ML20005B546

Effects on Populations of Exposure to Low Levels of Ionizing Radiation:1980. Certificate of Svc Encl

ML20005B546
Person / Time
Site:	Perry
Issue date:	12/31/1980
From:	NATIONAL RESEARCH COUNCIL
To:
Shared Package
ML20005B540	List: ML20005B539 ML20005B546
References
RTR-NUREG-1150-2-V1-2.41, RTR-NUREG-1150-2-V2-A.66 NUDOCS 8107080355
Download: ML20005B546 (50)

Text

.

4

- XPosure to yon?.

evels o^'

L s.

('

izing Radiation 19y

~(

NAS i-NA3

.OV

~

M 2

ExhiDit 1 l

?x.n

.__ sm m.___

n-av t

'hhe Effects on 77; l

Populations of

'.. T Exposure to It '

w,s Low Levels of Ion 1 Zing i

Radiation: 19807 i

pMMITTEE ON THE BIOLOGICAL e

EFFECT5 OF IONIZING R A DI ATIO N S, i

Division of Medical Sciences Assembly of Life Sciences National Research Council, Lg'BR EAR REGULATORY (y]%ylq WASHINGTON, D.C g STOP 555 b

APR 17 88I NATIONAL ACADEMY PRESS Washington. D.C.

1980

?

?

I

[

i

..s-s,y _ f n

m.

a :..

3g;G:,533

,.:.v

?'

s' 1

,Q

.v;e.wb%gn"

. _ ;~ :

~ r ~h v.;; j &gur., Em kyN

,I

...s f.N $:.

,...en s:=>..-

a t.

fth.

.c.k.[N. '

4

.w M t.t.S,. 6.

......4,.s.__,,

... _ O

'" JEBIEhm

.m this report mas approted by the Goserning July 22.19SO

.se members are drawn from the Councris of al Academy of Engineenng. and the Institute responsible for the report mere chosen for Afr. Douglas Costle appropnate hafance.

Administrator her snan the authors accordsng to procedares Environmental Protection Agency sisting of members of the Natmnal Academy 401 Af Street. S. W.

rirg. and the Institute of Medkine*

Washington. D.C. 20460

.hed by the National Academy of Sciences in

.cience and technology m.th the Academy's

., sing the federal go ernment The Council

Dear Afr. Cost /c:

r*cerrmed by the Academy under the author.

' '"* binh'd ' h' A c* d'm> ^ ' ' P'" * *' "*n -

I am pleased to transmit the report "The Effects on Populations of k,' o ien$e?and$,e na Acad$m Exposure to 1.ow I.evels ofloni:ing Radisation"prepmd under contran at a the gesernment the public, and the scien.

68 05 4.105 with TPA 's Office of Radiation Programs.

nnistered jointly t.y both Academies and the The report. familiarly known as DEIR 111lafter its authoring of Erismetring and the institute of Medncine Committee on the Biological Effects ofioni:ing Radiations!. has had a

. under the chart.t of the Natsonal Academ) troubled history. In hfay 1979. a version of the report was publicly released. But when it was learned tha; a significant number of mo ted by the off,ce of Radsat>on Programs.

committee members believed that the somatic effects section of the

.act ss.on. cog.

report did not adequately reflect thefull range of committee opinion generated by the admittedly incomplete data base, further distribution the typesenpt edition that mas released in was discontinued.

y the Commnttre (necessitated chiefly by cor' It is not unusualfor scientists to disagree on the interpretation of

"'P*#

data. Generally, the sparser and less reliable the data base. the more opportunityfor disagreement. In this case, there are sufficient data concernh;g the effects of exposure to high doses ofioni:ing radiation.

^* #

but little reliable information concerning the consequences of exposure

,3095.x to lower doses, especially those low doses to which a human population might be exposed. U?>on the issue of how one may extrapolatcfrom the high doses o the low, scientific argument turned on the question of how one may nulidly extrapolatefrom the measured effects of high doses to the most probable effects oflow doses.

The BEIR 111 report exhibits the range of opinion concerning how this extrapolation may be performed. Afany committee members believe that the data best support a linear quadratic modelfor estimating risk:

others. however. believe that the linear or pure quadratic models s

provided better estimates. The report presents all of these views. in balancedfashion. The committee as a whole, despite individual preferences, has agreed that the report treats ecch of the possible interpretations in afair manner. Two members have notfound it possible to endorse the report. The dissenting statement by Dr. Radford espouses the linear model: that by Dr. Rossifavors the pure lii

- ' lC s L+;&. ' ZW @6pir.m"" n

~

~'

f

[*

~W? ? %%CQ;jd&g.w g. w g w +% n&'@g,$,n " W

.f ff.

miniy'b(s' 'f f

'4$hfbf5 PM.? PW;a% m :x& w& ??W 6 '_ ?, :

' q;& ::?:^Y:t,kl<l.Q yi~:?:.y:

!!?":gW;E " 2' M ~ K '

'C

=*y W

' ' ' OZ

, % y.

4l; ;.f ti'h D,,','.;.y. % x... g l 5 Q' ~n i; y % 3 g'? y. ;~ _ *' &

,';y.e

..a, x,

' 2 7. -%,. : Y-1,,.

?

-=

n xh W

%.m.

g; " ;-

quadratic model. Both models are included. The polarity of these two views best illustrates the degree to which scientists disagree on this subject in absence of sufficient evidence to compel conclusion.

We believe that the report will be helpful to the EPA and other agencies as they reassess radiation protection standards. It prosides the scientific bases upon which standards may be decided after nonscientific social values have been saken into account. If social values dictate a conservative approach, the report's linear model risk estimates may serve as a guide. If one wishes to accept scientists' best judgment while recogni:ing that the data simply will not permit definitive conclusions, one may select risk estimates using the linear quadratic model as a guide. Other considerations may lead to use of the pure quadratic risk estimates.

We regret that the transmirual of this report has been delayed so long. The Acade ny believes that the delay was necessary to permit time for restating the report so as to display all of the valid opinions rather than distribute a report that might create thefalse impression of a clear consensus where none exists.

Sincerely yours.

Pw Lir H ANDLER President, National Acaderny of Sciences iv q#,,

3

'HE EFFECTS OF IONIzlNG RADIATION

,,,',', 'i'"l,,(hal nd atlon da'na a in rnan-Statement Concerning the Cutrent Version of Pa a ed Radiations. Cmnd. 9780. London: H'. M.

Advisory Committee on the Biological Effects of ino. 5. Tokuoka. H. Esaki. I. Nishimori. and

,ng aiomic domo, unison. Hiroshima and Ionizing Radiations (BEIR III Committee) t>2:134* lM9.19*9.

of gamma irradiation on the deselopment of EDW A RD P. R ADF0RD, M.D.

effects. Radiat. Res b0:303-316.1979.

the Effects of Atomic Radiation. Sources and Profe33Or of Environmenf af Epdemiology Graduate School of Public Health Univert,ity of Pittsburgh t to the G.neral Auembly. New York: United i

Chairman. BEIR III Cof'Imittee end and Welfare. Third Natonal Cancer Sursey.

d J. L. Young. Jr. N sti. Cancer Inst. Monogr.

Chairman. Subcommittee on Somatic Effects

, ington. D.C.: U.S. Gosernment Pnnting Of.

l n the pathogenesis of leukemia in anirnals and i

gs of Third Natonal Cancer Conference.

Sniffen. f rfluence of age at time of irradiation mors in er mice. Proc. Soc. Emp. Bml. Med.

The present version of the report of the Advisory Committee on the Mihailonch. K. V. Rao, and L. 5 Lombard.

Biological Effects of lonizing Radiations (the BEIR !!! Report) is a on r.-diaten caremogenesis in sanous tissuc of modification of the draft report approved by the Academy in April 1979 and released at a press conference at the Academy on May 2,1979. Subse-1 life shortening and premature agu.g. pp.

quent modifications of this approved draft have been prepared by a group

. Advonces in Radiation Biology. Vol. 5. N;.

iapointed by Dr. Philip Handler, President of the Academy, consisting of six members of the somatic effects subcon;mittee and one member of the genetic effects subcommittee. The modifications involve principally the section of the report summarizing cancer risk estimates (the third and final section cf Chapter V) and some of the conclusions that flow fiom this section. Cancer is a somatic effect of radiation, that is an effect on the body cells of individuals exposed, as distinct from effects on the germ cells or genetic effects. Thus, the sections at issue have been the responsibility of the subcommittee on somatic effects of the full BEIR !!! Committee. This subcommittee originally consisted of seventeen members w hose names are given in the front of the report. This r. amber has been reduced to fifteeit by the deaths of Dr. Benjamin Trimble in November 1977 and Dr. Cyril Comar in June 1979.

Dissenting statements prepared by indiv' dual rr.:mbers of a National Research Council com-mi~ee are not subject to the normal review processes of the National Academy of Sciences; not are they subject to ummittee or staff editing or renew. They appear exactly as the dissenting committee members prepare them. The nas. pac neither endorses nor takes responsibihty for the content of the statements.

227 1

l l

.a

..... v v.

~

228 THE EFFECTS OF IONIZING RADIAMoM The material prepared by the subcommittee on somatic effects was u rit-ten largely during 1977-1978, with occasional one or two day meetings of the subcommittee to review draft material as it was prepared. It is impor-tant to note that the Icst meeting of the full subcommittee was held on December 15,1978, a one-day meeting. The new material incorporated in the report since May 1979 has, therefore, not been approved by the sub-committee as a whole except by the process of asking for comments by mail. Perhaps because completion of the mE R ::: report has been delayed for such a long time, few members of the subcommittee have responded.

Nevertheless, the present version of the report includes very major change from the earlier draft and from the BEIR I report of 1972. That is the deci-sion to adopt the so called linear-quadratic model(excess cancer risk =

aD + bD2, where a and b are constants and D is radiation dose) as the f'

basis for calculating risk at low doses of low LET radiation for all cancers, 4

4 and not just leukemia as in the previous draft. In addition, risk estimates s

calculated from a model in which the excess cancer was assumed to be proportional to the dose squared (the so-called pure quadratic model, ex.

cess risk = bD2) were also included. The effect of adopting the linear-quadratic model is to reduce the risk estimates at low doses somewhat.

The pure quadratic model implies a very low risk at low doses.

The decision to use the linear (straight line) no threshold model (excess cancer risk = aD), which implies e risk directly proportional to dose at all levels, for all radiation types and for all cancers except leukemia was the f.

result of a vote taken in a meeting of the subcommittee in October 1977.

This vote has never been rescinded by action of the whole subcom nittee, and thus as chairman of the subcommittee, I cannot consider that the present version is in accord with the perceptions of at least several of its members.

The most serious consequence of this alteration in the conclusions of the earlier draft, however, is that all of the discussions and evaluations of the data on cancer risks that took place among subcommittee members as the draft material for the report accumulated during 1978, did so on the basis that the linear model would be applied. In this regard the subcommittee was adhering to a pr:neiple adopted by the BEIR I Comrs.ittee, and as an expedient measure, in v'ew of the limited amount of tina available, I had felt that we would not spend our time reviewing in detail the scientific basis for those conclusions which agreed with the BEIR I report. In short, the requirement to complete the report in 1978 imposed by the Academy staff meant that the extent of discussions of fundamentalissues had to be limited, particularly for matters that had been thoroughly presented in BEIR I. Thus, a detailed and critical discussion by the subcommittee of the

,y-s A

P

~

+-

et i

w 4,.-

g

' +-

s

HE EFFECTS OF IONIZING RADIATION Somdfic fffecfJ Cancer 229

>mmittee on somatic effects w as w rit-scientific basis of deciding whether one or another dose response model seasicnal one or two. day meetings cf was applicable to cancer risks was not undertaken, terial es it was prepared. It is impor-One exception to the abose statement was the data from the Japanese

' the full subcommittee was held on A bomb survivors. The results of the follow up of cancer exprience ig. The new materialincorporated in through 1974 in this important study population had been made available fore, not been approsed by the sub-to subcommittee members in page proof by Dr. Gilbert Beebe in 1977, but process of asking for comments by n this form it uas used primarily to proside an important source of data the BEIR ::: report has been delayed for the individual cancer risk sections being prepared by sescral members f the subcommittee base responded-of the subcommittee and now found in Appendix A of Chapter V. Bound ie report includes sery major change copies of this report (Life Span Study Report 8 Technical Report RERF TR trR report of 1972. That is the deci' l 77)were distributed by the Academ~y staffin mid 1978. The significance idratic model (excess cancer risk =

of this distribution was that for the first time all the members of the sub-ants and D is radiation dose) as the committee had, in an easily readable form, the latest information concern-of low LET radiation for all cancers-ing cancer risk in this population. At about the same tier we obtained sus draft. In addition, risk estimates the Oak Ridge calculations of factors by which kerma doses could be con-ie excess cancer was assumed to be i

serted to specific tissue doses for both gamma ray and neutron exposures so-called pure quadratic model, ex-n the two cities. Subsequently, a large amount of time during the remain-

. The effect of adopting the linear-ing few meetings of the subcommittee was spent in discussion of cancer i estiraates at low doses somewhat.

data from this report in terms of the tissue dose response relationships rery low risk at los doses.

that could be inferred from the data as presented. Since such a process ght line) no-threshold model(excess amounts at best to fitting theoretical lines to data points, in these discus-d directly proportional to dose at all sions the subcommitt.e did not address the fundamental scientific basis of all cancers except leukemia was the any of the models proposed to fit the Japanese data.

the subcommittee in October 1977.

In my siew, new data, obtained since the BEIR I report in 1972, strongly I cetion of the whole subcommittee, supported the decision of the nF R i committee to adopt the linea: no-imittee, I cannot consider that the threshold model for cancer induction by radiation.1) New human studies i

percept ons of at least several of its were available giving stronger evidence of effects in the 10 to 50 rad range, and these studies generally gase about the same risk of excess cancer per is alteration in the conclusions of the unit dese as the higher dose data had. 2) The range of exposure patterns e discussions and esaluations of the to low ttT radiation included more studies of multiple small doses which nong subcommittee members as the could be compared to effects of i ale dor.s. 3) Studies of individuals ited during 1978, dia so on the basis especially susceptible to cancer induction by radiation and other car.

d. In this regard the subcommittae c:nogens were being expanded (e.g., see Chapter II, the section entitled

>y the rEIR Committee, and as an

" Cell Mutation or Transformation"), and there was a possibility that ted amount of time available. I had these susceptible populations might be fairly large and not identifiable in te reviewing in detail the scientific advance. This possibility suggested at least that cancer risk estimates a:

red with the strR report. In short, low doses for this population subset could be somewhat higher *han would rt in 1978 imposed by the Academy be in' erred from studies of u selected populations. 4) Studies of on-ans of fundamentalissues had *.- 5>e cogenic transformations of human and animal cells in cuhure had been hnd been thoroughly presented in g,reatly expanded, with startling new results that challenged many of the scussion by the subcommittee of the traditiona; radiobiologi concepts that had formed a scientific basis for 1

v 2 "~

w.

.s

..c

..a g

%b.y3N;Qg ^,_y.. m ~ < [. w-#'*

"e

, m' 53; 4%rm. q

,4 230 THE EFFECTS OF IONIZING RADIATION extrapolatior of effects of higher doses of low LET radiation into the low dose range. These results suggested, for example, that DNA repair did not necessarily imply that low doses of low LET radiation would be less car-cinogenic per unit dose than high doses. 5) Finally, new evidence of cytogenetic changes observed in populations living in areas of high background radiation exposure had been obtained. At my suggestion this last esidence was not considered extensisely by the subcommittee, primarily for the same reason thy were not by the Bria e Committee; that is, the significance of these changes obsened in circulating lymphocytes in terms of human disease had not yet been defined. But these observations indicated that effects of radiation exposure at doses and dose ra'es moder-ately above background could be detected.

All of the above considerations indicated not only that the decision of the present subcommittee to reaffirm the applienbility of the linear no-threshold dose response relationship was the correct one, but also that such a decision was not so conservatise as had been thought at the time of the BEIR I report. That is, the cancer risk estimates for exposure to low doses based on the straight line extrapolation could be somewhat lower than might be found eventually to apply, especially to susceptible subsets of the population. Such an underestimation of risk, the subcommittee agreed, would be unlikely for low LET radiation, but the view that th; linear extrapolation greatly overestimated the risk of low LET radiation at low doses appeared to me to be equa!!y unaarranted. For high LET radia-tion, such as alpha radiation, the straight line extrapolation could underestimate th(.isk at low doses, but the evidence was not strong that such underestimation was Sery significant except in its theoretical in-ferences.

I now proceed to consider in some detail the scientific evidence perti-nent to estimates of cancer risk in human populations from low doses of radiation. Of special importance are two questions that have divided the subcommittee. First, what i> the experimental evidence to support the linear to threshold dose response relationship of cancer induction? Sec-ond, to what extent are the data from the Japanese A bomb sunivers con-cordant with all other human studies, and also consistent with linear or other dose response models? A problem related to this last question is the degree of concordance of results from the two cities, liiroshima and Nagasaki, and from comparison between the two cities the appropriate in-ferences to draw about the t'.;ative.4fectiseness of the neutron component of exposure in Eiroshima. (The type of bomb exploded in the two cities differed: both resulted in exposure to gamma radiation, but the liiroshima bomb had a sigaificant fraction of the radiation exposure from neutrons.)

m

., f. -

Y.-

q f

-y f f x }{

231 8omgric Effedfl Cancer x

i EFFECTS JF IONIZING R ADIAttON Some general comments are in order at this point. First, there was no

. of low trT radiation into the low disagreement cmong the members of the somatic effects subcommittee to r example, that DN A repair did not accept the linear no-thresnold dcae response relationship to define genetic

. trT radiation would be less car-effects of radiation at low doses, a position firmly taken by the BEIR !!!

ases. 5) Finally, new evidence of subcommittee on genetic effects (Chapter IV)in agreement with the aria t iutations lising in areas of high report. Based especially on the mouse siudies of William L. Russell a.2 en obtained. At my suggestion this a member of the subcommittee on genetic effects for both aria and i,rtR stensisely by the subcommittee, att, the subcommittee did recommend that for low LET radiation exposure

not by the atin Commi ee;that at low dose rates, the mutational risk per unit dose for radiation of the tt 4

,erved in circulating lymphocytes in male testis is probably less by a factor of three at low dose rates than for en defined. But these observations equivafent doses gisen at a higher rate,

,ure at doses and dose rates moder-In the present sersion of the report, there is an inconsistency between c'ed.

the conciusions of the two subcommittees with regard to the appropriate cated not only that the decision of

. dose response relationship to be applied for genetic and carcinogenic ef-the applicability of the linear no-fects of radiation. Consistency in esaluating these two effects of radistien was the correct one, but also that is reasonable because there is now wide agreement among the scientific as had been thought at the time of community studying cancer (for a summary of the evidence see Origins of risk estimates for exposure to Icx Human Cancer ) that a necessary condition for induction of cancer is pro-3 apolation could be someshat lower duction of one or more mutations in the DN A of one or more cells in a tissue.

i

)ly, especially to susceptible subsets l

imation of cisk, the subeva.nittee Thh rnutational change in somatic cells as a condition for carcinogenesis is the foundation of the use of short term testing of mutations produced by it radiation, but the view that the environmental agents as a screening test for carcinogenic potency

• ated the risk of low LET radiation at The entire process of carcinogenesis is a complex one, however, and an ly unwarranted. For high LET radia.

initiating esent, such as a somatic cell mutation, is not the only condition straight line extrapolation could required for cancer to akise, whereas a mutation in a germ cell that retains sut the evidence was not strong that its siabihty is the sole condition of. transmitted hereditary defect. For 6 cant except in its theoretical in-this reason one might anticipate that the dose response relationship for cancer induction could differ in certain ways from that of genetic muta-detail the scientific evidence perti-tion. But it is important to 'iote that the differences in the two processes iman populativns from low doses of arise because of host factors or other biological factors in cancer expres-two questions that have divided the sion that are essentially independent of the initiating event or events, thus parimental evidence to support the not necessarily related either in space or time to the dose of radiation. If, ationship of cancer induction? Sec-therefore. one argues from the above mentioned difference thit the dose-the Japanese A bomb survivors con-response curve for cancer induction should differ from that for genetic ef-s, and also consistent with linear or fects, such argument cannot be based on biophysical principles that relate

, em related to this last question is the to the initiating mutational event. Indeed, because we suspect that many om the two cities, Hiroshima and unrelated biological factors influence the probability of subsequent

.een the two citie s the appropriate in-development of human cancer after exposure to radation (see Chapter 11, fectiveness of the neutron component the section entitled 'dost Factors in Radiation Carcinogenesis"), it is far

! of bomb exploded in the two cities from obvious in which way one would postulate that the dose-response te to gamma radiation, but the curve should be modified at low doses. If evidence existed that a signifi-actior: of the radiation exposure from cantly large group were especially susceptible because of differences in

r~

20 THF F F FF r rs of loNIFING R A 9f AT10m some of the host factors rela'ed to carcinogenesis, we would expect any cancer initiator such as radiation could be more effectise per unit do at low doses than at high doses, where all or most of the susc could aircedy base cancer induced.

The fact. hat we do not yet understand all the factors governing deselopment in man was an important reason why the sul committee unanimousl: agreed to depend primarily on studies of human populations r-to define cancer risk from radiation exposure O

. The number of studies asailable is unpressise, about 50 insestigatmg cancer at sarious sites fr irradiatio for various reasons. In a fem instances the results are nega as one might expect on statistical grounds, or becauw epidemiologic criteria such as a sui table control population were difficu't to meet the remarkable fact is that the cancer rnk estimates der" Yet.

jority of the studies. insobing widely difft rent ethnic groups irradiated different ways for different reasons, show a conuderable agreement n

Chapter V. Appendix A), at least in the higher range of radiation dose where it has been possible to detect clear effects. The cancer mortaht data from the Nagasaki A bomb sunisors are perceised by some mem y

of the subcommittee as an exception. and this point will be d!scussed f

detail below.

EX PE RIMEN T AL B A 515 FOR DO5E-RESPON EE MODEL5 The present senien of the report has departed to some extent from the subcommittee de<.ision to depend primarily on human studies for cancer risk estimates in that adoption of the linear-quadratic dose-response model as the p-imary model to use fer extrapolation of low dose effects of low uT radiation has been strongly influenced by data obtained on laboratory animals. which usually show cancer dose respanse relation-ships cunilinear upward within. say. 200 rad. This influence is under-standable if one considers that the human evidence of cancer risk is sp for low radiation doses, but there are many reaso s wh:

animal stadies are of limited value. and indeed may be misleading. with regard to dose-re,ponse iniarmation for human cancers.

These reasons include: 1) Animal cancers at particular si'es may differ morphologically and in growth characteristics from human tumors at the same site. and for this reason initiating and pramoting processes could be quantitatively different. 2) The strains of experimental animals used for near!y all research are highly inbred, and for each strain susceptioihty to cancer induction is hkely to be more homogeneous than in manHuman 4

ii

~V

~E d(..

# 0 *d ) ' '

y) g j

• ~

I;(,.

g.

m

. ((Q :

gF C

3'~

' $, p j p., A' g. 7{

'7

,.4

.~

g O

- z;

.ff

.

• R' W L'..'_."*

l C-e THE FT*ECTs OF IONIZING RADIATION 8omGfic Effects: Cancer 233 a carcinogenesis, we would expect that p pulations have variable genetic makeup and it is known that genetic on could be more effective pet unit dose factors influence cancer susceptibility.5 This variability would have the iere all or mo;t of ti e susceptible group effect of making the response at low doses greater per unit dose than at I

higher doses where the proportion of cancer-sensitise groups affected d.

erstand all the factors governing cancer would be less. 3) The life span of most species such as rodents widely used portant reason why the subcommittee for experimental studies of cancer is short, generally two to three years, marily on studies of human populations and the latent period between exposure to radiation and onset of increased tion exposure. The number ei studies cancer incidence is proportionately a larger fraction of the life span in ivestigating cancer at various sites from these species than in man. 4) Because animals used for lifetime studies of a few instances the results are negative, cancer development are ker in artificial surrour. dings, on a fixco nutri-al grounds, or because epidemiologic tional regimen, and protected from intercurrent infections such as from population were difficult to meet. Yet.

siruses, exposure to a wide range of cancer-promoting or other factors neer risk estimates derived from a ma-w hich could modify cancer expression is thereby kept to a minimum. Such cly diilerent ethnic groups irradiated in exposure is considered to be the almost daily lot of human existence, and is, show a considerable agr eement (see may be an importa,t contributor to the very marked influence of age an in the higher targe of radiation doses incidence of most cancers in man.6 One consequence of this artificial ect clear effects. The cance mortality environment of experimental animals is that for any single chemical or Jrvivors are perceived by some members physical agent under study to lead to frank cancer. both initiating and ion, and this point will be discussed in promoting factors must be provided by the carcinogen;ii the parlance of cancer research, the agent tested must be a complete cccinogen. There are two important :onsequences d this condition: firs:, the latent perio'1 may be inversely related to dose,7 and second, one would expect that the cancer rate w uld more likely be proportional to the square of the dose.

.O R rather than to the first po*er of dose anticipated if only random initiating

'S events were required for cancers to appear. Both these reasons, as well as has depaned to some extent from the the longer latent period in proportion to the short life span of these primarily on human studies for cancer animals, lead to the dose response curve at any time after the onset of ex-of the linear-quadratic dose-response cess cancer being likely to be strongly curvilinear upward. That is, low e foe extrapolation of low dose effects of doses will appear to be less effective per unit dose than higher doses, esen

>ngly influenced by data obtained on if the probability of cancer initiation were random and followed a linear, ty shoc cancer dose-response relation-no-threshold relationship. It is significant that in human studies of say, 200 rad. This influence is under-radiogenic cancer w*iere an effect of dose on latent period was looked for human ;vidence of cancer risk is sparse

f. Appendix A), the inverse dependence of latent period on radiation dose tre nany reasons why animal studies are appears to be slight at most, consistent with the idea that the promoting step of radiation car nogenesis in man is ind pendent of the initiatirig o

y be misleading, with regard to dase-cancers.

event.

ial cancers at particular sites may differ For the above reasons, therefore. I believe it is unwise to rely on dose-aracteristics from hurnan tumors at the response data for cancer inductica in experimental animals to support use

.ating and promoting processes could be of any particular dose response model for human risk estimates from

rains of experimental animals used for radiation exposure at low doses, j

vd, and for each strain susceptibility to In the abrve discussion it is evident that the step of cancer initiation by ore homogeneous than in man. Human radiation is an important element in quantitative understanding of risks u

l fo

?-

2 4

~

u,

• ~ 9 -

a.

p r

.h --

2R Tut ut Fers on: IONIZING R ADI ATioN j-of radiati.:.n exposure. Because this proceu is believed to be a cellular a

.A o

phenemenon, albeit influenced ' tiwie and host factors. quantitatise

_Jt, T.

auessment of dose-response relationships for the proccu of oncogenic

--g

> f,pq - pp transformation of cells has been acSely pursued t.oth in theoretical terr,s and experimentally, especially since the atta i report. One of the mou

]. J fi d# ~

s v widely discuued theoretical concepts in recent years has been the

[

~

Kellerer Roni theory of dual radiation action." The cuence of this theory j

1-is found in Chapter II in the section entitled "Phpical Aspects of the Biologic Effects of Jonizing Padiat.on.

7 lt is important to rmte at the outset the fundamental auumption

^

underlying the theory. which is that pairs of sublesions, produced by I

radiation in critical sites in the cell combint to form lesions which are

~

esentualb expreued as a permanent change in the cel! such as a muta-tion or oncogenic transformation. This auumption is an extenuon of the theon of Lea.' deseloped to account for effects of gamma and neutron radiation in producing gross chromosomal aberrationt In thn particula case the auumption that two breaks f or sublesions) are required it pro duce the effect i< sery plausible. For chromosomal aberrations m human lymphoevtes a dose squared dependence of effects has been obsened for a.

[3 low LET radiation. 30 " consistent with 1.ea's theon. lo ex*end the auumption of two subleuons being required for other effects of r diation than grou chromosomal aberrations requires that experimental esidence g

2 of on effect proportional to the doseaguared he obsened for such effectv This experimental evidence, as referenced. is derned from uudies of chromatid aberrations in Tradescantia. the spiderwort plant '2 effects j

on bacterial spores. and radiation induced life-shortening m animals.

(This last effect of radiation would be expected to insohe non stochastic processes, in sharp contrast to cancer induction, marcoser, the subcom-mittee has coi.cluded on the basis of asailable human data, that no non-specific life shortening effect of radi. tion has been obsened m man.) Thn array of esidence ", far fran convincing justification of the assumption 3

that two sublesions are required to produce lesions in the DNA of mam malian cells that may lead. for example, to oncogenic transforrnat;on.

('

unien such transformation is consiste..:) auociated with gross chromo-somal aberrations.

if we follow the Kellerer-Rout formainm. nesertheless. on the further

+

anumption that the subles.ons interact to produce a lesien o e: a range of about 1 my in the cell. then the freq" enc? of rffects. E KtTD + DU.

f f.

i w here K is an arbitrary canstant and reta is e vaiabl< dependent on the frequency dntributien of specific energie. produced by singl(

!(

,.y gf-Kellerer-Rou, theory therefoi-leads to a linear-quadratic dependence esents The of effect on dose, a conclusion that is onuous from the fundamental

+

E k-; p

.i I

.(

gm

, g d^*"

~

s*#.

g

^ ~

,r g,4 1t M

' NM /g,,

w-M

.s

=

3 A

r%

... f y &.

g. &

• g Q k'.

p}_

(,'

l I"

E EFFECTS OF IONIZING RADIAT!oN i

assumption that pairs of sublesions are a necessary condition of ultimate rocess is believed to be a cell;lar effects. The theory has been applied to the problem of the relative biolog-sue and host factors, quantitative ical effectiveness of different type, of radiation at low doses,in which case hips for the process cf oacogenic both K and f are 5ariables which are used to fit the experimental data. Ex-y pursued both in theoretical terms

,,eriments of Cox et al.,85 in which mutation of HF19 human fibroblasts Se BEIR I report. One cf the most and V79 Chinese hamster cells by various radiations encompassing a wide in recent years has been the range of LET was examined, were analyzed by Goodhead in terms of the action.s The esseace of this theory Kellerer Rossi theory.86 Goodhead showed that the RBE Values predicted entitled " Physical Aspects of.he on the Kellerer Rossi theory were at considerable variance from those observed, and it was apparent that no col ;istent set of values for K and f tset the fundamental assumption in relation to LET could be derived from the data, nor were the derived pairs of sublesions, produced by

" constants" consistent for similar effects in the two species. Goodhead ombine to form lesions which are also pointed out that f, which is equivalent to the dose at which the linear hange in the cell, such as a muta, and quadratic terms are equal and which thus defines the dose range over i assumption is an extension of the which a simple linear fit to data is generally adequate, is sery markedly af-for effects of gamma and neutron fected by the diameter of the " interaction site," the locus within which the

.mtl aberration.. In this particular Pairs of sublesions are presumed to produce the lesion. For an interaction or sublesions) are required to pro-diameter of I m, Goodhead's calcu!ations indicate a value of f of about hromosomal aberrations in human 30 rad for Co-60 gamma rays, and about 100 rad for 250 kVp x-rays. For a ce of effects has been observed for l

more likely interaction dia' meter of 0.4 mg for cell transformation effects.

ith Les's tueory. To extend the the corresponding values are about 400 rad for both types of radiation.

luired for other effects of radiation These latter values are so high tnat one would conclude that over the range rquires that experimental evidence uared be obsersed for such effects.

of doses up to 200 rad, the Kellerer Rossi theory actually supports ap-P cation of the linear no-threshold dose response relationship for on-li rnced, is derived from studies of Ia. the spiderwort plant i2 effects cogenic transformation.

But even more significant than these theoretical considerations are the duced life shortening in animals.i4 results of recent studies of oncogenic transformation in marr malian cells expected to involve non stochastic by low do.es of x-rays. Borek and Hall first showedl? in hamster em-induction moreover, the subcom.

vailable human data, that no non-l bryo cells that split doses of 210 kVp x rays were more effectise in produc-ing tran> formations, and this result has been confirmed for coses below n has been observed in man.) This 18 and in A31-11 mouse BALB/3T3 100 rad in mouse 10TW cells ng justification of the assumption ibroblasts." Little and his colleagues 2o have pointed out the com oduce lesions in the DNA of mam, P exity of the role of DN A repair in these results, and have concluded from l

ple, ta encogenic transformation, studies in which a phorbol promoter or a protease inhibitor has also been atly associated wi:h gross chromo.

added to mouse 10TW fibroblast cultures that the DNA lesions and repair Process associated with cell killing and cell transformation are different.

alism, nevertheless, on the further This observation is especially important because the Kellerer Rossi theory to produce a lesion over a range of 21 also has ncy of effects. E = K (fD + DD.

has been mainly applied to studies of cell killing. Little Postulated that rapid DNA repair mechanisms are error prone, and result

' eta is a variable dependent on the in transformations. A slower, at least partially err -correcting repair pro-t es produced by single events The i

cess is also present, but if the cell undergoes DNA replication before this to a linear-quadratic dependence latter repair can occur, then the DNA afteration becomes " fixed" or is obvious from the fundamental I

P

..A

=

236 1HE EFFECTS OF IONIZING R ADIAT*0N 80mdfl

" stabilized" in a heritable form after one cell division. This change to dett I

~

becomes expressed as a transformation after e number of subsequent cell threshi divisions, the number influ:nced by w hether the cetis are exposed to other from l<

non transforming chemicals or agents during this nage.These results em-dicate.

^

phasize the importance of exposure to other agents affecting cell prolifera-an "ex tion in fixation and expression of transformational damage, a concept in The accord with much evidence concerning non specific factors in promotion type is I of human can er, nized t.

Work on this aspect of oncogenic transfotmation of cells is pregressing data 01 5

rapidly and can be expected to yield important new insights into the rela-the nu i I

tionship b-Neen transformations produced by low doses of all types of groups radiation u:o the process of carcinogenesis in animals and man. But the 50 99i f

~ ' ~ -

important point here is that the data in hand show clearly that biological than 2( -

factors such as DNA ?cpair mechanisms and exposure to other non*

vimrs i transforming agents markedly modify the probability of an oncogenic age 20 transformation, and the simp'e view that repair of initial damage pro-deaths duced by low LET radiation at low dose rates will inevitably r0 duce the and 63 -

subsequent probability of cancer induction when compared to the same fractio dose given at high dose rates, is cleatly untenable.

Report :

For both these biological reasons as well as the theoretical points rr.ade, c cess for example, by Goodhead I believe the Kellerer-Rossi theory is quite long rt.

unacceptable a having any relevance to dose-response relationships for will be -

human cancer. indeed, the cell transformation data suggest that the of 10/ !

m linear no-threshold dose-response curve as a basis for extrapolating car-The cinogenic effects frem high to low doses of low LET radiation could even rangei somewtn underestimate the low dose risk, as Miller and Hall 18 and in dose '

Borek22 have emphasized.

T here membi.

pretati

' place i DOSE. RESPONSE D AT A FROM EPIDEMIOLOGIC i

The STUDIES OF MUM AN POPULATIONS Nagas:

The above practical ~ :: theoretical problems thus refute the idea that Hirosh experimental evic'ence provides any basis for deciding on the particular difficu forms of the doe respor.se relationship in human radiatior car-for dei cinogenesis. This situation means that we must rely on epidemiologic popula evidence to estimate risks at low doses of low LET radiatio,, as the sub-smalli committee had concluded early in its deliberatioiss. Unfortunately, as the posed third section of Chapter V points out, good dose-response data in human bining populations of large enough size to provide statistically reliable risk tissue estimates in the range of doses less than 50 rad are very limited. Such data prose are needed if extrapolation to lower doses is to have any precision, or even Regar i

y...,

c.g p

W 3

- h.

c h k'-

., g g'y &g u n %' f.,b

, s > N

+

s,

s g.-%

J

~'

f*?((';.X**.

y l

.45 Q$)

b.V.}Y4.'.N&

l -;'- Q gg

237 80mutic fffects: Cuncer

' Ts OF IONIZING RADIATION h linear no-to determine whether the simplest extrapolation curve, t ei ks threshold model adopted by the subcommittee to est

ell disision. This change number of subsequent cell

! e cel;s are exposed to other dicate, use of the linear extrapolation can hardly l e considered to provide i

his stage. These results em-an " extreme" estimate of low dose risk.The only population st ents affecting cell prolifera-l type is that of the Japanese A-bomb survivorsJ lt is not generally r. cog-ional darnage, a concept in i

j nized that the strength of the Japanese data in epidemiologic terms lies n I

xcific factors in promotion data obtained for low doses, less than 100 rad kerma. The major part of dose tation of cells is progressing the number of survivors with significant exposures are in the two d

, it new insights into the rela-groups,10-49 rad kerma, or a mean tissue dose of about 11 rad, a i l by low doses of all types of 50-99 rad kerma, or a mean tissue dose of about 35 rad. For doses greater f

s animals and man, But the than 200 rad kerma, about 120 rad mean tissue dose, the numbers o sur-i show ciectly that biological vivors included in the Life Span Study October I,1950, and who were over age 20 at the time of the bombing (the grog in whic no exposure to other non-probability of an oncogenic i

epair of iaitial damage p*o-and 684 in Nagasaki, numbers that are small enough that if the dose s fractionated further into three dose categories, as has been done in nEnr

.es will inevitably reduce the Report 8, the results are likely to lead to statistically unstable estima when compared to the same 4thi excess cancer risk, especially in Nagasaki. Thus, it is f air to say that i nable.

' s ti,e theoretical points made, long run, a principal value of data obtained from this study population i

(ellerer Rossi theory is quite will be to permit estimation of cancer risk from acute exposures n a range

>se-response relationships for of 10-35 rad mean tissue dose.

tation data suggest that the The fact that the A-bomb survisors are the only large group with a wide a basis for extrapolating car' rage of who!e body radiation exposure makes them singularly important low LET vadiation could esen in dose-response evaluation of the carcinogenic effect of radiation in man.

There was general agreement for this position among the subcommittee 38 and k, es Miller and Hali members, and it was the reason that extensive debate concerning inter-pret. tion of the fo!!ow up data through 1974 from PERr Report 8 P ace up to the final meeting of the subcommittee.The are l

Nrgasaki data to evaluate effects of low ttT radiation. Because the EPIDEMIOLOGIC Hiroshima bomb led to a significant neutron exposure whose effect was TIONS tems thus refute the idea that difficult to ass ess independently, the Nagasaki data thus became the ba for deciding on the particular for defining low LET radiation effects. Unfortunately, the Nagasaki study ip in human radiation car-population is much smaller than the Hir:shima group, and is espec i f the ex-g.nallin the Zero dose category, the accepted control p]pulat on or ee must rely on epidemiologic posed populations. A better control population c

' low LET Fadiation, as the suh-

>erations. Unfortunately, as the i

tissue dose about 1.E rad), an approach which has been widely used to m

>d dose-ecsponse data in human prove the analysis by investigators reporting ovide statistically reliable risk O rad are very limited. Such data is to have any precision, or even

t 23b THE rFFECTs OF IONIZING R ADI ATION inesitablv show quite large statistical error ranges, especially at the higher doses.

Another important issue has been the relative importance of cancer mortality data frc.m the death certificate study compared to the results ob-tained from the Tumor Registries in the two cities. The resul s of the dose-t response analysis for both cities and for these two data sources are shown j

for all cancers except leukemia and bone cancer in Figures V-6 and V-7 of Chapter V. The mortality <iata in Figure V-6 are for the period 1955-1974 while the incidence data are for 1959-1970. The total nu of cancer 'ases in the two instances is about the same. thus the stat i

power ol.nalysis of results from the incidence and mortahty studies is also about the same.

The mortality data shown in Figure V-6 suggest from the fitted regres-sion l'nes that the radiation effect in Nagasaki was much less than in Hiroshima, thus implying that the neutron component in Hiroshima may l l

-e hase b(er. of major importance. But it is clear from analysis of the in-dividual data points that a major difference accounting for the low lope of the Nagasaki dose responce is the single point at about 120 rad (200-299 rad kerma). This point shows a quite high cancer rate in Hiroshima and low in Nagasaki. The data points for both cities are ;ow for the point at about 160 rad. At the request of the subcommittee Dr.

Charles Land ran the correlation for the data below100 rad (5 data points) and found that the results gase a reasonable linear fit with a dif-ferene: in slope between the two cities consistent with a constant RBE of about 5 for the nutron component.

While I do not suggest that this type of mathematical manipulation pro-vides a great deal of help in establishing firm conclusions I do beliese that it is important to understand that the apparent difference in response fcr the two cities indicated by the regression slope in ligure V-b arises because of differences observed at high doses, where the Nagasaki data especially are less reliable on state,tical grounds, rather than because of differences at low doses, where the data are somew hat j

mcre robust.

Moreoser, to attribute the difference entirely to a high neutron l

ffec-tiveness in cancer induction implies that an especially high RBE applies to high doses only, a conclusion entirely at variance with current siews of the effect of dose on the Rar of neutrons.

The results of the data from the Tumor Registries. Figure V ~ show a marked difference for the Nagasaki dose-respo,se compared to Figure V-6. and a concordance between the two cities that suggests a constant RBE for neutrons of about 5. It should be noted that the tumor incidence dose-responst data depend on the same denominator base of the Life Span Study population as do the rr.ortahty data One problem with the h

,r EFFECTS OF IONIZING RADIATION Tum r Registry data, however, is the fact that they have not yet been or ranges, especially at the higher "esaluated " that is, it has not been determined whether out migration 5

fr m the cities, which would lose cases at.d therdore provide a lower l 1e relatise importance of cancer estimate of risk, is randomly distributed by dose categories, and thus study compared to the results ob-w; uld not affect the slope of the dose-response curve. A random distribu-two cities. The results of the dose-ti n by dose category of loss to fo!!ow up from out migration occurred in these two data sources are shown the women studied for breast cancer incidence m the two cities.23 The

- cancer in Figures V 6 and V 7 of utdgradon was only 16%, despite the fact that the study

I rigure V-6 are for the period FPulan.on incMed m, M a large nunser of young umnen wb ndgM for 1959-1970. The total number e emete t mee becam of maMagd ne Tumor Restry data

.out the same, thus the statistical au ea antage, wever, that a high percentage of the ca;es have i lence and mortality studies is also either histologic or autopsy confirmat,on of the cancer diagnosis, and the i

Nagasaki Registry particularly is believed to be quite complete for the area

-6 suggest from the fitted regres-

• • und the city (Moriyama, I., personal communication to the subcom-dagasahi was much less than in on component in Hiroshima may

"" ' e death certificate data have an important defi-is clear from analysis of the in-at maj r ra. geme cancen are sigm andy undeMeponei ciency in mee accounting for the low slope Breast cancer in women is markedly under reported because breast cancer single point at about 120 rad has a relatively long survival tirr.e and thus death is often recorded as from vs a quite high cancer rate in an ther cause, and thyroid cancer is usually not fatal. Thus m both cases a points for both cities are low for y mp dant ra& gen cancen are not weH reponed m deam e

quest of the subcommittee Dr.

certificates-ut Psy studies have also confirmed that in the study popula-he data below 100 rad (5 data agnosed on deas ceMcates in mer haN de n lung cancer m

i reasonable linear fit with a dif, b

• onsistent with a constant ast of the major cancers induced by radiation are not accurately represented in the mortality data from death certificates, and for this reason, the advan-mathematical manipulation pro-tage of complete ascertainment of death records for the study group is irm conclusions, I do believe that largely I st. While it is unlikely that such under-reporting of cases could parent difference in response for by itself alter the dose response curve, it could have the effect of makmg

. ion slope in Figure V o arises the range of uncertainty at any dose greater.

doses, where the Nagasaki data in the final analysis, there are inadequacies for both the death cer-grounds, rather than because of tificate and Tumor Registry data, but when they are all taken together a 4

ta are somewhat more robust.

r-as<anable concordance appears. For all cases except the Nagasaki mor-stirely to a high neutron effec-ata Ge linear no-threshold dose response curve appears to be an ta y an especially high asE applies to adequate descr,iption of the results, although as the voluminous dicussion ariance with current views of the and tortured mathematics of the third section of Chapter V attest, it is Possible to fit a number of other curves to the data about as well as the r Registries, Figure V-7 show a I near fit. The Nagasaki mortality data are consistent with the rest of the oresponse compared to Figure results excePt for the two data points at high doses in Figure V-6. But the

) cities that suggests a constant chief point to he made at this stage is that mathematical constructs based noted that the tumor incidence n the Japanese data do not really contribute to decisions about the ap-denominator base of the Life Propriateness of any particular dose-response relationship. The data for ity data. One problem with the l

l l

l

p-.. r w,.,

y Q

1

=

24U THE Et t EcTs OF IONIZING kADIAT10N 3

~

all cancers are as yet too imprecise. and thus adoption of a particular i

dose-iesponse relationship remains an arbitrary choice.

The dose-response data for leukemia mortality from 1950-1974 in ig Nagasaki are based on only 22 deaths for those exposed above 10 rad kerma. and as anyone familiar with analysis of dose response is aware, it r.

is impossible to do much more than say that a significant effect of ex-posure exists with such a lirri ed number of cases Certainly these data are f

t totally inadequate to define the dose-response cune. Cases from the

[

leukemia registry results presented in RERF Report 8 are more numerous.

~

and suggest a cunilinear dose response relationship for both cities consis-tent with a constant RBE for neutrons of about 10.

J

' ~

In the present version of the report. the leukemia Registn data have l

r.

been used as a " guide" to define *he linear and quadratic coefficients (a and b abose) to be used in the linear-quadratic model apphed to a//

i cancers. In other words, mathematical adjustments to the coefficients.

necessary because the results of fitting the theoretical cunes to the Japanese mortality data led to unreasonable figures (all the coefficients derised from mortality " appeared out of line w ith the incidence estimates") were based on the leukemia " guide " On bological grounds the idea that dose-response relationships for solid tumor, must be similar l

.(>

to leukemia is far from reasonaole. First. of course, is the markedly dif-ferent time c >urse for induction of radiation-induced leukemias compared l

to the much more quantitatisely important so!id tumors. This fact sug-gests a major difference in the factors insobed in carcinogenesis, w hich by inference could affect the dose-response relationship Second is the obser-vation. thoroughly discussed within the full subcommmee t hat leukemias are the only human cancers in which distinct chromosomal abnormalities are consistently associatec' with the disease. In the case of chrome granu-locytic leukemia. qantitatisely a sery important type of leukemia induced by radia ion, the great majority of cases (-85%) have the Philadelphia chromosome abnormality present in the leukemic cells and there is agree-ment among cvtologists and hematologists that the abnormality is causally related to the disease.2n l

In contrast. consistent sisible chromosomal abnormahties in the early stages of solid tumors hase not been found. The imphcation is that the somatic muta: ions in these tumors either invoke point mutations or chromosomal changes small enough not to appear as rcadily si:ible j

3 translocations, deletions, or other abnormalities, or they are not associated with any particular chroraosome site. The importance in radiobiological terms of the association of specific chromosomal abnor-f malities with leukemia is that such abnormahties are well-known to be two-break esents and thus a dose 4quared dependence for at le.st part of i

.s n

'd g.

241 ECTs OF IONIZING RADIATION Somatic Effects: Cancer hus adoption of a particular the induced leukemias has a biological rationale. This is the main reason I trary choice.

accepted the linear-quadratic model for leukemia in the April 1979 draft.

nortality from 1950-1974 in Cuologie differences between leukemia and solid tumors such as those those exposed abovc 10 rad mentioned abose, support the view that the dose response curves may not s of dose response is aware, it be the same for all cancer types. This is an idea that Harald Rossi and I hat a significant effect of ex-both felt was an important contribution of the BEIR III Report; now of cases. Certainly these data are course in the present version it has been eliminated. In sum, the approach

>onse curve. Cases from the taken to " adjust" constants to provide risk estimates for solid tumors Report 8 are more numerous, based on the leukemia " guide" is arbit.3ry and in my view not scientifi-tionship for both cities consis-cally justified. The leukemia tailis still wagging the radiogenic cancer dog.

>out 10.

An important question is the extent to which the Japanese data are con-Leuhemi: Registry data have sistent with the data from all the other studies described in Chapter V.

and quadratic coefficients (a Appendix A, when expressed as an excess risk of cancer incidence per rad iadratic model applied to all per million persor, years, and roughly age-adjusted. In general, the con-cordance is excellent for the major cancers where seseral data sets exist

(

justments to the coefficients, I

the theoretiul curses to the I

such as breast, thyroid and lung cancer. Other sites show Sarious degrees

.le figues (all the coefficients of agreement. But the most important comparison is for total canccr in-f 4

of line with the incidence i

cidence coefficients derived for each sex from the Nagasaki Tumor guide." On biological grounds Registry data. From data presented in the April 1979 draft, these are r solid tumors must be similar found to be about 2/3 as great as the sum-of-sites coefficients summarized

>f course, is the markedly dif-in Table V-14. This degree of cencordance of results from human studies induced leukemias compared of a great range of exposure conditions, ethnic makeup and basis for t solid tumors. This f:ct sug-radiation exposure is truly remarkable. The relatisely small difference ed in carcinogenesis, whi + by could be accounted for in part by underascertainment of cases in the ationship. Second is the obser-Nagasaki data, and by a somewhat lesser susceptibility to cancer induc-subcommittee, that leukemias tion by radiation in Japanese as compared with occidental populations, a et chromosomal abnormalities reasonable conclusion because of the somewhat lewer total cancer rates in

. In the case of chronic granu-Japan compared with the U.S. The fact that the total excess cancer in-rtant type of leukemia induced cidence rate per unit dose in the Nagasaki A bomb survivors is quan-

-85%) have the Philade!phia titatively similar to the total excess incidence derised on the linear

~

kemic cells, and there i, agree-hypothesis from the aggregation of all the other available human studies hat the abnormahty is causally lends strong support to application of the risk coefficients from the data in Table V-14 for deriving cancer risk estimates from whole-body exposure mal abnormah.. ties in the early to low 1.ET radiation.

d. The implication is that the With regard to concordance of dose-response relationships between the r involve point mutations or Japanese data and other source: most of the other studies do not have a to appear as readily visible sufficient range of doses or sufficient numbers to permit comparison with ormahties, or they are not the Japanese data. For female breast cancer incidence vs dose, Figure A-1 me site. The importance in (Appendix A) shows good agreement of the three western studies cited specific chromosomal abnor-compared with the data from the A bomb survivors. (In this case the data malities are well-known to be for both Hiroshima and Nagasaki give a good fit to the linear no threshold dependence for at least part of relationship. with no evidence of an RBE for neutrons greater than 1.) For i

x u

242 THE EFFECTS OF IONIZING R ADIATION y

.'?

thyroid cancer Hempelmann's data in children ' do not agree closely 2

"PP' with those of Colman28 but taken together they are consistent with the

" h 'C '

E.

1.(.

linear model over a reasonably wide dose range. The lowest dose point at k"""

about 7 rad provided by Modan's results from examination of thyroid "O

cancer in 10,900 Israeli children given scalp irradiation for tinea capitis,

body

~

2 y

~

fits reasonably well with the linear extrapolation for the other two studies (see Appendix A). The lung cancer data for underground miners suggest that the dose-response curve from exposure to alpha radiation could be curvilinear downward. that is, low doses may be somewhat more effective in cancer inducSn per unit dose than high doses, a concept in accord with the idea that high LET radiation may show cell-killing effects at relatively low doses that would progressivelv reduce the cancer risk / rem as dose increased.

~

"""I Some members of the subcommittee believe that the " sum-of sites" method, used by the BEIR I Committee to estimate total cancer risks, oserestimates risks somewhat. because out of the numerous epidemiologic studies of radiation-induced cancer at individual sites presented in Ap-

""I pendix A. some would be expected by chance to yield higher than the true estimates, since in any study observed and expected cases hase an in-herent statistical sariability. For this reason selection of onl3 positise results would oias the risk estimates upward. To some extent mis problem low has been dealt with for seseral minor cancers by pooling risk estimates for them and striking a balance between high and low estimates. these sites

'l being particularly susceptible to the above problem because risk estimates for ther often were derived from a single study. But for two of the most importo a contributors to the total cancer incidence risk. thyroid and female breast cancers. there are several studies available for each that show excellent agreement, and thus the uncertainty of the risk coefficients tha: t is small and no selection of high values has occurred. For lung cancer there are also seseral studies, but only two involving low LET radiation. the ady Nagasaki Tumor Registry and British ankylosing spondyhtis studies frore These two studies show reasonable concordance. and are also concordantg with the studies of the underground miners on the basis, derised in-

indu, dependently from dosimetric and radiobiologic principles. that exposure indu to one Working Lesel Month is equisalent to a de,e of 6 rem to the basal cell layers of the proximal bronchial epithelium.

duu' The only sites contributing significantly to the totalin Table V-14 where the above argument could have some merit are those for the digestive esophagus. stomach and intestines. primarily large bowel Even m tract:

these cases there is reasonable concordance among the studies available, and the likelihood that selection of data has biased the risk estimates up-

'i ward is not great But this reason for rejecting use of the " summed sites" httk i

~ ~..

g.../'

Somar c Effects: Oncer 24.7 HE EFFECTS OF IONIZING RADIATION in children 27 do not agree closely approach to defining cancer incidence risks from whole body exposure gether they are consistent with the bscures two important points. First is that ionizing radiation is the only lose range. The lowest dose point at known human cancer producing agent that has been found to increase the

sults from examination of thyroid risk of cancer in nearly all the parenchymatous or epithelial tissues of the i scalp irridiation for tinea capitis,

body (see Appendix A). Indeed it is a reasonable conclusion that at high 2

r:polation for the other two studies enough doses, it should be possible to demonstrate a carcinogenic effect of its for underground miners suggest radiation on any human tissue. Therefore one may conclude that in posure to alpha radiation could be human studies where a small excess of cancer is found at a particular dose

. ses may be somewhat more effective f radiation but is borderline in statistical significance, it is prudent to in high doses, a concept in accord l

c nsider the effect may be real rather than to dismiss the study as m nry show cell-killing effects at negative.

siv:ly reduce the cancer risk / rem as See nd, as the follow up time of the human study populations in which many organs were irradiated is extended, evidence of excess cancers at tee believe that the " sum-of sites" many of the minor sites has emerged slowly over time because the excess is

.1ee to estimate total cancer risks, set against the usual variability of cancer arising from other causes. Thus

out of the numerous epidemiologic

" statistically significant" excess cancer in the irradiated population may t individual sites presented in Ap-n t occur for those sites where a lesser radiation effect is present until chInce to yield higher than the true many total cases at that site have accumulated. This phenomenon has d cnd expected cases have an in-been obvious from the successive follow up reports of the Japanese s re son selection of only positive A-bomb survivors, where the bulk of the cases are observed at relatively Iw doses. For this reason we must consider any quantitative risk n:rd. To some extent this problem tncers by pooling risk estimates for estimates, positive or negative, as tentative and could underestimate the high and low estimates, these sites risk until a lifetime follow-up is completed. For the above two reasons the

.ove problem because risk estimates idea that Table V-14 risk coefficients are biased upward by an effect of 1gle study. But for two of the most selection of positive results totally ignores the combined strength of the encer incidence risk, thyroid and evidence presented in Appendix A.

r:1 studies available for each that Another point raised by use of Table V 14 for estimating cancer risks is e uncertainty of the risk coefficients that it gives cancer incidence rates rather than cancer mortality. The deci-ues his occurred. For lung cancer si n to define cancer risks in terms of incidence rather than mortality was to involving low LET radiation, the adopted early by the subcommittee and constituted a significant change th enkylosing spondylitis studies.

fr m the BEIR I seport. This decision was based in part on the awareness icordance, and are also concordant that cancers of the thyroid and female breast are now major radiation-miners on the basis, derived in-induced cancers, and for these two sites mortality data give an inadequate obiologic principles, that exposure indication of risk. This change from BEIR I was also based on the con-lent to a dose of 6 rem to the basal sideration by the subcommittee that any radiation-induced cancer pro-pithelium, duces a major psychological, social and economic cost to the individual af-itly to the totalin Table V 14 where fected, whether or not the cancer is ultimately the cause of death. Thus mirit tre those for the digestive the Mea that cancer deaths alone are the proper measure of radiation im-ies, primarily large bowel. Even in Pact was rejected. Since the BEIR I report, new information was available lance among the studies available, which permitted better estimation of excess cancer incidence from radia-a his biased the risk estimates up-tion exposure to the thyroid and female breast; for other cancers there is ejecting use of the " summed sites" little incidence data except from the Japanese Tumor Registries, but

l 244 THE EFFECTS OF IONIZING RADIATION because most of the other important radiogenic cancers ir.cluding leukemia cre eventually fatal, mortality gives a reasonable estimate of incidence.

3 For this reason, the other coefficients in Table V 14 have been derived from mortality data.

I Because cancer incidence risk estimates are those intended by the sub-

/

committee, the amount of emphasis in the current version of the third sec-tion of Chapter V on discussion of cancer mortality data is unwarranted, and indeed the procedure of " indirect conversion of mortality estimates to incidence estimates" is clearly inappropriate for cancer of the thyroid and female breast. In my view the best basis for cancer incidence risk esti-mates from radiation exposure is Table V-14, because it draws on all the evidence available from Appendix A. much ofit obtained in American or 3

British study populations and on this basis more immediately applicable to risk estimates intended to be applied to the U.S. population. As pointed out above, it is supported well by the Nagasaki total cancer in-cidence data. These risk estimates applied to the 1969-1971 U.S. life table population are presented in Table V-30 of Chapter V, but it should be 2

noted that this table does not include the risk for leukemia and bone cancer incidence. To determine total cancer risk the data from Table V-30 must have added the data from Table V It,, where leukemia and bone cancer incidence are derived using the linear-quadratic model agreed by the subcommittee as appropriate for leukemia only (bone cancer is such a minor cancer that it contributes trivially to total cancer risk regardless of the model used). Failure to provide a single estimate of risk of total cancer incidence is another deficiency of the present version of the third section of Chapter V. Table V-30 gives a range of risk calculations for each s x ac-cording to the various exposure regimens. This range reflects our uncer-tainty about the appropriate model by which current estimates of risk are projected forward to a lifetime cumulative risk. The two projection models E

used are the so-called absolute and relative risk models (see Chapter II, the section entitled "Epidemiologic Studies as the Basis of Risk Estimates for Effects of Ionizing Radiation"). It is evident that these two projection methods give total risk estimates that differ by a factor of about 3 for the projections of total population exposures. There was general agreement among the subcommittee members that at least this degree of uncertainty applied to the estimates of lifetime risk in these instances. For the occupa-tionally exposed groups the two projections agree reasonably well.

in Table 1, I have combined Table V-30 with Table V-16 to give the best estimate of total excess cancerincidence derived for the exposure con-ditions used in the third section of Chapter V.

The exposure conditions adopted for illustration are unrealistic, in that it is extremely unlikely that 1,000,000 persons in the general population or

~

24S H' FFECTs oF ION!Zl!?G RADIATION Somatsc Effects: Cancer genic cancers including leukemia A E1 Estunates oRotal Lifetime Excess Cancer Incidence from tasonable estimate of incidence.

Exposure to Low.ET Radiation-Projections Based on 1969-1971 U.S.

a Table V 14 have been derived Life Table Population of One Million Persons at Start of Exposure.

According to Absolute Risk and Relative Risk Projection Models. Data

s are those.mtended by the sub-Th im T&s V-30 and V 16 of Chapter V e current version of the third sec-r mortality data is unwarranted.

Absolute. Risk Relative Risk nsersion of.iortality estimates to iate for caneer of the thyroid and Male Female Male Female is for cancer incidence risk esti-V-14. because it draws on all the I Sia:I' <1ro8=rr to 10 rad sch of it obtained in American or 1.M.000 penons of an agn 6 sis more immediately applicable Enpected lifetime cancers 285.000 260.000 285.000 260.000

. d to the U.S. population. As without radiation te by the Nagasaki total cancer in-j gacess cancersinduced by 2.600 5.500 8.800 16.200 radiation d to the 1969-1971 U.S. life table

) of Chapter V. but it should be

2. continuous exposure to i red'yr the risk for leukemia and bone to l.000.000 persons at outset

.cer risk the data from Table V-30

a. Layerime exposurefrom 6.rri.

V-16, where leukemia and bone Espected lifetime cancen 283.000 285.000 283.000 285.000

.ithout radiation

'incar quadratic model agreed by Escess cancen induced by 16,200 37.600 31.100 185.200 kemia only (bone cancer is such a radiation to total cancer risk regardless of

b. Exposure are 20-65 gle estimate of risk of totalr ancer Empected lifetime cancers 292.000 300.000 292.000 300.000

-sent version of the third section of

.ithout radiation risk calculations for each sex ac-Excess cancers induced by 11.100 25.400 15.000 49.500 ns. This range reflects our uncer-radiation

<hich current es. mates of risk are e

et e ca een 2%m 2%m 2%m 2%m i

is els e hap er without radiation lies cs the Basis of Risk Estimates Excess cancers induced by 6.800 14.700 7.600 17.500 radiation

, evident that these two projection iffer by a factor of about 3 for the

d. riposure ages 50-65

'es. There was general agreement Expected lifetime cancers 295.000 269.000 295.000 269.000

.ithout radiation ct letst this degree cf uncertainty Excess cancers indu:ed by 3.100 6.300 3.100 6.300

.n these instances. Fc he occupa.

'*dii "

ions agree reasonably well.

V 30 with Table V 16 to give the

!ence derived for the exposure con-ipter V.

illustration are unrealistic,in that ersons in the general population or

e 77%

P 246

,;_. g niE EFFECTS OF toNIZING R ADI ATION among radiation workers would ever be exposed either to a single dose of Ag 4 ;. [* ^

10 rad or to continuous doses of I rad per year. The numbers of radiation-induced cancers appear to be large in most instances. but it is important d

p MQ to note that except possibly for the case of lifetime exposure to one 1

.* ' % '1, rad year, esen with these unrealistically high exposures it would be very

(!

difficult to detect by epidemiologic methods that the excess cancers had

-Q7

[,(([.

occurred except for those particular sites which are especially sensitive to

(!

w.-. O..N...m cancer induction by radiation.

(

32;1 E g e

On the linear hypothesis, the data for the single exposure to 10 rad can 9 kpg,, A 2,

(

a 7

be conserted to consentianal " risk per rad" estimates by dividing by ten.

f i

,y " ;;. J p

This yields a range of 260 to 880 cases per rad per million exposed for y%" 1 males, and 550 to 1620 cases per rad per million exposed for female <

P.

. } t.d. ?

we adopt an intermediate valuc as more likely to obtain (that is. If

/l H

,the a

[.

relatise risk model will only partially be found to be correct), the risk per

- ~4 rad for cancer induction is about 500 cases per million for males and 1000 C

case < per million for females. These values are higher than the risk a

4~,

estir..ates from BEIR I. in part because incidence is conMdered instead of a

mortality. and in part because the new data indicate somewhat higher C

lifetime risk than was evident in 1972.

5 If one applies total cancer risk estimates obtained from the life table D

projections for a singie exposure in Table 1 to the Japanese A bomb Life a

Span Study population by use of the lii. ear hypothesis and the same C

b method as was done to produce the estimates in Table 1. some important limitations of the Japanese A bomb follow-up study become clearer d

f*

Life Span Study population has a greater proportion of younger people The than the 1969-19'l U.S. life tab!e population. a circumstance that means 5

the total radiation-induced cancers anticipated per number exposed will be somewhat greater than predicted from the model apphed to single dose d

exposure in Table 1. Nesertheless some approximate conclusions arejusti-F fied. First is that the number of excess cancers obsened to the present C

follow-up in 1974 constitutes only about one third of those that esentually n

will be expecteo if the tirne for expression of excess cancer risk is the

'I lifetime of those exposed oser the age of ten

^'

in other words, the follow -up period for the Life Span Study group is still too short to define total cancer ir risks adequately. Second. esen on the upper limit assemption that the lifetime re/arive risk model applies. no statistically sigmficant excess of all dl

.p.

N cancers will eser be bserved in the two lowest dose categories of the study",

'f y,

population in Nagasaki. that is at mean tissue doses of 2 rad and 10.8 rad.

at!

For Hiroshima the same statement can be made for the lowest hk dose ra category (mean tissue dose 17 rad) regardless of the RBE assumed for rd neutrons within any reasonable range For the next dose categon.10-44 i

rad kerma or a tr.ean tissue dose of 10 rad. if a statistically significant ex-ra r

"C

_k' y.

w HE EFFECTS OF IONIZING RADIATION 8omatic Effects: Cancer 247 be exposed either to a single dose of cess of total cancers is ob.terved in Hiroshima compared to the zero dose per year. The numbers of radiation.

group, such an observation will be consistent with an RBE for neutrons a most instances, but it is important greater than one, but because the mean tissue dose from neutrons is only case of life:ime exposure to one one rad in this group, the reliability of any numerical estimate of RBE tily high exposures it would be very derived from this escess will always be weak indeed. For the next dose

ethods that the excess cancers had category,50-99 rad i erma or a mean tissue dose of about 34 rad in each ites which are especially sensitive to city, a significant li'etime excess of total cancers will be easy to demonstrate in Hiroshima, but for Nagasaki the smaller sample size will ior the single exposure to 10 rad can Probably mean that the statistical significance of the excess will be a rad" estimates by dividing by ten.

marginal if the lifetime relative-risk model is found eventually nn.: to hold.

ses per rad per million exposed for This application of the current total cancer risk estims9s to the per million exposed for fema'es. If A bomb survivor populations again emphasizes the caution that must be more likely to obtain (that is, the applied in interpreting the data for excess cancer risk in this study group, be found to be correct), the risk per esPecially at low doses. Another implication of the abose analysis is that cases per million for males and 1000 an excess risk of cancers at particular sites which are sensitive to radiation

. values are nigher than the risk and have a high natural rate will always be easier to demonsnate, espe-e incidence is consicered instead of cially in Nagasaki, than will an excess for all cancers, beause the inclu-ew d:ta iadicate somewhat higher sion of a large number of cancer types with low or zero radiation sensitivity increases the random " noise" in the data. The above phenomenon is mates obtained from the life table already obsious in the analysis of breast cancer incidence up to the pres-ible I to the Japanese A-bomb Life ent. In sum, the fact that the Japanese data at any follow-up state may not e linear hypothesis and the same be strong enough in statistical terms to show a significant effect of low imates in Table 1, some important doses on total cancer risk does not prove that effects art not present; the Ilow up study become clearer. The excess cancer risk may be better evaluated by looking at particular cancer ater proportion of younger people sites.

ulation, a circumstance that means With regard to the appropriate RBE for high LET radial!on and its ticipated per number exposed will dependence on dose, the data for alpha radiation compared with x rays or m the model applied to single dose gamma rays give reasonable RBE Values of about 10 to 20 for lung and liver approximate conclusions are justi.

cancer (Appendix A). Comparisons of the Hiroshima-Nagasaki results do es cancers observed to the present not allow any definitive statement with regard to the RBE of neutrons for t one third of those that eventually the following reasons: First. the rates for total cancer incidence from the ssior. of excess cancer risk is the zero dose (control) populations are substantially higher in Hiroshima than

' tea. In other words, the follow up in Nagasaki, and thus the assumption that the neutron componert is the still too short to def'me total cancer sole factor accounting for differences in cancer dose-response is unten-upper limit assumption that the able. Second, at low doses, where the results are most important, excess etatistically significant excess of all cancer rates are not yet statistically strong enough to provide an appropri-lowest dose categories of the study ate estimate of the contribution of neutrons and in some instances are tissue doses of 2 tad and 10.8 rad.

likely never to be strong enough (see above). Third, neutron and gamma an be made for the lowest dose ray exposures were highly correlated for Hiroshima, and in the low dose gardless of the RBE assumed for range tissue doses for neutrons were only about 1/10 those for gamma For the next dose category,10-49 radiation, thus random differences in results greatly magnify the imputed ad, if a statistically significant ex.

neutron effects at low doses. Fourth, the dosimetry for gamma rays and

w m c m w &wn m M" m. +- '

Ti,O f

24A THE E FFE CTS OF IONIZING R ADI ATION neutrons is estimated to be good only to 130%. thus any consistent dosimetry erron could also greatly affect the analysis of neutron effects i i

[.....

y the comparison.

o 4g@ f... '.' g p;g -

It should be pointed wt that the auumption that the RBF for high LET compared to low LEr radiation increases as the dose decreases does not

[y i necessarily imply that the lose-response cune for low 1ET radiation must

+

[

be cunilinear upw ard at low doses it is equally possible that the dose-respimse cune for high LE T radiation is curvilinear dow nward The poin.

is that if we assume a fixed RBE independent of dose, we may underesu-mate somewhat the risk ef low doses of high LET radiation and oseresti-mate somewhat the risk of Icw doses of low LET radiation But the asail-able human epidemiologic da'a do not indicate to me that this degree of

44..

oser-or under estimation is.en great. that is, more than a factor of 2 4 y -.7 When we consider that cancer risk estimates may esentually base to take account of a significant subfraction of the population whose radiogenit f;-

cancer risk can be expected to be higher than the population at large. any consenatism arising from auumptions that may oserestimate the risk by a sma!! amount is justified at this time.

' 'h~

Pertment to this question of the relatne effectneness of high LFT radia-tion at low doses are the results of chromosome aberration studies m L

populations Ining in or otnerwise exposed to high background radiation.

In those situations where exposure has been especially to radon-222 the alpha radiation can account for these essentially two-break effects on the iromosomes

• H In the Brazilian population Ining in a village on monaiste sands. chromosome abnormalities were found elesated corn pared to a control group not so exposed '2 In this case. it was postu-lated that alpha radiation ' rom the Pb-212 daughter of Rn-220 reached the lungs or blooo and this exposure rather than the high background of gamma rays accounted for this effect On the other hand, the dose-related chromosorral aberrations obsened by Esans et al." in nuclear ship-yard w.,rk ers exposed to relatnely low cumulatise doses were from er

~

posures to "almost exclusisely gamma radiatiori ' It is of interest that 5 rad of acute x or gamin

'diation has produced in human lymphocytes significant chromosom urations. " " Luchnik and Sesankaes "

7 +

also obsened an ani "p 'teau" of effect at intermediate gamma Y

i4...

doses. s ery similar to ti..

bu

• for cell transformanons by Miller and Hall P an effect which me t extrapolations from doses of 50 to Y

k 400 rad would underestimate tr. effect at the 'owest doses The produc -

tion of chromosomal aberrations at low doses cannot be considere<' patho genic for ary disease as yet. as mentioned abose. but these obsenations 7

irdicate that cautmn is warranted in any assumptions about he relatne

• )I

HE EFFECTS OF IONIZING RADIATION 8omatic[ffecl5lConcer 249 nly to 30%, thus any consistent effectiveness of tigh and low LET radiation at cumulative doses of 10 rad fect the analysis of neutron effects in or less.

ssump%n that the RBE for high LET ases es the dose decreases does not SUMM ARY AND CONCLUSIONS 1se cur e for low LET radiation must j

It is equally possible that the dose-It is evident that adoption by the somatic effects subcommittee of the is curvilinear downward. The point linear no-threshold dose response model for defining radiation induced pendent of dose, we may underesti-cancer risks remains empirical at this time. There is no adequate theoret-of high LET radiation and overesti-ical model of human carcinogenesis that permits derivation of a dose-of low LET radiation. But the avail-response relationship from first principles. The fact that radiation-i at indicate to me that this degree of j

induced cancer risk estimates from a large number of human studies with at. that is, more than a factor of 2.

great variability of ethnic, cultural, and other environmental factors timates may esentually have to take capable of influencing the results are as consistent as they are when com-of the population whose radiogenic pared on the basis of the linear extrapolation, suggests that radiation acts ier than the population at large, any by increasing the probability of an initiating event, a somatic mutation.

is that may oserestiman the risk by Other environrnental factors which can modify the subsequent chance of ne.

neoplasia are sufficiently widely and randomly distributed in all human ative effectiveness of high LET radia-populations that the excess cancer risk is defined primarily by the prob-chromosome aberration studies in ability of oncogenic cell transformation by radiation exposure, if such a nosed to high background radiation.

transformation kvolves a radiation induced point metatien or other small as been especially to radon 222 the modification in the cell genorr.e. then on classic target theory the linea

- essentially two-break effects on the no-threshold dose-response curve is entirely appropriate. Until we know population living in a sillage on more about the process of cancer development in man, we cannot go fur-malities were found elevated com-ther with this problem.

tosed.32 In this case, it was postu-The new evidence concerning cellular mechanisms of radition car-

'b-212 daughter of Rn-220 reached cinogenesis available since the BEIR I report represents in my view a major rather than the high background of change in emphasis from the past. Whereas biophysical considerations, of On the other hand, the dose-related a hich the Kellerer-Rossi theory is an example, have previously dominated by Evans et al.33 ir. nuclear ship-the field and played an important role in concepts of effects of low doses of sw cumulative doses were from ex-the different types of radiation, it is apparent that much raore prominent a radiation." It is of interest that 5 now are biological variables that can involve the con ersion of an initiating as produced in human lymphocytes esent induced by radiation into a fixed or heritable cell transformation, 15.3*35 Luchnik and Sevankaev35 and the subsequent host factors that determine the probability of develop-1" of effect at intermediate gamma ing cancer. These biological factors include DNA repair processes and r cell transrormations by Miller and cellular mechanisms that modify them, the action of promoting agents extrapolations from doses of 50 to and conditions that affect cell proliferation, the influence of viralinfection et at the lowest doses. The produc-on transformed cell DNA, immune processes aficCling survival of

.v doses cannot be considered patho-transformed cells, and the effect of age on replication characteristics of oned above, but these observations the transformed cell or cells, any assumptions about the relative The above comments appear to be quite straightforward, and I believe

~

1 s

150 THE FFFECTs OF IONIZING RADtATION were the consensus of the somatic effects subcommittee du i when the subcommittee was continuing to meet r ng the period with that acbpted in the third section of Chapter Y of the pres. Contra The basis of the ratio of t'te linear and quadratic coeff' ients (a 'b ent sersion.

A-bomb survivors, data which do show a defin e

anese of the data to the leukemia results. Thus 1 ukem of dose-e t from other radiogenic c?ncer types is taken a er with re j

[

' i~~

number of important inferences for aII rtdiation-induced cancers overning a factors derised from leukemia are then used to fit the ob

-#f '.-

. These 4 i, cancer morrohry, which as has been discussed above are deficient in e

ata for k,y%k are then comened to incidence data by applying

.m-2' ' ' y.

)

ity by site show n in Table V-15 This approa h ortal-

~~

Japariese Tumor Registry data for total cancer incidence whi h fstudio c

Nagasaki and Hiroshima (with adjustn. tnt of an RBE c or both about 5) are in excellent agreement with the incidence data dern d f for ncutrons of dividual sites from the extensive international studies d e or in-dix A. and summarized in Table V-14 pnen-

~ ~

The roundabout approach taken abose in the present version in eff discaros all the human studies of radiation-induced cance ect Japanese data in defining cancer risk from lo s LET radiation It al xcept the the effect of reducing the cancer risk estimates sufficientl so has ble for the conclusion to be drawn that the er R six canc y that it is possi-estimates are about the same as were derned in BEIR.

~ ~

y risk y-nores. of course. the important step of changing to cancer incidt ce a

g-data, especially for cancers of the thyroid and fe as a

. g portn e dicate that as the follow up of human study populations has b

. w hich in-tended, esidence of cancer risk is increasing the dose een ex-base been observed hase progressisely decreased and the numb s at which effects ferent human cancers in w hich radiation exposure has show n a i

er of dif-been extended. The present version of the third section of Cha t n effect has failed to make these important points. and thrs has not prouded i p er V has siew. an adequate up-to-date scientific anessment of risk wh

. n my

.,e purpose for which the BEIR 111 Committee was established ich was the The fact that the numan epidemiologic data whict are relesar:t to th dose response issue are generally consistent with the linear n e

dose-response model remains the principal basis for use of this o-threshold should be emphasized that.nery effor in presenting epidemiologic model it eu-B-.-QRQ%.QQ,lN'=,A~

o WA w.%' b < L "?

t

+

P!

?s J T ha. '

' '7 'hjkNf..?

.e ;

a 4

</ -

'~

dence of cancer induction by radiation should be as carefully and EFFECTS Of IONIZING R AD.ATION s subcomm.ttee during '.ne period rigotously done as possible to take account of the dilution effect of non-i g to meet. Contrast tms position radiosensitive cancers, age specific adjustments, effects of confounding Chapter \\ of the present version.

variables and the influence of latent period. In Appendix A, and for the Japanese data in Figs. V-6 and V 7, efforts have been made to achieve this guadratic coefficients (a/b, in the ia registry data from the Japanese aim. The graph of Japanese data presented in Dr. Ilarald Rossi's dissent-a definite cursitinearity of dose-ing report has not been corrected for age, which is a major correction for for neutrons is taken from the fit cancer evaluation because of the sharp effects of age on cancer rates; the s leuhenua, a human cancer with Nagasak Life Span Study population is younger than the Ifiroshima ftar irradi tion differing markedly population at;d the age distribution varies by dose category. In addition iken as the paradigm gosermng a the period 1950-1954 has been included by Dr. Rossi for all cancers, w hen radiation-induced cancers. These we know that for all cancers except leukemia no excess risk is likely to used to fit the observed data for have occurred during this period. It is time to recognize that epidemiology gussed abose are deficient in im-is a rigorous discipline requiring special attention to detail thet charac-tive solid tumors. Mortahty data terizes any science.

applying factors of cancer mortal-Finally,I would like to take this opportunity to thank those members of 3 roach studiously avoids using the the full sommittee who have worked hard so produce those parts of the cancer meidence which for both currer.t sersion that provide a scientific basis for assessing somatic a:id nent of an RBE for neutrons of genetic risks. It is regrettable that tl., results of their hard work have been the, eidence data derned for in-so lorg delayed in being released for general use.

m

,ional studies describec: m Appen-ve in the present version in effect REFERENCES iation-induced cancer except the

1. Russell. W. L. L B. Russell and E. M. Kelly. Radiation dose rate and mutation fre-om low LET radiation. It r.lso has quency. Sc enu 1263.(461550. IM i

nd its apphcability to the estimation of the imates sufficiently that it h possi-un H.

ug t m us the BEIR !!! Cancer mortality risk genetic hazards of radiation. In Proceedings of IV International Congress of Radiation tved in BEIR t. Th.

Research. Evian. France. June 29-July 4.1970.

is conclusion ig-

3. Origins of Human Cancer. H. H. Hiatt, J. D. Wawn and J. A. Winsten. Eds. Cold h**" ting to cancer incidence as a Spring Harbor Laboratory.1977. See e.g. R. B. Setlom. F. E. Ahmed and E. Grist.

889-901.C.

X" d"~"* P gmentosum: damage to DN A is involved in Carrinogenesis. pp.

ie considerable body of supportive i

aid and female breast, which in-HeMehrgu. Scogenic inmfonnasn of rodent ceu unes by chennca carcinogent J. McCann and B. N. Ames. The Salmonella /microsome mutagen cits study populations has been ex-pp.1513-1519.

test: predictise value for animal carcinogenicity, pp. I U1-1449. M. Meselson and K.

asing, the doses at which effects 1473-1481.

Russell. Comparisons of carcinogenic and mutagenic potencv. pp.

leereased and the number of dif-

4. McCar.n. J., E. Choi. E. Yamasaki and B. N. Ames. Detection of carcinogens as CIposure has shown an effect has mutagens in the Salmonella microsome test: Assay of 300 chemicals. Proc. Natl. Acad.

he third section of Chapter V has c 230Mm n and thus has not provided. in m}-

5. Knudson. A. G.. Jr. Genetic predisposition o cancer In Origins of Human Cancer. H.

H. Hiatt. J. D. Watson, J. A. Winsten. Eds. Cold Spring Harbor Laboratory.1977. pp assessment of risk which was the 45-52.

tee was established.

o. Peto. R. Epidemiology, multistage models and shott. term mutagenicity tests. In Origins

ic data which are relevant to the f Human Cancer. H. H. Hiatt. J. D. Watson. J. A. Winsten. Eds. Cold Spring Harbor tent with the linear no threshold Mt ry.1971 PP.1403-1428.

pal basis for use of this model. It

7. Druckrey. H. Quantitative aspects in chemical carcinogenesis. In forential Car.

in presenting epidem. logic evi-io l

1

~

~

~

+.

s 252 Tur rrrecTs or tourzrNG RADIATION canogener Hazards from Drug: U 1 C C Monograph Series. Vol Berhn 196' pp 60 '8

'. Spnnger Verlag.

8 Kelleret. A M. and H H Ross:

Radiai Res Quar crh 8 k5-158.19'2The theory of dual radiarmn action Current Topics in 9

inn. D E Acnon of RaJwron on Lemne Cells 2nd Ed L'nn 1956 e-eruty Vreu. E ambndge.

10 Schmid. E. G Rimpi and M Raucninger Dow responw rela *mn of 4'

ratwns m human lymphocytes after m Utro irradiation with 3 MeV ele t chromowme aber Res P 22k-236.19'4 c rons Radiat 11

^

Kucerosa. M. A. J B Anderson K EBuckton and H J. Esans X ray-mduced chromosome aberratons m human penpheral bked leucocytes lesels of exposure in utn Int J Radiat Bmo 21359-396.19'2 the responw to lou y

eh 12 Nean. G J. J R K. Saiage and H J Evans Chromand ab pollen tubes indt.ced by monochromatic a rays of quantom energs 3 and 15 k erranons m Tradescanna Radiat Bol 8 l-19.1964 nt J 13 Pomen E L. J I Lyman and C A M,

Tobias Some effects of accelerated charged par y.-

twies on bactenal spores Ir.t J Radiat Bel 14 313-330. 1966 11 Sad..r. G A Dose. dose-rate, radiatmn quahty in Agmg Carcinogenesu

. and host f actors for radiatmn-mdaced hfe shortenmg Plenum Pubbshmg Corp N.Y and Radnanon Biologs k C Smsth Vd 19'S pp 493-51'

. f 15 t

Cos R-. J Thacker and D T Goodhead 3 aluminium characternue ultrauf t x-rauinactuanon and mutano

~

mahan cells b hamster and human diploid atis to alurr.inii.m n-rays aoj radianoIl Dose responses of '. hme int J Radiat Bel 31 Sol 5'6.19" ns of different t E r 16 Goodhead. D T Inactnation and mutation of cultured m mmahan ce charactensoc ultrasoft :

y aluminium J Radiat Biol 32 43 'O. 19 " rays !!! Imphcanons for theon of dual radiatwn action in.

17 Borei C.,nd E J. Hall Effect of spht doses of n rays on neop!zin smgle cehs Nature 252 499-501. 19'4 c transformanon of lb Miller. R. 4,nd E 1 Hall cogenic trarnformanons m vitro C3H 10T mouse embno aus NatureE 19 L."le. J B Qua~natne studies of radiation transformatior:

2'25960. 19 'h BALB 3T3 ceh hne Cancet Research with the A31 11 mouw Little. J B. H Nagasawa and A 39 14'4-1480. 19'9 20 R Kenneds tion effect of s irradianon.12-0-tetradecanoy! phorbol 13 acetateona repair and ma9gnant tra hibitors on transformatum and sater-chromand enchanges in mouse 101

. and protease m Res '9 241-255. I9'9 celh Radiat

21. Little 1 P Radiatmr of Human Cancer aremogenesn m utro imphcanons for mechanisms laboratory.19'?. pp 923-939h H Hiart. J. D Watson. J A Winsten Eds Cold spring Harbor in Orrams

22. Bottk. C. Neoplastic transformation follommg spht doses cf s ra g

50.845-o46. 10 "

- ys Bnt J Radiol

+.

23 Tokunaga M. J G Norman. M Asano. 5 Tokuoka. H Eraki Tsup Ma hgna n' breast T Nishimon and Y rumors among atomic bomb sunoors Hiroshima and Nagasak. 1950-19'd J. Nati Canar Inst 62 134'-59. 19'9 24 Belsk). J. L. K. Tachikana and 5 Jablon decade of esaminatum m a fixed pot utanon i rThe beahh of atomic tomb survnon A

25. Steer. A. C E Land. I M Mornama. T Yamamoto M Asam and Bion Med 46 284-296. 19'3 curaev of diagnous cf cancer in the JNilP Ance hie span study sample Rad anefup Ac-Research Foundanon Technical Report n 1 '5 Hiroshima ianon Effeca Foundanon.19'S Radianon Effects Pescarch 9

r 1

g

253 THE EFFECT5 OF IONIZING RADIATION Somatic Efjects: Cancer C. Monograph Series. Vol. 7. Springer Verla8'

26. Fialkom, P. J. The origin and development of human tumors studied mith cell markers.

i -

"8'

' ' 'I

• I icory of dual radiation action. Current top cs in

27. Hempelmann. L H., W. J. Ha!! M. Phi!!ips. R. A. Cooper and W. R. Ames. Neo-P asms in persons treated with a rays in infancy. Fourth survey in 20 years. J. Nat.

l 3 CEIL 2nd Ed. Unisersity Press Cambrid e-Cancer Inst. 55.519-530. 1975.

8 an M., L R. Simpson. L K. Patterson and L Cohen. Thyroid cancer associated er. Dose. response relation of chromosome aber.

mith radiation expmure: dose effect relationship. Proc. Symp. on Biol. Effects of Io*

sitro irradianon with 3 MeV electrons. Radiat' level Radiation Pertinent to Protection of Man and His Ensironment. GAEA SM202) em geng Vknna. N m Ebh nL c

E. Buckton and H. J. Esans. X-n.y. induced

29. M dan. B, E. Ron and A. Werner. ThyroM cancer follo*ing scalp irradiation. Radi-ipheral blood leucocpes: the response to los 0308Y 123:741-744 1977-8' it. Biol. 21:389 390, 1972.

Evans. Chrcmatid aberrations in Tradescantia

• 7 x rays of quan'um energy 3 and 1.5 kev. Int. J*

radioactisity. Radiat. Res. 80:61-80, 19'9.

31. Stenstrand. K., M. Annanmaki and T. Rytomaa. Cytoger. etic insestigation of people in l

Fints <! nsing household mater mith high natural radioactivity. Health Phys. 36:441 obias. Some effects of accelerated chargedPar.

. Biol.14:313-330.1968.

i quahty, and host factors for radiation. induced

^"

sis and Radnarion Biologv. K. C. Smith. Ed-L E. Castro.' Cytogenetic 'mestigation in a Brazilian population lising in an area of e.igh 3,493 337.

d. Inactivation and mutation of cultured mam-ultrasoft x-ravs. II. Dose. responses of Chinese minium x rays' a d radiations of different tar *

34. Schmickel. R. Chromosome aberrations in leukocytes esposed in vitro to diagnostic

• I tica. of cultured mammalian cells by aluminium cations for theory of dual radiatiSn action. Int' human lymphocytes.1. Dependence on the dose of gamma-rays and an anomaly at som doses of n-rays on neoplastic transformatien of ray dose fractionation on the induction of on.

T mouse embryo cells. Nature 272.59-60.1978.

liation transformation with the A31-11 mouse 9:1474-1480, 1979.

nnedy. DN A repair and ma!ignant transforma-leca'ioyl-phorbol 13-acetate, and protease in-omatid exchan6es in mouse 10T cells. Radiat.

vitro: implications for mechanisms. In Onsins atson, J. A. Winsten. Eds, r ld Spring Harbor o

lloming split doses of x-rays. Brit. J. Radio!.

2. S. Tokuoka. H. Ezaki. T. Nishimori and Y.

ig atomic bomb survivors. Hiroshima and ist. 62:1347-59,1979.

don. The health of atomic bumb survivors. A tion. Yale J. Bot. Med. 56:284-296. 1073.

i T. Yamamoto M. Asano and H. Sanefuji. Ac-occ life-span study sample. Radiation Effects a 175. Hiroshima: Radiation Effnts Research

V Separate Statement Critique of BEIR III t

I i

li A RALD 11. ROSSI 1

SUMMARY

The fdst report of the Committee on the Biological Effects of Ionizing Radiations (BEIR 1) has profoundly influenced gosernmental regulations and the public attitude towards radiation. It is to be expected that the i npact of the current report (BEIR III) will be equally significant. The Committee drafting that report has thus been faced with a heavy re-sponsibility because its findings are likely to affect national energy policy and the practice of medicine. In both of these areas overestimates sell as underestimates of the radiation hazard could result in serious at detrittent.

This is especially important with regard to the risk of radiogenic cancer which is frequently considered to be the major hazard of ion; zing radiation. This critique deals with this subject only.

BEIR !!! represents an advance over BEIR I in a number of respCCts:

I. The uncertainties of risk estimates are stated more explicitly and it I

is stressed that the so-called " linear hypothesis" is likely to result in l

overestimates of the hazarc' trom low-LET radiation.

l Dissenting statements prepared by individual members of a National Research Cousil com-mittee are not subject to the normal resieu processes of the National Academy of Sciences, nor are they subject to committee or staff editing or renen. The) appear exactly as the dissenting committee members prepare them. The NAs NRc neither endorses nor iakes i

resr> insibihty for the conient of the statements.

234

253 Somatic Effects: Cancer This has led directly or indirectly to further improvements,

11. It is acknowledged that it is probable that the cancer risk rises with absorbed dose at a rate that is higher than linaar and the preferred mathematical model conforms with this postulate.

111. Extrapolations to single whole body doses of less than 10 rads are eschewed.

IV. It is stated that the effects of annual radiation doses of the order of 100 mrads (Iow LET) are unknown and that it is unlikely that they can l

be demonstrated.

V. It is recognized that Ras is an important factor and itisfrequently assumed that it increases with decreasing lese's of effect. In most in-8 stances data from Hiroshima and Nagasaki are not pooled on the as-i sumptions of equal effectiveness.

BEIR III is howeser deficient in two ma;.ar respects:

1. Many of the risk estimates provided are still based on the " linear hypothesis" despite continuing and raounting contrary evidence from n the Biological Effects of lonizing radiobiology and epidemiology (much of it quoted in nEiR 111). Even influenced gosernmental regislations though these figures are given somewhat less prominence, they are likely iation. It is to be expected that the to assume primary importance for standard-setting bodies which, for the

!!!) will be equally significant. The sake of prudence, are likely to adopt the highest estimates.

. thus been faced with a heavy re-

11. BEIR !!! fails to present explicitly data that indicate risk factors re likely to affect national energy that are less than the lowest given in its report. This does not only again in both of these areas overestin ates tend to support excessive risk estimates far low LET radiation, but may iation hazard could result in serious also lead to, perhaps egen n ore important, underestimate = of neutron hazards.

l i regard to the risk of radiogenic

1 to be the major haz;rd of ionizing

"'I' DETAILED COMMENTS r EEIR I in a number of respects:

The inadequacies of the epidemiological information on radiogenic cancer in man permit a wide variety of iaterpolations and extrapolations ites are stated more explicitly and it of dats that are often uncertain, if only in the statistical sense. The r hypothesis" is likely to result in deduction of the most likely risk estimates can, however, be facilitated

-LET radiation.

by considerations of theoretical or experimental findings of radiobiology which make certain models more--and sometimes much more-plausi-nembers of a National Research Council com-ble~

ocesses of the National Academy of Sciences; Theoretical considerations permit def...initive conclusions on the dose-hting or resiew. They appear exactr3 as the effect relation for individual (autonomous) cells, but at this time they t The N45-mac neither endorses nor takes cannot be employed with any assurarse to determine this relation for I

+

256 Tile EFFECTS O'loNIZING RADIATION the complicated process of radiation carcinogenesis. They do, however, lead to the conclusion that the RaE of high. relative to low LET radia-tions should increase with decreasing Icvel of effect to va!ues which are very substantial and that this should be so not only for eutonomous cells, but also for interacting cell systems.

Experimental observations on higher organisms base confirmed this expectatien. In line with theoretical predictions, the RBE generally increases wita decreasing neutron dose, Du, according to RBE = K(DryD2, in a number of systems K has been found to be about 45 if D3 is ex-pressed in rads and the neutrons have energies comparable to the mean energy of the fission spectrum (-0.5 MeV). RBE values in excess of 100 have been observed at neutron doses of the order of 100 mrads wh are thus equivalent to gamma ray doses of the order of 10 rads.

While experimental radiobiology is in accord with theoretical predic-tions regarding the dose-RBE relation, it also discloses a wide variety of dose effect relations for carcinogenesis. Some of these even show a reduction of the natural incidence at moderate doses of low LET radia-tion (and esen for high LET radiations). This is only observed when the natural incidence is high; however, statistical limitations would not permit a clear indication of this effect when the natural incidence is low.

In most (but not all) instances, the curvature of the relation for lo.c.T radiation is positive indicating that in addition to any linear depender.ce on low doses (regardless of sign), there are positive quadratic and per-haps higher order terms in dose at intermediate doses. At high do',es, a reduction or even a reversal of slope is often observed.

in summary, radiobiological considerations lead to the expectation that if cancer incidence is related only to terms that are linear andhr quadratic in dose, only a rough approximation may be attainable a many instances. In such approximations the relative magnitude of lineas and quadratic terms is likely to differ depending on the type of neo-plasm involved rnd a summation for all neoplasms could have a particu-larly complicated shape. It would, however, also be expected that, in general, linear extrapolations from doses of several hundred rads lead to an overestimate of the effects of doses of the order of 10 rads. It would i

furthermore ce expected that because of the dose dependent RBE. the shape of any dose-effect relations is not the same for gamma and neutron radiations and in particular that they not both be linear above gamma ray doses in excess d about 10 rads or neutron doses that are 100 times less.

BEIR III employs three approaches to the analysis of epidemiological o-i gh _ _ _

g-

=

., 7

l.

r 257 THE EFFECTS OF lONillNG RADEATIONSumatic Efects: Cancer nesis. They do, however, data on radiation carcinogenesis: They involve the " summed sites" har the a E of h h 8 IVe to I w LEr radia-method, the mortality (i.ss) data for Japanese atomic-bomb survivors ith decreasing level of, f to values which are and the Nagasaki tumor registry data.

<at this should be

• ut n mous The " summed sites" treatment is based on estimates of the incidence

ting cell systems.

camers in individual organs as gisen in Appendix A. The input data tions on higher organisms h ith theoretical predictions Ib doses in excess of 100 rads. The Japanese data employed are essentially g erally i

g neutrori dose. Ds, accordi"8 '

all from Hiroshima with assigned asE values that vac between sites but are independent of dose for any of them. With the exception ofleukemia, ass = K(D )-

2, the " linear hypothesis" is employed throughout. This treatment evi-3 dently conflicts with radiobiological knowledge on several counts. There K has been found to b bout 15 if DS is ex-are further objections to these data largely obtained frmn diseased neutrons have ener mParable to the mean individuals of different ethnic backgrounds. For example, the spondylitic trum f-0.5 Mev) au s f 100 population was exposed to very high doses and these were applied only rutron doses of the order f rads which to tissues in or near the spine. This poses problems in the assessment of nma rey doses of the ord f

ds, the "awrage" dose. If the leukemogenic effect of large doses depends on fiobiology is in accord w h heoretical predic-the square of the x. ray dose (as in fact assumed in BEIR !!!) and if 40%

mE relation, it also disel $

wide variety of of the bone marrow (that located in or near the spine) is irradiated with carcinogenesis. Some of th $

v n show a a dose, D,, with the remainder receiving essentially zero dose, the effec-ncidence at moderate dos 5 f i

• LET radia-rive dose is not the mean 0.4 D, but instead 0.63 D,. Such discrepancies

r radiations). This is

however, statistical

% v d when the bec'me even more pronounced if the irradiated fraction of a tissue os

.ut tmns would not organ becomes smaller.

f this effect when th ral, eidence is low.

In the absence of other information, these estimates might be con-m Sces, the curvature of th f r low.LET sidered as crude upper limits of the true risk for individual organs.

-ing that in addition to I-dependence liowever, the utilization of their sum in the methods employed to deter-f sign), there are posit, dese at intermediate d dratic and per-mine the o$erall cancer risk is one of the principal deficiencies of the BEIR HI report. As was to be expected, it results in a substantially larger 8

of slope is often observed peal cons:deraticas lead to h e Pectation estimate may well be adopted by standard setting bodies who, in the related only to.tres that interest of caution, may select the highest estimate provided.

r and/or rough approximation n

be anam.able in The tss data are generally considered to be the most reliable source 2roximations the relat'

. de of linear of information on sadiogenic cancer in the Japanese cities. They also n

ily to differ dependi Pe of neo-pe mit a straightforward assessment of the cancer risk for a period of stion for afi neoplasms l

H w ng Madatie of a munal(aoch eenicaHy Mc0 8mM yr would, however, also b s from cose5 of several hundred E**"

E s of doses of the order of 10 rads. It would calculations, it masks the true dose e!fect relation of the over.all cancer st because of the dose d impact for which the statistical fluctuations are substantially less. Figure n

E, the lations is not the same I is a plot of cancer mortality per person year as a fraction of total o 8 nd rticulN that they not both be h,near above kerma at Hiroshima and Nagasaki. These curves are not corrected for if about 10 rads os neutror: doses that are sex or age, but it may be assumed that such corrections could introduce 1

nly min r changes.

voaches to the anal 5I5 Of epidemiological It appears that at Nagasakiit is impossible to detect an excess cancer i

258 niE Errscrs or ron zino RADIATION t

i s-

/

4 2

3<- - -?

1 _... _... -

[

i 3,_

4...p.... q ----' -~i l

I k

1 e

too aoo soo eco soo Totes Kerms/res l

EIGURE I Average for the period 1950-1974 of the mortality from all malignant neoplasms per person year sersus total free in-air tissue kerma at Ifiroshima (closed circles)

(

and Nagasaki (open circles). The bars represent I standard deviation.

i f

incidence at kerma values of less than absut 300 rads although the populations exposed in each of the low dose intervals were about 1,000 or more. It is also e-ident that in line with other radiobiological in'or-mation, the RBE of neo*rons was very high. It should be borne in mind that at a given kerma at Hiroshima, only roughly 10% of the total absorbed dose to deep lying organs was due to neutrons. At high kerma,

{

the Hiroshima data exhibit fluctuations which may be due to a variety of reasons, but the low kerma data con be approximated by s

Mu, = 2.2 x 10-3 + 8 x 10-6 (Kn,/ rad),

[

while the Nagasaki data conform to f

.li Ms, = 1.8 x 10-3 + 5 x 1 r'(Ks./ rad)2 h

Here M is the mortality due to all malig int neopiasms per person year, K the total free in-air tissue kerma and the subscripts stand for the two cities.

Because of the high RLv of neutrons and their virtual absence at Nagasaki, it may be assumed that at low doses, all cancers were induced by neutrons at Hiroshima and by gamma radiation at Nagasaki. Em-tl

mym,,

k h. -

$.6 f TEST TARGET (MT-3) 1.0 lll82 E4

_ EE E I 1.1 \\,[ EM

-)

I-

/

I.8 g

E 'I.25 1.4 1.6 dl

=

/

6"

++>+,

<>+#3 kV#f'3,, _ _

4},,:}{4

)

t s

o.

e_

=~MP eu r-@m V f/

+NNNi#S.

IMAGE EVALUATION TEST TARGET (MT-3) 0 g

1.0 E82m y @ EE o

~

1.1 M.'" E g

g I.25 1.4 ll i.6 i

/

6" s%

++

'N

$Ng, A

///

t a

_ _ _-_ 1

_eu :

' EFFECTS OF IONILNG R ADI ATION 259 Somatic Effe' v Cancer ploying the dose sersus kerma relations given in BEIR 111 one obtains approximately o

Ms = 1.3 x 10-' (Ds / rad) l

. *....---Y and

~~*'~~~......

M, = 1.7 x 10-s (D,/ rad)2, Seither of these relations (and especially not the second) should be applied to absorbed doses that are less than about 10 rads. The estimate

~

for gamma radiations is lower than any gisen in BEIR III. The neutron est mate is higher than any values that might be inferred from this rep >rt.

I The failure to explicitly provide the information in figure 1 and to f,

derise the above estimates is another major deficiency of BEIR Ilt. Dis-4 of the mortality from a!! m.lignant cussions within Committee did not produce substantise reasons for t ssue erma at Hiroshima (closed eirclesi rejecting the salidity of this analysis and while there may well be reasons for considering other approaches, it is apparent that this analysis is of considerable significance.

n about 300 rads although the The so-called L4 estimate for these data is not only scientifically contraindicated, but also lacks any foundation in the absence of a sensi-dose intervals were about 1.000 with other radiobiological infor.

ble linear component for Nagasaki. Efforts to approximate this curse by tigh. It should be borne in mind a linear and a quadratic dose term yield a negatise sign for the former.

only roughly 10% of the total in principle, there is no reason why this should not be so and mortality due to neutrons. At high kerma, data for lung cancer at Nagasaki point in this direction. The statistical which may be due to a variety of evidence for this possibility is nesertheless much too weak 'o provide approximated by significant support for the assertion that the natural cancer mortality was reduced by small doses. On the other hand, the LQ-L analysis is

< 10-6 (Km / rad).

based on the relatise magnitude of liccar and quadratic terms as derived from leukemia incidence data from the rumor reg l y and this is being justified by the objective of introducing a linear term into data :n which linearity cannot be found.

< 10-' (Ks./ rad)2 The mixing of data from the t.ss series and the tumor registries is also inappropriate because they appear to be discordant. The reasons for this nant neoplasms per person year, are not clear at this time. Although the t.ss data are generally considered the subscripts stand for the two to be more rel.2ble and cancer mortality may be deemed to be more relevant then cancer incidence, BElk III quite properly decided not to is and their virtual absence at ignore the registry data especially since they indicate higher risk factors.

Analysis in terms of all three models can be justified including that by l'

doses, a!! cancers were induced ma radiation at Nagasaki. Em.

the L L model since the Nagasaki data are best fitted by a linear rela-4

F l

i 260 niE EFFECTS OF IONIZING R ADIATION tion. Howeser the implied dose independence of EBE Casts further doubt on the va!idity of the registry data.

The arguments presented here lead to the conclusion that the mo -

pi:usible estimate of the cancer risk from IOH LET radiation is Ioner than any of the ones gisen in BEIR III. As a corollary BEIR III may moti-vate potentially dangerous underestimates of the hazards of high LET radiation.

??

4' t-l

, '.,, 4f pf 7 je#. %,.f :.' '. ' Y ' * ' ' h4, p ~ s

. f-'

- h,[a T

3

>,3 x

.[_,., #; j ' '* ~ M, gg -j.' ge, ' f f'-

' pigf(- 7.:g,,*7;g.?.

• j',.;,,

.-s r_3 _jy N,

1 l

rHE EFrECTs OF IONIZING RADIATION pendence of RsE casts further doubt Comments on Certain Divisive Issues Noted in BEIR 111 ad u the conclus on that the most i

s!t from low-LET radiation is lower it. As a coro!!ary BEIR til may moti.

EDWARD W. WEBSTER

• timates of the hazards of high-LET This commentary is not intended as a dissent from the principal findings of the Committee, but rather to illuminate some of the issues on which the Report notes divisions of opinion within the Committee.

Probably the most important charge to the Committee was to estimate the increased risk of cancer likely to be incurred as a result of low doses of low-LET radiation delisered to the whole body. A linear quadratic dose /effect relationship, defensible in the light of current radiobiologic findings, has been adopted by most of the Committee members as a reasonable basis for prediction of the risks of radiation induced cancer.

While subscribing to this important change in scientific viewpoint of the BEIR 111 Committee Compared to that of BEIR I(1972), I must express a number of caveats regarding the actual forms of the dose /effect relationships utilird in the BEIR li! risk estimates. I recognize, however, that the three estim;tes of mortality from solid cancer are not incon-sistent with the Nagasaki mortality data.

Dissentmg statements prepared by monidual members of a National Research Council com-mittee are not subject to the normal review processes of the National Academy of Scie.eces; nor are they subject to committee or staff editing or reuce. They appear enactly as the dissentirig committee members prepare them. The has-%nc neither endorses nor takes responsibility for the content of the statements.

• Dr. Ingram subscribes to this statement. Dr. Mays also subscribes to this statement.

mith the addition of the paragraph that appears at the end of the statement.

261 l

l

'~

e 202 THE F FF F cts OF IONIl!NG R ADioTION h o rna 1 On page 16' it is noted that in the linear quadratic relation fitted to the Nagasaki solid cancer mortality data (Figure V-6). the slope of the o.

the hnear component is about 0 4 esceu cancer per milhon per year per 0 015 rad. This slope depends on the auumption that the hnear term and the square law term are equa! for a gamma dose of 1 0.0066 or 116 rads.

d This particular knear-quadratic relation was re - cred by some Committee dedui members on two main bases a) the not is about 91 for a neutron dose 4

of 1 rad. b) the ratio of solid cancer to leukemia for gamma rays is exces 0 4 whereas the British ankylosing spondyhtis study for high doses of site r x-rays suggests a ratio of about 5. The relationship was thereupon ca na adjusted to melude the RBF for the leukemia LQ model sir. 23. T his arbitrary change caused the slope of the hnear component of the LQ E

relationship to be mereased from 0.4 to 1.4. that is, by a factor of 3.5.

cana The sohd cancer nsk estimates fmally propoundeo in Table V 19 and tha which are the " preferred" estimates, are based on this larger slope it is important to note that a) the retent study of leukemia in the A b mb ihr s

- o surmors by 1shimaru et al i estimated the asF for i rad of fission appo,>

neutrons at 4h based on a quadratic model for gamma response. this hau is similar to the alue of 45 proposed by Rossi on more general grounds.-

ser e and b) there no obsious reason w hs the ratio of solid cancer t"

(

leukemia shoum be 51. particularly in the low dose range The rano in N w ill depend on the specific shapes of the leukemia and sohd cancer imp dose response cu n es Thus m the animal studies by Ullnch et al '

o ra t...

the ratio of the modence of 3 sohd tumors (osanan, pituitan. and emat H a rdena n ) to the incidence of thymic leukemia saned from 2.4 at 100 rads to 0 b at 25 rads and 0 I at 10 rads The ratio was more nearly and con sta nt for neutrons Morcoser it is noted or p 209 that the solid ratht tumor leukemia ratio is "sery sensitise to the age distribution of the irtaa subjects under study and to the duration of followup ' For example.

the work of Stewart and K neale' on in g

utero e x posu re mdicates a ratio of 25 25 or 1.1. Thus. if the Hiroshima Nagasaki mortahts data is dose not adjusted for RBE m this arbitran fashion, the " preferred" risk int:

y estimates presented m Table V-14 would fall by a factor of about 3.

(

4 2 in the Report. the arguments on p 18~ summanied abose were du also employed to change the slope of the /mear dose effect rela tion g

employ ed for nsk estimation Whereas the slope of the best-fitting hne for gamma radiation data show n in Figure V.6 and Table V 9 was 4

140. the actual slope employed in Tables V-Il and V-20 was 3 4~

me:

an merease by a factor of about 2.5. Thus the knear model estimates of g,

cancer mortah ty presented are higher than those suggested by the g,

Hiroshima Nagasaki study by this factor.

y' 1 Ir 'he Report. the arguments on p.16' were a so used to change

.. ~ _ - Q, -.

- Q[ysf K j f gya;-.,' ' '

Q.. ~^ ' ;. mp,..,;:

4 N

y

..y.

s.: ;.

[Q, y i.

m

,.,. '. ;.,,'].y.

. ~_

.c 5

. ?. Q n'kyr

.. fy f

Qn

h.4,

W,-

.y yt ex y.

'gi Q.:,n wl:d?. %f.N

s P

'.7 M-

.., "f.* *,

Q

%^w

.x ex

- gr -'. -'w

... ' $ 6 c

1. R*g h ~

z.

- ~ y.'-t V,. %'a: v..

.,s q

x n

4 3-~

w-p

. A.

pO 2Dw}, n ZJA% WS -

.~ W

' M t h% QM ' ~W.b+"?:

HE EFFECTS OF IONIZING RADIATION Somatic Effects: Cancer 263 the linear-quadratic relation fitted the coefficient of the quadratic relationship from 0.0047 (Table V 9) to ity data (Figure V-6). the slope of 0.0184 (Table V-11 and V-21), an increase by a factor of 3.9. Again cess cancer per million per year per therefore t!'e estimates of excess solid cancer mortality presented for the nption that the linear term and the quadratic (square law) model are higher by this factor than would be ima dose of I/0.0086 or 116 rads.

deduced a priori from the Hiroshima Nagasaki data above.

on was rejected by some Committee

4. The Report fails to state explicitly that the linear risk estimate for RBE is about 91 for a neutron dose excess cancer incidence derised from Table V-10 (sum of the individual er to leukemia for gamma rays is site risks) is grossly incompatible with the linear estimate for excess

.pondylitis study for high doses of cancer mortality derived from the Hiroshima Nagasaki study (Figure The relationship was thereupon V-6). The aserage incidence risk from Table V-14 is 18 cases per million leukemia LQ model: viz. 23. This per year per rad, which is about 13 times greater than the 1.40 fatal f the linear component of the LQ cancer cases deduced from the J. panese study, or about 7 times greater to 1.4; that is, by a factor of 3.5.

than the incidence risk derived from the Japanese study using the ex-ily propounded in Table V 19 and pansion factors in Table V " his great difference seriously challenges are based on this larger slope. It is the credibility of the linear tu estimates based on the " summed sites" study of leukemia in the A-bomb approach of Table V-30. This writer belieses that these values not only sted the RBE for ] rad of fission hase " considerable upwurd bias" as stated in the Report, but cannot be c nodel for gamma response; this seriously considered in the light of the Japanese experience.

by Rossi on more general grounds::

5. It is stated on page 179 that "the data [for the site-specific estimates vny the ratio of solid cancer to in Appendix A] are reasonably firm for only a few organs." One of the in the low dose range. The ratio imPortant organs to which this applies is the lung. irradiated by low LET of the leukemia and solid cancer radiation. The risk estimates for lung derived in Appendix A are almost animal studies by Ullrich et al.3 entirely dependent on the epidemiological studies of miners exposed to d tumors (osarian, pituitary, and high-LET radiation in the form of alpha radiation from radon inhalation e leukemia varied from 2.4 at 100

) rads. The ratio was more nearly and on the lung cancer incidence in Hiroshima. The assumption of the is noted on p. 209 that the solid rather low RBE values of 10 for alpha irradiation and 5 for fast neutron irradiation exaggerates the effect of low levels of low LET radiation.

tive to the age distribution of the ation of followup." For example.

M re importantly the lung section fails to note that the Nagasaki mor-tality data (Iow-LET radiation) show a deficir of lung cancer cases at on in utero exposure indicates a doses up to 100 rads and this is also reflected in the Tumor Registry oshima/ Nagasaki mortality data is incidence data for low gamma ray doses.5 The risk estimate for lung ary fashion, the " preferred" risk camer imm low-LET radiation is almost wholly dependent on the high uld fall by a factor of about 3.

dose (200 rad) ankylosing spondylitis study and is likely to be consider-

)n p.187 summarized above were of the linear dose /effect relation ably less at low doses.

as the slope of the best fitting line Additional comment by Dr. Mays: "I support the thoughtful com-i Figure V-6 and Table V-9 was ments of Dr. Edward W. Webster, and am particularly concerned that ables V Il and V-20 was 3.47 an the risk coefficient derised from the sum of individual site risks exceeds hus the linear model estimates of by a factor of about 13 that derived directly from the A bomb life-span 1er than those suggested by the mortality data. I feel that the latter is more likely to be appropriate tor.

and that future efforts by the Scientific Community should be directed n p.187 were also used to change toward resolving this discrepancy."

l a v a;

.n

'{

(

y3 ygg'.

)

y y

+

?..

M. - - #

.y be.

.' W }! L.:

1 ggik, J Qg],_

y

.}

j

}

y

..a kr

%k, 204 THE EFFECTS OF IONIZING RADIADON REFERENCES wM.

I. Ishimaru. T., Otake. M. and Ishimaru. M.: Dose.responw relationhip of neutrons s

sc.[ICf

+d[

and gamma rays to leukemia incidence among atomic bomb sunivors in Hiroshima

'"l and Nagasaki by type of leukemia. 1950-71. Rad. Res. 77:377-394.1979.

2. Rossi. H. H.: The effects of smati doses of ionizing radiaten. Phys. Med. Biol 255-262. 1970.

15:

3.1.!!: rich. R. L.. et al.: The influence of dose and dose rate on the incidence of n dnease in erM mice after neutron inadiaton. Rad. Res. te.Il5-131.19'b.

4. Stew art. A. and Kneale. G. W.: Radiation dow effects in relation to obstetr?c a. rays and childhood cancers. Lannt 1:1185-1188. 1970.

5. Beebe. G. W.

Kato. H., and Land C. E.: Life Span Study Report 8: Mortahty esperience of atomic bomb sunivors 1950 74. Technical Report na ar in 1

77. Radia-tion Effects Research Foundaten. Hiroshima, Japan.1978, pages A202 and A350.

\\

ISSN 0036-805

. ?ch - < : -

c.. m s v, m c] a: m \\ n t J

' :r n.n.

. (re,.el ~ f r

..;. ~

r -}e Y...).,,g, oq r

( (o3 N:.w.. s ;

. I, 11.

n ta s.

'(,

af j

5-

~

i e

n4

-.. g v.s< %W." L.w r.a:::a A.L.b

-. k ~, Gn/ w '.L s

p yt r"1

=

%.,s M'r.m 212, W.1M O

P kJ t

LETTras Popper on Darwinism: R. E. Kofn/d; II. /ciscl; l'raud. Sticace, and W:

Safego:nds: E. D. Ncimarl; W. S.prirc; Evalu -tmg the Soe;al

,?.

Ssiences: A. Ala:nr; OS!!.\\ Pohcy for Labom:ories: D. Wall.cs. fr.;

L More Elistory of Calculuc W. Wu scr.

873 y 1

E01 roat?.t.

Research Equipment Acquisition: I.. M. Rianuumle.

877 s

%J 6

-p..

ARTICs.F.s Tectoriics and I'so!ution of Vent:c R. J. Phillips et al.

879 6

The Mutation Component of Geoctic Dam: ige: 1. I'. Came and C. Ocimimm.

BPS Conflb. ting Objcetis es ia Regul.' ting the Automobile: L. B. Isac.

833 tszws AM)conz97 New A llomb Studies Alter Radiation Estimat:s.

900 Techno'o;y Transfer Reappraised 032 Science Adsiser Post lias No.n: nee in View.

S33 j

Briefing: Monuons Rchel on MX: A First thush with Neu Broom at N AS; Breeder Stumb!es in llouse: UCl.A Designing Big Agent Orance Stml;; Court Gives r

934 j

Blessing to llospital of Faith lic i

News of Bone Research Causes Fractiac.

935 Point of ricw: 1. con Rosenberg on the "llumen Life" lhil.

937 s

n:13EncN t;Iws Cultura! Disersity Tied to Genetic D:tferences.

033 e

soon nsvizws The Sahara and the Nile and Prehistory of the Eastern khara. sciicwcd

/. W. R. liamnd; So,iet and Western Aniluopo'ocy. 51. J. L o't:;

lleas, R. E. Lcwis; Gene Expiession, R. D. Colc; ih>oks Reseis ed.

911 REPORTS Phpical Effects of Vehicular Disturban. on Aiiil 1.ard. cape.

R. AI. Irer4on et al.

915 1

7 PhotoperioJ;c Control and Effects of Mebtor.in on Nonshiurirn

'I Fermogenesis and Brown Adipose 'lissue: G. II< /dmaict et ;'

917 4 rf W e D.4'.L'Na C /.5v FP." w it m i n ^S t L.

L'i.s,*3 Os o W C;c t; W s.~ex *n0TT.LO Nu.;

e c+

m rn. r*

Pree v%t I bWJ.4. t..W.: t F w n. t 4 %,N

.w i s 3..m n nr e.a4 ress.

c om iswu e,

?!h"cs G.

• W

' tt " W V- *.

o n ' E.~T.* ? !

......., c_ c.,., q n.;

F.3 p' I* rr es, n,wy f gyy u.m e.

nm w, v

4t m u o c.

n c.%..

fv., A.ny it f.ccw n cerve, w) pour.c.w ror eJ. t <. % C m'a c pn:c'o w cc o c a

%.:s

c. m.,e.~. e,

~..u,

.w,,a r.n u-s c.m m c. u, om. em.

E';4W4 (C) r#.~ r J P < P F MAM'GTGx !. ts +,: L5 gj r.; F**MTP c.;* " u WG A ; ?

r/, L td 0 J.T.

• w.

H.

f.vw 4 P 5.r- - -..

C + +.a V/,

A-

r.ma

. r.,,e ww w.,

e..

,.o w. ca t j /g3 Oq3 ALA3M s D vraiG'4 PACFIC Div S'04 SfN Ds t ~' TITI WD 4f..t v tR % O.

7*

N*S s.M $m. trey

/.^f %.

f.* c P L. -

• t*t*.e.

J..#r k

. D,s. s e..

.e.m e,-

m, n.

t,t e-

.* A swc.

. Sr t** A.*v w u r ' of ar a 5 5 M ' t u a - t s 't s /.va S./ W u - e

."'1..

U ' ' O f 9 WD %, c. e f o, Fry.y,,c,f e. p,;,.* e en Cat. etw 1 y ? - A m r

. est C.

A;,

c,,

.~, ' ' 2 i 9 U F.', y

-,... m.. f *t i 1*.,

4 t'

i

.f/

s,.u,'.,..'**a

,.e**

.t 3 4 % "- 4

.r, s

.*,.or

-. P. ) k 4

m,.., r 4 <

mn m

r m w m.

., a

.w..,n.

w. a

.- w:

, t m pc,,,..,: n.ro.,a w ~. ~ o,.....mA...u. n..

.,n,,n

+...

. e n,..,,

~

mn:

_,...,.... ; n....

-..r ~ s.....

n ~ w..v.

n-,,,.,,.

e,n.. n n.

Exhibit 2

=

= - -. ~ - - -

.%vq and Comment New A-Bomb Studies Alter Radiation Estimates The basis of 15 years of radiation research may be in error; radiation toxicity may be understated Some of the most important data on

),

the efles15 of nu2Icar radiation on hu-6

~

mc.tw may l'c wr0ng, according to new

=

~ 7[wa g %we-%,AhiN%)'

resea:ch bei u done at the Lawrence s

• I"-

i f 4'W l'M" ':

Livermore wearans labo. ; tory in Cali-e forn a and tbe Oak Ridge National Labo.

i

, g,, A.,. j ' +c..,_; M -

• e i

f}~

.' h '.

ratory in Tennessee The new fmdings 7

3

)

I A-are far fiom wcicome, as one consultant o.

hd&

,. /

in this work sais, for all the resisions

.q Q Q{ N "are moving in 15, t rong d;rection"-a direction that w ill w orry the ad vocates of

~ ~

nudear power. Government physicists u s. A, rec.

base recalculated the da'a on the radia-Did it proctuce neutrons or mostly gamma rays?

tion E: Ids created by th2 atomic blasts at gjg,, g, g,,y3 gj,,, gj, gj,,gjmy Hiro hima and N,gasa'i and produced some unapected resu!!s.Their statistics factor of 6 to 10. Since the effects on epidemiologist at the Unisersity of pitts-show that most of the cancer caused by human health remain the ame,one must ' burgh. 't he risk estimates in the fin.d 4

i those bombs came fiem low LET gamma conclude that the gamma ra) s w cre more report of July 19S0 were not as high as cys.* supr.esting that this common type tosic than bad been thought.

Radford argued they should be nor esen of t diation is met. h, zardous than had if this research prr.ves correct-and it as high a<. those in the 1972 report.

Feen assusaed befine.

has survised a few peer cha!!enges al-Neither Radford nor Rowi endorse 6! the The impetus for the revision comes ready-it will necessnaic the r ew nting of document.

4 l

rrimari!y from Liuamore, where physi-man basic documents on the hazards of Rossi' concedes that the Livermore cist3 William Loev.e and Edgar Mendc!- radiation, including the.

f attempt to calculations may do away with the evi.

scan last year us.'d a car.puter to recon-derme such risks published in 1980 by dence for his theory that neutrons were struct the two evio6ns..beir findings the National Academy of Sciences. That responsible for the high cancer incidence are being checked and comp!ctncnte '. by stuJ. the work of the Committee on the in Iliroshima. But he does not expect to 3

j a group at Om Itidge led by George Biological Effects of loni/ing Radiation n!ter his general siew that the h...rds of Kerr. He bepn wo4 k on a similar project f ahe BEIR reporth was fianght with con, radiation are exaggera;cd. Radrord, in ja 1977, Alved it, and then returned to troversy on this very que tion.

contrast, says the new Iliroshima d.aa the ta<k in carnest when 1.oewe's data Although much of the BEIR report sindicate his position and invalidate Ros-I became known. Dean Kaul of Scienee was releat.ed to the press in May 1979, si's. Furthermore, Radford consiJea s the App!ications, Inc., in Chicago also car-the Academy decided to recall and re-BEIR 1980 report obso?cte and evpects i

rLi out some eavly calculations that write it because of dissension among the that the p.ohabilities it gives for ti. risk sparked interess in ll.c issue. Kerr.11 ul, authors. Some of them, led by Co!umbia of dying of cancer after exposure to and Jew Marcun: of Research and De-University biophysicist Harald Rossi, ar.

gamma radiation will be doubled, l.ike-vekipment Awociates in Santa Monica, gued that the paper overstated the can-wise, he thinks the probabib in for con-Cabfornia, have L n funded by the De-cer-causing effects oflow LET radiation.

tracting any form of cancer after irraJia-l fcase Nuclear Agency to esplore the Their arguments Icaned heasily on Japa-tion will be quadrupted.

problem and check some of the old as-nese data and particularly on the thesis The importance of the new rescaich is sumptions which have not yet been reex-that many of the cancers in Hiroshima that it completely changes the scheme of amined.

were produced by high LET neutron radiation doses that people are surpo,ed Ahhough they differ in some of the radiafon.

to has e reccised in Japan. particularly in deteih they st:ess, t!! of thew scientists Using the cht Iliroshima radiation data Hiroshima. Until now, it was thought ayee that the accepted figures for high as esidence, Rossi argued that the BEIR that the Iliroshima blast was unique in 3

LET (neutron) radiation at Iliroshima committee sho dd lower the caneer risk that it produced a large f%I of fast are p.mfy overstated. I'or exmaplc, the estimates pubbshed in an earlier BEIR neutrons a h'ph LET form of radiation.

neutron r.diation at a distance of 1180 report in 1972. Instead, the committec Neutron radiation is considered more meters frorr the epicenter of the blast raised the risk estimates. Rossi cor.siJ-dangerous than low LET radiation. a appears to have been overestimated by a cred this an a!>rmist move and withdrew category that includes + rays, clectrons.

l his support from the document. In the and gamma rays. Its singular paesence in

'Te ieum tm trr s.na wth 1.t.r" tro, end. the Academy felt compclied to,HiroAim was said to make the cancer hnc-enern tr.srcr,.crer to the rhnicar wt4 of weite a report that effectiscly split the risk found there anomalous. Most of the e

the ra). fr. LI r r dwn ta o rcL,mch httle er. era w.. trnes. tm m courw. maa mt.an differer;.c between Ros..s pomt of view radiation peopic encounter is not of th.is i

etunenw m.w. ran. and g.ran.thrh 1.1:1 r:.d. - and that of his chief adversary. the com-kind. The waues from nucleai reactors.

im tw crc,ry more r r.JI a n ireels. and nchoa Na.m or s.eun..nualrrotom.

rnittec chairman, Edward Radford, an for esample, emit gamma reys..Thus, a e

oor.uc5 niv::w.nlmo cor9.cht

• 1938 ^^"

SC3FNCILV"' 212 ??h!^Y3*l

I nmebqoNientists base atuap coiwid.

the number of neutrons thoopht to Itase muse sense foi the Depaitment of 17 pei.

, ereJ linoshinu a special high-tisk e.ne.

been present i, now so small that one gy or the Nuttcar Regulatory Commie and m stud ing the peacciime ha/ards of m at account for their effects by increas-sion to pay foi thi umL. and "the j

3 radi.aion. they h.ne di counted some of ing tne estimate of th,ir potenc>. The electric power people really should be the camer data fiom thet city res ultant kithng pouer of neulions is interested," accoiding to Auton, it is As it I appens, the c;meci mottahty "incrediblef Auton sap. Industrial impoitant that the new t esearch be credi.

data frem liiroshima are the most s alo-safety rules would h.ne to be iesised, tsle. Auten apecs that it would be l.est if able in ti e world. Unhkt the d..ta from reducing esposure limits for neutron ra-the spon or were an independent poup N.casaki. they me abundant enough te diction to or.e tenth of the present limits.

not awociated uith the weapons pro.

rncal a clear relationship between Aes 17er critical jobs, companies uould have gr.nu or the nucle.u industo.

of rdiatim receis ed anJ ill effects. 'I hat m

reNion hi;s is defined by a line.n equa-

[ g. ' e < ' %.

• +

tion: an increme in dose abose the na u.

.W' W r

..f y

?

<r tal backrom d sadiation correlates with

wQ ny / e

• 4' a proporGen.d inen ase in i!! cffe cts. ~1 he

'. [ K nM M'

c. '

c

p. a :n a g ests that any ineicase in

- V8l,. - [

d-

_7 radiation. no matte; how sm:dh direet!>

". NQ.

4 j

,4 increases the rbk of getting cancei. The

,y p

.m. f,.

j,

N y-

- s

, ".,. f,

,. n mortality data from Nagasaki are sketch-

%dQ,.&g g ?, ~

ici, nabg them su ceptible to a uriety

,1.

% *;.. > &, ;. v.

"r

,, [J.

?

. ~. > *., '; x

[s j of imerpietahont The significant point

^

. / Pc s'

~.

is that if the new immb calculations a e

,W r

1. ,*Q.C..W,,,'.

t accur.ite the data from Nage tki and i

i

'T f./

A.4

.i F ,

6 IliroQina car be combincJ and ticated g.

?

f ;W".-

,e as a sing'. coherent pattern of response to hm i 12.I' radiation it is too early to

' $ Qh'M,.J d", [ /, '.3 f

/

us aw say precisely u hat that p ttern will look hke. betause now the dose-must be Hi oshIma fN5 ru.Jculated fer each radiation sictim.

h w omortcfw h eiunsedt/vi>61 But mod of the resenrchei, w ho spcke to Sc ina r s dd the f.ew data w ould p:ob-to mn!o, n times as many people.

Arthm Upton, the former direttor of

, the health phpies the National Cancer Insntute at:J en W

ably increas the rNL est%f es for gam.

.nkn the Rossi prin-espert in raJiobiolon, has folho ed this ma rcJ>ation.

c, lh.

rJ. an advocate of this N. of cipi,:

. c that ne.aly all the contros er g closel> ance hc Ic.u ned of i

the nea bomb data last fd h is en s ie s c hi. s that the aym ' our IF-careers i: 1. odima were produccJ b3 roshima ar,J its m nhty data.as be:n ;

gamena rap. nc: neutrens. 'I hat new 3 important issue, he sap, and shou!d be dntr lion from tt.e main boJ) of scien-will not be v ek or either.

the subject of mme sesemth. vomored by a r.eutral scient fic organWtina such lif' aidcNe. H. sa: s the IM lifik kiton wishes P ntly that someone reps : t miseakulated in emphasizing else w cc funihm thh research, which he as the joint U.bjynese Ra Aten ht.

meht) data so heasil). for death cer.

thinh i-apostant for future health and fects kescarsh l'oundation. If the neu tificae-do not gise a ser) accurate eneigy poJcy. Ths effi; is doing it be-dose estimates are correct, Upton sap.

reaJi:, of the number cf canr. css or cause "no'. od> else w as interested."

"l am net su e enc an substantiak the cven cancer deaths in a con.nnmity. bd-

'!he centiou r.y has been breaing for at Roei thesis." Il n'ay remain impat.,.

I ford thinks it was a mistake to pay so lemt 4 > cars, for that is how b r4 h has for rathebiolon. for there..rc M.

been since a g.n et nrnent consulent fu st enees in the w ay 'h a p! mis nrJ.s. a nuh attemba to Rowii themy abo m deat!:s in ihroshima, for he clai ns the raised miou questions abut the valk!- respon t in the Iai, am to hi; '..e J fon j

thco:) is contraditted by "W nereent" it) of the IWshima data. Accordmg to I.fil radiation Wton avec with I:ad-i of the epidemio!ogical data on record.

Auton, havever, it was,t.rt 5 moi.ths ford that the new 6ta greatly sta nshen lie is p!cced that tF e Iliroshima data ago that hc was appro chcd by liarold the atgemsnt that these h no " safe" rrny now look cursi tent with all the Wycko'f chairman of ; eccial eommit. level of e sposure to radiation, in th..t tee ass r.ned to study t' n question for escry in;temental bit of esposure ic-

{

rest.

"Ihe implications are far reaching fer the Natunal Council on RaJiation Pro.

creases the c h.mtes of injto y.

hechh renc!ali m and n selear pow er in tection and Mcasarements. It i - private One of the curims aspests of this this coautry in.. a c t.J san DasiJ organization the. mllects and punishes rewarsh is the mann r in whish it was Auton, a phpicist in it effice of t.m radi:oion risk info:mation Smee no oth.

pubh Ld. The record ser.c, as a com and dance assessmer.t of the Defe er agency would fund the research, Au-p Ihnc argument lor dect Mfying as Mica Apener. lii> offi;c is fundm; ton say s, he apeed to hase the Defense much as possiNe of what is done at j

the re cuch at Oak I idge that may wn-Dep.utment pick up the tab fm w or k gas er nme r.t lab, fa many of the as-j f.im the rew dose estimates. b he de-bdog donc at O.L Ridge, and thus come sumptions in this case might bas e been sCIibes thC situation, the health ph)ves up with som. answa s for Wychof f. ~i be chaficcyd sooner ha.1 the underl ing 3

community faces a nasr> dder ma,if the funding began about a mont5 ago.

data been as ada' le for se:utin).

nm bomb data are accurat On one "lbis u ak is of marginal interest to

'lhe Rosetta stone et Japan se raJw h:*!. the Man !slactreis my adhcre to us ar.d we reall> can't afio:d to sps ".d tion dosimetry b know n as I MI). w hL h 1 ouiN princip5 which m dntains that s cry wh money studying civil el stands foi tentatis e dose evnn qcs com.

n

") of the canecis ; rodo.ed in Iliro-fccts," Amon says, but it n impoit.mt to rifcd in 19M. 't he hyurcs w er

assea, shi na were caused by int neutrons. I!ut resohe th. uncertainties it might mde Ned by phy sicht JoSn Ansier of 0.A

%I 2.' M AY swl 1

s

~- ~~

.we

-_.e e.a o e.

4..

Ri.tg in a painstaking analysis of mea-clawified because it described ia detail years of research on health effects in surements.made during and af ter the the makeup and radioactise output of the Japan, were nes er described in sletail. In J.4 ee blists, inten iew s with the Little Itoy (lliroshima) and l'at Man 1977, however, the goscenment pub.

bomb.udiers. and a test explosion in the (Nagmakit bombs. Ausier's methods of lished a quasi. technical nanatise b3

Nesade desent. Some of his work was computing the doses, which underlie 15 Ausier (lihihan, Encigy Research and Development Administration.

TID 2hkSU) sising some additional ir. forma.

tion on Ausier's methods.

As questions about these figures arose 1echilology fransfer Reappraised re r

iniheiate i e.tne Naii.mai Councd on Radiation Proicction (NCI:Ps asked Trwsfer of techr. ology from industriahzed countries to developing coun.

Auxier tojustify his estimates with more tria emesged in the 1970% as a highly charged inue in the so. called North, supporting information. After worLing Seuth dialo:.e. E ess-deseloped coentries protested that control of technob

' on this proje/ lor seseral months, Auv ogy by the ir.dastrialized North Lteps them in a state of technologreal ier explained that he cou!J not repro,hi;e depende iee.

all the data because some had been lost.

A tepar'*just issued by the Orranization for Economic Cooperation and lie explained to.%Icnce that when Oak Development (OECD) in Paris questions major assumptions on w hich the Ridge was reorganited in 1972. he was te:bnelogy transfer debate has been conducted. It argues that technology moved fiom one place to another, and I

t.asfer has been mutually benencial for industrialized and foi Jewtoping his old clawihed fi!es were left behind in mtries, or at least some' of them.

his laboratory. Auxier sap that the rec.

11 e report notes that tec hnology tna.fer has helped a group of"industri-ords daision.it Oak Ridge made a miu alizing" des cleping cotntries to p.u ticipate, on stronger terms, in the w orld take in shipping the fdes: the valuable trading sptem. These include llrazil, Mexico. South Krea. Taiwan, lion; data were sen: to the shredder.

I Kang and Singapore.

The NCRP contimted to a A for confir/

The report's main challenge to the notion of technological dependence is mation of the T65D numbers becauw its s.ssearion that " technological monopolies are te nporary," that chang is they had become important in the debate propclied by a " technology cycle." New technolog) introdneed in one on the hazards of radiation and because co'.mtry is transfened o' ter tight control first to other developed countria new data were becomin t available. In and then to lew.deve!vwJ countrics. As licen<ing and sale of the technolo.

1976. the Los Alamos Scientific Lebora-gs spreads, it b.omes itan tardi/ed.

to'y la New Mexico, a derpons desipn Proof that thn process is workian is seen in the rise in imports by centcr. rele.ned an estimate of the radew

.lustiial countria of manufactused goods from deseloping countries.

actise output of the liiroshima bomb for Alorcoser, some in tastriali/ing countrics are themsels es esportinp techno!.

the first time. ~1he 0gures u ere not pub-ogy. saastly in the foim of tornkey plants and equipment.

lished, but gisen in a private letter to C.

1 cedback from tec knofegy trmfer Ao affects industrial countries. The P. Knowles of Rescatch and Desefop.

b met t 6 been most conspic.ms in the decline of traditional industries, ment Awociates, who was trying to help h dothing, footwear, and hght manuf: cturing. that have fsced off-the Defense Nuetear Agency pin doan a+

she com?.tition. Low ofjobs has created a protectionist backlash that the precise espiosisc power of the Fat inclux!cs cuicism of technology transfer. [3ut. says the repast, technology

Man bomb. This is one of the key unecr-transfer has benefited the United States and other OECD cour tries by tainties in the recoid; some say the b!ast ating esport markets for their capital-goods industries duiing a period of equated the power of 12.5 kilotons r

+

o su v growth.

TN 1, and others say it may has e besu as fly focu dng on the industrblizing coentries, the report offers a selectise potent as 15 Li!viens. Sescral people ii siew of the prob!cm3 facin; dueloping co:mtries. It does note in passing the weapons arid biophysics cem nuait -

that for the poorest countries, the cou of importcJ oil, trade deficits, er.d soon obtained wpics of the letter, ir,-

forei;n debt make the outlook bleak. Esen for the industriali/ing countria, cluding Kerr at Oak Ridge and Kauf at the borden of energy costs, deficits, and debt hase " led to pewimism Science App!ications. Using the ne w regarding future financing of desclopment."

data and computer techniques not avaib The report was picpared by the staff of OECD. w hich is essentiall) a club able w hen Ausier did his sescawh. Kau!

of governments of western industrial nations plus Japan. OECD serves as a and Kerr in separate projects cami up data gathering and interi.overrunennd policy. planning organi/ation. It is, with numbers that we e at edds with th -

thenfore, not surprising that the repos t awewes technology transfer m. J)

T65D resul s.

t from :he sc!!crs' point of view, Kerr's labo... tory is the best equipped in bro. d serms, what the report's authors say is occmring is a major and best fundcd for this espensise com-restruaming of the internativn.d industrial system. For the induMrial puter work. Kaul says, and for that sea-countries an " adaptive strattyy" is counsc!cd. Wuh a two-way trade in son it has bcen given the piimar) respon.

indust,ial products now establish:.l. the North can ret..in i: comp;u atis e sibility for reviewing the o!J numbers.

adont.se only by keeping its "innosatory capaci'y" at a high lese!.

Kerr's task is complicated by the fact Prewure to transfer R & D activitio to developing countries will build as tim he is in a sense Ausieri succewor tivir scienti'ie infrastructures strengthen. The report bonows from I.cwis at Oak Ridge and woikCjnst down the Carro!! to obserse that industrial countries must "Leep running to stay in hall from this senior official whose woil the unae place."--JOHN Wat su he has been asked to resiew.

Ausicr, tucanwhile, says that his data wca..,4 7,, w!..,m m.1 w. t n s r.p,,o.u,,n A r,,.J. o,pwanon are the best available, not likely to be rm,ym. 7 t, A rar r onon.e c.vra.on.wr e changett much by the work of lattes da) w:

ro, w:, u om neo sa o cormlet e mt Aus setrNct. s OL 212. 72 si a IwI

-_c o n~ m y W_a-

~-ne

=-

^

~~

f *'l tesisienists. Ilis judgment is widely re-it. liaalier controversim has e L.ught him argues that there is no evidence showing j,pected' As the grand old man in this. to move cautiously in matters as impor. that neutrons were piesent in sign

}

tield. he is in a position to influence tant as this, and he still thinks there quantitics in fliioshim:-

f fanding decisions on new rescaich. Aus-could be sorne weaknesses in the new Loewe, Kerr. Ander, and others in a

ier told Scicner there is no need for an bomb data, this controversy will present their ay;u-independent resiew of the discrepancies This stalemate existed for ses cral ments at a meeting sponsored by the between his data and Kerri, expressing

>eais until the.umrner of 1980 when Radiation Research Society on 31 May in opinion which may base made it Loewe decided to rework the calenta. Minneapolis. Auton calk it "the begin-an dit6etdt to get the present review start-tions. lie stmted the pioject because the ning of an important dialogue," one

~

ed Autoa, the Defense Noelcar Agency old lliroshima data and Rossi's re:ent which he probably will not be able to

. i, off cial who makes the funding deci-warnings about the pote:;y of neutrons attend because the new A6ninistration i

sim. sap that he ha-great respect for woiried peop!c in the lab. Lisermore has reduced the bureaucrw >'s tras ci al-b.

Ansieri wo;k, a respeet based as much scientists are invohed in weapons re.

Iow ance s. liut Auton hoacs th: meeting hE on Au. der's stimding in the commumt) scarch and are frequently exposed to will Icad to a Feneral und independent t

a. en his abstity to "dng out corro'. ora-neutron radiation. They wanted to know resicw of the issues. "If the weapsms more about the dangers. Leewe's inse - folks" make it a strictly intem! project.

I

!~

tr.c J.it a '

Kerr h:4 neser pub'ished :my of his tigation. completed last October. found he says, "I ju<.t hase a concern that work outside the laboratory, he says both the Iliroshima data and Rossi's nobody will belicsc the resu!ts."

because he prefers to be " timid" about principle to be unsubstantiated. Loewe

-Et for M st<sn tr.

i It

?

Science Adviser Post Has Nominee in View i

Tho job, tumed down by several candidates, may now be offered to a man who is not a member of the science establishment lhe chcice of science adsiser to Presi-

" W7-WT ' ' r ' FN Such concerns are abruptly di missed by Key worth's supporters. Al. hough he 9

7

- q$.' '

G s' nt Reym ha b:en nanowed down to s

. sir gie canJM : George A. Ooy)

[/

" lacks obvious credentiah that doesn't mean he will not do a wperb job." s tys kyworth. 41-year-old ph> sicist from one. Agnew scoffs that "he has all th.-

a i

112 Los Ala.aos Scientine Laboratory.

']i.

right credentials-all he doesn't base is Nthough IN: job had no: formal!y been

~

k's W [ M ;;

20 years membership in the club." In a

,..f. fd 'A of:ere I to Kcyworth as of this uriting.

'l r 1 O D' telephone interview with.h icm c. Ag-AdrsJstration officials expcet an an-

/.

ss b new abo said that he thinks much of the

(.h3? - M N-- ;".a unease about Kesworth is simp!> due to neancement by the c'd of May, but cautien that somethi' could still go i awry even at thi'; late -toge of the selec-j; ~_-

V-the fact that he is an outsider "If you

~

i tien proecss.

2<, j.: %,.c. y' get a bunch of chief ens together and you

% ben Ke> worth's name came up as a S#l

'Z,"'

put in a new rooster. they sta t c!act.ing potential candiJare late in April. it drew

~ - T - (..- f f,.-

] and running ato'md." he re.naiks.

As for Keyw >rth's shortage of Unks to i

a mistute of su.;' rise and unease from,

.b!

.i

. the scientific establishment. Agnew sen Y

the wientific establishtrant.1 hc surprise

'h

.F c.-^H d / [,,',

,. 1 ;../ that " defense will be the thrust of this.

stems frorn the fa;r that Keyworth h sittually unknown outsid: his field. And

' Mj }

?..[. c-j Administration. and somebody who has the respect of the peo;d.:in the defense the tmme is relatol to the fact that his 'du'tsUsr ca. sune$se

- -- labs is needed." lie adds: "I'or the past candidacy was being vigorously support- 'r four sems, you have had a geologi :n cd by Ed'.atd Teller, the so-called "fa-charge, and the defense community has ther of the hydugen bon:b," and llaro!d Agnew, president of General Atomics "a very gooi scientist w ho is a Ia: broad-suffered."

.nd forrrerdirector of Los Ahunos.130th er than his background would indicate."

llow did somebody fro... outside the ice w c?! Lnon n for their haw Lish defense Ilis background does not, ho es er, traditional ranti, of a.ndidates for sci-include sersice on the.nud roaod of ence adsiser ce sehted? Kc) worth siev.s.

j

't hew who know Keyworth c'escribe government science comrnittees. llen;c says he was appinached about the job l

hhn as srnart and personabic. !!is re-he has little experience with federal sci-carly in A 3ril, and "it came as a surprise i,

search has been concerned mostly with ence pohcy and has made few links to to me.'"I he post was forru.t!!y offered in j -

nuelcar structm e and low-energy nocicar thc scientifie establishment. "He doesn't March to Arthur llucche, head of re-reactions. and fu. the i.ast 3 y ears he has proviJe any channel bctween the nation.

Search an f des elopment at General Elec.

directed the phpics dhhion at Los Ala-al Scientifiel community and the Wh;tc trie. but he was foned to Im n it dow n f or One seientific coli. arm, Arthur llouse " comp!ains onc veteran of sci-personal icasons. Sescrat other peopl.

Ker. nan of MI1 describes Keyworth as ence and gosernment affans.

were subsequently sounded out about mos.

W emmw m:7 neo 'on conrsht r ivM A AAs Kit.stL \\ ot 2e. 2: u Ay tvxt u w.m.

m_ au.muu,_ -er._._ --. - _.- --

esm-o 19 June 1981

' b s !, g!

?

S ' MNi R de h

w ~~~s w r~c a.] at n tT*u f U.

ze*wx?"

W i

S 2 %:# m. a dbi] b M %.A/d. M r

Volume 212, INo. 4.iOI i,

LETTENs Alternative to Peer Review?: 1. C. Liebman: R. L. Alt Crecry; Af. R. Kult; S. S. Easter. Jr.; H. Atorawet:: 1. 5. Jacobs; H. K. Bhnbamn; Af. W. Berns; R. Roy; Bridge De w..t W. Ha

e......

1335 sotTops.u. Coal Research.

1341 1

i At1TICLus Critical Data for Critical Needs: n. R. I.ide Jr..

1343 5-Meth>Icytc ine in Eukaryotic DNA: Af. Ehrilch and R. Y..H. Wany....

1350 Allocating Petro' tm Products During Oil Supply Disraptions: R. //. #c:dc4 and W. B..

r....

1357 t:EWS AND COMMA 7 New A. Bomb Data Shown to Radiation Expeits..

1304 1

1%6 Ilatch Batters NCI with Strr.us Case.

. But Straus Defends Himself in Boston...

13S7 e

}

Gold Pipettes Make for Tight Lips..

1368 1

A Manhattan Project Postscript..

1369 i

Briefing: Nuclear Policy on IIold Until After French Vote;lluman 1.ife flill s

Arouses More Opposition; World Bank Puts O!T Energy I

Lending Plans; FDA Plans Actioa on Sodium in Foods; It is lilegal to Say That One is Sane.

1372 3

Angiogenesis Inhibitor, Link Many DiseasIs....

1374 nsSIARCH M2WS Alecting Highlights: Nonmarbe Iridium Anomaly Linked to Extinctions; A liot Spot Found, An.ither Discarded; Are Ocean Islands Recycled Ocean Crust?

137G Let Us Now Praise Famous Boojums..

1378 ecox ;tsvisws The Social Context of Soviet Science,iniewed bv H. Baker; The D ing 3

Community. R. Mc C. Adams; Anima! Tool behavior, J. R. I.uncastcr; Photosynthe is, R. Atollin; Books Received.

1379

+

REPonTs The Mount St. Helens Volcanic Eruption of 18 May 19S0: Minimal 1383 Climatic Effect: A. Robock....

4 A

Gru AD CF DIMCTOM*k FMCE*"V MOSTELLER D ALLAN B.409tEY E. MsiP'LA GT IN&OCE EtOV E. Cuk NaNOZ L CNC Hatarq 3" e**. Cramet Pres.cet Prestant4W LD'/AW E. DAVO. Ja LAVfC 4 H".W3 C}l.O et? t S N.%N Q 9 8ATH5*...

' a l.m PhfSiCS (N C1tPMSTM (C)

A37AC?.oM / t03 g T.q ffA,ql Q y Wh P Me'A9 Go'r+at'er Rr.L r W Pa 'y Oe* U.*;.<n U

f.d.;h 4 3Cf;TIO G PSYCHot8AJ (J) 500AL ECONOM C. A%n PouTIC AL SCIENCES m 64t3r09 / AND PhTSOM4Y C# Scif t.C5 O 6.MWr *1 NO M *,

G.cre,e A War James G l*ru.

Heav M f' trami t. em

. Me

• Cram GAnntvet Oe*4 L Hai Ocnns i:

f.'. w.

b EDUCMICN Ci DEN rtsPY IN THA9"ACFUT CF. SC.ENCES (s)

P.500mT:ON. COMou r:*.C. AM COWON CM CN (D t

~ Ana C. Hm uv.s1 <. Hiae Arttmy P $.r+.P4 CeT;e W e ressed Rorc= G G Mad Hvn M. F#raer Ret *1 A WV/

l.tedMe M herVews D I W Y.s O N '*.

ALA3AA OtW$iCN PACIFIC Dtof$lOrt SOUTHW! STERN ANO POCO MOUNT A:N Ce; Jor.n ru,h T. ned De ens hatrce M S=ceney

1. • aE Leven Max P Drforj M Mc%*P'*

Piesdar't Esartwe 5, Wiry Prescari Dwtne D.recte*

Presee*

Fa ecwe C's.

--m m._-ay n,. Frliev. e au,r.~+= la I we..'s ;n Cec,e,mN e, *>y t,t** Am -r4,=._4 A ssoc,icio_n or tse._nvencem set r 5 erre..%,. cm.-

.U,N. W esh.q:on. D C.

S."-4C E *s wtrows..

f A

f 1515 6 s...euset...una

m. ~

,...w. a. :.,..~

, m ne m.

.. m

f. irk 1 d..'is73 ffP" ' + n.g.sn i f d.acrg.fr*. g51 assaesi 543 D rna.,t< enter.at E e.rt'r.' t 91 w *4 Sa's F& r p 5 it

• W < C *W M & o er Mve ** 9 I Tr IV*b PW ***

• t f 4 Dr8N -

J3. eoss% a, e a-rw ' m. ? n mt r'im. a *r *, W per. nrvi s:u ient rs.es on f>=qu' Nn.f* e.m 52 fu so af m49, t wa.*.es 5.1 r31tr rnah c' war t r 5 ci %+W rnance et r

e6.'l'*st' a sfu. F4..- ~.1

• es#. M hid e a *JdNt.* *. Vid kpvefbdp ashvarit r#@.rf P(.*.'J'14Sl*P. S. rad Fyrt W3 to Sc mee, Isis M*. sad u *s A.tr e. NW. W#@ % D C sW Fw= es u

W'**f' * ? 9 ed 4 w 3

1. +.

. wer.pfe*W L *.y en. D nJ grt Se ***Jd Sp'#:8-lied er Cean d,-

~'N 2.LT7:~ ~'.::;L;;;;r3*~~:3 1e nA.c.n.P':U C B :C% m u-v wrar

~:

r u.xn

_ _ _ _ _ _ - _ - _ - _ _ - - - _ - _ - _ - - _ - _ _ _ _ _ - - - - _ _ - - - - - __ _ - - _~.,. m.x.w 1-u. s. m-- - _-- - - ----- -

Nws pnd Comment

~

New A-Bomb Data Shown to Radiation Experts Conference goors are impressed with the revised picture of Hiroshima, but foresee littie chango in risk estimates Mi.mrupolis. Physicist William l.oewe Measurements (NCRP), said. "I would to complete the resear.h quickly and spake at the annuai na eting of the Radia-strongly digree with anyone using this shore up the $100 milfon investment in tion Rescanh Society here on 31 May data to detennine risk coefficients." It is.'apanese data.

and gave the first pun;e presentation of too caily to do that, he said.

Second, Jablon saio, "I think it's po-the uork bc and Edgar Mendd>ohn hae Loewe agreed that it would be wrorg - ing to be absolutely necewary in 'this dove at the Lawrence Lisermoic Na-to draw broad conclusions based on his murky situation that anydosimetry sys-tior.1i sovory. They has e drastically preliminary work, but he did tell the tem that is finally decided upon be rea-reviJed the estinntes of radioactive fail-Minneapolis Tribr%c that he thought the sonaNe in terms of biological influences out imm the Iliro3hima and Nagasaki new data will '. ave a negligiMc impact on that we know about.

. And since the atomie bombs. The most important sin-risk figures. Othen s. such as Warrcn Sin-prob'em is of wider scope than merely no clair, president of the NCRP and an phpics. perhaps it would be adsisable to y'e t'nding they reponed was thc.f r.cutron radiation of any statistical signif-organizer of the meeting, uere stronger consider addityt some biological exper-icance was present at Iliroshima, sug-in their denunciation of Radford. sug-tise to the (NCRP] tek force "

gesting that nearly all the t'emb-related gesting that the r.n Livermore data may Third, Jablon said, "I think that the cancers wm produced by ga nma rays.

even make radiation look less harmful way this whole problem devcleped is If co: rect, this means there aie no good than before.

s cry unfortunate. Most of us, cert ainly I, hun an Jata for judging the toxicity of If the sponsors of,the meeting ucre herrd abc,ut the pioblem.. by woid of neutron radiatioa.

unhappy with the way Loewe's werk mouth. The next thing was to rcecive The audience was receptive. and sev-was presented to the public, other mem-pieces of paper which were n ' for pubh-er.d o!d hands said they found Loewc's bcrs were as unhappy with the way the. cation, quotation, or citatio u..Iam work impressive. No general consensus information had been circulated (or not told the Japanese Ibt is about to hm-c a was reached on whether or not Loewe's circulated) within the community. Per-debate on the subject, and still thcu is i

data should replace the old estimates of haps the most outspoken was Seymour nothing published that one can point :

i 39mic rx.fiatim prep red in 1965 by Jablon, the National Academy of Sci-and rebut or accept or what'scr."

NCRP President.Sinclair responded John Auxier of the Oak Ridge National ences' staff officer for joint U.S.-Jea- _ by sayin.t that there wts aheady one Labo atory.

nese research on late effects of atomic Most of the participants a;rced on one radiation, lie is a veteran observer-biologist on the NCRP task force. and thin;. bowever: they were unhappy with Jablon rose during the general di>cus-that he would consider ;dding more the way the news of the possible revision sion to meke three points. The NCRP_ when an attempt is made to estrape' ate

. health eflects from the bomt data Or e of the physicists who has been at wo:k on the prob!cm the longest, Geor ge Kerr uGiven the unique experience at H.iroshima...

or oak ridae Nationai Laboratory, said t

it really is appalling to think that we that he thought the da'a had not h:en P"Ni'h'd '" " b""'*

• h' Y "' ~ ' ""'

i stand here 36 years later, debating or-stron; enough to stand up to peer re-ders of magnitudo in the doses," Seymour siew. crwo icievant pancis have now

""d* "'" " "'

l Jablon said.

b**"

$"? ""*d '", Revised dose esti-techmca! notes:

_. mates at liiroshima and Nagasaki," by Loewc and Mendelmhn, and 'lImplica-w as reported.and they were annoyed by has know n since 1976 that there mi;ht be tions of new flirosh:ms and Narasak; the interpretation sivca by Unisersity of flaws in the Japanese data, he peinted dose estimates: Can:cr risks and neutron e

Pithburgh epidemiologist Edward Rad-out. "Mmwhile, the EPA is busy set-RBE/" by Tore Stinume and R Lowry ford, wha has said that it may be neces-ting occupation:d radiation] standard <t Dobson.)

sr.ry to douNe or quadruple th: risk other people interested in standards hau Radford, who is not a member of the figures for retiing cancer at;er exposure been making noises. It really is urgent Radiation Research Society, skipped the l

to radiation (Scirm e. 22 May. pap 909).

that we get on with this job.

. Gisen meeting. Ile exprewed dNappointment.

S;saker after speaker cchoed the theme the unique experience at Iliroshima and however, at the attitude that "we can't sounded early in the meetiny. that not Nagasat - and the tens of millions of say anything until we has e everything in enough woik has been done to permit a do!!ars which have been spent trying to hand," as he descri6cd it. According to j

con:ladon such as the one Radford accumulate the human b:olorical data. it Radford, that attitude an be use I to 1

reached. !!arold WyckofT, chairm.m of a really is appalling to think that we stand - dclay reaching an-y cenclusion: "I t 's

. task force created in 1976 espressly to here,36 years later, debat:ng orders if what the tobacco industry did for years review this prohkm for the National magnitude in the doses." lie p4sded with the epidemiotopic.d evidence relat-Coun.il on Radiation Protection and with federal officials present to give aid ing cancer to smoking. They just said.

scar.Sct. YOL 212.19 mNE IW ma w.4omo oswinnso o c< rper i9iAAAs

r

~

which, he claims, is to minimite the paper, Rossi said he considtred it just l

% dh tpat Lo.1 study wasn't peifect, so "an intesestmg exercisc," no more. Ile dangen s of iadiation.

.w e'll iynon it.' '

liarald Rowi is a Col mibia l'nis e rsity believes that if th: Livermore data are l

'Ibe net etTeet of the new tescarch,

!bdf0:d insis% is not hard to summa-biophysicist who cha!!cnged Radford*>

co icet, they will make it impmsible to i

rize: the radiatin J tra fo. the tuo cities views as alumist aben both were serv-say anything con;futhe about neutrons of Hiroshin'a and Nagasaki are now hLe-ing on the liEIR committee. f Radford in Ifiroshima.

ly to come cut looking vez) similar.

was the chairman.) Rowi:ngucd that the An important eascat applies to all of i

"You can state that as a general prinei.

hazards of gammi radiation were esag-the recent wo L an udiation in Japan: it i

pie. sap Radford, "and I do state it.

gerated, and he cited the Japanese bomb does not indude corre,tions for changes That bei: the case, they confim the data to support his case. As p.ot of thk in the shielding prosiled by builihngs or thesis, Rowi put forward tbe idea that by body tissue. According to Jew Mar-1, that ; reduced the c;mcers, and that the many of the fatal cancers at liiroshima cum, a contractor for Oil Ridge for a i

fact that i. was primanif> ganwa rgs resiew of th: data, sip 6frant resisions b

had been caused by neutrons not ga n-n;" tron. ter all practima! purposs, con-tri' ated so htthe that they're not im vr-ma rap. Neunon radratior is found rare-of the Lisermore dose en, es may be I

l tent KaJferd beliacs that the i nerraore 1

i d.ca stre dhen his argument that a Im-According to Jess Marcum, significant car na tin eshotd modelis the coirect one for des.ribin;; the carcinogenic e!Teets of revisions Of the Livermore dose estimmes espesure tolow lesels of radiation. And may be necessary before one can reach if :his is concet. he says, the risk esti.

me published b: th: National Acade-a COf duSIOn about toxicity, l

say of S;iences in its 19W tepart on the IM'opeal Efn ets of lanizinc. Radiation I

1 risks foi contracting fatal cancer from ly in nature, and as a practical matter it is necer,ary htoic en: an teach a conch:.}

(BElm should be restated. He thinks t' -

t taJetion t he,uid be douh'cd. lie would of concern onf> to peep?c esposed to sion on toxicity. Mateum says he has fa the rM at 250 to 5nd excess deaths nuclear w eapons and the innards of oper-spent about !.nanth researching shield-per r:3 o' inevsed radiation per I mi!-

ating rmdear plants. Ro wi's work ing by buitdim s and has Geosered that the estiro Jes of gamma.'oses in t nny lien I 2.v le, not 100 to 250 deaths as he prompted the NCRi' to send out a spe-cases wii! La c tobeloweicJ In ti. mea cial advisoiy to weapons laboratories s

sy UEIR and other documents have warning them that their wfet> standards ofintccest, im0 to IW naters frxn the fned it. Radford wwld also hke to see epicenter of the Mast, Marcum c dsu-th: :r.s st.aed m terms of cancer inei-might be inadequate became neutrons lates thtt indoor gamma ra) do.;s wi:1 t

dence ~t mortality, so as to recognize aight be more dancerros than had been have to be reduced by a littb rwn than t'

real injary is done by c. meets wlach thoa;;ht. That was 3 y ears aco.

do not nc;cssarity kill Inch: ding these Loewe and Mend sohn were sucpt 60 ;<reent. ~l he net elbt. he 1.haes.

l figures, Radford say s, would ma'.c it into this abate in 1979 becau e they vill be to make gamm i dc^, fo uuvid-a i

uorked at thermore, a weapons lah, ich in Iliroshima about the sann as in and were concerned about the NCRp the old cuimates produccJ at Oa, g;Jpc n:cewty to further rake the main rhk t

Loe a e did rot discuss Radforf s inter-adsisory. Lheimore did not change its in 1%5 v bile the Lgasaki do e w di b.

l ce;3 icnt used in the UEIR m..t.

louer th n the 19M f>gures.

l safety standards, but it did finance some In addnion, Geo:g Keir of Oak Ridge pre'a %n et the meeting execpt to say computer work by Locue and Mende!-

that he could not underriund how such siews could be suppm tt d. Lo.:w e said sohn, who attacked the e',idence for h reca!,u ating the shielJmg eff ct of I e dd nat ce how one coald draw a Rossi's thesis. Their ea!colations, noa body tis <:e for cerVn " tar et" organs e

su;h.n th: breat,15,

'. coton, and sa I

sinW hoe tt,rmgh the old or new mJi-made pubhe, do not derwn trate Ib at stian efTeet J da. Indeed. tw u scientists neutrons, re safe. They simp!> sho,v that ca. Maroim c;p u ts sect. ! Far.d,F.cir from Li crmore who have been working neu rons v ere so scarce i the Japanese is in Europe) tha Nasts that one cannot measure their ef-fmal aJjustraent may be to produce no i

i I

in conjunetim with Loew e. Tore Straume fccts with accuracy. At the same time,change in the 1;ukemia risk 1;nta s for and R. Lowry Dobson, presen'ed a paper the Livermore work sigaif,cantly in-the two citics but to increase slighdy t:,.

suggestinp hat the new bomb data may uunn the estimate of gamom radiation rist for breast cancer, bringing the lattei low er the rhk cstiraates for low doses of in Iliroshima and slightly dn n am the into agreem nt with U.S. medicat data l

pamma radhtion. 'lhe), too. w ere skep-on breast canser caused by tray s. If 1:c;l of all : at Redford had said.

gamma radi:. tion in Nagnati.

So nmny variabl:s have been cited in Using this data, Dobson arJ Straume true, this is an "extraorJin..t > mnh thh cer.'roversy that it trmy be worth-have made preliminary nev. e.:inmies of sion," Marcum says because it uill ps e whi'e ur! uning just whbh data belong the tosi;ity of gamma and nectron radia credbihty to the research done by j

to wN:n. hadfa !, fh3t of tdh has done tica. Thcir paper con &d s. among oth.

Loewe, Marcuin, anl Kerr, as w cll as to to new re carch on this iwue. He is an eri

s. that if ene uu, the tot 2 conc er the Japanese epidemioiegka! d da.

j One ef the few things that i, clear in at!

cpider.hdejst with stron; opinions on d ata as a guide, low doses of gamma the subject, and he ha<. seized upon todiation i

>L kss harmful than befose.

tha is that I.ivermorcs rescurch h.s (Other statistical guides produce ddfer-irrcscisibly torp!cd the status quo. It Lee've's work as fresh esidenN to sup-ent results.) They aho sugge,t that it aho scems clear th;.t the fed.ral govesn-I part his siew that many documents un-may still be pos.iYe to Mame the smoi!

ment would be w ell adsisel to hnmee derstide the hazards of low-level rah-tion. Ldford aho says that in defendin.

nun,ber of ne itions i, Hirod hn i for the work nc. css.u) to bring a new esti.

this outlook he is uorLing against the many of the cancer Ahties. Asted mate of todiation daimetry into fa us as profes ional bias of heafth phy sicists, about thh part of the Str,aime-Dobson quickly as possiHe -Et ior Masu st i s

lW 19 JL NE 1%I

(

l

• • * -~m e yy - ~

x

"#~*

-Nr_-w-*ggp se_8_4_3'.; + {yr.Ptvr' wi'mp m=;- - - - - - _ -

a UNITED STATES 0F AMERICA NUCLEAR REGULATORY COMMISSION BEFORE THE ATOMIC SAFETY AND LICENSING BOARD i

In'the Matter of

)

)

CLEVELAND ELECTRIC ILLUMINATING

)

Docket Nos. 50-440 COMPANY, et al.

)

50-441

)

(Perry Nuclear Power Plant,

)

Units 1 and 2)

)

CERTIFICATE OF SERVICE I hereby certify that copies of NRC STAFF COMMENTS ON CONTENTI0ils PROPOSED AT SPECIAL PREHEARIllG CONFEREiiCE in the above-captioned proceeding have.

been served on tne following by deposit in the United States mail, first class or, as indicated by an asterisk, through deposit in the Nuclear Regulatory Commission's internal mail systen, this 6th day of July,1981.

• Peter B. Bloch, Chairman Donald T. Ezzone, Esq.

Atomic Safety and Licensing Board Assistant Prosecuting Attorney U.S. Nuclear Regulatory Commission 105 Main Street Washington, D. C.

20555 f.ake County Administration Center Painesville, Ohio 44077

• 0r. Jerry R. Kline Atomic Safety and Licensing Board Tod J. Kenney U.S. Nuclear Regulatory Commission 228 South College Apt. A j

Washington, D. C.

20555 Bowling Green, Ohio 43402

• Mr. Frederick J. Shon Daniel D. Wilt Atomic Safety and Licensing Board Wegman, Hesiler & Vanderberg

'~

U.S. Nuclear Regulatory Commission 7301 Chippewa Road, Suite 102 Washington, D. C.

20555 Brecksville, Ohio 44141 Jay Silberg, Esq.

Jeff Alexander Shaw, Pittman, Potts, Trowbridge & Madden 920 Wilmington Ave.

1800 M Street, N.W.

Dayton, Ohio 45420 Washington, D. C.

20036 Terry lodge, Esq.

Attorney for Intervenor 915 Spitzer Building Toledo, OH 43604 I

---==vnwm-

E o

_. 2 _

Atomic Safety and Licensing Board Panel U.S. Iluclear Regulatory Coimiission.

Washington, D. C.

20555 Atomic Safety and Licensing Appeal Board Panel U.S. fluclear Regulatory Conmission

'Jashington,,D. C.

20555 Docketing and Service Section Office of the Secretary U.S. iluclear Regulatory Conmission Washington, D. C.

20555

1. C Charles A. Barth-Counsel for tiRC Staff

.