I'm going to talk about the ah, difference between the official risk estimates and what we think should be the true situation. Ah, for many years I've been associated with ah, two studies which have come up with much lower ah, doses required to cause cancer than the official ah, wr--ah, story. So, this--you will find this work I've done if mixed up--

The story of what we've ah, done in the past will be interwoven with what is the official view. And ah, with all those threats about timing, I better start straight away. Is that better?

Ah, the official source of ah, risk estimates is, ah--comes from a long term follow-up of A-bomb survivors which started five years after the bombing of Hiroshima and Nagasaki and it led to the ah, repeated confirmation of the fact that this population of five year survivors had a normal non-cancer death rate and a cancer death rate was directly proportional to the dose. In addition, the cancer risk was greater for leukemia than for other forms of cancer. Ah, it was also greater for people who'd been under 30 years of age when exposed, than for older persons. And it h--showed that only persons who had received the equivalent of five rads had any increase in the cancer death rate.

Now, this was a very [conserving?] findings for people who immediately jumped to the conclusion that they had in the A-bomb survivor population an ideal situation for confirming what had been found in animal experiments and for deciding how to estimate the risk at much lower doses than of course the ones that had been given to the A-bomb survivors. So, the first part of the story is ah, they in--becoming increasingly confident than everything is all right, because other studies of high doses, mainly animal experiments were confirming what had already been found. But they--so confident did they become that they finally came to the conclusion that--how to use the data--I'll show you my first two slides.

Here is a--a summary of some findings up to a certain date which ah, ah, relate to the fact that there's no evidence of any extra lung cancer deaths. Although there were a th--fourteen thousand four hundred and five deaths from natural causes other than cancer, analysis of the whole material and its major components provided no support for the belief that diseases other than cancer are included. You will note that the [left span?] study population which is a special group had 80,000 people out of a total of 2--280 thousand people that were available.

I'm no good at doing this.

The final conclusion was that the use of A-bomb data for risk assessment is generally predicated on the assumption that the survivors apart from their radiation dose [are representative?] human beings. This comes from an annual report of the Radiation Effects Research Foundation and no clearer statement can be made that they thought that they had a perfectly normal population on their hands.

Now, this is wha--what worried me about this ah, ah, study: If that is true, then the survivors are representative of the non-survivors. They were also representative of the people in the country at large, so that they could ah, ah, be chosen to ah, have a method of risk estimation that applied within the group. The method of risk estimation within the group was that everything was linear--that the cancer is--was directly proportional to the dose--so you could come down to a low dose level on the basis of what had been found mainly at high doses among A-bomb survivors.

Now, if that was so, and this is what worried me, wh--what had happened to all those people who died before 1950. There'd been hundreds, thousands of deaths from ah, acute effects of the radiation and from devastation effects of the bomb. These risks were espec--these deaths were a special risk for infection sensitive people. What had happened to a population that it could get back to normal in less than five years. It didn't to me make sense.

But there was a further reason f--for me to be worried about this. My work on the children had led me to believe that ah, there was a cancer risk from a single exposure through diagnostic X-ray before birth. When this ah, estimate of similar exposures for A-bomb survivors--prenatal exposures--they failed to find this evidence. In other words, here was--why was data from the A-bomb survivor study being less efficient in disclosing a cancer effect of pre-natal irradiation than a study in which there'd been no question of any high dose exposure at all. You can see what was in--running through my mind--the possibility of the deaths intervened and stopped you seeing it.

But of course the official view was that we were wrong. This, by 1980, we had also discovered with the help of Dr. Mancuso that the risk estimates were workers in the nuclear industry, was producing evidence of cancer risk that this was supposedly safe. You will remember the original official view is there's going to be no cancer risk unless anybody had had 5 rads. And here was evidence of a cancer risk at a fraction of that dose, and worse still, this risk was getting worse the older the person. Two contrasts with the official view. One about the relationship with age. Our finding to me seemed more sensible, in other words that elderly people are more sensitive to all causes of death than young people. See me fumbling with my--why should suddenly radiation be an exception to this rule.

So this is the point at which I put forward the possibility that what had really happened was that the normal lung cancer death rate was an artifact. How could it be an artifact? That there really'd been selection against infection sensitive people which had left the population strongly biased in favor of infection resisting people, but contrary to what everybody'd thought some extra lung cancer deaths had continued. The idea was that of course that you wouldn't be able to find this cause there would be one particular cause but supposing that, ah these extra deaths were merely the result of having the immune system damaged--then there would be no specific cause of death--it would just be certain people who would be less resistant to infections and they would have been if they hadn't had the bomb exposure. That was my position and it was sustained to some extent by the fact that a big common cause of death in the early days following the epidemic of [marrow bone?] damage, had been deaths from a-plastic anemia. And these were continuing after 1950, but had all been ascribed to leukemia.

Now, this is my position in 1982. That all we were looking at was an artifact and that really the A-bomb survivor population was greatly prejudiced in favor of highly resistent people, therefore the risk estimate to the public would only apply to say, young adults, and you wouldn't be getting the true situation in regard to the two populations we'd been looking at--one, unborn babies and the other, workers who were over thirty years of age, and this is the story I have to tell you.

Now, the official t--view demands certain things. It says if these survivors are equivalent--have the same of infection sensitivity as the non-survivors then you should find that the--this is also true of the people in the survivor population who were most likely to have died. Who would those people be? Well, of course high dose would be the immediate thing, but no, we wanted something independent of those [...?...]. The obvious one to go for--how about the people who had had acute bomb damaging [risk?] compared to those who hadn't had it. Now here was a possibility, because in fact at the time when the A-bomb survivor population study was assembled, everybody had--who was in the population had been asked where they were standing, whether they were shielded and whether they'd had any radiation injuries. And these had been subsequently coded under four headings which you'll see ah, i--in the slide. So I knew that data existed. The testing that--that one requirement of the official hypothesis which was that inside the survivor population there should be no difference between people with than without, ah, injuries. I hope you can see this in the slide.

It says here is--this is the situation according to the current hypothesis. Here are the people in 1950 living without injuries. These--these survived, these died. Here is the situation today inside this population of survivors [who have been to?] group of non-survivors who is exactly the same as the others. Or my hypothesis which--which said the infection sensitive people will be--people who die will be more infection sensitive than the people who live. Here they're the same, here they're different. Therefore you should get a similar difference here.

Now for many years I had to press to try and get the use of these data. I wasn't able to do so, but, so I had plenty of time to consider these necessities for two theories. Ah, we've got here what I call the mortality requirements. You've got the LSS cohort and you've got a c--cohort consisting of people who had survived in utero exposure, which is repeatedly being used to study brain damage effects of radiation and cancer effects of people irradiation and it was this, of course--this cohort th--different from mine. Now, what requirements will you have for those to be ah, what there came to be, namely representing normal population. It would have to have a normal lung cancer death rate--the LSS cohort. It might have to supposedly been found. But there might be a weakness in that story, but that over and above that there should be no difference between survivors with and without injuries.

So far so good. What would you require for the in utero cohort. Now the in utero cohort consist of children who'd been conceived before the bombing. All human populations have a constant daily rate of birth, therefore if this [...?...] representative sample of the people who'd been conceived but not yet born, then there should be a const--it should have a constant birth rate. As it was, in fact, the in utero cohort, that first made me suspect that something was wrong, let's look at it and see whether it does agree with this requirement of normality that the same number of births in each month of this population. Now, ah, the population that has been described under the headings of the months of birth contains one thousand two hundred and eighty-eight people. Those of you who are familiar with the studies of the brain damage will notice that it's a bit smaller than the population they usually quote, which is fifteen hundred ah, but is all that you can get in published data.

Now here are the calendar months of birth--that's all they give you. This is of course the approximate people age of these people and here is the birth rate that was observed and what you will notice is that the month of January which is in the middle of the scheme has a birth rate very near the [eight?] births per diem and no other month has anything equal to it except the one immediately next door. So let us assume this was the quietest month between a period in which you could easily be getting extra still births because of the devastation effects. The first half's a period--period between August and the end of the year. A mo--the other half of the population which would have been most sensitive to in utero deaths from radiation. The dose required to kill an embryo is very small comp--compared with the dose required to kill a fully grown person. And here sho--in--I've included here what would have been the effect if there'd been a constant birth rate of eight per diem. And it says you will be missing 952 places from here. Mo--some of them will come from the period before and some after. So, here it fails to meet the first requirement of a constant birth rate and it fails to do so in a way that would very much favor my theory which was that this population was at risk of extra still births because of the general devastation and at risk of extra abortions because ah, ah, the foetuses are very sensitive to radiation effects. So, [why not to ask...?].

Right. What had--what was ["us/ask"?] consist of? We never--for a long time we had no access at all to A-bomb data and in then [the IRF?] released a thing called ah, LSS data on disk, in other words, a floppy disk with certain items of information about the population--not enough, of course to identify any individuals but enough to be--for an for an indiv--person who could buy this tape--you could--$50 would purchase it from Hiroshima--could study the causes of death under various headings. Sex, City, Dose, Age, Age of Death and Cause of Death. Now that looks generous, doesn't it--36 causes of death to choose from but in fact, one is all causes, so that reduces it to five, and 30 are different types of cancer and that leaves you with any five causes of lung cancer deaths. The--that was possible to study this.

Th--this is where I began requesting we have something added to this which was the data that I knew existed about the injuries. And what you're going to hear next is what did come when we finally managed to persuade the DOE to release--well, DOE had to persuade Japan who had to [...?...] that finally allowed us to have added to this information here--first of all an update to 1984, secondly the updated dose, and thirdly the addition of the injury data.

Here's the injury data. Here was the--I call it self-claimed injuries. You must understand that this was collected after the 1950 census, from people who were going to be included in this population by virtue of their--where they were when they were--the--the--at the time of the bomb and they gave an answer to the following questions [were coded?] under the four headings and they had burns, spontaneous breathing [puerperia?], mouth ulcers, throat ulcers, or had--they had temporary loss of hair. And in addition to giving the numbers that you hear in the order of frequency, I've told you how many denied having an--all four injuries. They have to have four denials. Who claimed one--a single injury (8,000), and who claimed more than 2,000. Now various doubts have been expressed about this ah, suitability of taking self claims, so we said, well, the person who denied all four injuries probably knew what he was talking about, and the person who claimed two or more injuries probably also knew what she was talking about, for the--so for the purpose of our study, we will compare those who denied having any injuries to people who claimed at least two out of a population fl 74,000 0-4-2. Those of you who are acquainted with [Bier 5?] will recognize this figure. It's actually about 2,000 less than the 7-5-9-9-1 cases which are the present basis of the cancer risk estimates on which ah, all the safety regulations now depend.

Now what does the [Bier?]-5 do? Bier 5 only recognize one cohort--the total cohort. But it gave for all causes of death--the various other causes of death--the results of what of a p--what progression--what [s...?] regression analysis, which allows you to recognize whether the number of deaths recorded [per given?] exposure age--age at the time of the bomb for six age groups, it allowed the [Bier-5] committee to discover--decide--whether there was any extra deaths. Now, the deaths here, it says--it's a no extra deaths s--sh-- so should be zero naught point naught naught would be that you were [bang on?] scale, minus would say you had too few and anything about that should be rated as a percentage.

For instance, here there was 70% of extra deaths in the youngest age group. These are needed for certain tests [...?...] but you can see straight away what all causes of death what total cohort was, as ah, officially recognized. There is a certain amount of in--evidence of increase, but it's mainly in the younger age groups and doesn't amount to very much in the way of ah, significance.

Now, what happens if we divide it into our groups. These are the 2601 cases with multiple injuries and here are the 63,000 odd with no injuries at all and there's the intermediate group. We will--here is one that, of course, is most like the original one because it contains 85% of the whole group and you will see sure enough, it comes down somewhat the same--the findings are very comparable. But what about the c--our case group?

And here you get an immediate, fantastic figure for a [m...?...] around small group-- people who were under ten when they were exposed. Now, I'm showing the whole of this rather carefully because this is what we're looking for. Is there any [...?...] group inside the survivors which was too small to have attracted any attention at the time, but could make all the difference to the final conclusion which is, do survivors represent non-survivors, or are they really a totally different group but just a small residue that's going to s--tell us what really happened. This is our small residue that--these are the people who are most likely to have--does that mean stop? [No, almost...not yet] Ah, is there any--now the word used is non-homogeneity of this group. Is it what it's-- telling us it its ah, is--it's a normal, single, homogenous population or has it got the special group in here which is really showing the characteristics ah, more closely of the people who died. So here's the whole group, here are the controls and here's the intermediate. Here's the [eight?] distribution, and this--[...?...] look at this more closely for other causes of death. Here we got [familiar with?] tumors. Again, you've got the total cohort--these figures are almost exactly what you would find in ah, [Bier-5] then this is where the reputation comes that exposure in the first half of the life span, under 13 years of age is more dangerous than exposures ah, after. You can see the figures. Here's a--a doubling of the risk for the first year--108% increase, 79%, 27% and much lower figures here: 31 and only 2 and 50 [...?...] ah, the--this we don't need--[this is for the test of s...?]. All right.

That again--it's repeated more or less with our control group here. High risk in first half, low risk in second. But is it repeated for our cases? Not at all. Here you find now that there's evidence of trouble not only in the youngest age group, but also in the oldest age group. These figures are fantastic--this is nearly 8,000 times too high. And this is, of course, the same figure. So there is a difference. We need go no further to say that we've established our point which is that it is not a ho--normal homogenous population. Of course the test also allows you to say is cancer the only cause of death after the ah, five years? And here we have a regression analysis and now [...?...] I've narrowed it down so that we see the--the--the control group in the cases are now called nil injuries or 2+ injuries and the same age here, and the same excess relative risk as before but just so you can see it now for cardiovascular diseases only. And here I've put the numbers in there so that you can get some idea of the [...?...] for this.

Ah, in the control group, there's no evidence of any--none of those [...?...] significant, any excess, but look at the ca--control group. A thousand fold increase and it goes on having a hun--over a hundred percent up to the age of 30. Here there's more difference. So here is evidence that there's extra deaths from pink label cardiovascular diseases. What about the other non-cancer deaths? And you get exactly the same story. Here there's--here's the negative findings for ah, the--the total group controls, and here are the positive findings.

Now, I put this diagram up to show how it is that we've acquired this reputation for being ah, out of line with the official view. You will realize that there's now a sort of c-point system that you're dealing with. StUdies of ah, exposures before birth. [Birth?] is represented here, and here is studies before birth. Studies of ah, the A-bomb survivors, going right through the whole age range, and studies of workers which begins about age 20, and go up to age 64, when you retire.

Now, what has really been happening is that our findings have not--ha--there's not been much overlap with the group which is the dominant group in the A--A-bomb studies. The people who were most likely to survive were the young adults. And I've cut this out because as you see, I'm already at the end of my time period. Ah, didn't--haven't shown you any of the details, but the these gross underrepresentation of young people in this population--old people--most of them are c--concentrated here, where we don't have much difference with the occupational [side?]. When I say much difference--these lines seem to meet--I may t ell you that only twice as high--low as it ought to be. But as soon as you come out to the--ah, our story about the prenatal and postnatal, then you get a totally different story. So I think I will finish the talk by saying that if you do allow--if we allow relaxation of radiation standards, the people who are going to suffer most are either the unborn children or the young children in this world or the second half of the age span--over 30. Thank you.