With the recent events in Japan, there have been a lot of diaries and comments discussing what can happen, what sort of worst-case scenarios there are, but I have yet to see a diary discussing radiation and its relation to cancer risk. This diary is an attempt to discuss the possible health effects of what is happening, and to try and put some of the numbers we're seeing in the news into perspective.
My background (seems to be a prerequisite these days). I have a PhD in molecular biology and have been studying DNA damage responses at the molecular level for over a decade now. I work with radioactivity (32P for aficionados) on an almost daily basis. I am not a nuclear physicist nor a medical expert, but I am capable of reading and understanding scientific literature, and this diary is meant to distill some of this information for those people where scientific writing may as well be greek.
This is not meant to be a primer on radiation, rather a primer on what we know about cancer risk in response to ionizing radiation. Samer has a pretty good recent diary on this subject, as well as earlier diaries on radioactive iodine, cesium, and strontium that are well worth reading.
To begin with, we need to understand that epidemiology is difficult. When talking about the effects of low-dose radiation, we are trying to measure small increases in the risk of cancer in a background where there is already a very large number of cancers reported. In the US, according to this fact sheet from The American Cancer Society (pdf!), over 1.5 million people were expected to be diagnosed with cancer in the US in 2010.
To put it another way, from Cancer risks attributable to low doses of ionizing radiation: Assessing what we really know (not sure if this is publicly available or not), a study published in 2003 in PNAS:
Compared with higher doses, the risks of low doses of radiation are likely to be lower, and progressively larger epidemiological studies are required to quantify the risk to a useful degree of precision. For example, if the excess risk were proportional to the radiation dose, and if a sample size of 500 persons were needed to quantify the effect of a 1,000-mSv dose, then a sample size of 50,000 would be needed for a 100-mSv dose, and ≈5 million for a 10-mSv dose...this relationship reflects a decline in the signal (radiation risk) to noise (natural background risk) ratio as dose decreases. (7, 8).
Most of the studies on the effects of radiation to humans come from just a few places, namely survivors of Hiroshima and Nagasaki, Nuclear Power Plant workers, radiologists, airline crews, and people who live in areas with naturally high radioactive background. Studies done to asses the risks of radiations are usually based on compilations of smaller studies done on these groups.
A more or less scientifically agreed upon conclusion (at least in my limited reading I haven't seen any disputes) is that when humans receive doses above 100 millisieverts (100mSv), there is a statistically significant increase in cancer.
Lets talk briefly about what a sievert is. Different types of radiation have different effects on the body. When we talk about how much radiation someone or something absorbs, we measure this in grays (or rads in the US). However, because different types of radiation affect biological systems differently, scales have been designed to measure the biological effects of radiation; this is what a sievert is (rem is used in the US). Conversions between grays and sieverts are based on what type of radiation is absorbed, where it is absorbed, etc.
Back to our discussion, what do we know about the increase risk of cancer from absorbing 100 mSv? When exposed to 100 mSv of radiation, the generally accepted increase of developing cancer is 0.5%. At higher doses, the increase in cancer risk increases proportionally, that is, their is a linear dose response relationship. 200 mSv is a 1% increase in cancer risk, 250 mSv (the new legal limit in Japan for disaster response), a 1.25% increase, 2 Sv, a 10% increase, and so on.
What about levels below 100 mSv, which is what most of us are concerned with. As I mentioned above, below this level we don't have the statistical power to measure a meaningful increase. Therefore, the linear risk that is observed above 100 mSv has been extrapolated, also in linear fashion, down to 0. So a 50 mSv exposure would result in a 0.25% increase, a 10 mSv exposure a 0.05% increase, and so on. This is what is known as the linear no-threshold model, which is the model accepted by most governments and scientific agencies (there are of course controversies explained below).
There is, of course, a handy dandy online calculator to calculate a cancer risk based on a given dose, and you can find it here . If you want radiation levels in Tokyo, you can find them here and for all of Japan, here(and if anyone has better websites, I'd be happy if you could link to it in the comments). The websites provided use nGy, but you can put those into the calculator as well. Based on the current numbers as of this writing, if the radiation levels remained as elevated as they are for the next 1000 hours in Ibaraki, the highest measured prefecture, the increased risk of cancer would be 0.003%, or 1 additional cancer per 36,000 people.
You can also use this calculator to see what effect there might be on the workers at Fukushima. If they do not exceed their newly allowed dose of 250 mSv, the increase risk of cancer would be 1.25%, or 1 additional cancer per 80 people.
Now, as I mentioned, the linear no-threshold model (LNT from here on out) is not universally accepted, the notable holdouts being the French Academy of Sciences and National Academy of Medicine, the Health Physics Society, and the American Nuclear Society. Generally, in their opinion, the lack of concrete scientific evidence for health effects below 100 mSv means we should not try and make models about risk below this level. The French Academies say that below 10 mSv there is essentially no evidence and therefore no support at all for a no-threshold model, they instead support a threshold model. That is, below certain doses, there is no increased risk of cancer.
In addition to the threshold model, there is also the model of radiation hormesis. In this model, low doses of radiation are actually beneficial to human health and would reduce one's risk of cancer. While there are plausible molecular mechanisms for this effect, there is little or no evidence to support it in humans (though I've read as many as 40% of animal studies show a hormetic effect).
Finally, lets talk about briefly about radiation poisoning. Once we exceed doses of 250 mSv, radiation poisoning becomes a serious risk. These levels are much higher than the population of Japan can expect at the moment, though workers in the area are legally allowed to reach this threshold. Doses between 1000 and 2000 mSv results in acute radiation sickness, doses above 8000 mSv are fatal.
What are the risks at the moment for civilians in Japan? Based on the LNT model, the increased radiation exposure will lead to a very slight increase in the number of cancers if the radiation remains elevated for long. Based on the other models, there might be no increase in cancer at all, or perhaps even a slight benefit.
Regardless of which model is in fact correct (we will likely never be able to prove or disprove any of the three), lets hope that the radiation levels remain relatively low and the situation at Fukushima is resolved soon.
Further reading for those interested:
National Academy of Science LNT pdf
French Academies LT pdf
A review on radiation hormesis