Perinatal mortality and congenital malformations following ... · Teratogenic radiation effects Perinatal mortality and congenital malformations following Chernobyl Alfred Körblein

Teratogenic radiation effects

Perinatal mortality and congenital malformations

following Chernobyl

Alfred Körblein Nuremberg, Germany

Teratogenic radiation effects: Adverse health effects to the fetus after exposure to ionizing radiation:

Congenital malformations, perinatal deaths

(i.e. stillbirths and early neonatal deaths), cognitive impairment

UNSCEAR on teratogenic effects:

• UNSCEAR 2008 (Chernobyl consequences):

A search in the document finds no hit for

“teratogenic” OR “perinatal” OR “malformation”

• UNSCEAR 2013 (Fukushima consequences)

(b) Other health effects, F14:

“Prenatal exposure was not expected to increase .. the incidence of spontaneous

abortion, miscarriages, perinatal mortality, congenital effects or cognitive

impairment.”

2

ICRP 90 (2003): Biological effects after prenatal irradiation (embryo and fetus)

• Experimental data on age-dependent patterns of sensitivity to malformation

induction re-inforce the view of maximum sensitivity during major

organogenesis, but with a true dose threshold.

For practical purposes, risks of induction of malformation at low doses may

therefore be discounted.

• (44) Induction of malformations […]

most pronounced during organogenesis (weeks 3-11 p.c.)

• (60) Most experimental data basically fit

sigmoid or shoulder-type dose-response curves [… ]

3

ICRP 90, Fig.1.1: Occurence of lethality and abnormality after prenatal radiation exposure of about 2 Gy

4

Shape of the dose-response relationship for stochastic and teratogenic effects

From: http://gd1.med.uni-giessen.de/ugm_2/deu/ugi_nuk/PDF/Rad_V2_Strahlentherapie.pdf

5

Results of epidemiologial studies

6

Perinatal mortality in Germany and regression line.

Excess rate in 1987: 0.36 per 1000 (P=0.018), 317 excess deaths from: Körblein A, Küchenhoff H. Perinatal mortality in Germany following the Chernobyl accident. Radiat Environ Biophys. 1997 Feb;36(1):3-7.

4

5

6

7

8

9

10

11

12

13

1980 1982 1984 1986 1988 1990 1992

pe

rin

ata

l mo

rta

lity p

er 1

00

0

7

1987

Perinatal mortality in Poland and regression line.

Excess rate in 1987: 0.57 per 1000 (P=0.015), 348 excess deaths

12

13

14

15

16

17

18

19

1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991

pe

rin

ata

l mo

rta

lity p

er 1

00

0

8

1987

Deviations of perinatal mortality in Germany and Poland from secular

trend (standardized residuals) and range of 2 standard deviations.

-8

-6

-4

-2

0

2

4

6

8

1980 1982 1984 1986 1988 1990 1992

sta

nd

ard

ized

re

sid

uals

Poland

Germany

9

Monthly perinatal mortality rates and trend line. The vertical line indicates the date of the Chernobyl accident.

10

Perinatal mortality in Germany

Deviations of mortality rates from undisturbed trend line (standardized residuals) and 3-month moving average. After Chernobyl, mortality peaks are found in

February and November 1987

11

Infant mortality rates in Poland and trend line with seasonal variations Lower panel: standardized residuals and 3-month moving average Peaks are found in January , April, and November 1987

12

Monthly infant mortality in Poland

Upper panel: monthly rates and regression line Lower panel: Standardized residuals and 3-month moving average Maximum in April 1987

13

Early neonatal mortality in oblast Zhytomyr

Trend analysis of monthly data

• Logistic regression with linear-quadratic time trend

• Seasonal effect modeled by dummy coding for individual months (dummy variables for Feb through Dec, Jan=reference)

• Cesium burden of pregnant women calculated from cesium concentration in cow milk (measured), beef, cereals, and pork

• Linear- plus quadratic cesium dependency to allow for a curvilinear shape of cesium effect on perinatal mortality

14

15

From: UNCEAR 2008, Annex D, page 95

Measured cesium-137 concentration in cow milk (black points) and calculated cesium concentration in beef, cereals, and pork, semilogarithmic plot

16

0,1

1

10

100

1000

1986 1987 1988 1989

cesiu

m in f

oodstu

ff [

Bq

per

kg

]

calendar years

milk beef

cereals pork

Measured cesium-137 concentration in cow milk (black points) and calculated cesium concentration in beef, cereals, and pork, linear plot

17

0

50

100

150

200

250

300

350

400

1986 1987 1988 1989

ce

siu

m i

n f

oo

dstu

ff [

Bq

pe

r kg

]

calendar years

milk

beef

cereals

pork

Cesium-137 concentration in cow milk (crosses, left axis)

and cesium burden in pregnant women (grey line, right axis), calculated with a biological half-life of 70 days

18

0

5

10

15

20

25

30

35

40

45

0

50

100

150

200

250

300

350

400

1986 1987 1988 1989

ce

siu

m b

urd

en

in w

om

en

[B

q p

er

kg

]

cesiu

m in c

ow

milk

[B

q/l]

calendar years

cesium in cow milk

cesium burden

Cesium-137 concentration in cow milk, beef, cereals, and pork and cesium burden in pregnant women (grey line, right axis), calculated with a biological half-life of 70 days

19

0

5

10

15

20

25

30

35

40

45

0

50

100

150

200

250

300

350

400

1986 1987 1988 1989

cesiu

m b

urd

en in

wom

en [B

q/k

g]

cesiu

m in f

oodstu

ff

[Bq

/kg

]

calendar years

Monthly averages of cesium burden in pregant women from ingestion of milk and other foodstuff

20

0

5

10

15

20

25

30

35

40

45

1986 1987 1988

cesiu

m

burd

en

[Bq

/kg

]

calendar years

other food

milk

Residual sum of squares obtained with regressions of German monthly perinatal mortality rates as a function of time-lag between cesium burden and mortality. Horizontal broken line indicates critical chisquare value. Best estimate of time-lag is 7 ± 1 months.

21

186

188

190

192

194

196

198

200

202

204

206

0 1 2 3 4 5 6 7 8 9

devia

nce

time-lag [months]

Standardized residuals of perinatal mortality in Germany and 3-month moving average.

Lower graph: cesium burden of pregnant women, shifted by 7 months

22

-8

-6

-4

-2

0

2

4

6

8

1985 1986 1987 1988 1989 1990

sta

ndard

ized re

sid

uals

Dose dependency Increase of perinatal mortality in Germany as a function of lagged cesium burden,

and regression line. The cesium effect on mortality is significant (P=0.0016, F-test). The dose-response is curvilinear: Linear term <0, quadratic term >0 (P=0.034)

23

Dose dependency Increase of infant mortality in Poland as a function of cesium burden, and regression

line. Cesium effect on mortality is significant (P=0.014, F-test). Dose-response is curvilinear: Linear term <0, quadratic term >0 (P=0.016)

24

Congenital malformations in Bavaria

• After Chernobyl, a study of congenital malformations in Bavaria was conducted by German Federal Office of Radiation Protection (BfS-ISH)

• The following isolated malformations were included: - neural tube defects, - anophtalamus, microphtalamus, cararact - limb anomalies - cleft lip and palate - atresia of gastrointestinal tract - cong. anomalies of the heart, cong. anomalies of abdominal wall - diaphragmatic hernia

• Malfomations with likely genetic origin: trisomy 21, malformation syndrom (multiple malformations)

• Malformation rates in Southern Bavaria (total cesium soil contamination: ~30 kBq/m²) were compared with rates in Northern Bavaria (~10 kBq/m²)

25

Method: Cesium burden cs(k, t) is defined as the product of (1) cesium soil deposition (kBq/m²) in district k (k = 1 to 96) cs(k) and (2) cesium concentration (Bq/kg) in pregnant women as a function of time t, cs(t) cs(k, t) = cs(k) * cs(t) [kBq/m² *Bq/kg] The model allows for a linear time trend and a linear-quadratic dose-response i.e. a nonlinear dependency of malformation risk on cesium burden. Result: The association of malformation rates with cesium burden is significant (P=0.014, F-test ). The dose-response curve is non-linear: neg. linear term (P=0.0092), pos. quadratic term (P=0.0023) from: Küchenhoff H, Engelhardt A, Körblein A, 2006 available at: http://www.ratical.org/radiation/Chernobyl/chernobylebook.pdf, pp 179-183

26

Combined spatial-temporal regression

Odds ratios (~relative risks) of malformation rates in Bavarian districts as a function of cesium burden, aggregated values and regression line. The error bars are standard deviations.

27

Summary

• In 1987, perinatal mortality is significantly increased in Poland and in Germany. Effect in Poland is 1.6 times greater than in Germany.

• 317 excess perinatal deaths in Germany and 354 in Poland in 1987

• Significant association of perinatal mortality in Germany and infant mortality in Poland with calculated cesium burden during pregnancy

• Curvilinear dose-response (negative linear, positive quadratic term)

• Significant association of malformation rates in Bavaria with cesium burden

• No “safe dose“ (threshold dose) for teratogenic effects

28