NEA W/S, 1993

From "Radiation Protection: on the threshold of the 21st Century", Proceedings of NEA W/S, 1993.

Low Dose Studies

Studies of populations exposed at low dose rates, which are more directly relevant to protection, are generally hampered by lack of statistical power and possibly also by confounding factors. However low does rate studies should increasingly provide support for the risks derived by extrapolation from high dose studies. The main studies of interest are on workers from areas of high natural background and a series of studies on domestic exposure to radon are underway. Some information may also become available from the population exposed as a result of the accident at Chernobyl.

Worker Studies:

Several studies have been conducted of nuclear industry workers. In the USA, Gilbert et al (1980) performed a joint analysis of data for 35,933 workers at the Hanford site, Oak Ridge Nat'l Labs and Rocky Flats weapons plant with a collective dose of 1140 mSv. Neither for the grouping of all cancers nor for leukemia was there an indication of an increasing trend in risk with dose.

A recently published study of just over 95,217 individuals on the UK's Nat'l Registry for Radiation workers (NRRW) with a collective dose of 3213 mSv has examined cancer mortality in relation to dose (Kendall et al, 1992a+b). For all malignant neoplasms, the trend in the relative risk with dose was positive but was not statistically significant (p=0.10). Based on a relative risk projection model, the central estimate of the lifetime risk based on these data was 10 10^-2 Sv^-1 (90% CI < 0,26) which is 2.5 times the value of 4 10^-2 Sv^-1 cited by ICRP (1991a) for risks associated with exposure of workers (4 10^-2 Sv^-1) although with confidence intervals which span the ICRP value. For leukemia (excluding chronic lymphatic leukemia (CLL) which does not appear to be radiation-inducible), the trend in risk with dose was statistically significant (p=0.03). Based on a BEIR V-type projection model (BEIR, 1990), the central estimate of the corresponding lifetime leukemia risk was 0.76 10^-2 Sv^-1 (90% CI 0.07, 2.4) which is 1.9 times the ICRP value for a worker population (0.4 10^-2 Sv^-1). There was also an indication of an increasing trend with dose in the risk of multiple myeloma (p=0.06), the estimated trend in relative risk was about 3 times that obtained from the Japanese survivor data under a linear dose-response model, with 90% confidence ranging just under zero up to 20 times the Japanese value. An increasing trend in multiple myeloma risk with dose was similarly found in the US study by Gilbert et al (1989) (p<0.05).

The NRRW therefore provides evidence of raised risk of leukemia and multiple myeloma associated with occupational exposure to radiation, but, like the combined study of US workers (Gilbert et al, 1989), is consistent with the disk estimates of low dose/dose rate exposures derived by ICRP (1991a) from the Japanese survivor data. In particular, combining the NRRW and US results produces central estimates for lifetime risk of 4.9 10^-2 Sv^-1 (90% CI<0, 18) for all cancers and 0.30 10^-2 Sv^-1 (90% CI<0, 1.04) for leukemia excluding CLL (Kendall et al, 1992b), which are similar to the ICRP risk estimates.

Although the workers studies performed to date lack sufficient power to determine risks with any great precision, this can be increased by pooling the data already published and by initiating further studies. This pooling of data is to be undertaken under the auspices of the International Agency for Research in Cancer (IARC) and will include a study at present underway in Canada covering nearly 420,000 individuals. Further studies are also being set up or are in progress in a number of countries and are also expected to be pooled by IARC.

Enhanced natural radiation:

Studies of exposed natural radiation (other than from radon) have generally involved looking for a geographical correlation with cancer rates. Such studies are difficult to interpret however owing to the effect of confounding factors such as socio-demographic variables and other factors that vary geographically and are unlikely to provide any quantitative data on cancer risks. A number of case-control and cohort studies have recently been set up to examine the consequences of domestic exposure to radon and its decay products. Again there is the problem of limited statistical power of the individual studies but pooling of the data will improve the power of any analyses.

Table 2

Dose limits for whole body exposure recommended by ICXRP and ICRP
Date Occupational dose Members of the public
1934 0.2 R per d
(~ 1.2 mSv per d)
1951 0.5 R per wk
(~ 3 mSv per wk)
1955 3 mSv per wk
1959 30 mSv per qtr
or 50 (N-18) mSv[1] 		5 mSv per y
1966 50 mSv per y 5 mSv per y
1977 50 mSv per y[2] 5 mSv per y[3]
1990 20 mSv per y
average over 5 y 		1 mSv per y[4]
[1]
N = age in years
[2]
Presumed that average dose will be lower
[3]
To critical groups, average doses much lower
[4]
Higher in single year allowed provided average over 5 y does not exceed 1 mSv per y.
Note: in the 1955, 1959 and 1966 recommendations limits given in rem (1 rem == 10 mSv).

Table 3

Risk coefficients for fatal cancer adopted by ICRP

Fatal cancer, 10^-2 Sv^-1
Organ or tissue ICRP 1977 ICRP 1991


Population workers
Bladder
.3 .24
Red bone marrow .2 .5 .4
Bone surface .05 .05 .04
Breast .25 .2 .16
Colon
.85 .68
Liver
.15 .12
Lung .2 .85 .68
Oesophagus
.3 .24
Ovary
.1 .08
Skin
.02 .02
Stomach
1.1 .88
Thyroid .05 .08 .06
Remainder .5 .5 .4
Total 1.25 5.0 4.0

Table 4

Numbers of deaths from all cancers other than leukemia among the Japanese atomic bomb survivors with a DS86 dose of .75 Gy or more (from Preston and Pierce, 1987)
Age at exposure
(years)
Time since exposure


5-25 25-40 5-40
<20 O[a] 14 44 58

E[b] 4.03 17.8 21.8

O/E[c] 3.47 2.47 2.66
20-34 O 26 48 74

E 13.0 24.4 37.4

O/E 2.01 1.96 1.98
>=35 O 119 99 218

E 86.7 68.9 155.6

O/E 1.37 1.44 1.4
All O 159 191 350

E 103.7 111 215

O/E 1.53 1.72 1.63
[a]
O - Observed number of deaths
[b]
E - Expected number of deaths in an unirradiated population, based on rates among those with DS86 dose < 0.1 Gy.
[c]
O/E - Relative risk.

Table 5

Projected lifetime risk for a Japanese population following a whole body \gamma exposure to 1 Gy (low LET) radiation at high dose rate (UNSCEAR, 1988)

Projection model Risk of fatal cancer, % Gy^-1
Total population A 4[a] - 5[b]

M 7[b] - 11[a]
Working population (25-64 y) A 4[b] - 6[a]

M 7[a] - 8[b]
UNSCEAR 1977 A 2.5
A -
additive
M -
multiplicative
[a]
age-specific constant relative risk
[b]
age-averaged constant relative risk

Table 6

Human populations available for risk estimation
Atomic Bombs Japanese Survivors

Marshall Islands
Medical Diagnosis Multiple Fluoroscopies (breast)

Prenatal Irradiation

Thorotrast Injections

Thyroid Disorders
Medical TherapyPelvic Radiotherapy (cervix)

Spinal Radiotherapy (ankylosing spondylitis)

Neck and Chest Radiotherapy (thyroid)

Scalp Irradiation (tinea capitis)

Breast Radiotherapy

Radium Treatment
Occupational ExposureUranium Miners

Radium Ingestion (dial painters)

Table 7

Summary of dose and dose rate effectiveness factors
Source DDREF
ICRP 1977 2
NCRP 1980 2-10
UNSCEAR 1986 up to 5
UNSCEAR 1988 2-10
BEIR 1990 2-10
ICRP 1991 2
[a]
2 for breast

Table 8

Estimated lifetime fatal cancer risks in populations (all ages, both sexes) associated with exposures to low LET radiation at high doses and high dose rates, based on a multiplicative projection model
Population Fatal cancer risk 10^-2 Sv^-1
UNSCEAR 1977-2.5[a]
BEIR III 1980USA2.3-5.0
UNSCEAR 1988Japan7-11[b]
BEIR V 1990USA7.9[c]
ICRP 1991Five Nations10.0[d]
Muirhead 1992UK11.8[e]
[a]
additive model
[b]
range based on age-averaged and age-specific constant relative risks.
[c]
see text (section 4.8)
[d]
average value based on US, UK, Japan, Puerto Rico and Chinese populations. Risk for workers 8.0 10^-2 Sv^-1.
[e]
risk to 40 years 4.9 10^-2 Sv^-1, based on age-specific constant relative risks.