Wednesday, 9 March 2016

Linear No-Threshold Model and Standards for Protection Against Radiation

From: Linear No-Threshold Model and Standards for Protection Against Radiation, Comment #: 582 (pdf)

This is a, sort of, re-blog of Mohan Doss' submission to the U.S. Nuclear Regulatory Commission, NRC, on radiation protection standards (link is above). Don't let that put you off reading it. It is a well-written summary of what we currently know about the harms of radiation and actual causes of cancer.

by Mohan Doss

I am responding to comments received from many members of the public who have raised major concerns regarding the recent petitions to discontinue the use of the linear no-threshold (LNT) model for establishing radiation safety regulations, and the proposed higher radiation dose limits for the public. These concerns are based on a misunderstanding that exists regarding cancer and radiation effects in our society.

1. Misunderstanding of the Primary Reason for Cancers

It is generally accepted that the primary reason for cancer is the transformation of a normal cell into a cancer cell through mutations. In fact, this transformation is generally referred to as carcinogenesis. A recent publication noted that almost everyone has cancer cells in their bodies (Greaves, 2014). However, everyone does not have cancer, as the lifetime probability of developing or dying from cancer are ~43% and ~23% respectively in the USA, for example (ACS, 2014). In an autopsy study, the presence of cancer cells was determined as a function of age (Imaida et al., 1997) (Figure 1). Whereas the percentage of patients having cancer cells was found to be relatively unchanged from age ~50 to age ~80, cancer mortality rates is known to increase by more than an order of magnitude between these ages (WHO, 2011) (Figure 2), indicating something other than cancerous mutations is the primary cause of clinical cancers. If cancerous mutations are not the primary cause of cancers, what is?

Figure 1. Percentage of autopsy patients having cancer as a function of age.
Figure 2. Cancer mortality rate in Japan as a function of age.

One hint is given by the observation that when the immune system is suppressed, e.g. in organ transplant or HIV patients, cancer risk increases by a factor of 3 to 4 (Oliveira Cobucci et al., 2012) (Figure 3). This is a huge increase, and indicates suppression of the immune system may be the primary cause of the cancers. Many characteristics of cancers are consistent with this point of view that cancers are caused by the suppression of the immune system. For example, (A) As we age, our immune system response reduces significantly (Levin, 2012) (Figure 4), and indeed cancer risk increases drastically with age (Figure 2).

Figure 3. Standardized Incidence Ratio for Cancers in immune
suppressed transplant recipients and HIV/AIDS patients.
Figure 4. Immune system response as a function of age.

(B) Increased rate of breastfeeding which boosts the immune system, and (C) earlier childcare attendance which subjects children to increased rate of infections and stimulation of the immune system, reduce the risk of childhood leukemias (Rudant et al., 2015) (Figures 5 & 6).

Figure 5. Odds Ratio for Acute Lymphocytic Leukemia as a function of breastfeeding duration.
Figure 6. Odds Ratio for Acute Lymphocytic Leukemia as a function of age at start of daycare attendance.

(D) Vigorous exercise, which causes increased DNA damage (Fogarty et al., 2011) (Figure 7) and boosts the immune system response (Simpson et al., 2012) (Figure 8), reduces the risk of cancers (Orsini et al., 2008) (Figure 9), consistent with the immune suppression model of cancer and contrary to the mutation model of cancer.

Figure 7. DNA damage from exercise as a function of exercise intensity
Figure 8. Effect of exercise on immune system response
Figure 9. Effect of exercise on cancer mortality rate

(E) In young growing age, when cells are dividing at the highest rates, they would be most susceptible to mutations, and so the accumulation of mutations would occur at the highest rates, as has been observed in an animal study, for example (DeGregori, 2013) (Figure 10). Based on this, the mutation model of cancer would suggest that cancer rates would reach a high level at a young age. The immune system response, on the other hand, is at the highest level at young age (Levin, 2012) (Figure 4), leading to the prediction that cancer rates would be at the lowest levels at young age, using the immune suppression model of cancer. The very low cancer rates observed in young age is therefore not consistent with the mutation model of cancer but is consistent with the immune suppression model of cancer.

Figure 10. Accumulated mutations and Lymphoma incidence as a function of age in C57BL6 mice.

Summarizing, though mutations are necessary for causing cancers, they are not the primary cause of cancers, and suppression of the immune system is a key factor that causes cancers.

2. Misunderstanding of the effect of low-dose radiation on cancer

Since radiation causes mutations, using the widely accepted mutation model of cancer, it is generally believed that even the smallest amount of radiation can increase mutations and so can increase the risk of cancer. This is the basis of the LNT model. However, we have seen that there are naturally occurring mutations including cancerous ones in almost everyone (Greaves, 2014). When there is a low level of radiation exposure, the damage caused by it would stimulate the defenses in the body, and these increased defenses would reduce the amount of mutations that would have occurred naturally (Feinendegen et al., 2013), resulting in less mutations overall. This has been observed in animal studies, e.g. (Osipov et al., 2013) (Figure 11).

Figure 11. DNA damage observed at 40, 80, and 120 days following chronic low-dose irradiation in mice.
Figure 12. Number of activated natural killer cells as a function of radiation dose.

Thus, there is no justification for the LNT model even using the mutation model of cancer, since a small amount of radiation does not increase but decreases the overall mutations. The major error in formulating the LNT model was that the defensive responses of the body and their effect on the naturally occurring mutations were completely ignored.

Another effect of low-dose radiation is that it stimulates the immune system (Yang et al., 2014) (Figure 12). As noted earlier, the boosted immune system would reduce cancers. This cancer preventive effect of low-dose radiation, known as radiation hormesis, was proposed in 1980 (Luckey, 1980) as a method of reducing cancers.

The above arguments suggest that low levels of radiation would reduce, rather than increase cancers as predicted by the LNT model. Which of these opposing predictions is correct? The scientific way of deciding between competing hypotheses is to perform experiments and compare the results with predictions from the hypotheses. However, since the LNT model was adopted without any evidence for its validity in the 1950s, it caused fear of the smallest amount of radiation, and so blocked prospective human studies of the cancer preventive effect of low levels of radiation. Though such prospective studies could not be conducted, there is indeed considerable amount of data from studies of population groups that have been exposed incidentally or accidentally to radiation. There have also been some clinical trials using low-dose radiation for the treatment of cancers. I will discuss some of these data now.

3. Studies of populations exposed to low levels of radiation

On several occasions in the past, populations have been incidentally or accidentally exposed to low levels of radiation over prolonged periods of time. The graph below shows data on cancer rates in some of these population groups, with radiation dose on the X-axis and estimated normalized cancer mortality rates (100 being the estimated cancer mortality rate for the un-irradiated control group) on the Y-axis. It is clear from examining the green data points that cancer rates have reduced following exposure to low-dose radiation. The red dashed line shows the effect of low-dose radiation on normalized cancer mortality rates, as predicted by the LNT model, using the radiation cancer mortality risk factors recommended by BEIR VII Report (NRC, 2006). It is clear the LNT model is not consistent with the data. The grey band indicates the range of normalized cancer rates that would be observed in the absence of radiation, using the example of cancer rates in Japan during the period 1960-1994 (FPCR, 2013) (black dots on the graph), a period during which cancer rates were relatively stable. The blue line shows normalized age-adjusted cancer mortality rates during the period 1960-2007 in the USA (Remington and Brownson, 2011).

Figure 13. Normalized cancer mortality rates according to the linear no-threshold (LNT) model (red dashed line) using radiation cancer risk factors recommended in BEIR VII report(NRC, 2006) are compared to evidence (green data points) as a function of radiation dose. Also plotted are normalized age-adjusted cancer mortality rates in USA as a function of year (blue solid line). Baseline cancer rate is displayed as a grey band to show the variability of cancer rates from year to year.

The blue dots in the band show the cancer mortality rates in Japan for the period 1960-1994, a period during which the cancer rates were relatively stable. Data plotted are from various studies:

  • Taiwan – Study of apartment residents in Taiwan exposed to radiation from contaminated building materials (Hwang et al., 2006);
  • NSWS – Study that compared cancer mortality rates between radiation workers and non-radiation workers in US nuclear shipyards (Sponsler and Cameron, 2005);
  • British Radiologists – Cancer mortality rate in male British radiologists who entered the field during 1955-79 in comparison to male general practitioners (Berrington et al., 2001);
  • Mayak – Cancer mortality rate in population evacuated from villages near Mayak nuclear weapons facility (Kostyuchenko and Krestinina, 1994). (See the report at https://goo.gl/rPghvr for more details on the data points plotted).

4. Harm from LNT model-based regulations

LNT model based regulations, by requiring reduction of radiation dose from 50 mSv (a dose someone might receive but for the regulations) to 1 mSv (annual dose limit for the public), claim to reduce cancer rates from A to B (Figure 13). In view of the year-to-year variation in cancer rates (grey band), this cannot be considered to be any reduction in cancer rates, since both A and B are within the grey band, the range of variation of cancer rates over the years. In reality, considering cancer outcomes observed in populations incidentally or accidentally exposed to low-dose radiation (green data points), these regulations have likely increased cancer rates by taking them from C to B and also prevented studies that could have reduced cancers to the level D (~30% reduction) (Figure 13). Cancer mortality rates in USA have remained stubbornly high at level E, there being only ~7% reduction in cancer mortality rates in 2007 in comparison to the rates in 1960 (Figure 13), with little progress in ~50 years (Remington and Brownson, 2011), in spite of tremendous investments and efforts in the war on cancer. This graph makes it clear the regulations have harmed the public by preventing the study and use of radiation hormesis for reducing cancers.

Let us now examine some additional data on the effect of radiation on cancers.

5. Atomic Bomb Survivor Data

It is widely recognized that the most important data for determining the health effects of radiation are the data from the atomic bomb survivors. For example, (Hall and Brenner, 2008) claimed that these are the gold standard data for estimating the health effects of low-dose radiation. The BEIR VII Report from National Academy of Sciences stated (on p. 141) that these are “the single most important source of data for evaluating risks” of low-dose radiation. A re these data consistent with the LNT model?

Figure 14: Excess relative risk for solid cancer mortality
in atomic bomb survivors as a function of dose.

The atomic bomb survivor data, after the latest update (Ozasa et al., 2012), no longer support the LNT model, because the dose-response data are not linear any longer, and have a significant curvature (see Figure 14). Ozasa et al. state: “The curvature over the 0-2 Gy range has become stronger over time, going from 0.20 for the period 1950–1985 to 0.81 for 1950–2003, and has become significant with longer observation”. This significant curvature, resulting from reduction of cancers when the radiation dose increases from ~0.25 Gy to ~0.5 Gy, cannot be explained by the LNT model. There is a possible explanation for the curvature using the radiation hormesis model, as explained in the publications (Doss, 2012, Doss, 2013). This idea, that the atomic bomb survivor data no longer support the LNT model, is being accepted implicitly by more and more scientists. For example, in the recently published debate in Medical Physics journal (Doss et al., 2014), the proponent of the LNT model did not use the atomic bomb survivor data to justify low-dose radiation cancer risk in his opening statement whereas in the 2009 debate in Radiology on the subject (Little et al., 2009), these data played a prominent role, as he stated "Most of the information on radiation-induced cancer risk comes from (a) the Japanese atomic bomb survivors (b) .....". Ian Fairlie, in his Comments to NRC http://www.ianfairlie.org/wp-content/uploads/2015/08/US-NRC-Consultation-4-1.pdf, did not refer to the latest atomic bomb survivor data to justify his concerns regarding low-dose radiation, after having referred to the atomic bomb survivor study as a major study in 2011 http://www.ianfairlie.org/news/are-radiation-risks-overrated/. In a recent article on the subject (Brenner, 2014), the author used older atomic bomb survivor data (Preston et al., 2007) to justify cancer concerns from low-dose radiation, ignoring the latest data (Ozasa et al., 2012). When he was challenged in a Letter to the Editor (Doss, 2014) for this omission and other errors in his publication, he had no rebuttal to the arguments. Thus, the change in the status of the atomic bomb survivor data is being recognized implicitly by authors of publications that have supported the LNT model, though they have not acknowledged it explicitly.

6. 15-Country Study of Radiation Workers

Another dataset that has been used to claim increased cancers from low-dose radiation are the 15-country study of radiation workers (Cardis et al., 2005). When the data for all the countries are combined, the analysis showed statistically significant increased risk of cancers among the radiation workers who had been exposed to an average of 20 mSv radiation dose (See Figure 2). As seen in the figure, the Canadian data show a much higher excess risk of cancer when compared to other countries’ data. The authors stated in the publication that if the Canadian data were excluded, the overall data would no longer show a significant excess relative risk of cancers. Canadian Nuclear Safety Commission (CNSC) investigated the discrepancy and found major problems with the data. As a result, CNSC withdrew the Canadian data from use (CNSC, 2011). A revised analysis indicates the Canadian worker risk estimates would be substantially reduced (Zablotska et al., 2014). Hence, the conclusion of the 15-country study would be that cancer risk is not increased following low-dose radiation exposures.

Figure 15. Excess relative risk/Sv for all cancers in radiation workers of different countries and for all combined in the 15-country study. Error bars indicate 95% CIs.

7. Studies of Cancers following Childhood CT scans

Two recently published studies, (Pearce et al., 2012) and (Mathews et al., 2013), have claimed increased cancers following childhood CT scans. These articles received high publicity in popular media, and have been cited in many publications to claim that cancers are caused by low level radiation exposure to children. Both of these studies have been discussed by several authors pointing out the major flaws in them, e.g. (Cohen, 2013), (Walsh, 2013), (Boice, 2013), (Walsh, 2014), (Socol, 2015), (Harvey, 2015). One of the main flaws is the likelihood of reverse causation, because the authors did not take into consideration the reason why the CT scans were performed. When the reasons for the CT scans were taken into consideration, two studies (Journy, 2015), (Krille et al., 2015) have shown no significant increased risk of cancers following pediatric CT scans. The Mathews and Pearce studies are examples of faultily designed studies that “demonstrate” increased cancer risk from low-dose radiation.

8. INWORKS study of leukemias and solid cancers in nuclear industry workers in France, UK, and USA

INWORKS study wanted to answer the question: “Does increased radiation exposure over an extended period of time increase the risk of leukemias and solid cancers?”, by studying leukemias (Leuraud et al., 2015) and solid cancers (Richardson et al., 2015) as a function of occupational radiation dose in nuclear industry workers during the period 1944 to 2004 in France, USA, and UK. If the occupational radiation dose were the most important part of radiation dose that the workers received, and other radiation doses they received during the study period (1944-2004) were relatively unchanged, then their study design would indeed be reasonable for answering the question they posed. However, as is well known, medical radiation dose increased considerably during this time period (NCRP, 2009), and occupational radiation dose in nuclear industry reduced drastically in the same period (Thierry-Chef et al., 2015). Therefore, during the later period of the study, the medical radiation doses the workers received, assuming they underwent similar amounts of diagnostic examinations as the average members of the public, would be significant in relation to their occupational doses. Not including the medical radiation dose into consideration is therefore a major error in the design of these studies. In addition there were other major flaws, rendering the conclusion of the studies not trustworthy (Doss, 2015b, Doss, 2015a).

9. Claims of increased thyroid cancers in Fukushima children

A recent publication (Tsuda et al., 2015) has claimed increase in the rates of thyroid cancers in Fukushima area children by a factor of about 30 in the years following the Fukushima nuclear reactor accidents in 2011, in comparison to the observed thyroid cancer rates in Japan. As described in the detailed analysis (Boisvert, 2015), the reported results are likely due to increased screening for thyroid cancers in the Fukushima children. It should be noted that the thyroid dose to children were low in the Fukushima children, as indicated in Table 6 of (UNSCEAR, 2013) (maximum of ~80 mGy). In another study, children who were subjected to median dose of ~1 Gy to thyroid from diagnostic administration of I-131 did not exhibit increased thyroid cancer risk (Hahn et al., 2001).

10. Claim of increased leukemias in children exposed to higher background radiation levels

A recent study (Kendall et al., 2013) has claimed that there were increased leukemias in children exposed to higher background radiation levels. Breastfeeding and childcare attendance are known to be important factors that affect childhood leukemias (Amitay and Keinan-Boker, 2015, Rudant et al., 2015). These were not considered as confounding factors in the (Kendall et al., 2013) study. Since relative risk per mSv was 1.07 (with 95% CI: 1.01-1.13) small changes in these confounding factors could make the excess risk not significant. Hence, we should await better studies which take into account the major confounding factors before considering the excess risk estimation from this cohort.

11. Note on publications claiming that low-dose radiation inceases cancer risk

The above described situation, that publications report increased cancer risk from low-dose radiation because of faulty data or analysis, is very common. Such publications receive high publicity in popular media, spreading the fear of low-dose radiation among the public. When corrections or criticisms are published showing the conclusions of such publications are no longer valid, these are not even mentioned in popular media, as they are not sensational. The misinformation therefore persists among the public.

As seen in the above examples, publications claiming that low-dose radiation increases cancer risk do not have a reliable record, with faults being identified sometime after they are published, negating their conclusions. I would suggest popular media and the public do not pay attention to such publications for several years to see if they withstand scrutiny by the scientific community.

13. Evidence for no increase in cancers following low-dose radiation exposures

Several studies have shown the existence of a large threshold dose, below which increased cancers do not occur following radiation exposure. The data from some of these studies are shown below.

Figure 16. Leukemias in Hiroshima
survivors (UNSCEAR, 1958)
Figure 17. Bone sarcomas in radium
dial painters (Evans, 1974)
Figure 18. Breast cancers in Canadian TB patients who
underwent fluoroscopies. (Miller et al., 1989)

Figure 19. Breast cancer mortality in Canadian TB patients
who underwent fluoroscopies. (Howe and McLaughlin, 1996)
Figure 20. Second cancers in radiation therapy
patients. (Tubiana et al., 2011)

14. Repeated low-dose radiation treatments improved survival of cancer patients, performing as well or better than chemotherapies

Several clinical trials involving cancer patients showed that low-dose radiation treatments (TBI=Total Body Irradiation, HBI=Half-Body Irradiation) resulted in patient survival as good or better than chemotherapies (COP, CHOP) (Chaffey et al., 1976, Pollycove, 2007), and such treatments, used to supplement standard radiation therapy, resulted in better patient survival (Sakamoto, 2004). (Figures 21-23)

Figure 21. Survival of lymphosarcoma patients treated
with low-dose radiation (TBI) and chemotherapy (COP)
Figure 22. Survival of non-Hodgkin’s lymphoma patients who were
given low-dose radiation treatments (TBI) or chemotherapy (CHOP)
Figure 23. Survival of non-Hodgkin’s lymphoma radiation therapy
patients who were given low-dose radiation treatments (TBI or HBI)

15. Concerns regarding low-dose radiation exposures to genetically susceptible individuals

Concerns have been expressed about the potential harm from low-dose radiation exposure to genetically susceptible individuals, e.g. (Brenner, 2015). In this presentation, Brenner quoted a study of meningiomas in patients treated for tinea capitis which showed that genetically susceptible people had increased risk of developing radiation-induced meningiomas (Flint-Richter and Sadetzki, 2007). However, this study involved high doses of radiation, and its conclusion cannot be extrapolated to low doses.

16. Concerns regarding higher sensitivity of children to radiation

Concerns have been expressed about children’s higher radiosensitivity with reference to the increased use of pediatric CT scans (Brenner et al., 2001). One of the reasons quoted is the increased lifetime cancer risk per unit dose among atomic bomb survivors who were exposed at younger ages (Figure 24 shows newer data on these risks). However, the excess cancers were observed in the survivors exposed to high radiation doses only, and only be using LNT extrapolation can these risks be transferred to low doses. Such an extrapolation does not have any justification since the LNT model does not have any valid supportive evidence, as seen above.

Figure 24. Excess absolute risk vs. Age attained, for
atomic bomb survivors exposed at different ages.

Another reason generally quoted for the increased radiosensitivity of children is that they have more dividing cells and so are more susceptible to radiation-induced mutations. This reasoning ignores the defensive response to low-dose radiation exposure which would reduce the overall accumulated mutations (Feinendegen et al., 2013, Osipov et al., 2013), and so would reduce the cancer risk using the mutation model of cancer. Using the immune suppression model of cancer, since children have the highest immune system response (Levin, 2012), and low-dose radiation boosts the immune system (Yang et al., 2014), children would face reduced risk of cancer following low-dose radiation exposures. Thus, the concerns expressed regarding higher sensitivity of children to low-dose radiation are not justified.

16. Opposing views on cancer risk from low-dose radiation

The evidence quoted in the pages above presents a coherent view of the health effects of low-dose radiation, that it reduces cancer risk. There is no ambiguity in this conclusion. Though many publications have claimed that low-dose radiation increases cancer risk, I have discussed several such publications and have shown why their conclusions should be rejected because of updates to the data, identification of major flaws in them, etc. Many scientists and advisory bodies have however reached an opposite conclusion, that low-dose radiation increases cancer risk, after analyzing scientific literature. One of the reasons for this difference is that these scientists and advisory bodies have failed to take into consideration published data such as shown above without providing any reasons. For example, the BEIR VII Report (NRC, 2006) and the BEIR VIII scoping meeting failed to consider the data discussed above. Similarly, scientists opposing the NRC petitions, e.g. Ian Fairlie, have also failed to consider these data but have utilized already discredited data to justify low-dose radiation cancer concerns. Since the conclusions from such incomplete analyses are more sensational, they receive more publicity in popular media, perpetuating the misconception among the public that low-dose radiation increases cancer risk.

17. Conclusion

In summary, we have seen that exposure to low-dose radiation in many different ways, including relatively high cumulative doses when received over an extended period of time, has resulted in no increased cancer risk or reduced cancer risk, completely contradicting the LNT model. Though there are many publications that have claimed support for the LNT model, they have major flaws in them raising major doubts about their conclusions. The most important data for estimating health effects of radiation, the atomic bomb survivor data, no longer support the LNT model with the recent update to the data, and the data are more compatible with radiation hormesis model. Thus, there is no longer any justification for the continued use of the LNT model. In view of the observed cancer preventive effect of low-dose radiation, there should be no concerns among the public regarding low levels of radiation exposures. A massive education program needs to be initiated to educate the public about the health effects of low-dose radiation and allay their concerns about the proposals in the NRC petitions.

Disclaimer: These comments are the personal professional opinions of the author and do not necessarily represent those of his employer.

References

Posted by at

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)