Guest essay by Roger Graves
In a recent post I discussed the exaggerated fears our society seems to have about nuclear power. One of the primary objections to nuclear power is the belief that all ionizing radiation, at whatever level of intensity, is harmful and carries a risk of cancer. This essay is concerned with the effects that ionizing radiation has on human beings, and in particular whether low doses are harmful.
First, let me say that, although I am a physicist, I am not a medical physicist and definitely not a cancer specialist. Many other, far more knowledgeable people have written on this subject, so what I write here should be considered largely as a summary of other people’s work.
…………………………………………………………………..
There are two schools of thought on the effect of ionizing radiation on human beings. The first holds that all ionizing radiation is harmful, and that any exposure to it, down to the smallest detectable amount, carries a risk of cancer with it. Doses are cumulative, so that only the total, lifetime dose is significant, not the rate at which it occurs. Cancer risks at low doses can be predicted by linear extrapolation from known risks at high doses. This is known as the linear no-threshold (LNT) hypothesis.
The second school of thought is that there is a threshold dose below which the cancer risk effectively disappears. Humans are thought to be largely insensitive to small doses of radiation, or even large doses received at low rates over a long period of time. Harm occurs only when high doses are received at high dose rates; dose rate and total dose are equally important. Furthermore, there is some evidence to suggest that moderately high radiation levels can have a beneficial rather than a harmful effect. This will be referred to as the non-LNT hypothesis.
The LNT hypothesis brings with it a number of practical problems. If any level of ionizing radiation is harmful, then we must undertake extraordinary measures to ensure that radiation levels are kept to an absolute minimum. This results in a general fear of nuclear power, together with massively increased costs for constructing nuclear plants because of the regulatory burden which accompanies this fear. In addition, it has created an atmosphere of distrust in medical procedures such as CT scans [1].
This essay is concerned with the validity or otherwise of the LNT hypothesis.
First, let us summarize the generally agreed facts:
- Ionizing radiation, by definition, has sufficient energy to strip electrons from atoms when it interacts with matter. If the matter in question is a human cell, the cell will presumably be damaged in some way.
- There is a correlation between large acute doses of ionizing radiation, of the order of 1 seivert (1000 mSv), and cancer rates. Acute doses are defined as significant doses received within a short space of time, and therefore at high dose rates as measured in mSv per hour. (A typical such dose would occur at rates of the order of several hundred mSv/hr or higher.) People who are subjected to large acute doses have a statistically greater chance of developing cancer at a later date than those who are not so subjected.
- Although cancer induction from radiation is believed to be a result of damage to genetic material (DNA) in cells, the process by which such damage then proceeds to cause cancers is not well understood.
- Large acute doses do not immediately give rise to cancer. (Very large acute doses will induce radiation sickness, but this is not cancer.) The onset of cancer from a large acute dose will not in most cases occur for several years or even decades. The relationship between ionizing radiation and cancer is therefore likely to be a complex one.
Organizational Survey
A detailed literature survey of the topic of LNT would be far too long for this essay, so instead I will provide an organizational survey. Major scientific and technical organizations can be presumed to thoroughly review any authoritative document released in their name, so one can assume that the whole weight of the organization is behind the opinions expressed in the relevant document. A representative list of organizations who have expressed opinions in this way is given in the table below. I do not claim that this list is exhaustive, but I believe it is reasonably representative of the major players.
| Organization | Relevant Publications | Abbreviated Names | LNT Stance |
| US National Academy of Sciences | Health Risks from Exposure to Low Levels of Ionizing Radiation (2006) | BEIR VII [2] | Pro |
| International Commission on Radiological Protection | Low-dose Extrapolation of Radiation-related Cancer Risk (2005) | ICRP-99 [3] | Pro |
| United Nations Scientific Committee on the Effects of Atomic Radiation | Report of the United Nations Scientific Committee on the Effects of Atomic Radiation to the General Assembly (2000, 2006) | UNSCEAR-2000 [4]
UNSCEAR-2006 [5] |
Neutral |
| Academie des Sciences (Paris) and Academie Nationale de Medecine (joint report) | Dose–effect relationship and estimation of the carcinogenic effects of low doses of ionising radiation (2005) | ASP/ANM-2005 [6] | Anti |
| Electric Power Research Institute | Evaluation of Updated Research on the Health Effects and Risks Associated with Low-Dose Ionizing Radiation (2009) | EPRI-1019227 [7] | Anti |
US National Academy of Sciences
The US National Academy of Sciences report on the biological effect of ionizing radiation, commonly known as BEIR VII [2], is the major proponent of the LNT hypothesis. (More precisely, the report is BEIR VII Phase 2, published in 2006. An earlier version, Phase 1, was published in 1998.) According to this report:
“The [National Academy of Sciences] Committee concludes that the current scientific evidence is consistent with the hypothesis that there is a linear, no threshold dose-response relationship between exposure to ionizing radiation and the development of cancer in humans.”
The BEIR VII report does not state categorically that the LNT hypothesis is correct but merely that the data appear to be consistent with the LNT hypothesis. The report does not rule out the possibility of a threshold, but considers it unlikely.
International Commission on Radiological Protection (ICRP)
The International Commission on Radiological Protection publication 99 [3] states that “ … while existence of a low-dose threshold does not seem to be unlikely for radiation-related cancers of certain tissues, the evidence does not favour the existence of a universal threshold. The LNT hypothesis, combined with an uncertain DDREF [dose and dose rate effectiveness factor] for extrapolation from high doses, remains a prudent basis for radiation protection at low doses and low dose rates.”.
While this is not exactly a whole-hearted endorsement of LNT, it does indicate support for it.
United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR)
UNSCEAR produces a regular series of Reports to the General Assembly of the UN. These are produced by an international team and are weighty, authoritative reports. The two reports that have the most bearing on the LNT hypothesis are the 2000 [4] and 2006 [5] reports. Their general tone appears to be neutral. For example, in UNSCEAR-2000, Vol II, Annex G we find:
“… even very extensive studies, which have taken considerable resources, have demonstrated that it is not practical to obtain information on radiation effects at doses much below about 20 mGy for chromosome aberrations, 100 mGy for cell transformations, and 200 mGy for somatic mutations. The exact form of the response for cellular effects at low doses must therefore remain unclear.”
I take this to mean that UNSCEAR, as an organization, has no opinion on whether the LNT hypothesis is correct or not.
[Note – for the present purposes, gray (Gy) and seivert (Sv) are equivalent units. There are significant differences between them, but these can be ignored for the time being. See https://www.ccohs.ca/oshanswers/phys_agents/ionizing.html for further information.]
Academie des Sciences (Paris) and Academie Nationale de Medecine (joint report)
The French Academie des Sciences (Paris) and the Academie Nationale de Medecine issued a joint report in 2005 entitled Dose-effect relationship and estimation of the carcinogenic effect of low doses of ionising radiation. According to the English-language abstract [6]:
“The aim of the Joint Report of the two French Academies is to discuss the validity of the linear non threshold (LNT) dose-effect relationship for assessing the detrimental effects of small doses such as those delivered by X-ray examinations (0.1 mGy to 20 mGy). The conclusion of the report is that extrapolation with LNT could greatly overestimate those risks.”
The abstract goes on to say:
“Epidemiology has not evidenced cancer excess in humans for doses below 100 mSv.”
and:
“Experimental animal data have not evidenced a carcinogenic effect for doses below 100 mSv. Moreover, dose-effect relationships are very seldom linear; most of them are linear-quadratic or quadratic. A practical threshold or hormetic effects [see below] have been observed in a large number of experimental studies.”
A longer summary of this report is available in an internal US Nuclear Regulatory Commission document [8] describing a presentation made to them by the French Academy. According to this document:
“The French Academy presenters stated that effects at low doses should not be extrapolated from effects at high doses because damage repair mechanisms at the cellular level can be quite different. Further, extrapolating observations at the cellular level to the tissue, organ, or organism level is also uncertain.”
Electric Power Research Institute (EPRI)
The EPRI released a report in 2009 entitled Evaluation of Updated Research on the Health Effects and Risks Associated with Low-Dose Ionizing Radiation [7]. The report was simply a literature study, but included more than 200 peer-reviewed publications on the health effects of low doses of radiation, with emphasis on the most recently available information. It came to the conclusion that:
1. “Recent radiobiological studies in the low-dose region demonstrate that the mechanisms of action for many biological impacts are different than those seen in the high-dose region. When radiation is delivered at a low dose-rate (i.e. over a longer period of time), it is much less effective in producing biological changes than when the same dose is delivered in a short time period. Therefore, the risks due to low dose-rate effects may be over-estimated.”
2. “From an epidemiological perspective, individual radiation doses of less than 10 rem [100 mSv] in a single exposure are too small to allow detection of any statistically significant excess cancers in the presence of naturally occurring cancers.”
The EPRI, having connections to the nuclear power industry, might fairly be called a biased source. Nevertheless, this is an authoritative report which cannot be lightly dismissed.
Report Comparisons
We appear so far to have a ‘he said/she said’ situation, in which one group of reports says one thing (LNT is true), while another says the opposite (LNT is false). Is there any way in which a decision can be reached, one way or the other?
One problem with this whole area is that we are generally dealing with very small cancer risks. According to UNSCEAR-2006 [5], “ … the lifetime risk of death from all solid cancers together following an acute dose of 1 sievert [1000 mSv] is estimated to be about 4.3–7.2 per cent, and for leukaemia 0.6–1.0 per cent.” The LNT model would then give solid cancer rates in the range 0.4-0.7% for a 100 mSv dose and 0.04-0.07% for a 10 mSv dose; presumably the non-LNT model would give significantly less, if any at all. Since these are excess rates, i.e. rates over and above natural cancer rates, there would need to be a very large population sample indeed before there was any chance of detecting them against the natural background rate.
In this context it is worth noting that twenty years after the Chernobyl nuclear accident, a major statistical examination of almost the entire population of Europe was unable to detect any additional cancers [9].
The BEIR VII report is, in effect, the standard-bearer for the LNT hypothesis. However, it has a number of flaws in it, and has been criticized by experts with far more knowledge of the subject than myself (e.g. [1][9][10][11]). What follows are the flaws in BEIR VII that occur to me as a generalist in this area.
First, limited data sources. BEIR VII relies for its data to a great extent on studies of atomic bomb survivors from Hiroshima and Nagasaki. Such people would generally have had large acute doses. Extrapolating from this to small, chronic doses is not necessarily scientifically valid. Furthermore, ethnicity and environment factors come into play, since natural cancer rates for 1940’s Japanese and 2000’s Americans are not necessarily the same.
Second, natural background radiation. BEIR VII deals very poorly, in my opinion, with natural background radiation. According to BEIR VII:
“ … the BEIR VII lifetime risk model predicts that approximately 1 person in 100 would be expected to develop cancer (solid cancer or leukemia) from a dose of 0.1 Sv above background … Lower doses would produce proportionally lower risks.”
The dose of 0.1 Sv (100mSv) is not specified as a single event, and therefore presumably could be an acute dose (all at once) or a chronic dose (received over a long period of time). On this basis, people living in a high background rate of, say, 10 mSv per year above average would receive this additional 100 mSv every ten years.
There are many places on our planet where the natural background radiation level is significantly higher than the average value of 2.4 mSv per year. Examples of this are Karunagappally in Kerala State, India, and Yangjiang in Guangdong Province, China, both of which have deposits of thorium-bearing minerals. Studies of cancer rates have been made at both places [12][13], and have concluded in both cases that cancer rates were no higher than in places with much lower background levels.
Both the Karunagappally and Yangjiang studies are mentioned in the BEIR VII document (table 9-4), yet are rejected as evidence on the basis that the studies were ecologic rather than epidemiological studies. To quote BEIR VII:
“These studies did not find higher disease rates in geographic areas with high background levels of radiation exposure compared to areas with lower background levels. However, these studies were ecological in design and utilized population-based measures of exposure rather than individual estimates of radiation dose. Thus, they cannot provide any quantitative estimates of disease risk associated with the exposure levels found in the areas studied.”
I have been involved in many studies in many areas of science and engineering in the course of a long career, and one thing I have learned is that one takes the available data in whatever form it is offered. Indeed, it is rare that real-world data will be in precisely the form one might want for the purposes of a particular study. Rejecting data simply because it is not in an optimum form strikes me as merely an excuse to reject data that does not fit whatever hypothesis the study is proposing.
Third, radiation hormesis. BEIR VII appears to dismiss radiation hormesis out of hand.
Hormesis is a biological phenomenon whereby a beneficial effect, such as improved health or stress tolerance, results from exposure to low doses of an agent that is otherwise toxic or lethal when given at higher doses. (Many prescription drugs exhibit hormetic behavior.) Radiation hormesis is the phenomenon whereby low radiation doses can have a beneficial effect.
BEIR VII states:
“At this time, the assumption that any stimulatory hormetic effects from low doses of ionizing radiation will have a significant health benefit to humans that exceeds potential detrimental effects from radiation exposure at the same dose is unwarranted.”
Radiation hormesis is a difficult subject to research on an epidemiological basis because one would not normally expect a civilized state to subject its citizens to large-scale experiments to determine the level at which radiation doses become harmful. However, quite by chance, this occurred some years ago in Taiwan.
Some time around 1982 a container of cobalt-60 was accidentally mixed with steel scrap and was melted down and made into steel reinforcing rods, which were then used to construct a number of buildings in Taipei City and nearby counties [14]. These included 180 residential buildings containing about 1700 apartments, plus various schools and small businesses. About 10,000 people occupied these buildings for extended periods. The apartment buildings in particular were occupied for a minimum of 9 years, with some residents living there for up to 20 years.
The radioactive state of the buildings was gradually discovered, beginning in 1992. People living in the most contaminated buildings were estimated to have received a mean annual dose of 525 mSv in 1983, and a cumulative dose of up to 4000 mSv over the 20-year period from 1983 to 2003. (Cobalt-60 has a half-life of 5.3 years, so the radiation level would decrease by a factor of two every 5.3 years.) The averaged dose for all 10,000 affected people was 74 mSv in 1983 and 600 mSv cumulatively from 1983 to 2003 [15].
According to the data models given in ICRP-99 [3], during this 20-year period there would have been an expected 232 cancer deaths from natural causes in this group of 10,000, plus a further 70 deaths from radiation induced cancer, for a total of 302 deaths. The observed number of cancer deaths in this group over this period was 7, or 2.3% of the expected death rate. Similarly, in this population group over 20 years there would have been an expected 46 cases of children born with some form of congenital malformation, such as Down’s syndrome or cerebral palsy. The observed number was 3, or about 6.5% of the expected rate [15].
While a reduction in the expected cancer death rate of, say, five or ten percent is within the range of normal statistical variation, a reduction of nearly 98% verges on the improbable. Similar considerations apply to the 93.5% reduction in birth defects. Almost certainly something intervened in this group to cause these startling reductions, and while one cannot definitively state that it was due to the elevated radiation background, it is difficult to see what else it could have been.
The Taipei data points towards some form of radiation hormesis: people exposed to moderately large doses of radiation on a chronic basis are healthier than those not so exposed. It is not conclusive proof, since this is a single (albeit large-scale) incident, but it is in direct contradiction to the BEIR VII linear-non-threshold hypothesis.
It is interesting that BEIR VII makes no mention whatsoever of the Taipei incident, even though it must have been known to at least some of the authors.
Conclusion
I, along with a number of much better qualified authors, do not find the LNT hypothesis altogether credible.
Possibly the best summary of the deficiencies of the LNT model is given by Sacks et al. [10]:
“The overriding fallacy embodied in the LNT model is that it ignores the fact that the body responds differently to radiation at high versus low acute doses and dose rates, as has been demonstrated in many studies: high-dose exposures are associated with inhibition of protective responses and extensive damage to the organism, while at low doses the body eliminates the damage through a variety of protective mechanisms, evolved in humans from eons of living in a world bathed in natural background radiation.”
A useful analogy here is our relationship to alcohol. While our bodies can metabolise a moderate amount of alcohol, too much at any one time will result in unwanted effects such as intoxication and hangovers, and a gross overdose can result in potentially fatal alcohol poisoning. Both dose rate and total dose are determining factors in the effect of alcohol. For example, although I am a moderate drinker, if I were to drink in one session all the alcohol that I normally drink in a year, I should probably die from alcohol poisoning.
Perhaps the last word should go to Sacks et al [10]:
“Radiation science is dominated by a paradigm based on an assumption without empirical foundation. Known as the linear no-threshold (LNT) hypothesis, it holds that all ionizing radiation is harmful no matter how low the dose or dose rate. … Belief in LNT informs the practice of radiology, radiation regulatory policies, and popular culture through the media. The result is mass radiophobia and harmful outcomes, including forced relocations of populations near nuclear power plant accidents, reluctance to avail oneself of needed medical imaging studies, and aversion to nuclear energy—all unwarranted and all harmful to millions of people.”
Roger Graves is a physicist and risk management specialist who, much to his chagrin, is not associated with big nuclear, big oil, or big anything else.
References
1. http://www.aapm.org/meetings/amos2/pdf/59-17320-63249-582.pdf
3. http://new.icrp.org/publication.asp?id=ICRP%20Publication%2099
4. http://www.unscear.org/unscear/en/publications/2000_2.html
5. http://www.unscear.org/docs/publications/2006/UNSCEAR_2006_GA-Report.pdf
6. http://www.inderscienceonline.com/doi/abs/10.1504/IJLR.2006.009510 http://www.epri.com/abstracts/Pages/ProductAbstract.aspx?ProductId=000000000001019227
7. https://www.nrc.gov/docs/ML0701/ML070160572.pdf
8. http://onlinelibrary.wiley.com/doi/10.1002/ijc.22037/epdf
9. http://www.rrjournal.org/doi/full/10.1667/RR13829.1?code=rrs-site
10. http://link.springer.com/article/10.1007/s13752-016-0244-4
11. http://www.radiation-scott.org/EMS_2005_Poster_Web_version_B.pdf
12. https://www.ncbi.nlm.nih.gov/pubmed/19066487
13. https://www.ncbi.nlm.nih.gov/pubmed/11142210
14. http://articles.latimes.com/1994-06-12/news/mn-3195_1_suburban-apartment