that iodine deficiency might have increased the risk of 131I-induced cancers, particularly in the youngest, but that’s not what Coulter said. She implied that iodine deficiency could account for the elevated incidence of thyroid cancer among those affected by the fallout. Much more about the health effects of the Chernobyl disaster can be found here. It should also be noted that most people who lived in the area were not exposed to that much radiation according to the United Nations-sponsored team investigating. Most were exposed to about 9 mSv, about 1/3 the equivalent of a CT scan of the chest, abdomen, and pelvis, once the short-term doses to the thyroid were subtracted

Poor Ann. That’s what you get for not doing a bit more research. Basically, every claim she makes in her article can be shown to be either mistaken, grossly exaggerated, or based on old evidence. She even cites Tom Bethell, author of The Politically Incorrect Guide to Science, as a source. Bethell is an all-purpose right-wing science denialist, who, besides viewing scientists as attention whores who trump up alarmist findings in order to secure more research funding and castigates science for its commitment to “materialism,” also denies evolution and anthropogenic global warming. He even rejects relativity and embraces “AIDS reappraisal,” while extending his view on hormesis to argue that hormesis actually protects us from toxic chemicals in the environment that, according to him, we don’t have to worry about nearly as much as environmentalists say we do. In fact, Coulter includes a paragraph in her article that is so unintentionally hilarious that I can’t help but cite it:

Although it is hardly a settled scientific fact that excess radiation is a health benefit, there’s certainly evidence that it decreases the risk of some cancers — and there are plenty of scientists willing to say so. But Jenny McCarthy’s vaccine theories get more press than Harvard physics professors’ studies on the potential benefits of radiation. (And they say conservatives are anti-science!)

I doubt that Coulter appreciates the irony encompassed by this paragraph, given that this paragraph is further encompassed by an article that uses many of the same deceptive techniques of argumentation that the anti-vaccine movement, as epitomized by Jenny McCarthy, likes to use. She then digs herself in deeper by correctly mentioning that Botox is a poison that is safe to use at high doses (Jenny McCarthy loves Botox, actually) and then pointing out the principle that many poisons are safe and beneficial at low doses but dangerous at high doses. If these arguments didn’t occur within the context of her spewing of misinformation, Coulter might actually be making some sense. Too bad she couldn’t resist adding:

Every day Americans pop multivitamins containing trace amount of zinc, magnesium, selenium, copper, manganese, chromium, molybdenum, nickel, boron — all poisons.

They get flu shots.

Perhaps Coulter has more in common with Jenny McCarthy than she would like to admit. Actually, there’s no “perhaps” about it. Coulter will also say whatever fits her political viewpoint. Last week she was ranting about how radiation is good for you. Back in November, she was complaining that the new Transportation Security Administration scanners do pose a “radiological danger.”

These scanners result in a dose of 0.001 mSv for about 5 seconds of full body exposure, and even frequent fliers would be exposed to much less radiation than Coulter is claiming to be just fine. Indeed, Ann Coulter should be lining up to be scanned. After all, that little radiation is good for you!

Is hormesis a real phenomenon?

Despite my irritation, I was rather grateful for Coulter’s article. It did remind me of a rather fascinating debate in radiobiology over what model best describes the biological effects of radiation. Hormesis might indeed be a real phenomenon in humans, but it’s been very difficult to demonstrate. Even one of the best review articles I’ve found that argues for the existence of hormesis as a phenomenon, an article by Tubiana et al entitled The Linear No-Threshold Relationship Is Inconsistent with Radiation Biologic and Experimental Data doesn’t exactly argue for hormesis. Rather, it argues that the LNT model is inconsistent with the data and needs to be modified to more of a threshold model, in which doses below a certain threshold are probably harmless but above a certain threshold start to increase the risk of disease. Arrayed against these sorts of arguments are scientists like Rudi H. Nussbaum and Wolfgang Köhnlein, who call hormesis and the zero-risk threshold dose “scientifically refuted, but stubborn myths.” They even argue that in some cases the risk of low level radiation exposure might well be underestimated. Not surprisingly, in her article Coulter used nearly every myth that Nussbaum and Köhnlein deconstruct in their paper.

Hormesis is clearly an area of science that is as yet controversial. The reason is because it’s difficult to demonstrate definitively one way or another whether hormesis occurs in humans in response to low dose radiation. As I mentioned above, the signal-to-noise ratio for studies of low dose radiation is very low. Moreover, studies of low dose radiation have been conflicting, although we can say with a fair amount of confidence, based on my review of the literature, that, if hormesis occurs, it probably occurs only below doses of 100 mSv. Remember, 30 mSv is the dose received from a CT scan of the chest, abdomen, and pelvis and can be estimated to increase one’s lifetime risk of a fatal cancer by 1 in 1000 to 1 in 500 in pediatric patients, while most people receive around 3 mSv per year from background radiation. To put this all into context, XKCD has a very useful chart that describes how much radiation we receive from various sources. Another good perspective comes from a recent AP article on the topic, which takes a much more balanced perspective.

The bottom line is that we just don’t know whether hormesis is a real phenomenon for radiation response in humans. Lacking that knowledge, we do know that the LNT model is a reasonable approximation for purposes of regulation because it is simple and defensible. Even so, different professional organization bodies have started to question it. For example, the French Academy of Sciences and National Academy of Medicine published a report in 2005 that stated:

In conclusion, this report raises doubts on the validity of using LNT for evaluating the carcinogenic risk of low doses (< 100 mSv) and even more for very low doses (< 10 mSv). The LNT concept can be a useful pragmatic tool for assessing rules in radioprotection for doses above 10 mSv; however since it is not based on biological concepts of our current knowledge, it should not be used without precaution for assessing by extrapolation the risks associated with low and even more so, with very low doses (< 10 mSv), especially for benefit-risk assessments imposed on radiologists by the European directive 97-43.

The Health Physics Society’s position statement, revised in July 2010, states:

In accordance with current knowledge of radiation health risks, the Health Physics Society recommends against quantitative estimation of health risks below an individual dose of 5 rem in one year or a lifetime dose of 10 rem above that received from natural sources. Doses from natural background radiation in the United States average about 0.3 rem per year. A dose of 5 rem will be accumulated in the first 17 years of life and about 25 rem in a lifetime of 80 years. Estimation of health risk associated with radiation doses that are of similar magnitude as those received from natural sources should be strictly qualitative and encompass a range of hypothetical health outcomes, including the possibility of no adverse health effects at such low levels.

Again, we just don’t know. My guess is that hormesis, if it occurs in humans in response to radiation, is not nearly as potent a phenomenon as its adherents claim. My further guess is that the way hormesis is invoked as a scientific explanation for homeopathy doesn’t help its reputation. Be that as it may, until science settles the question, I do know that, contrary to what Coulter claims in her nonsensical arguments, low dose radiation is not a magical “cancer vaccine.” At the very best, low dose radiation might not hurt you or might have some very slight benefits. At worst, it might actually hurt you more than the current scientific consensus accepts. That’s too wide range of possibilities and too much uncertainty to be laying down a barrage of misinformation as intense as Coulter’s.

ADDENDUM: Here’s an amusing little takedown of Coulter’s nonsense, for your edification. I had a good chuckle at this comment by Gordon Bloyer, who writes:

Schultz telling anyone about science, LOL. Coulter wrote an extensive column using back-up from science. Sgt. Ed should learn to read.

Coulter didn’t give her “theory” she cited science. She is right and O’Reilly just shoots off his mouth before he lets others complete a sentence. In this case he had NO IDEA what Coulter was talking about.

I invite Gordon to read my discussion of hormesis and see that the “back-up from science” that Coulter used is anything but. I do thank Gordon, though, for a moment of hilarity in a painful day of grant writing. Ditto felixw, who comments:

I know that O’Reilly went to Harvard. And Ann Coulter graduated cum laude from Cornell, where she founded The Cornell Review before getting her law degree at the University of Michigan, where she edited the law review. Then she clerked for the United States Court of Appeals for the Eighth Circuit.

Given how much a law school education has to do with science, by this logic, I should be able to confidently and definitively make pronouncements on the law! After all, I graduated from the University of Michigan, too, just like Ann Coulter! :-)

Posted in: Cancer, Science and the Media

Leave a Comment (58) ↓

58 thoughts on “Ann Coulter says: Radiation is good for you!

  1. Cosmo says:

    I was in Japan last week. I received low doses of radiation exposure while at the observation deck at Narita airport. Within 2 days I felt better than I have in years. Now I know why. Thank you Anne Coulter.

  2. squirrelelite says:

    A good article, David.

    I’ll try to give it a more detailed look later, especially the article by Bernard Cohen.

    I don’t take anything Ann Coulter says very seriously, but I think her over the top arguments do more harm than good to those trying to support a reasonable, science-based assessment of the risks of low-level radiation doses.

    As a nuclear engineering professor, Bernard Cohen was one of the more aggressive in arguing against the supralinear model (curve A) which got a lot of press in the 70’s and 80’s and against claims that plutonium was the most poisonous chemical of all.

    Unfortunately, it sounds like his JPANDS article may have been written more as a political argument than a serious scientific assessment.

    I read a few years ago about some plans to do some tests on rodent health in extremely low-dose environments (deep underground salt mines, I think), but don’t remember seeing any results. Perhaps one of our commenters can help.

    In the mean time, I am keeping a watch on Japanese efforts to restore power and maintain the safety of the reactors at Fukushima.

    Here is a link to some striking before and after satellite photos:

  3. Fredeliot2 says:

    Is it possible that “6.1 per 104 person years” should be “6.1 per 10^4 person years?”

  4. pdxjoe1966 says:

    Why is it that non-science-minded people put more confidence in something that “sounds” logical over what *is* logical? It seems like producing evidence and reproducible measurements just increases someone’s likelihood to dig in their heals and listen to someone like Coulter because it’s just easier to accept a lie than it is to learn about something.

    I’m always glad to see an impartial ripping of Coulter, so thank you, Dr. Gorski!!

  5. windriven says:

    Coulter, near the end of the clip, admits that she has only ever read the word hormesis and was uncertain as to the correct pronunciation. She has thereby earned the Maynard G Krebs award for journalism. How can a widely published pundit write a screed suggesting that low level radiation might have salutary effects on human health without even speaking directly with a single physician or physicist?

    What do you get if you cross credulity with laziness?

  6. David Gorski says:

    As a nuclear engineering professor, Bernard Cohen was one of the more aggressive in arguing against the supralinear model (curve A) which got a lot of press in the 70′s and 80′s and against claims that plutonium was the most poisonous chemical of all.

    Unfortunately, it sounds like his JPANDS article may have been written more as a political argument than a serious scientific assessment.

    Uh, Cohen published in JPANDS, dude. That’s associating himself with some serious ideologically-formed pseudoscience and/or denialism. There are two most probable explanations. Either Cohen is ignorant of just how loony JPANDS is, or he’s loony enough himself that he doesn’t mind being associated with JPANDS.

  7. qetzal says:

    On top of the difficulty of studying the effects of very low doses, there’s the added challenge that ‘radiation’ is not just one thing. There are different types of nuclear radiation – including alpha and beta particles and gamma rays. Then there’s the added complication of exposure ‘route.’ You can be exposed to radiation from a distance, e.g. if you are near the Fukushima reactors. But you can also ingest radioactive particles, e.g. if a reactor explodes and sends material into the atmosphere, or if it gets into the drinking water.

    I suspect it’s useless to argue whether there is hormesis from radiation unless one specifies a lot more detail about the type of radiation, the route & duration of exposure, etc.

  8. Joe says:

    Thanks for a great article, everything we wanted to know about hormesis but were afraid to ask. One wonders about Coulter’s point; is she in favor of nuclear power? Or did she just learn an unusual, new word and had to crow about it. Maybe we could get her a dictionary keep her occupied with something other than her usual fare.

  9. David Gorski says:


    Good point. I had wanted to get into that a bit, but my post, as happens far too often, was already way longer than I had originally intended, and I desperately needed to get back to grant writing all day Sunday. For instance, there’s at least one paper I know of that shows that alpha radiation can actually cause adverse bystander effects in nearby non-irradiated cells.

  10. Watcher says:

    Just a note to all the people who are interested in this. Xkcd has a great chart up that compares all doses from sleeping next to someone (lowest) to standing next to the Chernobyl reactor after meltdown (highest).

  11. David Gorski says:

    You didn’t read my post very closely, did you? I included a link to that chart in the body of my post. :-)

  12. passionlessDrone says:

    Hello friends –

    No surprises that Coulter is clueless.

    Tangentially, from todays xkcd re: radiation.

    – pD

  13. David Gorski says:

    True, but her cluelessness gave me an opportunity to discuss hormesis, a topic I’ve been wanting to cover for some time on SBM, given how often it shows up in alternative medicine circles. :-)

  14. Watcher says:

    Aww, i did miss it :)

  15. Harriet Hall says:

    I heard a report that increased radiation had been measured in California: one-millionth of the background level! Hardly enough to make you want to rush out and buy potassium iodide.

  16. Mojo says:

    @Harriet Hall

    Hardly enough to make you want to rush out and buy potassium iodide.

    If anyone does, though, I’m sure there will be plenty of people willing to sell it to them at a special price.

  17. Joe says:

    If you go to and search “chernobyl” you will find some interesting stories. My favorite concerns a biologist who is studying animals that have been thriving there since the disaster 20 years ago. Moose and wolves seem to be fine.

    Most striking were his observations of voles. Those little guys produce a lot of generations per year and one would expect to see accumulated genetic damage; but he hasn’t found any such signs (yet) despite the voles making Geiger counters “scream.” He has no idea why there is no apparent damage.

    David Gorski on 21 Mar 2011 at 11:41 am noted that sCAM loves hormesis. When I told my brother about this story a couple days ago (PZ Myers linked it) his first comment was “Homeopaths must love hormesis.”

  18. Here’s an interesting chart that shows the spectrum of radiation exposure….

    nah! just kidding.

    I do dig that chart though.

    I also think it’s funny that Coulter was on about cancer but, it sounds*, like she never mentioned risks of birth defects.

    *I can’t watch the video…I have a strict no Coulter consumption policy. She gives my nightmares, indigestion, paranoid hallucinations of a future America, sort of thing.

  19. vicki says:

    As far as I can tell, the reasonable “are those airport scanners safe?” questions are based either on not trusting the TSA claims about radiation levels, or concerns specifically for the low-level, poorly paid employees who run the scanners.

    On the first point, the TSA has admitted to not checking things closely. At this point, it would be reasonable to wait for independent audit data; the TSA has little if any credibility.

    The second concern is about improper shielding of the machines, and cumulative dosage for an employee who runs a hundred scans a day, standing near the machine each time.

    W.r.t. hormesis, even if there is such an effect, the natural dosage some of us get might already be as much as is good in that sense.

  20. Jan Willem Nienhuys says:

    This article is very interesting, but keeping track of all these vastly different amounts is difficult. The chart at

    show effects over a range of one billion (9 orders of magnitude). This makes any discussion hard. Lert me explain.

    On the quoted site
    the article ‘Radiation Effects at Low Doses’ states conveniently:

    risk of eventual fatal cancer: 0.05 per Sv

    But the chart says that the normal daily dosis from natural sources (10% = the potassium in our bodies, cosmic rays are another source) is only 0.000 01 Sv. So what is the risk of living 110 years? That’s about 40,000 days. The total dose of 110 years living is 0.4 Sv. Applying the above rule gives the risk of eventual cancer due to radiation from natural sources as 0.02.

    The actual risk of cancer is much larger, something like 0.5. So in effect this statement seems to claim that all natural sources of radiation together cause only a small percentage of cancers. How can they know? Can it be that this computation is wrong? Or does this apply only to doses to only part of the range between 0,01 times the normal daily background and 100,000 times the normal daily background? Is 13 years living in Denver, Colorado rather than in Los Angeles really comparable to a single chest CT scan?

    It is unclear how doses lower than the natural background can give rise to hormesis. It is not that the DNA repair mechanism is turned on, because it is all the time working already.

    One can imagine that the no threshold linearity hypothesis (LNT) fails at very small doses. After all, falling from 1 meter is not 1/10 as lethal as falling from 10 meters. Taking one tablet tylenol does not have 1% of the lethality of taking 100 tablets. So in many situations the LNT fails in the region corresponding to the normal range that the body can deal with.

    It is quite thinkable that the LNT holds for ionizing radiation, but I would like to see more proof of that.

    If scientists can spend millions to catch neutrinos in caves deep below ground (shielded from cosmic rays), then it should not be too hard trying to see what happens to animals that spend their lives in places shielded from normal background radiation.

  21. daedalus2u says:

    The effects of hormesis are real. The effects of radiation hormesis are real. But as qetzal points out, radiation isn’t “one thing”, and what radiation does isn’t “one thing” either.

    We do know that what ever hormesis is doing, it is doing through physiology. If physiology could have adapted a certain response to a certain radiation dose, it could have also adapted a higher response to the same radiation dose. Evolution did not configure physiology that way. Why is that? Presumably because there are adverse effects associated with the seemingly positive effects of hormesis.

    Without knowing the physiology behind hormesis, it is not possible to understand it. Without understanding it, it cannot be the basis for regulating exposures.

    LNT is very likely not correct for any type of radiation. All radiation is stochastic. A photon, electron, neutron or alpha particle hits something or does not. An ionizing trail generates free radicals in certain locations or not. A DNA strand is broken at a certain place or not. The damage is repaired or not. None of these processes are linear, the cumulative effects of many events can’t be linear.

  22. qetzal says:


    I saw some of that same work. Interestingly, the researcher in question found that voles around Chernobyl seemed to have high levels of variation in their DNA sequences. At first, they thought that was due to increased mutation rates caused by the high background radiation levels (link). However, they later found that was wrong. Neighboring populations of voles in non-radioactive areas had similarly high variability (a href=”″>link).

    Another complicating factor – various mammals appear to be thriving around Chernobyl, relative to their previous abundance, but the data suggests that’s mainly because there aren’t many humans around there anymore.

    I found all of that really interesting, in part because it further emphasizes how difficult it is to determine cause-and-effect relationships in the real world, and how confounding effects can easily mislead even experienced scientists.

  23. Jan Willem Nienhuys says:

    @ daedalus2u on 21 Mar 2011 at 6:15 pm

    The effects of hormesis are real. The effects of radiation hormesis are real.>/blockquote>

    I don’t see where the data about radiation hormesis come from. David Gorski wrote that

    if hormesis occurs, it probably occurs only below doses of 100 mSv

    I find even that hard to believe. 100 mSv is about 10,000 times the daily background of 0.01 mSv (i.e. 10 microsievert). That is huge, it represents about 30 years of background radiation. Such a dose probably should be spread out.

    It just might be possible that a few times the normal background doesn’t tax our cells’ capacity for repair, but that anything more for extended periods might be good for you strains credulity so much, that there should be solid proof for it.

  24. Charon says:

    LNT is very likely not correct for any type of radiation. All radiation is stochastic.

    This… is a pretty silly argument. It’s like arguing that water isn’t really a liquid, because a single molecule of water won’t take the shape of the container you put it in.

    We know radiation is quantized. So is matter. So, in all likelihood, are space and time. But continuous approximations are really quite good. And I say this as someone who’s day job is largely about counting single photons. Yes, I have to use Poissonian models, but I’m counting UV photons from across the universe, and we get something like one source count per five minutes, sometimes. And it’s still a pretty darn good approximation that twice the exposure time = twice the source counts. (For comparison, one count every five minutes is what you get from ~190 micrograms of U238. Or half an attogram of I131.)

  25. Charon says:

    I also say that as someone who can’t properly distinguish between “who’s” and “whose”, apparently. Seriously, when are we going to get the ability to edit our comments?

  26. The Blind Watchmaker says:

    Since when does Ann Coulter care what scientists say? She rejects scientific consensus left and right when it is inconvenient to her dogmatic beliefs about creationism and global warming. She cherry picks some obscure science references that even Dr. Gorski had trouble digging up to support her pro-nuclear stance.

    Don’t get me wrong, I think that there will be lost potential for safe modern-generation nuclear energy over this, but I do not think that extrapolating fringe, misunderstood quasi-scientific ideas into dangerous public health advise is warranted or responsible.

  27. Steve Packard says:

    I’m no fan of Ann Coulter but if what she is saying brings some well deserved attention to the evidence that has mounted against LNT then it may do some good.

    The whole LNT curve is based on the presumption that you can extrapolate the dose result ratio all the way down to zero and it’s going to be a straight line. There’s simply no empirical evidence for this.

    Of course, if you follow this idea, then the only “safe” thing to do is for us all to eat special food that has been isotopically separated to remove all carbon-14 and potassium-40 and then to move into lead-lined caves, far from anything organic and all contact with the atmosphere or any kind of mineral that might contain a tract of thorium or uranium.

    We should never ever fly in an airplane, eat a banana, cook with natural gas, watch a CRT television, get a dental x-ray, associate with other human beings, live in a granite building or go outside, because all these things increase our level of radiation exposure, and that’s dangerous (THERE IS NO SAFE LEVEL!)

    Now aside from the fact that this is completely absurd, there’s also the fact that no actual effect has ever been shown to occur bellow about 100 mrem, and not only that.

    There is, however, some interesting evidence to suggest that this is not the case. Not only statistical studies of persons living in higher background radiation areas but also some microbial and in vitro studies that were done at the Chalk River Laboratories in the 1950’s. If LNT were valid, one would expect there to be slightly higher levels of cancer in Dever Colorado than another city of comparable demographics at a lower altitude – that is not the case!

    Studies have been done on a population in Taiwan that was inadvertently exposed to levels of ionizing radiation well above normal average (though not extremely high by most standards, higher than normal) due to cobalt-60 contaminated structural steel. They actually had a small but statistically significant reduction in cancer.

    The most famous case is the area around Ramsar Iran. This is a region which has extremely high background radiation as the result of a unique local geology that causes radium to be found in springs. There’s high levels of radon as well. The leves are pretty damn high. Not high enough to give radiation poisoning, but high enough that if a nuclear reactor malfunction produced them, there’d be a congressional investigation and people would be tearing their hair out in panic.

    Living in Ramsar Iran gives you a dose of about 2.6 REM per year. That’s a hell of a lot. That’s would be considered to be absolutely intolerable for any nuclear plant worker to be exposed to. Again, not enough for acute radiation poisoning, but still a lot.

    So the population of Ramsar has a high cancer rate, right? wrong. They have a lower than average cancer rate.

    Yes, there have been efforts to account for every confounding factor possible. Demographic comparisons, breakdowns looking for the “healthy worker effect” Nope. There’s no other explanation out there for why the rate should be higher, but it’s not. it’s lower.

    There are several other cases. Cancer rates amongst those exposed to fallout from nuclear weapons tests in the Marshall Islands. Yes, the highest exposed persons had a high cancer rate, but those exposed to radiation above normal levels but bellow a certain level had lower.

    So we’re left following a model as if it were known to be true, despite the fact that it was cooked up in the 1940’s for lack of any other data and to provide a “worst case” to make estimates. Yet many hold to LNT the way some hold the bible: it’s an article of faith and unchallengeable!

    None the less, the LNT model, which was formulated for lack of any other data, has been used to attack every type of exposure to radiation regardless of how small. We’re told that CT scans are killing thousands. is there any empirical evidence of this? Of course not! but you multiply the presumed number of tumors per millirem by the exposure of a CT scan times the number of scans performed and there you go, you have your alarmist headline.

    (meanwhile the public is sent running in fear of life saving procedures ranging from x-rays to thyroid uptake tests because of that radiation – NO LEVEL IS SAFE! RUN FOR YOUR LEAD LINED CAVE!)

    In reality, there’s a legitimate question about whether there is benefit. It’s admittedly very hard to measure an effect which is so tiny, but based on the available evidence there’s at least as much scientific data to back up that it is.

    Of course, nobody will deny that very high doses are very bad for you. However, in the case of the Fukushima nuclear plants, we’re not talking about that. It’s not the kind of thing that is going to expose the public to multiple REMS of equivalent dose. It might (at worst) expose them to a few millirems.

    So the question to ask is this: If the average anual exposure to ionizing radiation is 360 mrems, with some people being on the low end (maybe 150 mrems) and some being on the high end (maybe 500 mrems) who is better off? One could make a pretty good argument that, on average it’s the higher one. You might be better off with a dose of 2 rems, which is way way way higher than what Fukushima is going to expose anyone to.

    Dr Ted Rockwell said of LNT that “a lot of money gets spent on trying to reduce even the lowest radiation levels. That money doesn’t go down a rat hole. It goes into a rat’s pocket”

  28. JMB says:

    @ micheleinmichigan

    I attended to “Medical Effects of Nuclear Weapons” course about 30 years ago. In regards to the question of birth defects in Hiroshima and Nagasaki, the statement was made that the incidence of birth defects decreased. The decrease was attributed to the increase in spontaneous abortions related to the physiologic stresses. An unhealthy fetus was more likely to not survive the stress. However, it is unclear to me whether microcephaly was classified as a birth defect at that time.

    This is a reference to the course, but there is no data presented on the website to affirm my memory.

    It should be noted that environmental stresses included fires, physical hazards (broken glass, etc), and breakdown in sanitation, not just radiation. I believe more people died of the breakdown in sanitation than eventually died of radiation exposure.


    The xkcd chart puts many things in perspective, but may add to some of the confusion over cited measures in Sieverts. Some calculations of Sieverts consider tissue weighting factors and radiation quality factors, others do not. The following points out the use of tissue weighting factors and radiation quality factors in calculation of risk of radiation exposure.'Connell.pdf

    Note that the radiation quality factors will differentiate alpha, beta, and gamma radiation (as well as the spectrum of gamma radiation).

    The chart does not highlight the point that radionuclides that are inhaled or ingested may produce radiation to organs for the lifetime of the person, not just the hours of exposure. In spite of the low numbers for radiation exposure in one day for cities in the vicinity of Fukishima, monitoring of the food and water supply is still important.

    A minor second point is that the quoted measures of 3 mSV for a mammogram exam is somewhat inconsistent with other estimates of effective dose of CT exams, chest x-rays, and background radiation. The following citation lists calculation of radiation risks considering both quality of radiation and tissue weighting factors,

    From that cited source, the effective dose of the mammogram exam is 0.4 mSv. The effective dose of a CT chest is about 7 mSv. The effective dose of a single view CXR is 0.1 mSv. The effective average background radiation dose is 3 mSv. The only significant variance is with the mammogram dose.

  29. Jan Willem Nienhuys says:

    @ Steve Packard on 22 Mar 2011 at 1:53 am

    Firstly, all these different units are very confusing. Please try to stick to one unit, say the sievert (1 rem = 100 sievert).

    Secondly, please try to read the post you discuss. You remark:

    Studies have been done on a population in Taiwan … due to cobalt-60 contaminated structural steel

    but that is what Coulter extensively discussed, and which was rebutted by Gorski, who wrote:

    So, basically, Coulter is completely wrong about the Taiwan incident

    giving detailed reasons. In the part he quoted you can see that in Taiwan there was no single average dose, but

    The average excess cumulative exposure was approximately 47.8 mSv (range < 1 – 2,363 mSv).

    (I have corrected a copy-paste error) Note the upper limit which corresponds to a dose listed as causing ‘severe radiation poisoning’ in that convenient table. 400 mSv in a single dose is already very dangerous, but if you spread it out it is plausibly having less effect.

    There is a lot to be said against the LNT hypothesis, namely that the fact that a single dose is much more dangerous than the same amount spread out over a longer time. This effect is seen in many different circumstances. It may even be the basis for the hormesis illusion.

    Ionizing radiation damages DNA. That is why it is carcinogenic. But our DNA is being damaged all the time. I have read estimates of 10 lesions per second in each cell of our body (not all due to radiation, free radicals are another source of damage and radiation actually has the effect of creating many free radicals). It seems plausible to me that permanent damage increasing the risk of cancer has a large chance of occurring when the repair mechanisms are overwhelmed, analogously to the fact that the body can handle small amounts of poison or injury, but not big amounts.

    So it is entirely possible that 1 chest CT scan 5800 (microsievert) has much more effect than a 150% increase of the background, which according to the LNT hypothesis amounts to the same.

    If you try to estimate the effects of high doses received incidentally and then linearly extrapolate to low continuous doses then you may find that the low continuous dose seems to protect.

    If this conjecture (a high dose at once has more effect than when it is spread out) is correct, you will have to adjust the way you calculate the cumulative effects from both incidental doses and continuous doses. It won’t be easy, I guess.

    That is maybe the reason why the Health Physics Society (as quoted by Gorski) said that they advise against trying to express risks into numbers if one is talking about doses lower than (1) 15 times the background accumulated in one year or (2) 40% more than the background accumulated over 80 years.

    It is all quite complicated and it doesn’t really help if Coulter says ‘in Taiwan there were only 5 cases of cancer’ and Gorski rebuts that, and if the Taiwan case then is brought up without any reference as

    They actually had a small but statistically significant reduction in cancer

    The same holds for the story that living in Denver is so healthy. Gorski mentioned the Nussbaum – Köhnlein article (not dated, but most recent references from 2001), which explains that after you correct for various factors nothing is left of the health effects of living in Colorado.

    Maybe one thinks that Nussbaum and Köhnlein are incorrect, but then you should explain why, not just asserting again

    one would expect there to be slightly higher levels of cancer in Dever Colorado than another city of comparable demographics at a lower altitude – that is not the case!

    without any reference.

  30. Jan Willem Nienhuys says:

    “1 chest CT scan 5800 (microsievert) has much more effect than a 150% increase of the background,”

    make that:

    1 chest CT scan (5800 microsievert) has much more effect than a 150% increase of the background during a full year,

  31. Nescio says:

    I find the late John Gofman’s arguments about low dose ionizing radiation interesting. He asserted that even a single electron track through a cell can cause damage to DNA that can lead to cancer. Since DNA repair mechanisms are not 100% effective, “Some carcinogenic injuries are just unrepaired, unrepairable, or misrepaired.”

    I think it’s worth quoting Gofman’s conclusion from the link above in full:

    ” It is factually wrong to believe or to claim that no harm has ever been proven from very low-dose radiation. On the contrary. Existing human evidence shows cancer-induction by radiation at and near the lowest possible dose and dose-rate with respect to cell-nuclei. By any reasonable standard of scientific proof, such evidence demonstrates that there is no safe dose or dose-rate below which dangers disappear. No threshold-dose. Serious, lethal effects from minimal radiation doses are not “hypothetical,” “just theoretical,” or “imaginary.” They are real.”

    That said, the carcinogenic effects of low dose ionizing radiation are very difficult to differentiate from background cancer rates. The wildly differing estimates of the number of excess cancers that have been and will be caused by Chernobyl are evidence of that. Greenpeace suggested, “nearly 100,000 fatal cancers”. The IAEA suggests 4000 deaths in total.

  32. JMB – in terms of birth defects and radiation. I was thinking of areas such as Semey (Semipalatinsk) Kazakhstan where the Soviet Union ran nuclear weapons testing*. (abstract only article in russian)

    I do realize that is not an appropriate parallel to Fukushima, but I think, to the average person, birth defects are just as much a concern as that of cancer, so it’s funny to me to see Coulter only talking about cancer. It just reinforces that she is spinning some carefully selected scientific claims to advocate an ideology rather than seriously attempting to balance risk/benefit.

  33. daedalus2u says:

    Charon, the premise of the LNT model is that the effects of radiation can be extrapolated. There is a big difference between interpolation and extrapolation. If you are in a fairly linear region, even highly non-linear effects can be interpolated, but only if you know you are in a fairly linear region. Extrapolation outside a region where you have data is not justifiable unless you understand what is going on pretty well and know that the underlying stuff is linear.

    In the case of water, we know that water is composed of molecules that interact, we know the size of those molecules and the length scale at which they interact. We know when we can use a continuum model of water (as a liquid) and when that model fails.

    We know that there are many effects of radiation, and that many of those effects are not understood. Without knowing the physiology behind those effects we can’t know how they change with dose, and dose rate.

    If there is such a thing as radiation hormesis, it is mediated through physiology and that physiology takes time to occur. It isn’t “magic” that some doses of radiation are beneficial and some doses are not; it is that there is physiology that adapts the organism to an exposure and that adaptation has other effects.

    It is like exercise. It is not the exercise per se that has beneficial effects, it is the compensatory pathways activated by that exercise that have beneficial effects. If you used an LNT model for excercise-like exposure, running two marathons back-to-back every 2 months would have the same health effects as running the same total distance at the same pace a little bit every day. We know that doesn’t work.

    A LNT model doesn’t work for exercise and we know it doesn’t work because we understand the underlying compensatory physiology is non-linear. We should expect a LNT model for radiation to not work either, but because the underlying physiology is not well understood, and the dose and dose rate in individual tissue compartments can’t really be measured, regulators have defaulted to a LNT.

  34. Scott says:

    I like this short version:

    LNT is almost certainly inaccurate to some extent. However, we don’t know what the right answer is so we have to make some assumptions.

  35. Joao says:

    David, I think you are dismissing Bernard Cohen’s studies way too quickly. On this web page where he has some of his publications available, I found that right on the first paper he addresses both the ecological fallacy and the radon vs. smoke inverse correlation.

    Regarding the ecological fallacy, my understanding of what he’s saying is that if we know that a LNT model applies for any substance, let’s say aspirins, and we know how much of this substance is a Lethal Dose, let’s say 100 aspirins, than we can conclude that if a million aspirins (10.000 lethal doses) are taken by 10.000 people (1 LD/person) or by one million people (1%LD = 1/100 chance of dying /person) or even if you distribute unevenly the million aspirins over a population, the resulting number of deaths will always be equal to the number of lethal doses taken. This implies that one cannot commit the ecological fallacy under a LNT model. Of course, as Bernard Cohen notices himself

    “…I have never claimed that my data mean that radon reduces lung cancer; that would be an application of “the ecological fallacy”. If LNT fails, I can’t logically interpret my data in quantitative terms. If LNT is assumed to be correct, its predictions are grossly discrepant with my data. Putting these two things together, I conclude that LNT fails”.

    On this same subject, there is a video presentation at (first video) where he explains his findings in some detail to a live audience. I would encourage everyone to take a look.

  36. David Gorski says:

    I’ve read Cohen’s studies, and I’ve read some of his rebuttals to his critics. I remain underwhelmed, but explaining why would probably require a separate post. Brief version, I don’t think Cohen fully understands the ecological fallacy. No surprise there. He’s not an epidemiologist. He also strongly implied, at the very minimum, that radon gas is nothing (or at least very little) to worry about.

    Also, Cohen strikes me as being a just wee bit disingenuous when he claims he’s not arguing for hormesis, given some of the statements I’ve read by him. Actually, if all he had argued for was a threshold model rather than the LNT model, there would have been little for me to complain about, whether I disagreed with him or not, but time and time again he appears to imply, if not explicitly argue, for hormesis. Of course, his publishing in JPANDS didn’t do much to burnish his scientific credibility on this issue, and I really wish he hadn’t done that. He didn’t need to.

  37. Steve Packard says:

    @Jan Willem Nienhuys

    I’m an American, which means I think in rem’s, which are still considered the standard unit by the NRC. And really, thinking in rem’s is the least of my problems because I also think in pints, quarts, inches, feet, miles, Fahrenheit and pounds.

    So every time I have to talk in seivets I am converting in my head, which granted is fairly easy with sievert, but still.

    But I’ll go one further: I dislike the sievert as a unit, even despite the fact that it is considered the Standard International unit, it’s an inferior unit to the rem.

    There was no reason to create the sievert, as the rems works fine and also conforms to metric prefixes (as in millirem). The REM was established and has been the unit that health physics instruments have read out in for decades. The sievert was concocted in 1979. The fact that it’s simply 100 REM makes it totally redundant and unnecessary.

    What makes it truly inferior is that it’s too large. Rems are the perfect unit division to measure radiation exposure because single rem’s are well indexed to what might be considered “high” radiation exposure and millirems are well suited to measuring radiation exposure from relatively small things like X-rays or annual exposure.

    You will never really encounter a situation that requires the use of kilorems (because that’s beyond survivable) or microrems (except perhaps in evaluating short-term ambient dose level).

    Sieverts are too large to use in any but extreme circumstances. Even then, they don’t offer enough precision unless used as fractions. Therefore, in nearly all cases, sieverts must be used as microsieverts or millisieverts. This is where the inferiority of the unit is most acute, because the two divisions are VERY easy to confuse, especially by those without experience.

    Therefore, in conclusion, while most of the time SI units are superior to the traditional English/Imperial/Traditional American units, in this case it is the opposite. The sievert is inferior to the rem and I will continue to use rems, as will the entire US nuclear industry and the US government.

  38. Jan Willem Nienhuys says:

    First of all, I think it is a matter of politeness to adapt your choice of units to the context. Here the context is gray and sievert. In this case the original article had already a confusing collection of numbers. Especially for people used to rem and rad the unit cGy (1 centigray = 1 rad) was used.

    Of course, if someone is suffering from megaarithmophobia (fear of large numbers) then nothing can be done. This is an incurable psychiatric ailment.

    The reason to use the SI units throughout is that if one takes as basic units the meter, the kilogram, the second, the coulomb (a very impractical large unit for static electrical charge, but 1 coulomb per second = 1 ampère) and the kelvin, then one doesn’t have to keep track of conversion factors in computations. So the natural unit for absorbed radiation dosis is 1 joule per kilogram.

    Many people don’t realise this is THE main advantage of using SI. A simple example: it has been raining, and the rainfall is recorded as 13 mm. How much water fell on an area of 2 by 3 km?
    Easy: 13 x 10⁻³ times 2 x 10³ times 3 x 10³ = 65 x 10³ m³ ; as 1 cubic meter of water happens to weigh one ton or 1000 kilogram, the answer is also 65 x 10⁶ kg.

    Now do a similar computation with as data: 5 inches of rain, area is 1 by 2 miles. Answer please in gallons and pounds. No sneaky detour allowed, using things like 1 inch = 0.0254 meter and 1 mile = 1609.344 meter and 1 US liquid gallon =0.003785411784 cubic meter and 1 avoirdupois pound = 0.45359237 kg .

    Undoubtedly huge amounts of time is wasted in US elementary schools on learning how to do this type of problems, or worse, it is skipped because too difficult. (In European schools, the simple trick of shifting the decimal point already requires a lot of practice.) The result is that when I see Randi explaining about homeopathy, he first has to explain what the funny little 3 in 10³ means. BTW, do people in the US nuclear industry and government understand exponents?

    It is the interconnectedness of all units that is the main advantage of the decimal system, and that is also the idea behind the SI-principle that each derived unit should be 1 if expressed in kilogram-meter-seconds-coulomb-kelvin.

  39. daedalus2u says:

    Steve, no. I agree with Jan, the main reason for using SI units is that the units are all made up of “the same” sub-units, so when you calculate dimensionless numbers no conversion is necessary because the “units” all cancel.

    For example the Reynold’s Number is D*V*rho/mu (diameter*velocity*density/viscosity). If I use the SI units for everything, then the units cancel and I get the Reynold’s Number with no conversion.

    I am an engineer, and often use “engineering” units like btu (mostly because that is what the data in my reference books is in), but whenever I calculate a dimensionless number I always use SI units.

  40. Jan Willem Nienhuys says:

    Oops! make that:

    Easy: 13 x 10⁻³ times 2 x 10³ times 3 x 10³ = 78 x 10³ m³ ; as 1 cubic meter of water happens to weigh one ton or 1000 kilogram, the answer is also 78 x 10⁶ kg.

  41. squirrelelite says:

    The Department of Energy has an ongoing Low Dose Radiation Research Program spending about $20 million per year.

    I did a few searches looking for population studies in long term ultra-low dose environments (below surface average of about 4 millisieverts per year), but came up empty.

    Most of the studies seem to be looking at various effects of short-term (prompt) exposures.

    Some are finding that there may be adverse effects at levels previously assumed to be safe.

    For instance, Kleiman et al note that:

    Preliminary data suggests that there may be an even lower threshold dose for cataractogenesis, if one exists at all. Extension of the presumed radiation cataract threshold in animal models to even lower doses is likely to be important to the development of appropriate guidelines for radiation risk.

    Thus, new data from animal models and from exposed human populations suggests that lens opacities occur at doses far lower than those generally assumed to be cataractogenic and these observations are consistent with the absence of a dose threshold. Given that all national and international risk standards for ocular exposure are predicated on a relatively high threshold, current guidelines for ocular radiation safety require reassessment.

    Whereas Edouard I. Azzam has done several studies that suggest a possible hormesis effect:

    exposure to low dose/low dose rate γ-rays can protect normal human and rodent cells against oxidative/clastogenic damages induced spontaneously or by a subsequent challenge dose of ionizing radiation

    Either way, these results will be important for any future changes in radiation protection standards.

  42. JMB says:

    In regards to the debate about rems and Sieverts, for many applications 1 Sv = 100 rem. However, rem was calculated by the radiation exposure times a quality factor that was based on the type of radiation (alpha, beta, or gamma, and the spectrum of gamma). Within the SI definition of Sieverts used for effective dose, Sv may be calculated from Gy using both a factor for the type of radiation, and a factor for the body parts included. When the entire body is exposed to the radiation, then 1 Sv = 100 rem. When only a part of the body is radiated, then the tissue weighting factors can be used with the results expressed as Sieverts. Here is a reference for calculation of Sieverts, note the observation that 1 Sv is estimated to cause 5 excess cancers per 100 people in a lifetime.

    The definition of Sievert is conducive to numerical representation of radiation risk, rem is more directly tied to tissue absorbed energy. Unfortunately, you have to be careful to read the methods sections of papers to determine if the paper is reporting the effective dose, or the absorbed dose when they are expressed in Sieverts. 1 Sv is inconveniently large because it represents the approximate threshold of when radiation induced cancer can be detected by epidemiology studies.

    Most radiation exposure from background sources or nuclear accidents are considered total body radiation. Radiation from ingestion of radionuclides resulting from the accident may be calculated differently (see committed dose in the above reference). When the effective dose is used, there is a quick estimate of radiation risk… multiply mSv’s times 5 in 100000. So if the references in the xkcd chart were selected for effective dose, then those numbers can be converted into risk of lifetime cancer. Note that the 3 mSv figure for mammography is not the accepted effective dose as I pointed out in a previous comment.

  43. squirrelelite says:


    I appreciate your comments. I’ll try to follow up on this discussion by looking for good illustrative cases where there is an important difference between dose calculated in Sieverts and dose in rems.

    The addition of tissue weighting factors may have been part of the original definition of the Sievert, but that seems to have largely been dropped as noted in Wikipedia:

    Historically, the weighting factors for radiation type and tissue type were separated out as Q and N respectively. In 2002, the CIPM decided that the distinction between Q and N causes too much confusion and therefore deleted the factor N from the definition of absorbed dose in the SI brochure.

    Like the Sievert, the rem uses a radiation quality factor to convert from the absorbed dose in rad’s or Gray’s to a number that more directly reflects the radiation risk.

    A lot of the residual inertia to keep using rem instead of Sievert is probably related to the entire structure of protective regulations in the U.S. being developed and put in place before the Sievert was established as a unit. I noticed that my health physics textbook was printed about the time the Sievert was first being proposed as a unit.

  44. squirrelelite says:

    Another little illustration of radiation risk from Wikipedia.

    The radiation dose from sleeping next to a human every night is twice the maximum dose from living near a nuclear power plant.

  45. daedalus2u says:

    squirrelelite, yes, but the “energy” from some people is more “toxic” than from others. ;)

  46. tmac57 says:

    Yes,I’ve heard that sleeping next to Ann Coulter is equivalent to getting a full body ‘Catty Scan’.

  47. Jan Willem Nienhuys says:

    MB wrote:

    illustrative cases where there is an important difference between dose calculated in Sieverts and dose in rems …

    When the entire body is exposed to the radiation, then 1 Sv = 100 rem. When only a part of the body is radiated, then the tissue weighting factors can be used with the results expressed as Sieverts.

    There is no difference between rem and sievert (small s), except for the factor 100.

    However, if for example a woman receives a chest x-ray, and lungs are viewed, then one computes the effective dose as follows (at least that is how I understand it):

    the x-ray intensity corresponds to 10 mSv (hypothetical amount, just for the purpose of showing how the computation goes)

    tissue weight factor lungs 0.12
    tissue weight factor breasts 0.05
    tissue weight factor skin 0.01 * A
    tissue weight factor remaining tissue (heart, esophagus, ribs, muscle) 0.05 * B

    (A = fraction of total skin irradiated, B= fraction of all tissue not explicity
    mentioned in the tissue factor list irradiated; I assume that the stomach was

    Add. Say that this gives 0.19 .
    effective dose is 10 * 0.19 mSv = 1.9 mSv

    I am quite sure if my example is correct, but it seems the only possibility to make sense of the information that all tissue factors should sum to 1, i.e. for each tissue type figure out what fraction of that type is irradiated.

    So a milligray is the absorbed dose (in millijoule / kg) and it is equal to 0.1 rad; with additional weighting factors for the type of radiation it is a millisievert (then it is 0.1 rem), but the same unit is used to measure the seriousness of the received radiation in comparison with 1 millisievert received uniformly over the whole body.

    Nominally sievert and gray both are a dosis absorbed energy (in joule / kg),
    in the sievert there is a quality radiation factor extra (then the outcome is called equivalent dose) or two quality factors (radiation and tissue, then the outcome is called effective dose).

  48. JMB says:

    My main point may have gotten lost in the discussion of sieverts ( I will use the lower case s) and rems. My main point is that several of the cited radiation exposures in the xkcd chart for medical radiation exposure appears to have been calulated using the quality factor and the tissue weighting factor. The 3 mSv factor for mammography in the xkcd chart would not appear to have included the tissue weighting factor in the calculation (it’s unclear how it was determined).

    Sieverts can be used for describing equivalent dose (correct term instead of my use of absorbed dose) or effective dose, but in a single chart, it would be better to stick with one or the other for consistency.

    The second point I was making was that the scale factor of 1 Sv was based on the threshold at which an increase in cancer could be detected. It is an inconvenient scale for day to day use in the nuclear power industry or the healthcare industry.

    Here is a reference to a draft report from the ICRP in 2007 discussing tissue weighting factors,

  49. JMB says:

    @# Jan Willem Nienhuyson 25 Mar 2011 at 12:44 pm

    A health physicist would be more familiar with the conventions used in the calculation of effective dose than I, but I would include the bone marrow in the calculation of the fraction of the tissue weighting factor for a radiation exposure of the chest. It is the red marrow that is significant, not the yellow marrow, and the chest xray includes between 1/3 and 1/2 of the red bone marrow. Also, I think you have to add 0.05 for each component in the tissues (muscles, thymus, esophagus). Consequently, I think your tissue weighting summation factor 0.19 is lower than typical for an exposure of the chest.

  50. clgood says:

    Ann can be a terrific polemicist. I haven’t read any of her recent books, but the ones I read a few years ago were spot-on when she spoke of her field of expertise, which is constitutional law. I mean, it’s hard to beat her summary of the entire document for both pith and accuracy: “Politicians are bad.”

    However, when she speaks to other subjects, her lack of expertise often shows. As in this case, and her creationism.

    Humans are interesting animals.

    Still, I’m kind of glad she brought this up because I knew only enough about hormesis to be dangerous. This discussion has been very educational for me.

Comments are closed.