ML20099J481

Applicant Exhibit A-145,consiting of Aug 1982 Rept, HEALTH- Marc: Health Effects Module in Methodology for Assessing Radiological Consequences of Accidental Releases

ML20099J481
Person / Time
Site:	Limerick
Issue date:	05/22/1984
From:	Hemming C AFFILIATION NOT ASSIGNED
To:
References
NRPB-M78, OL-A-145, NUDOCS 8411290075
Download: ML20099J481 (27)

Text

__ _ _ _ _ _ ___ _ _

n I t%c; i

1

~

,. l s:

,R:5-1'-

EG El -t\R C -e . e 1. - er e' , m c; e

. n ne wv. . r. - u,ess :z :e 71110.241 3. '. n 5 8 ' ' .e"CeS s I d ', 1. 7 9 _ r 13d9 1

I

= k -

l l

l l

D

'4l [ ;'90( > 7 'i R 4 0 5 l' '

Uhh ADory o c,o n o 3 %;

PDU

_ _ -aa A.- _ _.a .. - a

}

I l

l .

d

(

\

\

h x

s

.\

\\ s\yx gx' x

\x ', p,4r ,'.

l l

g , s's\

r .

M' s;. g 1

s

• o,s' t>

\ / \

h, ,

., \

%#wp 'es. .s t (3.y'yrs,,\ 't, s ~4,

... .- w S 't,

/ \

NRPB-M78 HEALTH-MARC: the health effects module in the methodology for assessing the radiological consequences of accidental releases C R Hemming ABSTRACT A methodology for the assessment of the consequences of accidental releases of radionuclides from nuclear facilities has been developed. The methodology m consists of a suite of computer programs which predict the transfer of activity from the point of release to the atmosphere through to the population. The suite of programs is entitled MARC -

_ Methodology for _ Assessing _ Radiological C,onsequences. This report describes the health effects models currently incorporated into the module HEALTH-MARC. Models are included to estimate the early and late somatic effects in an exposed population and hereditary effects in their descendants.

The models in the MARC procedure for accident assessment are under continuing review. This memorandus records the models currently included in HEALTH-MARC; additional models and improved procedures will be incorporated, as appropriate, in the future National Radiological Protection Board Chilton Didcot i Oxon August 1982 l

i l

i

- - . . . . . , .,, , . - , . . . - . , . , , . . - , - . . _ . ~ -,- . _ - ,

CONTENTS Page No.

f}.

y) 1.- INTRODUCTION 1

2. EARLY AND CONTINUING SOMATIC EFFECTS 1 2.1. Death due to irradiation of the bone marrow 2 2.2 Death due to irradiation of the lung 4

=

2.3 Death due to irradiation of the GI tract 5 1 2.4 Morbidity due to irradiation of the lung 5

. 2.5 Prodrosal vomiting 5 2.6 Total effects 6

3. LATE SOMATIC EFFECTS 6

' 3.1 Linear model 7 3.2 Non-linear models "

8 3.3 Total numbers of cancers 9 4 HEREDITARY EFFECTS 10 4.1 Linear andel 10 4.2 Non-linear models 11

- 4.3 Total numbers of effects 12

5. SQ9fARY 12

6. REFERENCES 12 TABLES (see list below) 14 FIGURES (see list below) 16

' TABLES

1. Risk coefficients of the increased incidence of cancer for use in 14 the linear model

2. Modifying factors to be applied to the risk coefficients for 15 the incidence of cancer for doses delivered in various periods following the intake of activity ,

3. Modifying factors to be applied to the risk coeffidients for the 15 incidence of hereditary effects in all future generations for doses

• delivered in various periods following the intake of activity

[

FIGURES f

( 1. Scheastic diagram of the, major components of MARC l -

2. Examples of dose-nortality relationships for bone marrow irradiation j used in other studies

3. Dose-nortality relationship for bone marrow irradiation ,

4(a) Pattern of accumuistion of lung dose for dogs exposed to various inhaled insoluble radioactive aerosols 4(b) Dose-mortality relationships for lung irradiation l

l 5. Dose eartality relationship for GI tract irradiation

6. Dose morbidity relationship for lung irradiation

7. Dose-norbidity relationship for prodrossi vomiting

8. Esamples of the models for estinating the increased incidence of cancer O 9. Incidence of cancers following ar. increment of radiation

et As from 1 April 1978 3R25 adopted the Intez:iational System 'of thits (SI).

The relationship between the new SI units dich are used in this report and the previous units are shown in the table below.

named t U*"#8A*" #8"t**

bantity a

""A '

Exposure -

CV l rentgen(R) i c V ~ 3876 R Ab star (0)7 J V' rad (rad) 1 or = 100 rad

• g g simrt(Sv) J V' rem (rea) 1 Sv - 100 res Activity becquerel (3g) s curie (ci) 1 sq ~ 2 7 x 10-l' c1 G

~

I I

l .

l l

i 9

s t

1. IlfrR000CTION i An integral- part of any assessment of the risk presented by nuclear

( installations is the evaluation of the radiological consequences of potential l accidental releases of radioactivity identified by safety analyses. Studies of

? the radiological consequences of accidental releases of radioactivity also provide as important input into the develoyeest of sittag and design criteria and pre planning of emergency arrangements. In order to undertake assessments of the (

radiological conseq===*== of accidental releases of radioactivity to the stasephere, a series of interlinked models designated MARC _ Methodology for gasessing _ Radiological Consequences -

has been developed. The overall methodology and detailed descriptions of some of the various modules asking up ,

the methodology have recently been published (

~

}

i- .-

. h methodology is shown schenstically in Figure 1. This report describes the health effects models that have been incorporated into the module, EEALTE-MARC.

The health effects module, EEALTE-MARC, evaluates the incidence of each of l

-the enjor types of health effect from the distribution of dose in the esposed l

i population, af ter taking due account of the application of protective actions.

The module is intended to be flexible. Options have been provided to change the i f

values of parameters in the undels from those used by default and, in sous cases, I a choice of andels is offered.

l ,

The pathways of esposure considered in MARC are as followet (a) external 3, y dose from the cloudi l (b) esternal y dose from the deposited activityi l

(c) internal dose from activity inhaled during the passage of the cloud; (d) internal dose from the inhalation of resuspended activity; j l (e) internal dose from the consumption of foodstuffs contaminated by activity [

. deposited from the cloud. l The estiastion of the distribution of doses in the esposed population via these exposure pathways is described in DOSI-MARC , and PROT-MARC describes the protective actions that any be invoked to limit the exposure.

f Of the assy deleterious effects which any result from esposure to ionising I radiation following an assidental release of radioactivity, it is sufficient, for  !

.:* the purposes of risk analysis, te limit consideration to three enjor categories of biological effects. These are early and continuing seastic danese and late seastic damage in the esposed population, and hereditary effects in their j descendants. Each category is coseidered in turn. l

2. EARLT AND COIrrtNUING SOMATIC EFyECTS

[ [

( Early and continuing seastic effects involve mortality and sorbidity. hy  !

are usually associated with large acute doses of radiation and of ten occur within days or weeks of esposure, although sometimes a year or so any elapse. In

• general, there is a threshold dose below which any significant clinical effect is j

^

malikely. For accidental releases from auclear installations, the irradiation of l

l 1

i I

L the bone marrow, GI tract and lung, would be the major causes of early death and each is considered in HEALTH-MARC.

Non-lethal radiation effects may also occur soon af ter exposure. These may require medical attention and some any result in a significant reduction in the quality of life and life shortening. They include lung fibrosis, prodronal vomiting, fertility impairment and pre-natal damage, skin damage, cataracts and hypothyroidism. The most serious effect for postulated releases of activity from reactor accidents is respiratory impairment from irradiation of the lung. This may have a mch lower initial impact than early mortality, but any be of considerable importance in the longer ters, because of the continuing health care that any be necessary and the existance of a large localised group of people suffering from a similar affliction. The incidence of lung morbidity is evaluated in HEALTH-MARC together with that of prodrosal vomiting.

Dose eartality and dose-morbidity relationships for the effects considered are discussed in more detail in the following sections. In using these relacion-ships, it is essential that the doses to which the relationships are applied are compatible with the doses for which the dose effect relationships have been derived. Account must also be taken of the relative biological effectiveness (RRE) of the radiation. Appropriate values of R3E for the incidence of early effects are discussed in more detail by Kelly et al( and Smith . Values of 1 and (0, respectively, are adopted for 8/y .and a radiation as default values in HEALTH-MARC, but other values say be substituted.

2.1 Death due to irradiation of the bone marrow

The human data on the incidence of death within 60 days of the exposure of the whole or a subetantial part of the body to penetrating radiation have recently been reviewed ( '

. The human data are sparse and dose mortality relacionships derived from ant =al experiments cannot be used directly to evaluate

, such relationships for aan owing to marked species variation. The evaluation of l

a dose mortality relationship for aan must therefore be associated with consider-

[

able uncertainty and reflect a significant measure of personal judgement.

The dose mortality relationships used in the US Reactor Safety Study (US-RSS) 10) and in a comparable Cernan study (IU are shown in Figure 2. Two 1

i curves are shown in the case of the American study; the first assumes no treatment following exposure and the second assumes the provision of simple i '

supportive medical treatment (eg, antibiotics, blood and fluid transfusions).

The latter dose-mortality relationship was adopted in the US-RSS as the most appropriate in the circumstances. The provision of simple supportive medical treatment was also assumed in the formistion of the dose mortality relationship in the Geraen study. The evidence in favour of simple supportive medical treatments enhancing the survival probability is, however, limited: the quantitative data energe from a few animal experiments but are supported by some I ,

clinical experience gained in the treatment of leukaemia and other malignancies.

L

.s The extent to which simple supportive treatment will influence the survival at f' varying levels of dose (in particular for the irradiation of a heterogeneous

)

\ - population, including the 'old, young and sick whose sensitivity any be greater is a matter of continuing debate and awaits further resolution. h availability of such treatamat is an additional consideration in the choice of the most appropriate dose-mortality relationship for large ac:idental releases where the' number of people potentially exposed to doses in the lethal range any be large.

h default value for the LD50 for bonewrrew irradiation used in EEALTR-It lies midway between the MAAC is 4 Gy. This value any be changeil on ini;ut.

likely bounds of uncertainty in the LD50, which reviews of the data (indicate 'N to be between about 3 and 5 Gyl furthermore it is similar to the dose-effect relationship, assuming no supportive undical treatment, as proposed in the US-ass (10),

' N slope of the dose-nortality curve is also uncertain and, again, the value chosen reflects a significant usasurement of judgensatt the similarity in the slope of the dose-mortality relatiosahips observed in a wide raege of asiaal experiosats is helpful in this respect but, again, there are reservations as to the applicability of such data to the exposure of a heterogeneous huana population. N shape of the curve has been approsiasted in MARC by a esiti-step linear function, which is illustrated in Figure 3, for an LD10 *I ' Of '

One final point warrasting further consideration is how protraction of esposure influences the probability of early death. N data and dose mortality relationships shown in Figures 2 and 3 are for brief espesure (up to a few hours). In general, where the espesure is protracted (greater than a few hours or days) a higher level of dose will be required to lead to the same probability

, of death. N enchantaes by which protraction of espesure influences survival are well understood but little effort has been directed towards the develoyeest of a, generally applicable quantitative framework that can be used to determine the influence of protracties for the wide variety of patterns of exposure that l

might be encountered is accidents. A relatively staple, robust and conservative l procedure was foreslated for use in the US Beactor Safety Study (10) and has been adopted without cheese in a moaber of similar studies ( 1, 2) h dose used in I' conjunction with the dose-mortality relationship is as follwe, l'

' 1 D Dose accumulated the first 7 days inf'

• T

• keseaccumslatedfroak day 8 to 30 1

1 This assumes conservatively that exposure actuaalated in the first 7 days is equally effective as a single brief exposure, while that accumulated from day 8 to 30 is half as effectives sereover, esposure accumslated beyond 30 days is regarded as ineffective in this contest. hee is no strong radiobiological l

justification for this relationships it merely represents an,empiri, cal judgement.

The foreslation was arrived at with knowledge of the likely veristion of dose with ties following reactor accident releases and recognition that doses

-3.

I

accumulated at dose rates of several tens of nGy d*l would not significantly influence the survival probability. In general, the dose races af ter 30 days to those individuals exposed in the lethal range (and uno had not already accum-lated sufficient dose to cause death) would be, at most, of this same order and typically be very m ch less. In practice, it is more likely that the exposure of such individuals, at least from external exposure, would be terminated by evacuation within a short time of the release.' In such circumstances the question of the precision of the above formlation for the effect of protraction becomes somewhat academic. This formulation has been adopted in HEALTH-MARC, -

while recognising its simplicity and potential limitations.

2.2 Death due to irradiation of the lung There are no human data on which to base estimates of dose mortality relati:nships for varying patterns of irradiation of the human lung. Appropriate animal data must therefore b"e used. The most relevant animal data, obtained from experiments with dogs , have been reviewed ( ' and a series of dose-mortality relationships derived for various temporal patterns of dose accumulation. The dose-mortality relationship is very dependent on the pattern of dose accumulation; the more rapid the rate of accumulation, the smaller the dose required to produce a given risk of death. The pattern of accumulation of dose and the corresponding nortality data are shown in Figure 4. The[ data for yttrium-90 and the combined, data for strontium-90 and cerium-144 probably encompass the upper and lower extremes of possible dose-effect relationships for accidental releases of radioactivity that might be encountered from nuclear installations. Dose-effect relationships based on these data are also shown in Figure 4 The form of these relationships shown in the figure is the model adopted in HEALTH-MARC.

In any application of HEALTH-MARC, it is important to ensure that the dose-effect relationship adopted is compatible with the pattern of dose accumlation for the release being considered. Iu the US-RSS(10), , dose effect relationship between those for yttrium-90 and yttrium-91 was chosen as the most appropriate for the releases considered in that study. For those accidental releases of radioactivity from reactors, where death due to irradiation of the lung is ,

important compared with that due to irradiation of the bone marrow, it is likely that the pattern of dose accumulation would be more similar to that of yttrium-

91. Consequently, the dose-effect re'Istionship appropriate to this pattern of dose accumulation has been adopted as the def ault option in HEALTH-MARC, although this can, and moreover should, be changed, depending en the release being considered.

The above dose mortality relationships were derived using the dose ,

accumulated to 365 days. The dose used in conjunction with these dose-sortality relationships in HEALTH-MARC is therefore taken to be that accumulated in the first year.

l 2.3 Death due to irradiation of the GI tract  !

hre are few human data on the incidence of death due to irradiation of the GI tract on which to base a dose-effect relationship. Reviews '

of the f

relevant maimal data, again obtained from experiments with dogs, have proposed , ,

I the dose-effect relationship shown in Figure 5. It is a simple linear function and has been adopted in this form in EEALTR-ttARC. h slope and threshold .of this relationship have also been adopted as the default values in HEALTE-tfARC, t

1 although alternative values any be substituted.

4 Death from irradiation of the GI tract would - only be of importance in .j j

" comparison with death from irradiation of h bene-earrow for releases of radioactivity which result' in a large internal dose to the GI tract. In this i

. . case, because of the moraal clearance processes that take place in the gut, the t majority of the dose would be delivered within the first 7 days. h dose used in the dose-mertality relationship for GI tract irradiation is therefore that [

! accumslated withia 7 days.

2.4 Marm ity due to irradiarias of the luna i

A further biological effect which needs to be considered following lung irradiation is the insidence of lung damage, particularly fibrosis, which any not i i be fatal but which any have serious consequences for the individual. Reviews of f i

the esperimental data on the incidenee of radiation-induced fibrosis have been I carried out

' and dese-ef fest relationships proposed. The medel used in Cs} , eh is I - EEALTE-MARC is hased on the eimple linear function of Wally et al shoua La Figure 4. N defanit values for the slope and threshold dose are also these proposed in that repo rt  ; it is recognised, however, that this f

a relationship is based on limited data, asch of it from asiaal expertannes for i which there is a lack of questitative information on the degree of fibrosis and l

hesee the degree of health impairnest.

espeeed to low-LIT radiation have indicated that Studies in ==i==1= ,

protraction of the dose reduces the degree of lung fibrosis. h effect of this protraction can be accounted for"by use of a dose equal to the dose accumulated

{ in the lues in the first 7 daye plus one quarter of that accummiated between I

7 days and 1 year . This procedure is adopted in EEALTE-ftARC. For a irradiation, the asiaal data suggest that protraction of the' dose does not reduce the degree of fibrosis (I , and the dose accumslated in the lung in the first 5 ye'ers is used with the dose-estbidity relationship in EEALTI-ttARC.

2.5 Prodroaal vomitina i I

Following sufficiently large whole-body espesure, prodroaal vomiting would occur sad would be the cause of temporary discoefort. It is unlikely to recur or i be h source of pernement injury, but the number of people affected would, in general, exceed the mamber that any esperience other early effects, including early death and lues fibrosia. N data on the incidence of vomiting following f' espesure, asseg huesa actients undergoing treatment for cancer, have been

-S-r

~-,m ~wn, ,m-ne.n-w-,am. w,,, _w., m -

, W

reviewed ( and a dose-response relationship, based on the incidence within 48 hours5.555556e-4 days <br />0.0133 hours <br />7.936508e-5 weeks <br />1.8264e-5 months <br />, has been proposed. The model used in HEALTH-MARC approximates this relationship by a two-s tep linear function, which is shown in Figure 7. The default parameter values of the function are those shown, but alternative , values may be specified.

Since the above dose-response relationship was based on the effects occur-ring within 2 days, it was assumed conservatively in the US-RSS that the dose used in this relationship should be that accumulated within 2 days. This procedure has also been adopted in HEALTH-MARC. ,

2.6 Total effects If each cause of early death was considered separately, the numbers of deaths could be overestimated, since the total probability of early death cannot exceed unity. Assuming that there are no synergistic eff ects between sub-lethal doses to several organs, the total probability of early death, Pg. is given by P

g =P + (1 - P g)P 2 + ( 1- P g)(1 - P )P3 2

where P.

g P

2 and P 3 are the independent probabilities of death caused by irradiation of the bone marrow, lung and GI tract, respectively.

Stailarly, the number of cases of lung morbidity will be influenced by the incidence of death. If f isg 'the independent probability of lung morbidity, the actual probability f is given by . .

t ft- (1 - P ,) ft -

Prodroaal vomiting, on the other hand, is likely to occur before the death of q

those sufficiently exposed; the number of cases of prodronal vomiting has therefore been assumed to be unaffected by the incidence of sortality.

3. LATE SOMATIC EF7ECTS The most important late somatic effect to be considered following an accidental release of radioactivity is the increased incidence of fatal and non-f atal cancer, and both are included in BEALTH-MARC. Fatal cancer is used to denote those cancers for which the cure rate is low, whereas non-f atal cancer is used to indicate those cancers for which the cure rate is high, but for which there any be physical or peychological reasons for the quality of life being reduced.

Various reviews have been ande of the increased incidence of cancer in irradiated populations ( ' ' '

'. Much of the human data derives from exposures at organ doses of between about 1 - 10 Gy. It has often conservatively been assumed that the risk at lower doses may be estimated by extrapolating the observed incidence at high doses linearly down to zero. The results of experiments with animals would, however, indicate that a dose-response relationship with an additional dose-squared term any be more realistic for low-LET radiation. Both linear and linear quadratic models are incorporated into l

HIALTR-MARC, and examples of these models and of the pure quadratic model (an extreme- of the linear quadratic) are shown in Figure 8. Values of the coefficients used in the different .nodels for the types of cancer considered are discussed in more detail in the following sections.

As with early eff ects, the relative biological effectiveness of the radia-tion anst also be taken into account. The default values used in EEALTH-MARC are i

-1 and 20 for $/y and a radiation, respectively, as proposed by Clarke and

$aith( , although alternative values any be specified.

3.1 Lineer model i b . A review of estiastee of the increased incidence of cancer in irradiated populations has been ande by Clarke and Smith ( and risk coefficients proposed

• f or the risk of excess cancers following irradiation of the whole body and important body orgass. These risk coefficients are shown in Table 1. The values are those used by default in EEALT54tARC, but alternative values any be substituted. Following irradiation the consequent health effects will appear over a prolonged period. The ridk coefficients in Table 1 are applicable to a

' cohort which will live long enough for the total risk to be espressed. For I calculating realistic expectatione of health effects in irradiated persistions, Clarke and Smith ( ' have proposed a log-normal distribution of the incidence of ,

cancer with time af ter irradiation, to allow for the observed delay between dose

^

and effect. These distributions of the incidence of cancer with tias after irradiation are shown in Figure 9. For the incidence of leuksesta, the proposed tian distribution has a median cias of appearance of 12.5 years sad a standard I deviation of 0.8; for the incidence of solid tumours it has a median cias of appearance of 25 years and a standard deviation of 0.4 In EEALTE-MARC there is the option to use the risk coefficient of Table 1 unmodified, or to take account of this proposed time distribution as described below.

l Secause of the time distribution of the incidence of cancer, the ties inte- <

! grated risk coefficients, r,, for the incidence of cancer in an exposed population with an age distribution typical of that of the UK, will be less than I these gives in Table 1, r, by a modifying f actor, M, is, t

[

t r, = Mr ,

f

,. [N(A) [ (F .g(A) g - P g(A)) [i-A 9 (t) dc DA *

....(1) where N = *

[N(A)

A i

where 5(A) is the number of individuals of age A in the population, Fg (A) is the probability of individuals of age A surviving to age i in the Absence *of the dese, k

i 7

i

1 9(t) is the normalised time distribution of the incidence of cancer, as shown in Figure 9.

This modifying factor will be applicable to the external y-dose from the passing cloud and the external y-dose from deposited activity, since in these cases it can be assumed that the age distribution of the exposed population is that of the population as a whole. In the case of external radiation from deposited activity, which may continue over extended periods, the modifying factor is strictly only correct if the age distribution of the exposed population group is assumed to remain essentially constant with time, although the individuals asking up the population group will be continually changing due to births, deaths and movements. The number of cancers estimated in this way will include those appearing in the initially exposed population as well as those subsequently exposed.

In the case of internal irradiation from inhaled and ingested activity, however, further modifying factors are required. Some nuclides taken into the body any have an extremely long residence time; in such cases the dose to some organs any be delivered over tens of years following intake, a period which is comparable with the appearance of late effects and with life expectancy. A dose delivered at spas eine T years after intake will be delivered to a population group that has aged by T years, and the appropriate modifying factor will be. .

[ (Pg ,g(AW - gP (AW0 f

[N(A)i>A+T 9 (t)dt MT) = A

[N(A)

A A staplified procedure has been adopted in HEALTH-MARC for the evaluation of this expression. The dose accumulated within the first year af ter intake has been assumed to be delivered at the time of intake. The dose accumulated in subsequent 10 year periods has been conservatively assumed to be delivered at the beginning of that period, and a modifying factor appropriate to the beginning of the period applied. The periods considered and corresponding modifying factors are given in Table 2. These are based on the age distribution and life expectancy of the UK population in 1977(I } . The modifying factors are particular to the model chosen to represent the time of appearance of cancers.

'and different values are given for solid tumours and leuksemia. Their sensitivity to the pattern of appearance is unlikely to be large, at least for .

patterns that are consistent with the existing data on the incidence of radiation-induced cancers in aan.

3.2 Non-linear models Both the linear quadratic model and the pure quadratic model are incor-porated into HEALTH-MARC. In the linear quadratic model, the increase in the D

.g.

risk of cancer, R. due to a dose, D, is given by R = aD + $D 2 i

In the pure quadratic model, they are related by the expression

, R = qD 2 The quadratic model any be thought of as one extreme of the linear quadratic model with a = 0, but has been included separately. Either model any be used instead of the -linear model to estimate the increase in the incidence of cancer in an irradiated population.

To use these models, appropriate values of the parameters a and S. for the linear quadratic undel, and q, for the pure quadratic model, need to be specified for each of the organs listed in Table 1. If the risk coefficients, r, of

1. Table 1 were based on observations of increased incidence of cancer at a unique dose, D,, for each organ, this process would be relat.ively straightforward, since the risk given by the three models could be equated at this dose, is, aD + SD 2 = qD o

2. r0 o o o This equation would determine q and provide one constraint on the values of a and S. The remaining problem for the linear quadratic model would then be of the relative angnitudes of a and S. In reality, however, the risk coefficients are based on an increased incidence of cancer, with large uncertatuties, from exposure over a wide range of doses. Although some estimates of th'ese parameter values have been ande , the above procedure involves a significant asasure of personal scientific judgement.

In addition, the use of non-linear dose-respese models for late effects introduces other difficulties which remain unresolved. These include non-uniform i irradiation of particular organs and protraction of exposure. Because of these difficulties, simplifications have been incorporated into the non-linear dose-response models currently included in HEALTR-MARC. No account is taken of the time of appearance of cancer following irradiation; instead, it is assumed

conservatively that all the risk is expressed in the exposed population.

L j

Any application of the non-linear models in HEALTR-MARC is , therefore, likely to involve anny uncertainties, and the models are, as a result,

~

. deliberately simplistic in nature at this stage. Sensitivity analysis using l

• these simple models( vill be used as a first step in identifying the areas where further refinement any be required.

3.3 Total numbers of cancers

! If the risk of death from cancer for each exposed organ were considered separately, the incidence of fatal cancer from large releases of activity could be significantly overestianced, because the probability of death from cancer cannot exceed unity. The actual probability of death from a radiation-induced

. - cancer, CT, where the independent probability of death from cancer in the nth organ is C , is given by 9

[

i .

l

- - - _ _ - _ _ . . . ~ . . . _ , _ _ - - _ . _ _ _ . _ . . _ _ _ , . , _ _ _ . _ _

C T

=Cg + (1-Cg)C2 * . . . . . . . . + ( 1-C )(g 1-C )2 . . . . . ( 1-Cn-1} n This expression applies where no other factors contribute to death. For releases of activity which give rise to early death, the actual probability of death from cancer will be further reduced due to prior depletion of the population by radiation-induced early deaths. In this case the probab'ility of death from cancer will be (1-PMT)CT where PMT is the probability of early death. Similarly the actual probability of a particular non-f atal cancer has been assumed to be (1-PMT)UW where CNF is the independent probability of a particular non-fatal cancer. This expression neglects thr possible effect of prior depletion of the population bv death from cancer.

4 HEREDITARY ETTICTS The final important category of health effects considered in HEALTH-MARC is that of hereditary effects. Unlike somatic effects, which appear in the exposed individuals, hereditary effects manifest themselves in the descendants of these individuals. The serious hereditary disorders considered include those due to both gene mutations and chromoscas anomalies. These will appear in succeeding generations with about half occurring in the first two generations. Only the total numbers of hereditary effects that occur in all subsequent generations are, however, evaluated in EEALTH-MARC.

There are no human data on the incidence of radiation-induced hereditary effects; estimates of their increased incidence are based entirely on animal studies. Although most estimates of the risk of hereditary effects have been based on a linear model( 0,14) , there is again evidence in favour of a linear-quadratic model for low-LET radiation, and both types of model are included in HEALTH-MARC.

As with other health effects, the relative biological effectiveness of the radiation sust be taken into account. The values used by default in HEALTH-MARC are 1 and 20 for $/y and a radiation respectively, as recommended by Clarke and Salth( , although alternative values may be used.

4.1 Linear model Estimates of the incidence of hereditary effects following a dose of radiation have been reviewed by Clarks and Smith (  ; a risk coefficient of 2 x 10-2 cy-1 (low-LET) for gonad irradiation is recommended for individuals who are irradiated before the start of reproductive age and subsequently have the average number of children. This figure includes the risk of hereditary affects in all generations, with about half occurring in the first two generations. This risk coefficient is that adopted by default in the linear model of BIALTH-MARC.

The population averaged risk of hereditary eff ects, r,, can then be expressed as e

e r, = Mr

[' where r is the risk coefficient for, individuals who have the full opportunity to  !

.t b produce children, and the modifying factor, M, is given by .

[ N(A) [A Cg Pg(A)

M= ..... (2)

([ N(A)) ([ Cg)

A A where N(A) is the number of people in the standard population of age A.

• Pg (A) is the probability of surviving to age i, given survival to age A, Cg is the probability of becoming a parent when age 1.

The denominator in the above expression gives the total number of children that would be born if everyone in that population had the average number of children.

As for the incidence of cancer, this modifying factor applies to the external y dose from the passing cloud, and the external y dose from deposited activity, where the age distribution of the exposed population can be assumed to i

be typical of that in the UK . In the case of internal irradiation from inhaled and ingested activity, however, the dose to the gonads me; a delivered over an extended period depending on the characteristics of the rad.onuclides inhaled. A dose delivered at some later time af ter intake will have a reduced effect due to ageing of the population and an overall reduction in child expectancy. For a 7 dose delivered T years after intake, the modifying factor becomes

[N(A) [ Cg Pg(A W M(T) = A

([N(A))(.[C)

A A A

A similar procedure to that used to evaluate the modifying factor for the i~ incidence of cancer has been adopted in evaluating the above expression. The

, ;ose delivered within the first year has been assumed to be delivered at the cine of intake, and the dose delivered in subsequent ten-year periods has been assumed to be delivered at the beginning of that period. The modifying factors used and corresponding periods are given in Table 3. The factors are based on the age a

distribution and child expectancy of the UK population .

4.2 Non-linear models As with the estination of cancers, both linear quadratic and pure quadratic f models are included in HEALTR-MARC to estimate the increased incidence of hereditary effects. The problems described in the application of non-linear models to the estimation of the increased incidence of cancer apply equally to hereditary effects.

It has been assumed, for simplicity, in the use of thase models, that all of the dose is delivered to a population with an age distributioe. and child l

. it .

?- w ar- - -eae -vae- + - w v- -ex-y9 *-w-eg--g ,rv wi'-,,t'--*w--*wt wn w w u-w'-w9-a----m-m-**-=*weweey-++wvgs---*wv&Nt*-e*=-?*--w-'-- "*W-< -=w+eW'M#- wN'*-T-***

1 I

expectancy typical of that of the UK. Thus, the risk estinated using these l

models is modified by the factor M given in equation (2). For accidental ,

releases typical of those postulated for nuclear reactors, a large proportion of the dose to the gonads is from external esdiation, and this assumption is, therefore, not unreasonable. Other simplifying aspects of these models will be the subject of further analyses.  ;

4.3 Total numbers of effects It. 'avaluating the incidence of hereditary eff ects, account is taken of their possible reduction by radiation-induced early deaths. No account has, however,

~

been taken of the possible influence of early morbidity or 1ste effects on child i

! expectancy. For late effects, in particular, this is likely to be minimal.

5. SUMMART The MARC suite of modules has been developed to provide a comprehensive methodology for . the evaluation of the radiological consequences of accidental releases of radioactivity. This report has described the models curreocly incor-porated into HEALTH-MARC, the health effects module in the methodology. Options on the values of parameters in the models adopted and on the choice of models have been provided throughout HEALTH-MARC to enable the user to select those most appropriate for any intended application. A ouaber of simplifications are cur-(

rently included in some areas of HEALTH-MARC; these are the subject of continuing i analyses, and refinements will be made, where appropriate, in the future.

6. REFERENCES l 1. Clarke, RH and Kelly, G N, MARC - the NRPB sethodology for assessing l radiological consequences of accidental releases of activity. Chilton,

NEPB-R127 (1931). (London, HMS0).

l Lf 2. Jones, J A and Charles, D, AD-MARC: the atmospheric dispersion module in the l methodology for assessing the radiological consequences of accidental releases. Chilton, NRPB-M72 (1982).

3. Brocafield, M E and Es11am, J, MET-MARC: the meteorological sampling module in the anthodology for assessing the radiological consequences of accidental releases. Chilton, NEPB-M73 (1982).

l l

l 4 Charles, D, Crick, M J, Tell, T P and Greenhalgh, J R. DOSE-MARC: the j dosimetric module in the astuHiology for assessing the radiological consequences of accidental releases. Chilton, NEPB-M74 (1982).

l S. Broomfield, M E, Simmonds, J R and Chapman, T A, POP-MARC and AG-MARC:

population and agricultural distributions for use in the methodology for assessing the radiological consequences of accidental releases. Chilton,

! NRPB-M75 (1982).

i .

l 6. Linsley, G S, siamonds, J R and Haywood, S M. FOOD-MARC: the food chain

! transfer module in the methodology for assessing the radiological l

consequences of accidental releases. Chilton, NRPB-M76 (1982).

l-

7. Hallan, J, Henating, C 1, Simeonds, J R and Kelly, G N, PROT-MARC: the i

countermeasures module in the methodology for assessing the radiological coe. sequences of accidental releases. Chilton, NRPB-M77 (1982).

( ,

8. Kelly, G N, Simmonds, J R, Smith, H and Stacher, J W, The radiological consequences of notional accidental releases of radioactivity from fast breeder reactors: sensitivity to the dose effect relationships adopted for (N ec.rly biological effects. Harwell, NRPB-R87 (1979). (London, HMS0).

9. Smith, H, Radiation-induced damage in aan. IN The Handbook of Occupational Hygiene, Instalaant 1, p 2.2.1-01. London, Kluwer Publishing (1980).

10. USNRC, Reactor Safety Study: An assessment of accidental risks in US commercial nuclear power plants. Washington DC, US Nuclear Regulatory Consission, WASE-1400 (1975).

11. Bundesministeriums fur Forschung und Technologie, Deutsche Risikostudie

-- Kernkraftwerke: Eine Untersuchung zu den durch St5rfE11e in Kernkraf twerken verursachten Risiko. Verlag TUV Rheinland (1980).

12. Kelly, G N, Jones, J A and Hunt, B W, An estimate of the radiological consequences of notional accidental releases of radioactivity from a fast breeder reactor. Harwell, NRPB-R53 (1977). (London, HMS0).

13. Smith, E and Stather, J W. Human exposure to radiation following the release of radioactivity from a reactor accident: a quantitative assessment of the biological consequences. Harwell, NRPB-R52 (1976). (London, HMS0). ,

14 Clarke, E H sad Smith, H, Calculation of the incidenet of stochastic(London, health effects in irradiated populations. Harwell, NRPS-R102 (1980).

HMS0). 1 London,

. 15. Central Statistical Office. Annual abstracts of statistics 1979.

HMSO (1979).

16. NAS/NEC, BRIE III, Report- of the advisory couaittee on the biological effects of ionising radiations. Washington DC, National Acadesy of N) Sciences / National Research Council (1980).

17. Bessning, C R, Kelly, G N and Charles, D, An assessment of the sensitivity of the consequences of actional accidental releases of radioactivity to the use of non-linear dose-response relationships for late effects. Chilton, NRP3 (to be published).

e S

0 O

-~, - . w- , . - - , , , - - - - - - . . - , . , . - , - . ,-,,--,w. -,,.,,-,,..--._,,n.--,-.. ,_,,.,,_--, . . _ . --,,,--n.,.g.. , - , .pn,,,.,-

s Table 1 Risk coefficients of the increased incidence of cancer for-use in the linear model Risk coefficient Organ (Gy-1) low-LIT a) Fatal cancers l Ereast 2.5 x 10~3 Red bone marrow 2 x 10~3 Lung 2 x 10-3

1. ' Thyroid 5 x 10 *

. Bone surface 5 x 10*

Liver 1 x 10-3 LL1 1 x 10~3 Remainder tissues 3 x 10-3 Skin 1 x 10*

b) Non-fatal cancers l

' Thyroid i x 10-2 .

Skin 1 x 10-2 Breast 2.5 x 10-3 c) Eereditary effects 2 Canada 2 x 10-2 Notes:

1. The risk coefficients are applicable to a population which will live long enough for the total risk to be expressed.

2. The risk coefficient is applicable to individuals irradiated before reproductive age and who subsequently have the average maaber of children.

I

% a y

e.

i

l. -

l t

s , w w ~ , s-,. .,--,-n-,_n.. .-.,m.nc,,,---..,n , ..

m .-nam ,,,-,,m,,,,,e n.e,-we,,n - w,w m e,w w ,w n ee.e4v.- w.-e,ww wr- s e. e r e-ww w w es.v-, v--' w'

Table 2 Modifying factors to be applied to the risk coefficients for the incidence of cancer for doses delivered in various ceriods following the intake of activity Period, y 0-1 1-10 11-20 21-30 31-40 41-50 51-60 61-70 1.euksemia 0.76 0.75 0.62 0.49 0.37 0.25 0.15 0.06 Solid cusours 0.63 0.61 0.49 0.36 ' O.24 0.13 0.05 0.01 Table 3 Modifying factors to be applied to the risk coefficients for the incidence of hereditare effects in all future generations for doses delivered in various periods following the intake of activity Period, y O

0-1 1-10 11-20 21-30 31-40 41-50 51-60 61-70 Modifying factor 0.40 0.39 0.26 0.11 0.02 2x10-3 2x10-4 0.0 G

9

~

Sou=ce to n h,

LD.4 R C MM m a Sampling of DM ^ '

meteorological

1. • M "

'f'N --

If condiziens

.p._

/ dist=1butions of:

t-Deposited /

N A1.h activitr motivity

/

Dogs MAac sh'R0

• • w 2 ym t:nnster sin S.

• If ac-Maac

> Food p=oduction 1f '

mm 1f par-Maac 305E-;rRC population > i Dosimet:7

- in8'*tian d.ist=1butian 1f If stor-MAac Countammeasures (tod and people) me.s e 3 p pl. ef t.d br ooan--ee JZeas of Doses in exposed M and population i fod i restrictice.s II

. 1f ERAZ55-MA3e EC50-MARC

, Isale effects > 4 !and u*.*.lisation and ooanamics g:14 If If

f

_ _ .e

1. / ~d ~_

~<

ad O - - - - - - - - - - - - - - - - - -

g

< o __

F aure 1 Schematic dias:ma of the as or ccurpenen'.s of FR

• i

' 1.0 -

/. *

, ~ f./ . .

/ .. ... ....

0.6

-.- US -RSS(10) ,

l ',.

(-)

s .

E

• 0.5 -

........ German Risk Study (11)

.j-. ' [./ ..

f.-

3 / l 2

20.4 - -

l *'l'

.C

.8 l ./.

~  ! 02 _ .

/ .

l

/ /

/ '/

n n .A. . .c " ' . -/s , , , ,

0 1 2 3 4 5 6 7 8 Absorbed dose. Gy. Iow LET Figure 2 Examples of mortality relationships for bone marrow irradiation used in other studies t

O

! 10 08 . Likely bounds of l

! x uncertainty E -

3o 06 -

.. x .

GL .

.8

e -

Q- 02 -

l l' ' " ' ' ' '

0 4 6 7 O 1 2 3 5 f- ,,

Absorbed dose . Gy. low L E T LO Figure 3 Dose mortality relationship for bone morrow irradiation u

. . , , _ . , , ~ . _ . , . _ . . _ , , . . . _ . , - . . . . - . . . . , _ _ _ . - . ~ . . . _ . _ - - . .

100 -

80-T x 83Y

~~

o~ 6 0 -

~

en g

" Ce >

0 40 -

• ~ 80 Sr '

- - 20 -

E.'

Q.

. . 0 -' ' ' ' ' ' '

0 100 200 -

300 350 Time, days Figure 4a Pattern of occumulation of lung dose for dogs exposed to various inhaled insoluble radioactive aerosols O

1.0 -

r--- ,

/

Y ^

/ ,

/

S /

/

'O '

.h Yr j y 0.1 -

/ 91 7 ,

g i ' [14 Ce 90 Sr e j ,

/

o I j'

I , .E /-

..- p

• i /

r

• a01 e t/ '

2 3 I 10 10 10 10 '

Absorbed dose. Gy. Icw LET O Figure 4b Dose mortality relationship for lung irradiation

,er ..--,---...-.---.,--a --e.,,.,.w _.,,,w,-m-a.--..,,-,-n- c,--,,- m..e- e vee-,.,,,w,,,.mww-w%---w,wwwr,nemw,,m-w.,w'r* Wet"r-*------*-*-"='

1.0 - .

0.8 -

r^s 10.6 o

?

] 0.4 -

.8 e

' 02 -

Se

. I , I

! 'o ,

20 40 60 Absorbed dose. Gy. low LET o Figure 5 Dose mortality relationship for G I tract irradiation i

O t0 , , , , ,

Y 3 08 -

8 E

06 -

04 -

4. g i .

02 -

0 O 10 20 30 40 50 60 70 Absorbed dose , Gy, low L E T O sisure e oose morsioity reiotiomenig ior iums irreeietiam

- , - ,v.-- ,c.- , ,,- + ,-,m,,-,e,w, - , , - ..,,_,y,-.,,,_,.,-,.,-,,,,_wwww,w,,-,,.-, .,_g,,.,-n,,mwv,-,e,w,,w,--,,.,,ww,c,., , - . -

1.0 - .

l

~ 0.8 -

3 a6 -

l y '

=

f 0.4 a.

e:

. 0.2 -

r,

,.' 0

-0 1 2 3 4 Absorbed dose. Gy. Iow LET Figure 7 Dose morbidity relationship for prodromal vomiting _

O

. /

Linear /

- ./

--- Linear-quadratic -

./

u - Pure quadratic .

g - .

... - / -

o / -

.- a / .

- .e / .-

• Gr. / =

,/ ..-

/,'/ . . - -

x.T......

Dose Figure 8 Examples of the models for estimating the increased incidence of cancer

-.-.....-.-.---,,._,..-...-,-._..,,,..v., ..._,,-.,,ey.,m,.~.-.r.y, m_,,,, . , , , , . , _ ,m_,.r,,m.v._-,_.,,-,w.,,,---

1 g

O ..

. . e, O .1- .

O 1

i i

006 , , , ,

1

} Leukaemia, median 12Sy , e = 0 8 1

s -

t 0 05 -

4 0 04 - Solid tumours , median 25y, e =0 4-o 1 C -

l 3u 003 -

l C

}

l 0 02 -

)

i 0 01 -

l 1

0 O 10 20 30 40 50 Time ,. y Figure 9 incidence of concers following on increment of l Irradiation i

i l

.. -. - ..