Japanese AJapanese A-bomb Survivor Databomb Survivor Data ...dels.nas.edu/resources/static-assets/nrsb/miscellaneous/Shore... · Japanese AJapanese A-bomb Survivor Databomb Survivor

Japanese A-bomb Survivor DataJapanese A bomb Survivor Data and Studies of Low-Dose Effects

Beebe SymposiumBeebe SymposiumNational Academy of Sciences

December 1, 2010,Roy Shore,

K. Ozasa, Y. Shimizu, K. Kasagi, K. Furukawa, W-L. Hsu,D Preston K Neriishi A Suyama K KodamaD. Preston, K. Neriishi, A. Suyama, K. Kodama

[email protected]

Wh i th A b b h tWh i th A b b h tWhy is the A-bomb cohort considered a benchmark for Why is the A-bomb cohort

considered a benchmark for assessment of radiation risks?assessment of radiation risks?

22

RERF A-bomb Related Cohorts

Residents in Hiroshima and Nagasaki at Explosions : 500,000

In Utero Exposed3,600

In Utero Clinical

Residents at 1950 National CensusA-bomb survivors : 280,000 In Utero Clinical

1,600 (1978-),

Life Span Study120,000 (1950-)

F1:Offspring of Survivors

77,000Adult Health Study 22,000 (1958-)

, ( )

Adult Health Study

F1 Clinical12,000 (2002-)

22,000 (1958 )Adult Health Study17,000 (1958-)

3

A-bomb Studies as a Benchmark for Radiation Risk Assessment

A-bomb Studies as a Benchmark for Radiation Risk Assessmentad at o s ssess e tad at o s ssess e t

Large prospective cohort ( 120 000) consisting of all ages atCharacteristics of the radiation-exposed cohort: Large, prospective cohort (~120,000) consisting of all ages at

exposure and both sexes Wide range of radiation exposure with relatively accurate

estimated dosimetry Cohort is unselected with respect to medical conditions,

occupational fitness, etc.p , High rates of follow-up and disease ascertainment for ~60

years f f f Information on potential confounding risk factors Adult Health Study – biennial clinical examinations and

biospecimen collection for a subset of about 17,000 – study

44

biospecimen collection for a subset of about 17,000 study correlates and mechanisms of radiation-related disease through biomarkers, etc.

Life Span Study (LSS) Cohort

50Cohort

(120,321 people)

30

40

(%

)

20

ropo

rtion

10

P

oseU

NK

NIC0 00

5Gy

5-0.1G

y1-0

.5Gy

0.5-1G

y

1-2Gy

>=2G

y

0

55

Dos <0.0

0.005 0.

1 0. >

DS02 Colon Weighted Dose (Gy)

Are the excess risks of cancer at low doses proportional to those at high doses? – i eproportional to those at high doses? – i.e.,

Is there dose-response linearity, less than or more than linear risk at low dosesless-than or more-than linear risk at low doses,

or a dose threshold?

66

LSS dose response: Solid-cancer incidence

No evidence of non-linearity in the dose response Significant dose response on 0-150 mGy

L d l i i h f ll Low dose-range slope consistent with full range

150

(LSS Incidence, 1958(LSS Incidence, 1958--1998)1998)

100

150

ativ

e R

isk ERR/Gy= 47% (95%CI: 40-54%)

Dose-threshold: 40 mGy(CI: <0, 85 mGy)

50Exce

ss R

ela

Fitted linear dose response at age

00 1 2 3

% Fitted linear dose response at age

70 following exposure at age 30Smoothed non-parametric dose response

77(Preston D et al: Radiat Res 168:1-64, 2007)

0 1 2 3Colon Weighted Dose (Gy)

LSS Dose-Response for Solid Cancer Mortality, 1950-2003

Gy

Lin-Quad(<2Gy)sk

per

G

Linear

elat

ive

Ri

xces

s R

e

Colon Weighted-Dose (Gy)

Ex

8

Colon Weighted-Dose (Gy)

(Ozasa, Shimizu, et al, Unpublished, 2010)

LSS Mortality Estimates of Relative Risk at 1 Gy for Various Dose Ranges (0 to Plotted Dose)

2.52.5

y

1.5

2.0

1.5

2.0sk

per

Gy

0 5

1.0

ERR

/Gy

0 5

1.0

elat

ive

Ris

0.0

0.5Full range

0.0

0.5

Exce

ss R

e

-0.50.01 0.1 1 10

Colon dose (Gy)

-0.5

Colon Weighted Dose (Gy)

E

(Ozasa et al, Unpublished, 2010)9

LSS Leukemia Dose Response

15

Ris

kFitted linear-quadratic dose response at age 70 following exposure at age 30Smoothed non-parametric dose response

10

elat

ive

R

5

cess

Re

0

Exc

0 1 2 3

Linear-quadratic fits better than Linear model.

0 1 2 3

Bone Marrow Weighted Dose (Gy)

10

qDose-threshold estimate: 80 mGy (95% CI: 30, 190 mGy)

(Hsu et al, Unpublished, 2010)

Which organs are at risk of radiation-related cancer?radiation-related cancer?

11

Excess Relative Risk (ERR) per Gy for VariousSolid Cancers (LSS Mortality, 1950-2003)

# Deaths

All solid cancers 10755

LiRectum

ColonStomach

Esophagus

1485419609

3103334

BreastLung

PancreasGall bladder

Liver

3161535495410

1485

BladderProstate

OvaryUterusBreast

177125155544316

Excess Relative Risk per Gy0 50% 100% 150%

177

12

Note; Estimates standardized to age 70 after exposure at age 30 and averaged, where appropriate, over sex.

(Ozasa, Shimizu et al, Unpublished, 2010)

D t i b h tDo certain subgroups have greater risk of cancer from radiation

exposure?

13

LSS Cancer Incidence: Site-specific Excess Relative Risk (ERR) Estimates by Gender

2.0

1.5

k pe

r Gy

1.0

ativ

e R

isk

FemaleMale

0.5

xces

s R

el

0.0

Ex

14

Breast Lung Bladder Stomach All solid Thyroid Liver Colon

Preston et al, Radiat Res, 2007; 168:1-64

Excess Rates of Solid Cancer Mortalityby Age at Exposure and Attained Age

Age at exposure45

50

-Gy

Age at exposure

35

40

45303020201010

,000

PY-

25

304040

s pe

r 10,

15

20

s D

eath

s

0

5

10

Exce

ss

1515Attained Age (Age at Risk)(Ozasa, Shimizu et al, Unpublished)

20 30 40 50 60 70 80 900

How great is the cancer riskfollowing in utero exposure?following in utero exposure?

16

Very large risks from in utero exposurehave been predicted

Based mostly on the large Stewart-Knealet l t d f f t l di ti

p

case-control study of fetal radiation exposure and childhood cancer, it was predicted that an additional 6% of persons would die fromadditional 6% of persons would die from cancer after 1 Gy of in utero radiation.(Doll & Wakeford, Br J Radiol, 70:130-39, 1997)

6% lifetime cancer mortality risk per Gy is: ~3 times as large as the corresponding estimated g p g

A-bomb risk after exposure at age 10, or ~6 times as large as estimated risk after exposure

t 30 (Preston et al Radiat Res 160:381 407 2003)

17

at age 30. (Preston et al, Radiat Res, 160:381-407, 2003)

Solid Cancer Risk Patterns for In Utero andChildhood Exposure, A-bomb Survivors

12 In-utero exposure 80

100

Age at exposure:0 P

Y-G

y

6

9

Age at exposure:ERR

per

Gy

40

60

Age at exposure: 0 – 5years

es p

er 1

0,0

00

P difference = 0.13

0

3 P difference = 0.2

Age at exposure: 0 – 5years

0

20

In-utero exposureExce

ss c

ase difference

10 20 30 40 50 60

0

Age

10 20 30 40 50 60

Age

18(Preston et al, J Natl Cancer Inst, 2008; 100:428-36)

In utero ERR/Gy= 1.0 (95%CI: 0.2, 2.3)

Leukemia after In Utero or ChildhoodA-bomb Exposure

Childhood exposure, ages 0-5

p

39 cases, 22 with estimated bone marrow doses >500 mGy

Steep dose response (ERR/Gy = 15 95%CI: 6 36) Steep dose response (ERR/Gy = 15, 95%CI: 6, 36)

In Utero Exposure 4 cases, all with estimated bone marrow doses <40

mGy N d (ERR/G 0 9 %CI 0 ) No dose response (ERR/Gy = 0, 95%CI: <0, 7)

19

(Kasagi, Ozasa, et al, Unpublished, 2010)

I th it d f di ti i kIs the magnitude of radiation risk altered by other environmental

exposures?

20

Models of Lung Cancer Risk: Radiation Dose and Smoking

151 Gy Non-parametric1 Gy Generalized multiplicative

**

10

1 Gy Simple additive0 Gy Generalized multiplicative

ve R

isk

*

0 Gy: Smoking effect

5ss R

elat

ivEx

ces

0

0 10 20 30 40

Cigarettes Smoked per Day

21

Cigarettes Smoked per Day** Gender-averaged excess risk relative to unexposed never-smokers

(Adapted from: Furukawa et al, Radiat Res, 174:72-82, 2010)

Liver Cancer Risk from Hepatitis C Virus (HCV) and Radiation

40

20

30

s R

atio

HCV-,1 Gy radiation

HCV+, No rad

HCV+ 1 Gy rad

10

20

Od

d HCV+, 1 Gy rad

0

2222

(Sharp et al, Int J Cancer, 103:531-37, 2003)

Are there noncancer risks from radiation exposure at low doses?radiation exposure at low doses?

2323

Radiation and Heart Disease Mortality

•• Clear evidence of heart disease risk at doses below 4-5 Gy • Dose-response for heart disease mortality appears linear, butthere is considerable ncertaint belo abo t 0 5 G

Heart Disease Mortality

there is considerable uncertainty below about 0.5 Gy.

Corroborative Clinical Evidence for Radiation Effects

30

40

50 ERR = 14% per Gy†

( 95% CI: 6, 23%) LLQ

(ER

R) P

erce

nt

for Radiation Effects

• ↑ Circulatory systeminflammation – numerous markers

10

20

30

s R

elat

ive

Ris

k (

of inflammation are ↑• Blood lipids – ↑ totalcholesterol triglycerides; HDL

0.0 0.5 1.0 1.5 2.0 2.5 3.0-10

0 L: linear

LQ: linear-quadratic

Weighted Colon Dose (Gy)

Exce

ss cholesterol, triglycerides; HDL cholesterol• Cardiovascular risk factors –

24

Weighted Colon Dose (Gy)†Adjusted for gender, age at exposure, attained age, diabetes, etc.

(Shimizu et al, Br Med J, 340:193, 2010)

↑ blood pressure andcalcification of arteries

• Radiation protection agencies had long believed there was no risk for

Cataract-Surgery Incidence, 1986-2005• Radiation protection agencies had long believed there was no risk for vision-impairing cataracts below about 5 Gy and set safety standards accordingly. More protective safety standards for the eye are now being considered.

35

40

Year

ERR/Gy = 21% (95%CI: 9-36%)

considered.

20

25

30

P < 0.001

0 Pe

rson

s pe

r Dose-Threshold: ~0.5 Gy (95%CI: 0.1-1.0)

10

15

20

Cas

es p

er 1

000

-5

0

5

Exce

ss C

25(Nakashima, Neriishi et al, Unpublished, 2010)

0 2 4

(Adjusted for city, gender, age at exposure, attained age and diabetes)Eye Dose-Equivalent (Gy)

How large is the risk to offspring from parental gonadal irradiation?from parental gonadal irradiation?

26

Relative Risk for Non-cancer and Cancer Mortalityin 41,000 Offspring of Atomic Bomb Survivors,

1946-2003

• To date, the frequencies of cancer and other diseases in theoffspring are unrelated to parental radiation dose, but 20-30

1946 2003

1 5

offspring are unrelated to parental radiation dose, but 20 30more years of follow-up are needed to provide definitiveevidence.

y 1 0

1.5 Non-cancer Cancer

lativ

e R

isk

at 1

G

0 5

1.0

Rel

0 0

0.5

27(Suyama, et al, Unpublished, 2009)

*Adjusted for potential confounding factors: Age, gender, city, birth year, parent’s ages at exposure and childbirth order

Mother’s Father’s Mother’s Father’sDose Dose Dose Dose

0.0

28Thank You