Search This Blog


Saturday, October 29, 2011

Not Hormesis Again!

Here we go again with the manufactuversy over low level radiation hormesis.

Hormesis is the proposed general beneficial effect of low level radiation.

This concept has been around for decades and was scrutinized by the National Academy of Sciences BEIR VII report of 2006.

There is no convincing evidence of this phenomenon.

Yet, here is a Dr. Calabrese, who has done work on chemical hormesis "discovering" that radiation might have hormetic effects.

In the first few minutes he describes homeopathic medicine (the pseudo-science that very dilute solutions have medical efficacy) with linear no threshold doses of toxic substances.

In medical efficacy tests, we are testing a drug for a predicted (or unpredicted) medical effect compared with a placebo.  The tests are very well controlled and the statistics can be quite good.  We need both a positive medical effect and the absence of adverse effects.

In toxicology, there is no good effect.  You are testing how much of a toxic substance a person has been dosed with, versus an observed effect.  Since researchers are not going to intentionally expose people to toxic substances, the researchers have to employ the best field research they can.   The statistics are not good, because there is a lot of uncertainty in the data.  Typically, we can easily observe the effects of high doses, but the observed effect trails down as the doses get lower.  There isn't enough statistical power to ever observe a teenie-weenie effect at a very low dose.

Most of the rest of the interview concerns the claim that an early radiation biologist may have lied about some early data.  No physical evidence is presented to support this claim.  But even if true, it would be irrelevant considering how far the science of radiation biology has advanced in the intervening decades.  It's not like we would shut down the food industry if we learned Pasteur lied about some data.

And of course, he offers no direct physical evidence of radiation hormesis.

He does describe what's called adaptive response which is described in the first link on page 50.  But those types of responses are not generalized, just specific to unique circumstances, and disappear at low doses.  We've known about this for decades.

Did you count how many times he mentioned the BEIR VII report?

Notice that the YouTube video poster is associated with LaRouch PAC, which has an anti-government regulation ideology.  I trust you see the connection...if people can be convinced hormesis is true, maybe regulations can be relaxed and/or delayed.

As long as the evidence is invisible...hormesis is invisible.


  1. Hi Bob. Now that you've gone to the trouble of establishing your own blog on radiation information, perhaps you can find time to look into that report you said you didn't have time for before:

    Happy blogging!

  2. I'll be watching for that too.

  3. Thank you! As my first commenter you have earned the ability to post all future comments for free!

    Before we go to your report I want to ensure you understand the philosophy of science a bit.

    If you throw a ball up in the air, real high it will reach its apex and start to fall. You can watch it fall for a few seconds and estimate its lower height, wait a few seconds and estimate its height (you have a strong arm).

    Now, suppose I tell you to close your eyes.

    Based on what you've observed, should you conclude the ball stops (threshold), falls to the ground (LNT), or reverses (hormesis)?

    Based on what you've already observed, in the absence of any new information the answer is LNT. It would take significant new evidence to think threshold, and extraordinary evidence to think hormesis.

    That is our conceptual starting point.

    Regarding comparing the two would take me months to do a side-by-side detailed analysis. I can only point out some obvious larger scale differences.

    The purpose of the BEIR VII report was to determine which model (of the 3 above) was the best model that fits the data for LET radiation only.

    I don't know what the purpose of the French report was (what triggered it). It seems to consider both LET & HET.

    I have no idea how the authors were chosen for the French report nor what their backgrounds are. It looks like there was 9 from 3 organizations. But the 9 could actually be part-timers in the organizations and be more independent. A good study would say.

    In the BEIR VII report there 17 authors. They are all independent and from diverse backgrounds. I can't write all the details here, but the selection process is very formal.

    There is no sign of secondary review with the French paper. The BEIR VII, after completion, is sent to over a dozen independent, diverse reviewers.

    The French paper lacks many's 58 pages compared to about 400 pages in BEIR. The number of reference studies looks to be about a third, but I didn't count.

    The most obvious technical difference is that BEIR VII actually performed a detailed risk analysis (with full details). The French paper only discusses other papers. It is intellectually lazy.

    The BEIR risk analysis shows that there is an excess risk at 0.1 Gy of 2,100 cancers.

    The French paper states, on Page 22, "However, for low doses (below 200 mSv) and a fortiori below 20 mSv generally encountered within the context of radioprotection, epidemiology can neither confirm nor refute the existence of an increased incidence of cancer".

    Since the French paper lacks the detail of BEIR, it doesn't provide the quality of evidence to accept either hormesis or threshold.

    For example, the French paper, on Page 11, says:

    "At low doses and low dose rates of ionizing radiation, the pro-apoptotic effect dominates
    and the damaged cells, of which there are only a few, can be eliminated or controlled."

    But BEIR VII doesn't see it that simply...they have evidence of Locally Multiply Damaged Sites, which are independent of dose rate.

    For most of the technical issues described in the French report, you find a more detailed discussion in BEIR. So, it's hard to know if the French considered things to the level of BEIR or not.

    All in all...BEIR is the gold standard. The French report around tin.

  4. Actually, something else struck me.

    BEIR VII has the scientific purity to not consider consequences. In other words, as far as we can tell, they close their eyes to the consequences of their result, and focus on the purity of what they are doing. The consequences are whatever they are.

    But the French paper discusses consequences of their study on Page 36.

    That is troubling.

  5. Part 1 of 2

    @Bob - I recognize that I am in a weak position to argue about the selection process because I have promised to protect my information source. Therefore, what I have to say is only hearsay and would be inadmissible in any court of law. It is even inadmissible in most circles of scientific inquiry because scientists like to attempt to keep themselves so far above politics that they do not admit that politics happens and influences most of what humans do.

    However, I will repeat again what I learned while serving as a mid level staff officer in Washington for 9 years. One of the people with whom I developed a professional relationship with was a public health professional who was working for the EPA on radiation protection issues. He worked with some career public servants in the EPA Office of Radiation and Air/Office of Radiation and Indoor Air/Radiation Protection Division. That is the office that has control of the funds for the atomic bomb survivor life span study (LSS) that is the primary source of all truth as accepted by the BEIR VII study.

    During numerous conversations over several years with two leaders in that office who happened to share the same last name (or might have been related, that point was not clear to me) he learned that they had agreed many years ago that they needed to cooperate to keep defending the LSS funding so that the study could be completed.

    As long serving DC staffers, they recognized that there are periodic forays against every budget line - in that city, there are whole armies of financial analysts who are looking to add someone else's money to their budget line. They will often agree to accept that person's "requirement" as well, but those agreements often last a very brief time and the money, being infinitely fungible, goes to do something else. There are thousands of examples of worthy projects and programs losing their funding and never being completed.

    Anyway, the LSS associated PhD's determined - and were quite open about their decision among colleagues in the hallways and lunchrooms - that they needed to maintain the primacy of the LNT in order to make sure that their study remained important enough to defend its funding year after year - until the very last atomic bomb survivor passed away.

    End of part 1 of 2 (please see part 2 of 2 for the conclusion)

  6. Part 2 of 2

    Having served in at least three separate budget related offices over my time in DC and having worked on 8 different budgets with dozens of different senior civilians, I can tell you just how powerful and stubborn people can be about defending "their" money and keeping "their" programs going. I know how long serving staffers influence decisions.

    I also understand how government research money providers - like the EPA - can influence committee selections, study funding, etc. The offices where I served funded a ton of research. I worked for the US Navy; the Department of Defense provides something like 50-75% of all government research money.

    Though Bob will point to academic tenure as proof that professors are immune from monetary or commercial considerations, any academic will testify that there is a difference between someone who merely has tenure and a job and someone who obtains sufficient research grants to become a department leader, fund grad students, fund the publication of peer reviewed papers and pay to attend various conferences.

    In the US, especially during the past 20-30 years, the government serves as the PRIMARY source of research funding. That might seem okay to some, but it means that POLITICS and bureaucratic maneuvering has a strong influence on the science that gets funding and on the conclusions that it reaches.

    Sorry for taking up so much of your space here Bob, but I figure that we can just pay each other rent if space is an issue. I am not sure what your moderation policy is, but I do not have any links to share because my primary source has a career to protect, a family to feed, and an expensive home to make payments on. I also cannot link to any documents that describe the "sausage making" process of federal budgeting and research dollar allocation, but I hope that at least some of your readers will have had similar experiences in their careers and will back up my description of the process.

    Bottom line - the BEIR selection process has been biased to favor those researchers who are already defenders of the LNT. There is a reason that it chooses to gloss over or outright ignore any study results that actually challenge the assumption and continues to point to the LSS as the gold standard study against which all others pale in importance or quality.

  7. Rod -

    Your arguments are just as silly as those from other deniers. I'll have a post on that later.

    You are a conspiracy lover. You imagined the Russians might have intentionally caused Chernobyl. The earlier gov't experience you described above has stuck in your head, and you think there is some evil conspiracy.

    No one has ignored any made that up. I didn't ignore Cohen's talk, which included his study. I showed why it was bad. You ignored that.

    We don't want people on high level meta-studies who think we should follow faulty science, we want those who follow robust science. We don't want a flat-earther on an important plate tectonic study, nor do we want a faith healer on an important medical study.

    Since the French study included HET/LET radiation, we KNOW they got it wrong (BEIR only looks at LET).

    We can compare the French conclusion to actual high quality radon studies which show LNT down to household levels of radon:

    The French have gotten their derrieres kicked.

    This is settled science, not conspiracy claims.

  8. We want the LSS to continue to get funding because we want better science. You seem to be suggesting we shouldn't get more accurate results.

    It's like counting a low level radioactive sample. The longer one counts it, the more confidence one has in the results.

    So of course those people you describe want funding. Anyone who is against that is trying to obfuscate accuracy.

  9. "We can compare the French conclusion to actual high quality radon studies which show LNT down to household levels of radon:"

    What is the dose, in mSv, from these "household levels of radon" over the observed period?

  10. The actual dose depends how long someone remains in the house at those average concentrations.

    Obviously, people don't sit in their houses all year. Going outside lowers the actual dose they get (for the vast majority of people).

    The point is, that if we observe excess cancer, IF people stayed in their homes all year, we are seeing LNT down in the realm of everyday life.

    And we are seeing LNT with the best LET-only study on the planet.

  11. Actually, I should clarify...we are seeing excess cancers (that is what we observe) even though people are NOT staying in their homes all year.

    If they stayed in their homes all year, we would see more cancers than we've seen.

  12. "we are seeing LNT down in the realm of everyday life"

    But are we seeing LNT at low doses?

    My point is we know that LNT applies at higher doses. If these studies are showing an increase for say 200mSv over the period then this is nothing new. (This is not a "low" dose, even though it may be a normal "everyday" exposure in a moderately high radon area over several years.)

    If on the other hand they show LNT corellation for a few mSv in total than it is useful data lending weight to LNT for low doses.

    However, given the press release says the cancer excess is 11%-21% it suggests to me that the doses must be fairly high.

    So what is the total dose in these studies?

  13. I think you're missing the big picture.

    The only reason we care about low level radiation effects IS BECAUSE it causes cancer. We aren't really interested in some academic attempt to determine a relationship.

    People get excess cancers from just living in their houses (it's considered a low dose no matter what the number is because it's happening where people live their normal lives -that's "low"). We have seen a linear relationship.

    We can work out the the low end of concentration is 3pCi/L, assume a person is home 10 hours per day everyday of the year, the dose is 33 uSv.

    But radon daughters are HET/LET radiation and differs from only LET radiation which is what the 200 mSv is about. But we really don't care about that number either. We see cancer linearly decrease down to that number (actually BEIR goes down to 100).

    Over time, we may be able to go lower and lower (that's why we want to fund the LSS), but there will always be a statistical limit to how low we can go.

    So when we could only go down to 200, that doesn't mean "insert hormesis" in the 0-200 range. And now we've gone down to 100, that doesn't mean "insert hormesis" in the 0-100 range. And if/when we go down to 50, "insert hormesis" in the 0-50 range).

    It's like a God-of-the-gaps argument...the more we learn about science in general, the less likely "God" explains things. The usage of "God" as an explanation has and will continue to shrink as an explanation for anything.

    With the LSS study, with more time, the more hormesis shrinks. Not the opposite. With more time we'll see the effect at lower and lower doses. Hormesis wasn't credible in the 1970's and it is less credible today, not more.

    And tomorrow, it will shrink some more.

  14. We "care about the numbers" because, in relation nuclear power, 200mSv is far above the extra exposure that most members of the public are likely to ever experience.

    The public dose limit is 1mSv per year, and the average exposure is generally orders of magnitude lower than that.

    200mSv is a relatively high excess lifetime dose even for the public around Chernobyl.

    That is why it is important to have a model that works in the range of a few mSv rather than a few 100mSv.

  15. No, that is "number fixation".

    We have a typical LNT dose vs. effect curve.

    If the subject was mercury, secondhand smoke, or whatever we'd all be quite satisfied with it. Everyone would just say, "Duh, that's LNT if I ever saw it! What's the issue?"

    But because some people are in the nuclear industry, they think we need to resolve the risk at those low levels of dose. (The anti-nukes do too, they expect super-linearity).

    We won't likely ever have a study that goes down that low at the 95% confidence level. You will be terminally disappointed if you expect that...kind of like a Creationist seeking every single fossil of every historical species transition over billions of years in order to prove evolution. Ain't gonna happen.

    We might get lucky on the molecular biology side and work our way up and find a threshold. But cancer is more than one disease and each involves multiple mutations and it gets pretty messy as far as our current state of knowledge. We might also be shocked and find super-linearity is the best curve at low doses.

    The current risk estimates are pretty low and reasonable. Things like diesel exhaust and wood dust are also carcinogens. And hopefully, no one sits and argues about them for days.

  16. "The current risk estimates are pretty low and reasonable. Things like diesel exhaust and wood dust are also carcinogens. And hopefully, no one sits and argues about them for days."

    There probably isn't the same level of excitement over particulate air pollution because
    a) the levels that people are typically exposed to that we need to regulate are well within the range where epidemiology can show an unequivocal dose response effect
    b) the target levels for regulation are also well within the range of observable effect

    Under these circumstance, nobody cares if there is a threshold or a sublinear response at low exposures, because we know people are dying from normal exposures of man-made particulate pollution (in cities etc) and that is where the focus is.

    For radiation however, nobody (outside of radiotherapy or nuclear war) is exposed to the levels of artifical radiation where we can show a definitive effect via epidemiology.

    The regulatory limits for radiation are set way below the level of obervable effects.

    So if there is a threshold, it actually matters.

    The middle of the LNT curve is fine (which is where we see dose-response for air pollution). But we know real-life dose response curves are not linear at the top end or at the bottom end. LNT is a convenient approximation for estimatimng effects in the middle of the graph, but it is almost certainly going to be wrong for low levels (and high levels).

  17. Not true, my friend.

    See my blog of today, regarding Singer.

    You have to recall the the old days, we had no rad limits...we got lots of cancers. Then we set high limits, got some, but fewer cancers.

    Today, if we saw a lot of cancers....we'd lower the limits again...and the LNT deniers would look around and say, "there must be a threshold, epidemiological studies don't show any cancer at these low levels!".

    No kidding, if they did...we'd lower the limit again.

  18. “You have to recall the the old days, we had no rad limits...we got lots of cancers. Then we set high limits, got some, but fewer cancers.”

    With respect, from my reading of the history of radiation protection, regulatory limits have never ever been set on the basis of observed cancers in the population they are protecting.

    The history is summarised here:

    Before the 1950s, protection levels were based on observed deterministic effects (which meant much higher limits than today).

    Since the 1950s, (note this spans the entire history of civil nuclear power) limits have been based on the cancer studies of atomic bomb survivors, which extrapolated the observable increase in cancers from high acute doses, and assumed a proportionately lower risk at low doses. This is LNT.

    But changes in dose limits were never based on lowering the limits until we “saw fewer cancers”. Since the 1950s the limits have always been below the level where cancer increases are observable via epidemiology. The risk is entirely based on the LNT assumption, not on observed cancers, at these very low doses.

    Even the maximum lifetime dose from occupational exposure only results in 3% mortality increase, which is imperceptible against the normal spontaneous cancer rate. The public limit is an order of magnitude lower, and real-world exposure tend to be a couple of orders of magnitude lower still.

    The notion that “we saw lots of cancers so we lowered the limit” is fundamentally false. The truth is we set the limit based on the estimated risk from the LNT model, even though no increase in cancer is empirically observable at that level or anywhere near it.

    By comparison, the lifetime mortality risk from particulate air pollution at the EU limit (20 ug/m3) is about 12%! And that is a public risk which is exceeded regularly in many European cities (and is even higher in China).

    That is my point. The regulatory limits for ionising radiation are set at levels below any observable harm; and the actual real-world doses are relatively easily reduced below these limits. In contrast the hazard from normal levels of inner-city air pollution cause measurable harm, and the limits are regularly exceeded.

  19. I'm not saying we saw the cancers at the bottom end of the scale. As we saw more and more cancers generally, the calculated risk/dose has increased. As that increases, limits get reduced to minimize the chance of seeing cancers at the low end, because of the overall shape of the response curve, which is LNT.

    It' like lowering a speed limit due to numerous accidents. The accidents don't have to happen at speeds right at the speed limit.

    And this make it even harder for epidemiology to ever find increased cancer rates at the low end.

    "Even the maximum lifetime dose from occupational exposure only results in 3% mortality increase, which is imperceptible against the normal spontaneous cancer rate. The public limit is an order of magnitude lower, and real-world exposure tend to be a couple of orders of magnitude lower still."

    Let's just say the public limit is 0.3% mortality increase. There are about 300,000,000 people in U.S., that's 900,000 people.

    If the above is correct, it is observable. An epidemiological study may not discern which victim died from radiation rather than from diesel exhaust at the 95% confidence level. And a medical doctor won't be able to tell the origins of the cancer. But there will still be a dead body.

  20. "I'm not saying we saw the cancers at the bottom end of the scale. As we saw more and more cancers generally, the calculated risk/dose has increased."

    You're saying we saw more cancers "generally"? Who? Who was exposed to radiation doses that were not at the "bottom end of the scale" (say 100mSv+)?

    No, they changed the limits because they reassessed the data from atomic bomb survivors. They did not change the limits because of perceived cancer increases in the general population of the US or any other country.

    "If the above is correct, it is observable."

    A big if. If the probability predicted by LNT is correct then there will be bodies caused by radiation. If it is wrong there will still be bodies (everybody dies of something) just not caused by radiation.

    The point is that there is no empirical evidence of what the effect of 1mSv (the annual public dose limit for man-made radiation) or even 10mSv is on mortality.

    UNSCEAR specifically does not use such risk models to predict cancer mortality for large populations exposed to low doses (up to and including Chernobyl) for exactly this reason. The uncertainty is so great that such predictions are meaningless.

  21. I realize you (and probably others) have this fixation with the low dose end. I will blog a series explaining what's going on. The A-bomb study has cancers below 0.1 mGy (what's effects below that particular level??? Therefore threshold!!!).

    Just kidding!!!

  22. Understandably the "fixation" is on low doses because those are the doses that people typically experience.

    Nobody (apart from those getting radiotherapy and the like) receives doses in 100s of mSv. So the range of doses where LNT is demonstrably correct is actually fairly useless for any practical purpose.

    People typically receive doses in the sub 10mSv range, but LNT tells us little about these.

    Even the a-bomb survivor studies show no excess risks of cancer or other diseases among survivors with doses below 0.005 Gy (5mSv).

    UNSCEAR states "No direct evidence of human health effects" for range 0-10mSv.

    Looking at the plot of dose vs risk for atomic bomb survivors (at the foot of this page):

    it is noticable than two thirds of the points below 0.1Gy (100mSv) show zero risk. Even some of those near 0.1Gy. The few points that do show risk in these ranges are well above the LNT line, and its error bounds. Obviously there is large uncertainty over what is going on at these levels sub-100mSv.

    It seems perfectly plausible that there could be some feedback from repair at these levels which would reduce the risk to sub-linear levels.

    This is not to say the risk would be zero.

    But it seems extremely bold to suggest, as LNT does, that 100,000 people receiving 1mSv will result in the same cancer incidence as 100 people receiving 1000mSv.

  23. Stay tuned to my series, and try to see things more the same context as secondhand smoke, pesticides, etc.

    Because that is the context public health people live in. They are consistent in their approach.

    But you're only posting on radiation...hence, "fixation". Are you equally concerned about other LNT models as applied to other toxics?

    No need to respond....go with the series so I can try to ensure we share the same background history.

  24. Note the homepage is: