© International Epidemiological Association 2001
International Journal of Epidemiology 2001;30:125–129
Printed in Great Britain
Patterns of acute leukaemia occurrence
among children in the Chernobyl region
Andrey G Noshchenko,a Kirsten B Moysich,b Alexandra Bondar,a Pavlo V Zamostyan,a
Vera D Drosdovaa and Arthur M Michalekb
Background The Chernobyl nuclear accident of 1986 released large quantities of radioactive
material causing widespread contamination. In the Ukraine alone, more than 4
million people were exposed to radiation. The exact health consequences of this
exposure are still being assessed.
Methods
To ascertain the effect of in utero radiation exposure and the development of
leukaemia, a review was undertaken of leukaemia sub-types occurring among
children born in the year of the accident (1986) and followed 10 years postexposure. A comparison was made of leukaemia cumulative incidence rates among
children from both an exposed and unexposed oblast.
Results
Rate ratios (RR) for the all cell types grouping of leukaemia revealed that rates in
the exposed Oblast were significantly elevated for females, males and both genders
combined. Rates of acute lymphoblastic leukaemia (ALL) were dramatically
elevated for males and to a lesser extent for females. For both genders combined,
the RR for ALL was more than three times greater in the exposed compared to the
unexposed region.
Conclusion
Study results suggest that the increased risk of leukaemia and acute leukaemia
among those children born in 1986 and resident in radioactively contaminated
territories may be associated with exposure to radiation resulting from the
Chernobyl accident.
Keywords
Chernobyl accident, radiation, leukaemia, incidence
Accepted
21 June 2000
The Chernobyl radiation accident is undoubtedly the greatest
environmental catastrophe in the history of mankind. Vast areas
of Europe were contaminated by radioactive fallout. In the
Ukraine alone it is estimated that more than 4 million people
were exposed to radiation.1 While this accident happened over
a decade ago (26 April 1986), it will be several more decades
before the exact health consequences are known.
Cancers of the thyroid and leukaemia have been found to
be associated with radiation exposure. Those thought to be at
greater risk of untoward health consequences from exposure
are people under the age of 20, particularly those exposed in
the first decade of life.2–4 Data from both the Hiroshima and
Nagasaki (Japan) atomic bombing and also from follow-up studies
of nuclear tests in Utah (USA) indicate that the maximum
health effects occur during the first 12 years after exposure.3,4
a Research Center for Radiation Medicine, Academy of Medical Sciences
of Ukraine, Scientific Institute of Blood Transfusion, Ministry for Health
Protection of Ukraine.
b Roswell Park Cancer Institute, USA.
Correspondence: Arthur M Michalek, Dean of Educational Affairs, Roswell
Park Cancer Institute, Carlton and Elm Streets, Buffalo, NY 14263, USA.
E-mail: [email protected]
We assessed acute leukaemia cases occurring among children
who were in utero at the time of exposure to determine to what
degree, if any, these events were associated with the Chernobyl
accident. This paper presents data on acute leukaemia occurring
in such children in two regions of the Ukraine; one presumed
to have received radioactive contamination (Zhitomir) and one
presumed to have been unaffected (Poltava).
Methods
The Zhitomir and Poltava regions were selected on the basis
of radio-dosimetry studies conducted by the Research Center
for Radiation Medicine of the Academy of Medical Sciences of
the Ukraine. Zhitomir is situated to the east of the Chernobyl
nuclear power plant (Figure 1) and is rich in mineral resources.
It comprises 29 900 km2 and has a total population of 1 507 000.
Exposure of the Zhitomir population during the 10 post-exposure
years reached 11.15 thousands person Sv.1 Poltava is situated
on the left bank of the river Dnipro, hundreds of kilometres
southwest of the Chernobyl nuclear power plant. It covers
an area of 28 800 km2 and has a population of 1 771 000. The
control population in Poltava was unaffected by radiation contamination from Chernobyl.
125
126
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
occurring among children born in 1986 and living in the observed
territories for the period from 1986 to 1996; and Σn(t) is the
sum of annual population (children born in 1986) in observed
regions for the period from 1986 to 1996.
Cumulative incidence rates (CI) for the period were calculated according to the formula:
CI = Σij/Σnij
where: CI is the cumulative incidence rate; ij is the incidence of
acute leukaemia for each year in the observed period; and nj is
the number of children born in 1986 living in the given territory
for each year of observation.
Relative risk of disease was calculated as: R = CI1/CI2
Figure 1 Map showing oblasts under study in relation to Chernobyl
nuclear power plant (NPP)
Children born in the year of the Chernobyl accident (1986)
and living within these regions comprised the study populations. The 1986 birth cohort in Zhitomir was 24 231 and in
Poltava it was 23 567. The case group included children born in
Zhitomir during the year of the Chernobyl accident (1986) and
diagnosed with leukaemia 1986–1996.
Cases of leukaemia were identified through an exhaustive
manual search of records contained at the regional hospitals and
oncological centre archives, as well as the register of children’s
cancer-haematological pathology of the Institute of Haematology
of Ministry for Health Protection of Ukraine. Record-based
histopathological confirmation of disease status was available for
all cases included in these analyses. Acute leukaemia diagnoses
were further confirmed by an international group of expert
haematologists who reviewed the original slides as part of a larger
case-control study. The same methodology of case ascertainment
was employed in both oblasts to avoid bias. Moreover, members
of the Leukaemia Diagnostic Working Group were blinded to
child’s place of birth. We are thus confident that all leukaemia
cases were identified and assigned the correct diagnosis.
Data on annual average population numbers within the
regions were obtained from the Ministry of Statistics of Ukraine.
Cumulative effective radiation doses for the study populations
are those accumulated between 1986 and 1996 and were calculated using the methodical approach described by Likhtariov.5
Exposures to a variety of sources were considered in the calculations of the annual average effective radiation dose. These
included: external gamma-radiation from both the radioactive
cloud (at the early stage of accident) and radioactive fallout on
the ground; and internal radiation from radioisotopes of caesium
(134Cs, 137Cs), strontium (89Sr, 90Sr) and also transuranium
elements (238–240Pu, 241Am) transmitted via food consumption
and inhalation. The accumulated collective radiation dose during 11 post-catastrophe years is sum-total of annual collective
effective radiation dose for each year of observation.
Incidence rates (P) for the period of observation (1986–1996)
were calculated according to the formula 2:6
P = i/Σn(t) · 100 000
where: P is the incidence rate level from 1986 to 1996 (per
100 000 person-years); i is the incidence of acute leukaemia
where: R is the relative risk of leukaemia; CI1 is the cumulative
incidence rate of acute leukaemia in the given year of observed
period among children born in 1986 and living in the contaminated territories of the Zhitomir region; and CI2 is the
cumulative incidence rate of acute leukaemia in the given year
of observed period among children born in 1986 and living in
the control territories of the Poltava region.
The 95% confidence intervals for incidence rate ratios (RR)
were calculated as described by Rothman.7
Results
Between 1986 and 1996 a total of 21 cases of leukaemia were
observed in Zhitomir occurring among children born during
1986. A total of eight cases were observed in the largely uncontaminated Poltava region. The majority of leukaemias, as
would be expected in this age group, were acute lymphoblastic
leukaemias (ALL). These accounted for 13 (62%) of the
leukaemias in Zhitomir and 4 (50%) of those in Poltava.
According to our estimates the collective effective exposure
dose for the studied population, accumulated to 1996 is around
107 man-Sv. More than 80% of that dose was accumulated
during the first 5 years after the accident. Individual effective
exposure dose in the studied population is estimated in the
range of 0.1 and 200 mSv.
Data on the distribution of leukaemia cell types by region
and gender are presented in Table 1. The RR for the all cell
types grouping of leukaemia cases indicated that the Zhitomir
rates are significantly elevated for both females (2.3) and males
(2.7) and for both genders combined (2.7). The only distinct
type of leukaemia found to be significantly elevated in the
Zhitomir region was ALL. The Zhitomir rates compared to those
in Poltava for ALL were dramatically elevated for males (4.1)
and to a lesser extent for females (2.2). For the genders
combined, the RR for ALL is more than three times greater in
Zhitomir compared to Poltava. The relative excess of the all cell
types of leukaemia grouping is the result of significant differences
in the distribution of ALL.
In addition to the overall excess in leukaemia incidence in
Zhitomir compared to Poltava these differences also persisted
over time. Figure 2 presents patterns of cumulative incidence
ratios for all types of leukaemia combined, for the period 1986
through 1996. While there is some resemblance between the
patterns, there are obvious differences. Rates are higher for every
CHILDHOOD LEUKAEMIA AROUND CHERNOBYL
127
Table 1 Cumulative incidence rate ratios by acute leukaemia cell type, sex and region (1986–1996)
Region
Zhitomir
Acute leukaemia cell type
Lymphoblastic
Myeloblastic
Monoblastic
Undifferentiated
All leukaemia cell types
Poltava
Sex
No. of
cases
Cumulative
incidence rate
No. of
cases
Cumulative
incidence rate
Rate
ratio
95% confidence
intervals
Female
5
4.0
2
1.8
2.2
0.3–4.3
Male
8
6.2
2
1.5
4.1
1.2–14.0
Both
13
5.1
4
1.5
3.4
1.1–10.4
Female
2
1.6
1
0.9
1.8
0.2–19.9
Male
3
2.3
1
0.7
3.3
0.3–31.5
Both
5
2.0
2
0.7
2.9
0.614.7
Female
0
0.0
0
0.0
0.0
Male
1
0.8
1
0.7
1.1
0.2–7.8
Both
1
0.4
1
0.4
1.0
0.1–7.1
Female
1
0.8
0
0.0
0.0
Male
1
0.8
1
0.7
1.1
0.2–8.2
Both
2
0.8
1
0.4
2.0
0.2–31.8
Female
8
6.3
3
2.8
2.3
0.6–8.6
Male
13
10.1
5
3.7
2.7
1.6–4.7
Both
21
8,2
8
3.0
2.7
1.9–3.8
Incidence rate per 100 000 person-years.
Figure 2 Cumulative incidence rate for leukaemias and genders
combined in Zhitomir and Poltara regions
Figure 3 Cumulative incidence rate for acute lymphoblastic
leukaemias, genders combined in Zhitomir and Poltava regions
year in the Zhitomir region. Moreover, rates rise until about
5 to 6 years after the catastrophe and then begin to fall. Statistically significant differences for ALL among those children born in
1986 and living in the Zhitomir and Poltava regions are observed
in the period from 1990 up to 1996.
Figure 3 presents temporal trends by region and gender for
ALL. Significant differences in rates for lymphoblastic leukaemia
are observed for the years 1990–1996 for all children (sexes
combined). Among males, statistically significant differences
were observed for the years 1990, 1995 and 1996.
Table 2 presents cumulative incidence rates for leukaemia in
general and ALL by region and across two 5-year intervals. It is
of interest that while the rates in Zhitomir dropped between the
two periods, rates still remained appreciably higher than those
rates observed in Poltava region.
without such exposure. While this study is ecological in nature,
the results are highly suggestive of a radiation effect on children
born in the year of the Chernobyl accident. A number of studies
have been conducted to study possible associations of exposure
to Chernobyl and childhood leukaemia. Results, for a variety of
reasons, have been somewhat equivocal but some have yielded
results similar to ours. An observed increase in leukaemia cases
in Belarus was reported for the 7-year period following the
Chernobyl accident.8 Rates were 1.2 times higher after the
accident compared to the pre-accident period among the population living in territories with a level of radioactive pollution
exceeding 555 kBq per m2. This study also was ecological in
nature and exposure was based on that of the territory’s population and did not take into account age-sex distribution and
estimation of individual radiation doses.
The temporal trends from time of exposure to diagnosis of
leukaemia observed in our study correlate well with data reported
from Hiroshima and Nagasaki, and the US.3,4,9 The Life Span
Study sample of the Radiation Effects Research Foundation
for Hiroshima and Nagasaki observed an increased risk of
leukaemia 1 to 3 years after the bombings with peak occurrence
Discussion
Results from this study demonstrate a difference in incidence
rates of all leukaemias combined and for ALL between a region
with Chernobyl-related radiation contamination and a region
128
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Table 2 Cumulative incidence rates per 100 000, rate ratios (95% confidence intervals) for leukaemia all types and acute lymphoblastic
leukaemia by region and time, sexes combined
1987–1991
1992–1996
Rate ratio
11.2
4.4
2.6 (0.9–7.3)
5.7
0.8
7.1 (0.9–58.0)
1.9 (0.8–4.8)
5.5 (0.6–47)
Zhitomir
8.6
1.8
4.8 (1.1–22.6)
Poltava
3.3
0.0
–
Rate ratio
2.6
(0.8–8.3)
–
Leukaemia, all types
Zhitomir
Poltava
Rate ratio
Acute lymphoblastic leukaemia
6 to 7 years from exposure.3 In addition, younger age at time
of exposure was associated with greater risk of leukaemia. Our
results are remarkably similar with peak rates of leukaemia
occurring in 1991 with attenuation thereafter. However, these
data alone are not all that remarkable and may be partially
explained by the natural history of ALL which is known to peak
in early childhood and then decline by age 10. What is remarkable is the consistent difference between regions with Zhitomir
demonstrating higher rates (overall RR = 3.4).
Stevens et al. in their study of leukaemia and radioactive
fallout from the Nevada test site observed a significant association for acute leukaemias among those individuals age 20 or
younger at time of exposure.4 However, in further subgroup
analyses no association was observed for individuals exposed
in utero or during the first year of life. Conversely, Petridou reported
that in utero exposure to ionizing radiation from Chernobyl
resulted in more than a twofold risk of leukaemia in exposed
compared to unexposed children.10 Akiba et al. reported increased
infant mortality from leukaemia among children resident in
areas with high background radiation (1.7–2.0 mSev/year).11
However, this study was based on only three deaths and there
is some question as to the comparability of their reference
population. Michaelis et al. 12 attempted to replicate Petridou’s
study using data from western Germany, but after completing
detailed time-trend analyses failed to observe any effect of
radiation from Chernobyl on childhood leukaemia. Petridou13
speculated that this lack of replication might be due to nondifferential exposure misclassification, sparsity of data and questionable correspondence between environmental measurements
and personal exposures. Hjalmars et al.14 reported no significant
increase in the incidence of acute childhood leukaemias in
Sweden after the Chernobyl accident. Sali et al.15 report results
of a series of studies carried out in the exposed populations of
Europe. They too report no observed increases in childhood
leukaemia that may be attributed to Chernobyl. It is difficult to
assimilate these various reports into one meaningful summary
given that it is sometimes difficult to discern the validity of
individual reports owing to differences in cancer registration
systems, means of case ascertainment and very real differences
in Chernobyl-related exposures across countries. Moreover all
of these studies, including our own, are descriptive and as such
subject to the ecologic fallacy. Darby and Roman16 point out
some of the challenges in the conduct of leukaemia studies in
general and international comparisons resulting from Chernobyl
in particular.
The major limitation of the present study is that it is ecological.
While reliable estimates of presumed exposures are available
for the regions included, individual burdens through internal
and external exposures are not considered. Therefore, while our
results may support an association of perinatal exposure to
ionizing radiation and increased risk of leukaemia, other factors
may be at work.
What other ecological factors could influence leukaemia risk
among people living in radioactively contaminated territories?
One could imagine that other industrial exposures via pollution
may have had some effect. However, in the years immediately
after 1986 many industrial enterprises have reduced or even
ceased operation due to a series of economic crises which still
exist at present. Another bias could occur if there was significant
out-migration of exposed individuals (i.e. those presumably
with greater risk of ALL). However, this might have differentially increased the rates of ALL in the unexposed oblast and
thus reduced the magnitude of the difference that we observed.
At the very least, our observations might therefore be a conservative estimate of the effect.
In summary, study results suggest that the increased risk of
acute leukaemia among those children born in 1986 and resident
in radioactively contaminated territories may be associated with
exposure to ionizing radiation resulting from the Chernobyl
accident. The increase in rates of acute leukaemia was primarily
due to the lymphoblastic cell type. This effect was more evident
among males than females. The observed pattern of acute
leukaemia occurrence from 1986 to 1996 is similar to findings
from the A-bomb survivors’ study and that of residents downwind from nuclear test sites in US. Finally, while risk of acute
leukaemia among the exposed population was distributed nonuniformly during the 11 years after the Chernobyl accident,
perhaps due to sample size and relative rarity of the event, the
peak risks were predictably noted approximately 3 to 6 years
post-accident.
References
1 Ten Years After the Accident at the Chernobyl NPP //National Report of
Ukraine, 1996. Kyiv: Ìinchernobyl, 1996.
2 International Commission on Radiological Protection. Radiation
Protection 1990: Recommendations of the International Commission on
Radiological Protection (ICRP). Publication 60. New York: Pergamon
Press, 1991 (in Russian).
CHILDHOOD LEUKAEMIA AROUND CHERNOBYL
3 Shimizu Y, Schull J, Kato H. Cancer risk among atomic bomb
survivors. JAMA 1990;264:601–04.
129
10 Petridou E, Trichopoulos D, Dessypris N et al. Infant leukemia after in
utero exposure to radiation from Chernobyl. Nature 1996;382:352–53.
4 Stevens W, Thomas DC, Lyon JL et al. Leukemia in Utah and radio-
11 Akiba S, Quanfu S, Zufan T et al. Infant leukemia mortality among the
active fallout from the Nevada test site. Case-control study. JAMA
1990;264:585–91.
residents in high-background radiation areas in Guangdong, China.
In: Wei L, Sughara T, Tao Z (eds). High Levels of Natural Radiation.
Amsterdam: Elsevier Science, 1977.
5 Likhtariov IA et al. Retrospective prognosis for doses of population
exposure and general 1997 dosimetry determination for the settlements of Ukraine affected by radioactive contamination as a result of
the accident at ChNPP. Generalized data for 1986–1997. Issue 7. RCRM
AMSU (in Ukrainian).
6 Breslow NE, Day NE. Statistical Methods in Cancer Research. Vol. 1. The
Analysis of Case-Control Studies. Lyon: IARC, 1994, pp.129–36.
7 Rothman KJ. Modern Epidemiology. Boston: Little, Brown and Co.,
1986.
8 Konoplya EF, Rolevich IV (eds). The Chernobyl Accident Consequences in
the Republic of Belarus. National Report. Minsk, 1996, pp.45–46.
9 Yashimoto Y. Cancer risk among children of atomic bomb survivors.
JAMA 1990;264:595–600.
12 Michaelis J, Kaletsch U, Burkart W, Grosche B. Infant leukemia after
the Chernobyl accident. Nature 1997;387:246.
13 Petridou E, Trichopoulos D, Dessypris N et al. Response to Michaelis
et al. Nature 1997;387:246.
14 Hjalmars U, Kulldorff M, Gustafsson G. Risk of acute childhood
leukemia in Sweden after the Chernobyl reactor accident. BMJ
1994;309:154–57.
15 Sali D, Cardis E, Sztanyik L et al. Cancer consequences of the
Chernobyl accident in Europe outside the former USSR: a review.
Int J Cancer 1996;67:343–52.
16 Darby SC, Roman E. Links in childhood leukemia. Nature 1996;
382:303–04.