Back to the basics: how radiation affects our health

radiation-damaged chromosomes

radiation-damaged chromosomes

Radiation is ubiquitous; an inescapable part of life on Earth.  Background radiation reaches us from outer space, from the rocks and soils we walk on, and from naturally radioactive potassium in our own bodies.  Through its entire history, organisms on Earth have been bombarded by radiation, and this will continue for as long as the Earth exists.  Today, the average person in the US is exposed to about 300 mrem each year from natural background radiation – about 1 mrem a day – and this level of radiation exposure seems to have no ill effects.  Of the estimated 600 or so mutations that occur in each of our cells each year (about 900 in those cells exposed to UV radiation), only about 5 are due to the effects of background radiation.  In short, environmental radiation is a mutagen, but it is not a major source of DNA damage. 

At higher levels, however, radiation can cause damage.  Continual exposure to low levels of radiation may cause a mutation that can initiate cancer.  Brief exposure to high levels of radiation can cause skin burns, radiation sickness, or a number of radiation-induced syndromes.

Radiation Damage to Cells

Radiation can damage cells by directly striking the DNA and causing damage such as single- or double-strand breaks or point mutations.  It’s more likely, however, that the radiation will interact with molecules in the cell’s cytoplasm, splitting them apart and forming reactive molecules called free radicals.  These free radicals, then, go on to cause DNA damage.  Free radicals are caused by more than just radiation – our mitochondria leak free radicals all the time, metabolizing our food can create free radicals, and even dissolved oxygen in our cells can cause DNA damage.  All of this damage is indistinguishable, with the exception of double-strand DNA breaks – we can’t “look” at a point mutation and tell if it was caused by radiation or mitochondrial free radicals.

When radiation passes through a cell the effects can range from non-existent to profound.  There’s a chance, for example, that a gamma ray will pass right through a cell without interacting at all or that the free radicals produced will simply recombine or be scavenged before they can reach the DNA.  If radiation (or the free radicals it produces) do interact with the DNA, there are only a few possibilities – either the DNA will be damaged or it won’t.

If the DNA is damaged, we have a few further possibilities – the damage may be beneficial (e.g., evolutionary advantage), harmful, or neutral (neutral damage is damage that has no effect on the cell – it may be in non-coding part of the DNA, or to a gene that’s not expressed in that particular cell, for example). If the damage kills the cell there’s no problem – in reality, the only way to cause problems is to have DNA damage that’s not fatal to the cell and that affects one of the handful of genes that can cause a cell to become cancerous.

However, the possibilities do not stop here, because our cells have DNA damage repair mechanisms – think of them as being like a spell checker; as long as they repair the damage properly then it’s as though it never occurred.  Although these mechanisms are very effective, they are not perfect.  This means that any bit of DNA damage may be repaired properly, may be repaired improperly, or might not be repaired at all.  It is at this point that DNA damage may become a mutation – a mutation is what happens when damage to our DNA becomes “fixed” and is able to be passed on to the next generation of cells.  As with DNA damage, mutations may be good, bad, or indifferent (neutral), and the detrimental mutations may be lethal or sublethal.  And, as before, it is only the sublethal damage that’s of interest to us, and then, only if it can cause the cell to become cancerous.

I’ve taken several paragraphs to describe the different possibilities of radiation interacting in a cell.  Part of this is for the sake of completeness, but it’s also to help drive home an important point – radiation is a weak carcinogen.  If we sum up all the possibilities above, I count over 20 different possibilities.  Of these, only 1 (sublethal damage that is misrepaired or unrepaired and causes a cell to become carcinogenic) have a chance of causing cancer.  Radiation is a carcinogen, but it’s not a very good one – not compared to many of the chemicals we work with.

In the next few sections, I will talk a little more about the effects of both acute and chronic radiation exposure on the organism, instead of the individual cells.

Acute exposure

If we are exposed to high levels of radiation in a short period of time, we will suffer from the effects of acute radiation exposure because the damage accumulates faster than it can be repaired.  If the dose is to a limited part of our bodies, we may end up with skin burns, sometimes severe.  There have been many instances requiring amputation of fingers, or even entire limbs.  High levels of radiation exposure to the whole body can lead to radiation sickness or death.

The effects of high radiation dose to limited parts of the body may range from no observable effects (if the dose is low enough) to blistering, burns, or necrosis depending on the dose received.  The effects of whole-body acute radiation exposure can be a bit more complex, and they are summarized in the following table.

Acute whole body dose (rads) Effect
1-10 Chromosomal changes (fragments, dicentric chromosomes, etc.)
25-50 Blood cell changes (depressed red and white cell counts)
100 Radiation sickness in about 10% of those exposed
~400 Lethal dose to 50% of the population without medical treatment
~800 Lethal dose to 50% of the population with medical treatment
1000 Lethal dose to 100% of the exposed population

Chronic exposure

The primary concerns with chronic exposure to relatively low levels of radiation are that we will develop cancer.  There are two competing hypotheses on this matter, and the matter is still far from being settled.


The linear, no-threshold (LNT) hypothesis suggests that all radiation exposure is potentially harmful (the “no-threshold” part), and that the risk of getting cancer from radiation is directly proportional to the dose received (the “linear” part).  LNT is the most conservative radiation dose-response model in that it predicts the highest risk from a given amount of radiation exposure.  This is one of the reasons that the LNT is the foundation of radiation regulations virtually everywhere in the world – since we really aren’t sure how we respond to low levels of radiation exposure, it makes sense to control dose (and risk) according to the most conservative model.

One problem with the LNT is that it can be used to predict cancer risks down to vanishingly small levels of exposure, and so it has been used to calculate expected cancer rates from exposure to radon, “dirty bombs,” and medical x-rays.  For example, say that the risk of getting cancer from a given radiation exposure is 5 additional cancer deaths for every 10,000 person-rem.  That means that exposing 10,000 people to 1 rem each should result in an extra 5 cancer deaths among those people.  Or, exposing 1 million people to 10 mrem each should also lead to 5 added cancer deaths.  It’s easy to see that we can use this model to predict added cancer deaths from any level of radiation exposure, no matter how trivial, if enough people are exposed.  By analogy, we can also say that, since a 1000 kg rock will crush someone, throwing a million one-gram rocks at a million different people will crush someone.

This doesn’t make much sense, and both the Health Physics Society and the International Commission for Radiation Protection have advised against this misuse of the LNT model.  In fact, we just don’t know what happens at such low levels of exposure, and we can’t make any such predictions for very small levels of exposure.  According to the Health Physics Society, in two separate position papers (which can be found on the HPS web page at, we simply can’t calculate a numerical risk estimate from any exposure of less than 10 rem, so even the first calculation runs afoul of HPS recommendations.  In a similar vein, the ICRP has suggested that, when looking at the risk from collective dose, if the most highly exposed individual receives a trivial dose, then everyone’s dose should be treated as trivial.

Threshold/Hormesis models

Virtually all harmful substances exhibit some level below which there are no apparent harmful effects.  This is part of the idea behind the No Observable Adverse Effects Level (NOAEL) – below a threshold dose you simply don’t see any effects from exposure to a substance.  There are those who feel that radiation probably behaves similarly – that there is a level of exposure below which there are observable effects from radiation exposure.

There are also those who think that exposure to low levels of radiation may be beneficial.  This is called hormesis and, although it sounds implausible at first blush, there are plenty of examples of hormesis in the world.  Two examples are vitamin D and selenium.  Both of these substances are vital nutrients, and both are acutely toxic in sufficiently high doses.  Low doses of aspirin can help to stave off heart disease (not to mention the beneficial effects on fever, pain, and inflammation), yet high doses of aspirin can be fatal, and people can also die of excessive salt intake or even water intoxication.  In short, the idea of hormesis is not outlandish; only the application of hormesis to radiation exposure seems unusual because we are all so steeped in the idea that radiation is uniformly bad.

The idea behind assuming a threshold in our response to radiation exposure is that, given the variations in Earth’s background radiation field, it makes sense that our cells should be able to adequately repair DNA damage from slightly elevated levels of radiation.  And, let’s face it; radiation is not one of the major environmental mutagens (it accounts for about 1%-5% of background DNA damage).  Our biochemistry contains very effective mechanisms for repairing DNA damage, and it is thought that these mechanisms are able to accommodate some level of added damage, such as would result from exposure to low levels of radiation.

The thinking behind positing hormesis effects is that, by presenting a continuing challenge to our mutation repair and tumor suppression mechanisms, they are kept at peak operating efficiency.  They are better able to contend with the ordinary, garden-variety damage that is always cropping up in our genome and, as such, our DNA is better protected than if this radiation exposure was removed.

The best way to test these hypotheses, of course, is to perform epidemiological studies of exposed populations, and many such studies have been performed with equivocal results.  Researchers have looked at radiation workers, residents of natural high-background areas, radon concentrations versus lung cancer rates, radiologists, and atomic bomb survivors, among others.  Some studies show that risks are slightly higher, some show no effects at all, and some show fewer cancers than expected in the study populations.  Part of the problem is that the effects are often smaller than the error bars, and this makes it very difficult to pick out what is actually happening.  Unfortunately, there is not yet a “gold-plated” study that everyone can point to and agree that it was properly done, controlled for all confounding factors, and shows a significant result.

Given this degree of uncertainty, many health physicists and most governments feel it is best to control radiation exposure under the risks of the highest-risk model, LNT.  The thinking is that, if we maintain risks at a low and acceptable level under LNT, then whichever model is correct, we will be at no more risk than we have agreed we can accept.  The only problem with this model is that, if one of the other models better represents reality, we will have spent a lot of time, effort, and money controlling illusory risks, and these resources will have been taken away from more effective risk-reduction measures.  So this question needs to be answered, and we will hopefully be able to do so before too much longer.

Tags: , , , ,

14 Responses to “Back to the basics: how radiation affects our health”

  1. Bob Applebaum February 28, 2013 at 11:20 AM #

    Radiation hormesis is psuedo-science. LNT is NOT the highest-risk model, a super-linear risk model is more conservative. Natural radiation doesn’t seem to have no ill effect. That radiation can mutate DNA, so it has an ill effect. It just that the effect is extremely weak and can be overshadowed by other effects.

  2. Dr. Y February 28, 2013 at 12:32 PM #

    I think you’re being a litle too facile with your dismissal of hormesis and equally glum about your assessment of the risks from environmental radiation. There are a number of papers on hormesis – some from reputable sources and some from sources whose objectivity I question. There was also an UNSCEAR report and a report from the French Academy of Science that summarize literature showing that hormesis is not proven, but that it is also not implausible. So to dismiss hormesis as “pseudo science” is a bit harsh. I agree that hormesis is certainly not proven and is unlikely, but at present there is no definitive evidence for or against it.

    Natural radiation seems to fall into this category also, although the evidence here seems to be a bit stronger. Cancer rates in the American high-background states are substantially lower than in low-background states (although the presence of confounding factors really muddies the relationship between background dose and cancer rates). There is no evidence of elevated cancer rates in Ramsar Iran, and scant evidence of elevated cancer rates in Brazilian, Indian, and Chinese high-background areas. Here, too, I have seen no definitive papers either for or against health effects (positive or negative) among people living in high-background radiation areas.

    You are correct that a supra-linear model for radiation carcinogenesis is more conservative than LNT. However, I’m not aware of any papers that show such a response outside of cell cultures. And many of the studies that have been performed on DNA damage from low-level radiation have neglected to include the countering effects of DNA damage repair.

    I think the bottom line is that, at the lowest levels of exposure, we simply do not know what the effects are. Papers have been published showing hormesis, LNT, and threshold effects – all to the 95% confidence interval – and in all cases the effect is so subtle that it is virtually meaningless. Thus, I think that your final comment is likely the most appropriate if paraphrased slightly:

    whatever effect there might be from low-dose radiation is extremely weak and can be overshadowed by other effects.

  3. Bob Applebaum February 28, 2013 at 12:48 PM #

    Hormesis is only plausible if a photon cannot cause a DNA mutation. But as you have written, there is always a risk that it can (only about 30 eV to cause an ionization, ionizing photons have much higher energy). There is no physical mechanism underlying hormesis, usually just examples of adaptive response are given.

    Using general background rates to discern differences in high background areas is to commit the ecological fallacy. In epidemiology, a cohort is usually chosen from a defined population. The epidemiology is more difficult when comparing the effects of natural background because there are at least two populations involved. So there are greater differences in genetic background, diet, pollution, etc. then when one only studies an excess radiation-exposed cohort compared with a non-excess radiation-exposed cohort from the same population.

    I fully stand by my original comment.

    A climate change analogy. A single molecule of CO2 is capable of trapping thermal radiation. So every molecule we add to the atmosphere increases the risk of global warming. Just because we can’t measure the temperature increase from a single molecule, doesn’t mean that magically the molecule is behaving differently.

    Likewise, a singe photon can damage DNA. Just because we can’t measure a cancer increase at low doses (due to statistics), doesn’t mean the photon/DNA is magically behaving differently at low doses.

  4. Jeff Walther February 28, 2013 at 2:27 PM #

    Other than to correct the record for lurkers, you waste your time engaging Applebaum on this topic. He is a well-known anti-nuclear agitator and no amount of evidence in the real world will ever change his stated opinions.

    • Bob Applebaum February 28, 2013 at 2:49 PM #

      I’m not anti-nuclear, just anti science denial.

      No evidence has been provided, just ad hominem.

  5. Dr. Y February 28, 2013 at 3:10 PM #

    Ultimately science is about what we can predict, measure, and that can be shown to be false. Bob’s comment that a single photon can cause genetic damage is correct – but so is the comment that this damage can be repaired. Thus, a single photon might – or might not – cause damage that can ultimately lead to cancer.

    It is also well-documented that the body produces additional DNA damage repair proteins when exposed to DNA-damaging agents. So the question is whether these repair proteins repair less damage than was caused, more damage than was caused, or the same amount – this is the difference between LNT, threshold, and hormesis effects.

    Thus far I am unaware of any experimental work that conclusively answers this question. I am also aware that epidemiological tools are inadequate to detect a definitive “signal” of either LNT, threshold, or hormesis effects at the low levels of exposure we are talking about. Thus – with our current understanding – arguing about exactly what happens at very low levels of exposure is not a scientific question since we cannot make predictions that can be tested reproducibly. This means that this discussion is – at present – a philosophical discussion and not a scientific one. As our epidemiological techniques or our biochemical tools become more sophisticated this may change.

  6. Bob Applebaum February 28, 2013 at 3:33 PM #

    Nope, that is ignorant. I mean that in the literal sense, not the insultive one.

    We predicted global warming decades ago, long before we measured it, because we understood that a single molecule of CO2 can trap heat. We realized (some of us) it was only a matter of time until we could detect a temperature increase.

    The history of heatlh physics worked out differently. We observed excess cancers before we understood why, but had we understood why first, we would have made a similar prediction like that of global warming….that there is an a priori detection threshold, but give enough people enough dose and the detection threshold will be exceeded.

    We don’t IGNORE the fundamentals.

    And yes, DNA damage can be repaired, but rarely is it repaired perfectly. If it was, then the risk WOULD BE ZERO.

    But it’s not. We can’t IGNORE that.

    So the overall risk cannot be zero or less. That is logical conclusion based on the evidence.

    I can’t measure our personal gravitational attractions. That doesn’t mean it might be repulsive or non-existent.

    It is a science question, my comments reflect the scientific consensus, not philosophical consensus. Not just because I’m parroting it, but because I understand it.

    • Dr. Y February 28, 2013 at 4:06 PM #

      I would suggest that we will not resolve our differences on this matter. It seems that you are dead-set against hormesis – based on your comments here as well as on your blog – which means that you are arguing from your convictions and that you view the scientific evidence from this perspective.

      You do have some incorrect statements, however:

      DNA damage repair is highly accurate, as evidenced by the fact that we experience millions of DNA-damaging events per cell per year and only a handful of them go on to become “fixed” as a mutation. In fact, the great majority of DNA damage comes from the presence of dissolved oxygen and the leakage of free radicals from mitochondria. You are correct if you are referring ONLY to the form of DNA damage due to alpha particle irradiation – double-strand breaks, for example. But for damage form beta and gamma radiation your comment is simply wrong.

      You are also incorrect in stating that we can’t measure the gravitational attraction from two bodies. Physicists have measured the gravitational attraction from much smaller objects (on the order of a kg or less).

      Dan Beninson, in his Sievert Award lecture stated that “just because we don’t see an effect does not mean it doesn’t exist” which is true. But if we cannot measure an effect it means that we cannot state as a scientific fact that it exists. In actuality, you are stating your belief that there is no threshold for radiation effects, acknowledging that we cannot measure this effect, stating that a lot of scientists agree with you, and asking us to accept your statement as a scientific statement based on an assumption that we will be able to detect this effect at some point in the future. In reality, until we can test your statement it is a hypothesis with as much weight as the threshold and hormesis hypotheses.

      You are also inaccurate in stating that there is “scientific consensus” with regards to the effects of low-level radiaiotn exposure. If there were such a consensus then we would not be pursuing this discussion thread. I was at a meeting in which the respected health physicist Bo Lindell (a very early proponent of LNT) stated that “LNT is almost certainly wrong, but it makes the book-keeping easy.” In fact, even ICRP and NCRP acknowledge that the available scientific evidence does not exclude other dose-response models – including threshold or hormesis effects – but that for the sake of conservatism it makes sense to continue using LNT.

      In any event, again, I doubt that we will come to agreement on this matter and I will leave it at this.

  7. Bob Applebaum February 28, 2013 at 4:40 PM #

    HA! DNA repair is highly accurate, otherwise we wouldn’t be here. But if it were 100% accurate (no risk), we wouldn’t be here either (no mechanism for evolution).

    Since it is inherent in DNA that it is mutable and since it is inherent in radiation that it can ionize DNA, there is a positive cancer risk to radiation exposure.

    That’s just reality!

    One can argue about what the risk is and what model best illustratesit. Our strongest, most expensive study suggests the dose response can be modeled either by LNT or LQ, there is no statistical difference. The risk to Americans is about 1% incidence per 10 rem.


    I didn’t say “we can’t” measure gravitational attraction, I said “I can’t”. I’m not wrong about that. I can’t. But from a “we” perspective, if we can measure it at around a kg, there’s no reason to expect gravity to magically transform below about a kg. The burden of evidence is on the person who claims gravity behaves differently below about a kg (or pick any value).

    For radiation health effects the burden of evidence is on someone claiming non-LNT (or non-LQ). Nothing is excluded from consideration by fiat; not hormesis, threshold or even super-linear effects. But the evidence is lacking for those other hypotheses, whereas LNT has been the consensus theory since around the time of the theory of relativity (gravity).


    Citing an individual is committing the fallacy of cherry-picking. My comments reflect the scientific consensus.

    Have a good day.

    • Leslie Corrice March 5, 2013 at 1:02 PM #

      Mr. Applebaum…radiation-induced mutations are but one of the foundations of evolution. There a multitude of environmental mutagenic and non-mutagenic factors related to evolution. Not radiation alone. You are looking st this in isolation and not taking the entire picture into account. One of the weakest of the mutagens is low level radiation. Since radiation exposure amplifies cellular and genetic repair, and most of the other factors do not, it follows that the non-radiolgical mutagenic-effect repair mechanisms improve due to the radioactive stimulus, as found by Lawrence Berkeley Laboratories last year.

    • Mohan Doss April 28, 2013 at 4:59 AM #

      Mr Applebaum,
      Recently I became aware of journal articles reporting concerns about DNA damage caused by normal activities that we all do daily, such as exercise and thinking/learning. These published concerns inspired me to write a satirical piece (emulating Colbert Report) regarding “Deadly DNA Damage” which you may find humorous, but also instructive, because it shows how unwise, misleading, and unhealthy such concerns are. Here is the link to my article in Google Groups.
      Doss Report # 1:!topic/doss-report-1/sYCY7yAIvZI

  8. K L Mills March 12, 2013 at 11:18 PM #

    Should there be a “no” in the last clause of this first paragrah?

    Threshold/Hormesis models

    Virtually all harmful substances exhibit some level below which there are no apparent harmful effects. This is part of the idea behind the No Observable Adverse Effects Level (NOAEL) – below a threshold dose you simply don’t see any effects from exposure to a substance. There are those who feel that radiation probably behaves similarly – that there is a level of exposure below which there are observable effects from radiation exposure.

  9. Dr. Y March 15, 2013 at 10:13 AM #

    Good catch – and thanks! You’re correct that, if there is a threshold before we seen adverse health effects from radiation, then there is a level of exposure below which there are NO observable effects. Thanks for pointing this out.

  10. Ian Soutar November 11, 2013 at 4:02 PM #

    Radiation Hormesis is real, at least in my own case. I literally embrace radiation and wear a few radioactive jewellery pieces and also a system for having bimonthy radon baths for my arthritis.

    I am grateful for the healing effects, as is most of the world through Healing Hot Springs which are embraced in all of Europe and Asia. These hot springs are rich in radioactive radon gas. Beside my bathtub is a radon generator system I designed.

    Here is a site listing the positive evidence as well as listing most of the science papers of note on the subject.

    Ian Soutar
    Microsec R&D Inc.
    Vancouver Island

Leave a Reply