Moved from 1 mSv/d to 2 mSv/d, 2024-01-23. 2 mSv/d is the correct 1934 ICRP limit. There is a properly referenced PDF of this piece at gordianknotbook.com. This is true of most of the Gordian Knot News posts. Figure 1. US Dose Limits through time
You and I agree that dose rate is critical, which puts us both solidly in the anti-LNT camp. And the dose rate associated with xrays, which is in the tens of mGy/s is many orders of magnitude above 1 mGy/d (1.2e-5 mGy/s).
But I have a problem worrying about tiny doses delivered in a very short time, eg 0.5 mGy in 0.01 seconds for some xrays. We know the DSB repair period is tens of minutes to about 10 hours depending on the amount of damage the cell has to deal with. So it looks to me that there is little difference from the cell's point of view whether it gets hit with 0.5 mGy in less than a second or 0.5 mGy in a few minutes.
But yeah the 1 day integrating period is probably too long. Maybe the rule should be tightened to a limit of 0.04 mSv/h. But in most nuclear power plant releases there would be no practical difference for the public.
In any event, as you point out, there is no necessary conflict between your position on mammograms and a 1 mGy/d dose rate limit in an NPP release. Whether X-rays are as dangerous as you believe, is logically a separate subject, which I will let others debate.
True, but that damage cannot propagate to the entire organism in the same way unrepaired DNA damage can.
I should have added that there is no conflict PROVIDED the regulatory period is short enough so it reasonable to assume the dose rate does not change a lot within that period. For nuclear power plant releases, a day will usually be short enough, an hour certainly will. But a year is nonsensical and could be dangerous. The current US nuke worker limit is 50 mSv/y. It is possible to abide by this limit and exceed the 1 mSv/d limit be a factor of 50.
An interesting contrast is that the ‘safe’ levels set for air pollution (PM2.5 and PM10) are already associated with a 2% increase in mortality. See https://dx.doi.org/10.1056%2Fnejmoa1817364
Hi Jack, been reading through your substack for a video I'm doing on nuclear technology and nuclear power. Would be excellent to chat with you about this all if I can!
Interested in your opinions on radiation emitted from CT scans. The exposure can be anywhere from 2mSv to 50 mSv depending on where is scanned and how much of the body is scanned.
The SNT mortality risk from a single acute dose of 2.5 mSv is 1.6e-6. For LNT it is 1.2e-4.
A factor of 100 difference. For 50 mSv acute, the numbers are SNT, 1.0e-3 and LNT 2.4e-3,
a difference of only a bit more than a factor of 2. Of course, these are implicitly whole body doses, and when we start talking about organ specific doses, it is time for this naval architect to shut up.
The fundamental difference between SNT and LNT is not in how they handle single acute doses; it is how they handle repeated chronic doses. LNT justs adds all the doses up assuming no repair. SNT assumes a repair period, estimates the risk (aka unrepaired damage) from each period, and combines these daily risks probabilisticly. This leads to differences of factors of 1000 or more when dealing with elevated dose rates over prolonged periods.
Yes that makes sense. I just ordered your book btw. Super interesting. I’m a big proponent of ultrasound and have some concerns of excess radiation exposure from CT scans. Especially in women and kids. From the data I’ve seen the effective doses are often a lot higher than generally reported and would significantly increase cancer risk based on current accepted models. For example https://pubmed.ncbi.nlm.nih.gov/20008690/
Thanks. The authors in the paper you cite used LNT to come up with their cancer risk. I repeat LNT is nonsense, but what counts is the dose within the repair period. In the case, of CT scans which take maybe 15 minutes, all the dose is received within a single repair period. In that specific situation there is not that much difference between SNT and LNT for the higher end of the CT doses. (If the CT dose was much above 100 mSv, SNT risk would be higher.)
But the "current accepted model" is still nonsense because it denies our ability to repair radiation damage. This shows up big time when the dose is received over many repair periods as in a nuclear power plant release. In bumper sticker terms, dose rate is far more important that cumulative dose.
I see. Interesting. While I have your attention and expertise what about a study like this in 680,000 people from Australia exposed to CT scans vs controls. Seems to be dose related and location related inc in cancer risk? Maybe you cover this in your book. https://pubmed.ncbi.nlm.nih.gov/23694687/
There's almost no discussion of medical exposures in the book. Probably an oversight that needs to be corrected. The book is concerned with NPP releases so focuses on cohorts
that have received large doses over long periods.
One issue with medical exposures with maybe the exception of mammograms is that healthy people don't get Xrays. This is especially true of young people and CT's. Something was wrong with those kids. And something wronger in the kids that got mutliple CT's. Does a one year lag turn these kids into healthy kids? On top of this childhood cancer is rare, so statistical power is poor. And a few extra cases generates a much larger ratio when compared to this small number. Doctors should use the absolute risk in balancing risk vs benefit. Finally, as you point out, the dose associated with a CT is uncertain and couldbe much larger than the authors' unmeasured estiamte of 4.5 mSv. All in all I don't think we can make much of the Australian study.
I should have skipped the bunker sticker comment. Dose is important but what is more important is how that dose is delivered. It makes a big difference whether a dose is delivered in a series of spikes or the same dose is acquired evenly. In the CT case, SNT says the harm is linear in the number of scans assuming each scan has teh same dose and is spaced at least a repair period apart. LNT says it makes no difference whether you get 50 CT's in a day or
I should have added the absolute risk numbers. Out of 10.9 million kids and young people 60,674 were diagnosed with cancer. That's an absolute risk of 0.0056 reflecting the rarity of childhood cancer. The CT group incidence was 24% higher an absollute risk of 0.0069.
The difference is 0.0013, a little more than 1 in a thousand. How much of that one in a thousand is due to the rad dose, and how much is due to whatever caused the doctor to order the CT for a child?
If we assume all the extra cancer was due to radiation and assume incorrectly all the CT kids got only one CT, then according to LNT that dose would have had to been 14 mSv and per SNT that dose would have had to been about 38 mSv. If we assume a significant portion of the increased cancer was selection bias and your observation on CT doses, I don;t think we have a dramatic conflict with either model, although SNT points to more selection bias than LNT. The argument against LNT lies elsewhere.
It could be that 1 mGy/day is harmless, but it is completely absurd to consider that for this reason 1mGy/second is harmless.
Daniel,
You are a nuisance, but a valuable one.
You and I agree that dose rate is critical, which puts us both solidly in the anti-LNT camp. And the dose rate associated with xrays, which is in the tens of mGy/s is many orders of magnitude above 1 mGy/d (1.2e-5 mGy/s).
But I have a problem worrying about tiny doses delivered in a very short time, eg 0.5 mGy in 0.01 seconds for some xrays. We know the DSB repair period is tens of minutes to about 10 hours depending on the amount of damage the cell has to deal with. So it looks to me that there is little difference from the cell's point of view whether it gets hit with 0.5 mGy in less than a second or 0.5 mGy in a few minutes.
But yeah the 1 day integrating period is probably too long. Maybe the rule should be tightened to a limit of 0.04 mSv/h. But in most nuclear power plant releases there would be no practical difference for the public.
In any event, as you point out, there is no necessary conflict between your position on mammograms and a 1 mGy/d dose rate limit in an NPP release. Whether X-rays are as dangerous as you believe, is logically a separate subject, which I will let others debate.
A high dose rate may also have effects independent of DNA repair. Free radicals may be managed by the cell when the dose rate is low.
True, but that damage cannot propagate to the entire organism in the same way unrepaired DNA damage can.
I should have added that there is no conflict PROVIDED the regulatory period is short enough so it reasonable to assume the dose rate does not change a lot within that period. For nuclear power plant releases, a day will usually be short enough, an hour certainly will. But a year is nonsensical and could be dangerous. The current US nuke worker limit is 50 mSv/y. It is possible to abide by this limit and exceed the 1 mSv/d limit be a factor of 50.
Of course, what matters is DNA damage. What I mean is that any protection mechanism is overwhelmed by a high dose rate.
The limit for a worker is problematic: if you get your 50 mSv/year in one second in a year, it's not the same as if it's spread over a year.
I fibbed by claiming a PDF of this piece was available at gordianknotbook.com. It's not up yet. Should be up by tomorrow. Mea culpa.
Dan Shannon, your spam checker is rejecting my emails.. Claims untrusted IP.
An interesting contrast is that the ‘safe’ levels set for air pollution (PM2.5 and PM10) are already associated with a 2% increase in mortality. See https://dx.doi.org/10.1056%2Fnejmoa1817364
Hi Jack, been reading through your substack for a video I'm doing on nuclear technology and nuclear power. Would be excellent to chat with you about this all if I can!
Lawrence,
Send me an email at djw1 at thorconpower dot com.
Interested in your opinions on radiation emitted from CT scans. The exposure can be anywhere from 2mSv to 50 mSv depending on where is scanned and how much of the body is scanned.
Larry,
The SNT mortality risk from a single acute dose of 2.5 mSv is 1.6e-6. For LNT it is 1.2e-4.
A factor of 100 difference. For 50 mSv acute, the numbers are SNT, 1.0e-3 and LNT 2.4e-3,
a difference of only a bit more than a factor of 2. Of course, these are implicitly whole body doses, and when we start talking about organ specific doses, it is time for this naval architect to shut up.
The fundamental difference between SNT and LNT is not in how they handle single acute doses; it is how they handle repeated chronic doses. LNT justs adds all the doses up assuming no repair. SNT assumes a repair period, estimates the risk (aka unrepaired damage) from each period, and combines these daily risks probabilisticly. This leads to differences of factors of 1000 or more when dealing with elevated dose rates over prolonged periods.
Yes that makes sense. I just ordered your book btw. Super interesting. I’m a big proponent of ultrasound and have some concerns of excess radiation exposure from CT scans. Especially in women and kids. From the data I’ve seen the effective doses are often a lot higher than generally reported and would significantly increase cancer risk based on current accepted models. For example https://pubmed.ncbi.nlm.nih.gov/20008690/
Thanks. The authors in the paper you cite used LNT to come up with their cancer risk. I repeat LNT is nonsense, but what counts is the dose within the repair period. In the case, of CT scans which take maybe 15 minutes, all the dose is received within a single repair period. In that specific situation there is not that much difference between SNT and LNT for the higher end of the CT doses. (If the CT dose was much above 100 mSv, SNT risk would be higher.)
But the "current accepted model" is still nonsense because it denies our ability to repair radiation damage. This shows up big time when the dose is received over many repair periods as in a nuclear power plant release. In bumper sticker terms, dose rate is far more important that cumulative dose.
I see. Interesting. While I have your attention and expertise what about a study like this in 680,000 people from Australia exposed to CT scans vs controls. Seems to be dose related and location related inc in cancer risk? Maybe you cover this in your book. https://pubmed.ncbi.nlm.nih.gov/23694687/
There's almost no discussion of medical exposures in the book. Probably an oversight that needs to be corrected. The book is concerned with NPP releases so focuses on cohorts
that have received large doses over long periods.
One issue with medical exposures with maybe the exception of mammograms is that healthy people don't get Xrays. This is especially true of young people and CT's. Something was wrong with those kids. And something wronger in the kids that got mutliple CT's. Does a one year lag turn these kids into healthy kids? On top of this childhood cancer is rare, so statistical power is poor. And a few extra cases generates a much larger ratio when compared to this small number. Doctors should use the absolute risk in balancing risk vs benefit. Finally, as you point out, the dose associated with a CT is uncertain and couldbe much larger than the authors' unmeasured estiamte of 4.5 mSv. All in all I don't think we can make much of the Australian study.
I should have skipped the bunker sticker comment. Dose is important but what is more important is how that dose is delivered. It makes a big difference whether a dose is delivered in a series of spikes or the same dose is acquired evenly. In the CT case, SNT says the harm is linear in the number of scans assuming each scan has teh same dose and is spaced at least a repair period apart. LNT says it makes no difference whether you get 50 CT's in a day or
1 CT a year for 50 years.
This is really helpful. Thank you for your insight! Excited to read your book
I should have added the absolute risk numbers. Out of 10.9 million kids and young people 60,674 were diagnosed with cancer. That's an absolute risk of 0.0056 reflecting the rarity of childhood cancer. The CT group incidence was 24% higher an absollute risk of 0.0069.
The difference is 0.0013, a little more than 1 in a thousand. How much of that one in a thousand is due to the rad dose, and how much is due to whatever caused the doctor to order the CT for a child?
Once you get me started , I can't stop.
If we assume all the extra cancer was due to radiation and assume incorrectly all the CT kids got only one CT, then according to LNT that dose would have had to been 14 mSv and per SNT that dose would have had to been about 38 mSv. If we assume a significant portion of the increased cancer was selection bias and your observation on CT doses, I don;t think we have a dramatic conflict with either model, although SNT points to more selection bias than LNT. The argument against LNT lies elsewhere.