RADIATION RESEARCH 162, 505–516 (2004)
? 2004 by Radiation Research Society.
All rights of reproduction in any form reserved.
Lung Cancer in Mayak Workers
E. S. Gilbert,a,1N. A. Koshurnikova,bM. E. Sokolnikov,bN. S. Shilnikova,bD. L. Preston,cE. Ron,aP. V. Okatenko,b
V. F. Khokhryakov,bE. K. Vasilenko,dS. Miller,eK. Eckermanfand S. A. Romanovb
aRadiation Epidemiology Branch, National Cancer Institute, Bethesda Maryland;bSouthern Urals Biophysics Institute, Ozyorsk,
Chelyabinsk Region, Russia;cStatistics Department, Radiation Research Foundation, Hiroshima, Japan;dDepartment of Radiation Safety,
Mayak Production Association, Ozyorsk, Russia;eDivision of Radiobiology, University of Utah, Salt Lake City, Utah; andfLife Sciences Division,
Oak Ridge National Laboratory, Oak Ridge, Tennessee
Gilbert, E. S., Koshurnikova, N. A., Sokolnikov, M. E., Shil-
nikova, N. S., Preston, D. L., Ron, E., Okatenko, P. V., Khokh-
ryakov, V. F., Vasilenko, E. K., Miller, S., Eckerman, K. and
Romanov, S. A. Lung Cancer in Mayak Workers. Radiat. Res.
162, 505–516 (2004).
The cohort of nuclear workers at the Mayak Production
Association, located in the Russian Federation, is a unique
resource for providing information on the health effects of
exposure to plutonium as well as the effects of protracted ex-
ternal dose. Lung cancer mortality risks were evaluated in
21,790 Mayak workers, a much larger group than included in
previous evaluations of lung cancer risks in this cohort. These
analyses, which included 655 lung cancer deaths occurring in
the period 1955–2000, were the first to evaluate both excess
relative risk (ERR) and excess absolute risk (EAR) models
and to give detailed attention to the modifying effects of gen-
der, attained age and age at hire. Lung cancer risks were
found to be significantly related to both internal dose to the
lung from plutonium and external dose, and risks were de-
scribed adequately by linear functions. For internal dose, the
ERR per gray for females was about four times higher than
that for males, whereas the EAR for females was less than
half that for males; the ERR showed a strong decline with
attained age, whereas the EAR increased with attained age
until about age 65 and then decreased. Parallel analyses of
lung cancer mortality risks in Mayak workers and Japanese
A-bomb survivors were also conducted. Efforts currently un-
der way to improve both internal and external dose estimates,
and to develop data on smoking, should result in more accu-
rate risk estimates in the future.
? 2004 by Radiation Research Society
During the early period of operation of the Mayak nu-
clear facility, which is located in the Chelyabinsk region of
the Russian Federation, many workers were exposed to in-
haled plutonium at levels much higher than those consid-
1Address for correspondence: Radiation Epidemiology Branch, Divi-
sion of Cancer Epidemiology and Genetics, 6120 Executive Blvd., MS
7238, Rockville, MD 20852; e-mail: email@example.com.
ered permissible today. A large number of these workers
were also exposed to doses of external ? radiation that were
substantially higher than current occupational dose limits.
Although workers exposed to plutonium in facilities in the
United States and United Kingdom have been studied (1–
4), the level of the exposures and the small number of
workers who have received such exposure greatly limit
what can be learned about plutonium-related health effects.
Because of these limitations, quantitative estimates of risks
from exposure to plutonium have been obtained either from
studies of persons exposed to other ?-particle-emitting ra-
dionuclides or by applying a radiation weighting factor to
estimates obtained from Japanese atomic bomb survivors
exposed to low-LET radiation (5, 6). The Mayak worker
cohort is a unique resource for providing information on
the health effects of exposure to plutonium as well as the
effects of protracted external dose.
Several papers have evaluated cancer risks in Mayak
workers. Recently Shilnikova et al. (7) analyzed risks of
solid cancer and leukemia with emphasis on the effects of
external exposure. From both human and experimental an-
imal data, it is known that the lung, bone and liver receive
the largest doses from inhaled plutonium (5, 6). Koshur-
nikova et al. (8) and Gilbert et al. (9) conducted a prelim-
inary evaluation of risks of bone and liver cancer in relation
to plutonium exposure, and several previous papers have
evaluated lung cancer risks. These include both case–con-
trol (10, 11) and cohort studies (12–14). The cohort studies
focused on a relatively small subgroup of male workers
initially employed before 1959 for whom quantitative es-
timates of doses to the lung from plutonium were available.
The current paper provides a comprehensive evaluation
of lung cancer risks with detailed attention to both internal
dose to the lung from plutonium and external dose. Unlike
previous papers on lung cancer risks, these analyses include
all workers who were initially employed in the period
1948–1972 in either the main or auxiliary plants, explore
both excess relative risk and excess absolute risk models,
and evaluate modification of risk by gender, attained age,
age at hire, and time since exposure.
GILBERT ET AL.
Number of Mayak Workers and Lung Cancer Deaths (in parentheses) by Plant, Plutonium Monitoring Status,
Sex, Year of Hire, Age at Hire, and External Dose: Mean External and Internal Lung Dose by Sex, Year of
Hire, Age at Hire, and External Dose
Radiochemical and plutonium plants
lung dose among
Total21,790 (655) 2582 (54)4493 (131)8856 (252)5859 (218)0.26 0.80
By year of hire
By age at hire
By external dose
aAmong those monitored for external dose.
MATERIALS AND METHODS
This record-based epidemiological study required no contact with the
cohort members. The project was reviewed and approved by the Insti-
tutional Review Boards of the Southern Urals Biophysics Institute and
the Radiation Effects Research Foundation. It was exempted from such
review at the National Cancer Institute (NCI) because NCI used only
The Study Population and Follow-up
The Mayak worker cohort and methods of follow-up have been de-
scribed in detail elsewhere (7, 15). The main plants of the Mayak nuclear
facility, which began operations in 1948, include nuclear reactors, a ra-
diochemical plant, and a plutonium production facility, but only workers
in the latter two facilities have the potential for exposure to plutonium.
The original cohort included about 18,000 persons who were initially
employed in one or more of these main plants in the years 1948–1972.
About 2800 persons who worked only in the auxiliary plants (water treat-
ment facility and mechanical repair plant) were recently added to the
Mayak cohort to expand the number of workers with little or no radiation
exposure. Table 1 shows the distribution of the 21,790 workers and 655
lung cancer deaths in the expanded cohort by sex, year of hire, age at
hire, and external dose, and also by plant and whether or not they were
monitored for plutonium. Also shown are mean internal lung doses for
those monitored for plutonium (see below). About 24% of the workers
were female, 43% were hired before 1954, and 37% were under age 20
at hire. Vital status is known for 90% of the cohort. By December 31,
2000, 8493 workers had died, and the cause of death was known for 8213
(97%) of these deaths.
The Mayak Workers Registry includes annual external doses estimated
from individual film badge monitoring data, maintained since 1948 by
the Radiation Safety Service of the Mayak plant. The external monitoring
program was designed to include all workers with a potential for external
exposure, and about 80% of workers in the registry have external mon-
Monitoring for plutonium exposure has been carried out at the South-
ern Urals Biophysics Institute (SUBI) since the late 1960s. Estimates of
body burden and of doses from plutonium used in this paper are based
on plutonium levels observed in urine collected over a 3-day period after
a period of at least 30 days with no potential exposure (usually after
vacation). This procedure avoided the influence of short-term clearance
processes on the urinary excretion, but it results in some distortion of the
plutonium kinetics (16). Although about 500 workers were monitored
before 1970, the median date of monitoring was 1982. As of the end of
1995, about 31% of those who worked in the radiochemical or plutonium
production plants had been monitored using this approach; extensive ef-
forts beginning in 1996 to monitor additional workers have raised this to
40%. Scientists at SUBI formulated a mathematical model of the behavior
of plutonium in the body based on their measurements of plutonium ?-
particle activity in urine and in body tissues at autopsy. In addition, the
model considers the worker’s occupational history and the physiochem-
ical form of the plutonium aerosols (17–19). This model has been used
to estimate annual equivalent dose to organs for each year of follow-up.
Analyses in this paper are based on the estimated dose to the lung, which,
for simplicity, is often referred to as ‘‘internal dose’’. As a result of
collaborative work of Russian and U.S. dosimetrists, improvements have
been made in the original SUBI model (20), and lung dose estimates used
in this paper reflect these improvements.
LUNG CANCER IN MAYAK WORKERS
Number of Mayak Workers, Number Monitored for Plutonium, Distribution by Estimated Internal Lung Dose
from Plutonium, Mean Internal Lung Dose, and Mean Body Burden, by Plutonium Surrogate Categorya
Plutonium surrogate category
Not monitored for plutonium
Total monitored (percent)
By estimated internal lung dose
Below detection limit
Mean internal lung dose (Gy)
Mean body burden (kBq)
aPlutonium surrogate categories are as follows: 0 ? reactors and auxiliary plant workers hired in any period; 1 ? main departments of the plutonium
plant, hired 1964–1972, or auxiliary departments of the plutonium plant hired 1959–1972 or radiochemical plant, hired 1954–1972; 2 ? main
departments of the plutonium plant, hired 1959–1963 or auxiliary departments of the plutonium plant, hired 1950–1958 or radiochemical plant, hired
1948–1953; 3 ? main departments of the plutonium plant, hired 1954–1958 or auxiliary departments of the plutonium plant, hired 1948–1949; 4 ?
main departments of the plutonium plant hired 1950–1953; 5 ? main departments of the plutonium plant hired 1948–1949.
bCategories 4 and 5 were combined in dose–response analyses due to the small number of workers in these categories.
To make it possible to use the full cohort including workers in the
radiochemical and plutonium plants for whom plutonium monitoring data
are not available (for the purposes of evaluating the effects of external
exposure), we developed a categorical surrogate index of plutonium ex-
posure based on occupational history data, including work locations,
starting dates, measured body burden values, and expert knowledge of
working conditions at various times in the different facilities (7). Table
2 shows the definitions of the plutonium surrogate categories used by
Shilnikova et al. (7) and in this paper along with the mean internal lung
doses and body burdens among workers in each category who were mon-
itored for plutonium. The table also shows the distribution by lung dose
of monitored workers in each category. Both lung dose and body burden
increase with surrogate category. However, the distributions by lung dose
make it clear that there is considerable variability among monitored work-
ers within a given surrogate index level. Because workers thought to have
been at risk of exposure to the highest levels of plutonium were more
likely to be selected for monitoring, the mean lung dose for monitored
workers within a surrogate category cannot be regarded as a representa-
tive value for all workers in that category. Thus unmonitored workers in
surrogate categories 1–5 are not considered to have internal lung doses
that could be estimated and are not used for quantifying the plutonium
dose response. However, the surrogate makes it possible to include the
full cohort for evaluating the effects of external dose and for estimating
the attributable risk from plutonium exposure.
The statistical methods employed in this paper are similar to those used
in recent analyses of the Mayak worker data by Shilnikova et al. (7).
Analyses were based on Poisson regression methods, where it is assumed
that the number of deaths from the cause of interest is a Poisson variable
with mean given by the product of the person-years and the cause-specific
mortality rate for each cell of a multi-way person-year table. Analyses
were implemented with the AMFIT module of the software package EP-
Person-year tables were classified by plant (auxiliary plants, reactors,
radiochemical plant, plutonium production plant), gender, plutonium sur-
rogate index categories (see Table 2), attained age (5-year categories),
age at hire (5-year categories), birth cohort (1885–1914, 1915–1924,
1925–1934, 1935–1955), calendar period (1948–1972, 1973–2000), plu-
tonium monitoring status, cumulative internal dose to the lung, and cu-
mulative external dose. Supplementary analyses with finer stratification
on birth cohort and calendar period also were conducted. Attained age
and calendar year refer to a worker’s age or calendar year in a specified
follow-up interval, and a given worker can contribute to several catego-
ries as he/she is followed over time. Plant, the plutonium surrogate, plu-
tonium monitoring status, and cumulative doses were also treated as de-
pendent on time. For plant and the plutonium surrogate, person-years
were classified according to the most dangerous plant (in the order listed
above) or the highest surrogate category the person had ever worked in
5 years prior to the time at risk. Because of indications that some workers
were monitored for plutonium as a result of suspected diseases, person-
years were classified as unmonitored until 2 years after the initial mon-
itoring date. Of the 5860 radiochemical and plutonium plant workers (218
lung cancer deaths) indicated in Table 1 as monitored, 176 workers (29
lung cancer deaths) were monitored in the last 2 years of follow-up and
thus considered as unmonitored in dose–response analyses. Most analyses
were based on dose received 5 or more years before the time at risk.
There were 14 categories for lagged cumulative internal lung dose, a zero
dose category and 13 other categories with boundaries of 0.2, 0.5, 0.75,
1, 1.5, 2, 2.5, 3, 3.5, 4, 5 and 6 Gy; unmonitored person-years were
classified according to the plutonium surrogate. The same categories were
used to classify cumulative external dose with an additional category for
unmonitored workers. Because there was no evidence that workers who
were not monitored for external exposure had risks that differed from
workers with zero dose, these categories were combined into a single
zero dose category for most analyses. Some analyses included evaluation
of internal dose in exposure ‘‘windows’’ of 5–15 years, 15–25 years, and
25? years prior to the time at risk; categories for these dose windows
were the same as those for the 5-year lagged dose.
Analyses based on both excess relative risk (ERR) and excess absolute
(EAR) models were conducted, although the ERR model analyses were
more extensive. The age-specific risk, ? (a, s, b, c, z, d), where a is
attained age, s is gender, b is birth cohort, c is calendar year period, h is
age at hire, and d is dose, is defined as follows for the two models.
Excess relative risk model:
GILBERT ET AL.
?(a, s, b, c, h, d) ? ? (a, s, b, c)[1 ? ERR(a, s, h, d)]
Excess absolute risk model:
?(a, s, b, c, h, d) ? ? (a, s, b, c) ? EAR(a, s, h, d)
The logarithm of the baseline hazard ?0(a, s, b, c) was modeled as a
linear-quadratic function of attained age and gender, birth cohort (four
categories), and calendar period (two categories). In addition, a parameter
that allowed for differences in mortality for male main and auxiliary plant
workers was included since there was evidence that male auxiliary plant
workers had higher baseline rates. This model for the baseline risk was
chosen after exploration of several alternative functions, including the
use of gender-specific functions of attained age, finer categories for birth
cohort and calendar year, and, for the ERR models, stratification on age,
gender, and birth cohort or calendar period. These alternatives gave re-
sults for the ERR and EAR that are very similar to those reported pre-
ERR is the excess relative risk function and EAR is the excess absolute
risk function, in which d involves both external and internal lung doses.
The dose response model that is emphasized is as follows
? ? I
where ?plu,Sand ?ext,Sare sex-specific parameters describing the respective
external and internal dose–response slopes. The first two terms in Eq. (1)
show linear functions of the time-dependent lagged internal lung (dplu)
and external doses (dext) given in grays. With ERR models, the parameters
?plu,Sand ?ext,Sindicate the ERR per gray, whereas with the EAR models,
these parameters indicate the excess absolute risk expressed as excess
deaths per 10,000 person-year Gy (PY-Gy). We also explored linear-qua-
dratic functions, but in no case did the addition of quadratic terms for
either internal or external dose significantly improve the fit of the model.
For some analyses, conducted for descriptive purposes, categories of dose
replaced the continuous variables. We also conducted analyses in which
the term ?pludpluwas replaced by the sum ?5–15d5–15? ?15–25d15–25?
?25?d25?, where d5–15, d15–25, and d25?indicate the internal dose to the
lung in the respective exposure windows: 5–15, 15–25 and 25? years
from exposure. In these analyses, the sex ratio was assumed to be the
same for all three windows.
The factors exp(?k?plu,kzk), exp(?k?ext,kzk), and exp(?k?jkwjk) allow
for modification of exposure effects by variables (zkand wjk) that included
the logarithm of attained age, the square of this variable, and age at hire.
Although tests of the need for each of these variables were conducted,
final models were more selective as described in the Results section.
The ?jdesignate the excess relative or absolute risks for categories
(indexed by j) of the plutonium surrogate index among the unmonitored,
with categories 4 and 5 (Table 2) combined. The surrogate index was
used for periods during which there was no plutonium monitoring data
(Iunmon? 1) while internal doses to the lung (dplu) were used during post-
monitoring periods (Imon? 1) for monitored workers. As noted above,
persons are treated as unmonitored for the first 2 years after the initial
monitoring date. Unmonitored workers in the auxiliary and reactor plants
were treated as monitored with dplu? 0 throughout their follow-up. Pa-
rameters associated with plutonium surrogate categories were constrained
to be non-negative, except for the purpose of showing confidence inter-
vals for these parameters. We also evaluated whether the coefficient ?ext
depended on plutonium monitoring status; this was done by replacing the
term ?extdextwith ?ext1dextImon? ?ext2dextIunmonand testing whether ?ext1
? ?ext2. For these analyses, it was assumed that the external dose coef-
ficients were the same for the two sexes.
In all cases, parameter estimates were computed with maximum like-
lihood methods. Hypothesis tests and confidence intervals were based on
likelihood ratio tests and direct evaluation of the profile likelihood. Two-
sided P values are used throughout.
In addition to parameter estimates, we present estimates of the expected
and excess cases, with the excess apportioned between internal and ex-
ternal exposures derived from the fitted models. These are calculated as
described by Shilnikova et al. (7).
For the purpose of comparing findings from this study with those based
on the Life Span Study (LSS) cohort of Japanese atomic bomb survivors,
we analyzed LSS data for the follow-up period 1950–1997 using the data
set that forms the basis of analyses in Preston et al. (22) and made avail-
able by the Radiation Effects Research Foundation (RERF). The LSS
cohort includes a large proportion of atomic bomb survivors who were
within 3 km of the hypocenters at the time of the bombings, and a similar-
sized age- and sex-matched sample of people who were between 3 and
10 km from the hypocenters and is described in more detail by Preston
et al. (22). For comparability with Mayak workers, our analyses were
restricted to persons exposed between the ages of 15 and 60. Baseline
risks for the LSS cohort were modeled as described by Preston et al.
Table 3 shows the results of fitting models of the form
shown in Eq. (1), in which the ERRs (column 2) or EARs
(column 3) are linear functions of internal lung dose and
external dose. Figure 1 depicts the dependence of the ERR
and EAR on attained age. Because the ERRs and EARs are
expressed on different scales for internal dose, external
dose, and the plutonium surrogate, these are shown relative
to their values at age 60. Both models indicate highly sig-
nificant associations of lung cancer risks with both internal
and external dose. No significant improvement in fit was
brought about by adding quadratic terms for internal or ex-
ternal dose (P ? 0.4 for both models and both doses). Both
the ERR and EAR for internal dose depend on gender and
attained age, although the nature of these dependences was
different for the two models. Among those whose internal
doses could not be estimated, risk increased with the or-
dered plutonium surrogate categories. The sections that fol-
low discuss findings based on each of the models.
Results of Fitting Excess Relative Risk (ERR) Model
Lung cancer risk was found to be significantly associated
with internal lung dose for both sexes (P ? 0.001). The
ERR per gray for females was estimated to be about 4.0
(95% CI: 1.9; 8.8) times that for males. There was strong
evidence of a decline in the ERR with attained age (P ?
0.004), which is depicted in Fig. 1. Risks at ages younger
than 60 would be larger than those shown, while those at
older ages would be smaller. For example, at age 50 the
ERR per gray would be about 1.8 times the values shown,
whereas at age 70 the ERR per gray would be about 0.6
times the values shown. We also evaluated categories of
attained age. The male ERR per gray for internal dose for
attained-age categories ?55, 55–64, 65–74 and 75? years
with their 95% CI were respectively 7.5 (3.9; 13), 5.1 (3.3;
7.5), 2.5 (1.2; 4.3), and 0.9 (?0; 5.4). There was no evi-
dence that the ERR per gray depended on age at hire (P ?
LUNG CANCER IN MAYAK WORKERS
Estimates of the Excess Relative Risk and the Excess Absolute Risk for Lung Cancer Mortality with 95%
Confidence Intervals (CI) for Internal Dose to the Lung from Plutonium, External Dose, and Plutonium
risk (ERR) model
Internal lung dosea
ERR per gray at attained age 60 Excess deaths per 104PY-Gy at attained age 60
Main effect per gray:
ERR per gray (all attained ages)
Excess deaths per 104PY-Gy at attained age 60
Main effect per gray:
Plutonium surrogate categoriesb,c
Category 4: Males
Category 4: Females
Ratio of female and male coefficients:
0.17 (0.052; 0.32)
0.32 (?0; 1.3)
ERR at attained age 60 (both sexes)
0.025 (?0; 0.17)
0.25 (0.08; 0.50)
0.54 (0.20; 1.0)
11 (4.7; 25)
2.4 (0.56; 4.4)
0.43 (?0; 1.6)
Excess deaths per 104PY at attained age 60 (malesd)
2.8 (0.1; 7.5)
8.3 (2.6; 17)
0.44 (0.2; 0.9)
aBased on persons who were monitored for plutonium or worked only in reactor or auxiliary plants.
bBased on persons who worked in radiochemical or plutonium plant and were not monitored for plutonium.
cSee footnote to Table 2 for definitions of the plutonium surrogate categories.
dEstimates for females are 0.44 times these values.
eExcept for category 4, there was no evidence that these ERR differed by gender.
FIG. 1. Attained-age effects for internal lung dose, external dose, and plutonium surrogate. The panel on the left shows ERR per gray (ERR for
surrogate) for lung cancer mortality as a function of attained age relative to its value at age 60. The respective functions for internal dose and the
plutonium surrogate are exp[–3.2 log (a/60)] and exp[–7.3 log (a/60)], where a is attained age in years. For external dose, the function would be exp[–
0.16 log (a/60)], which is nearly indistinguishable from the function with no modification by attained age depicted in the figure. The panel on the right
shows analogous functions of the EAR expressed as excess deaths per 104PY-Gy (PY for surrogate). The respective functions for internal dose,
external dose, and the plutonium surrogate are exp [2.6 log (a/60)?22 log2(a/60)], exp [5.7 log (a/60)], and exp [–1.5 log (a/60)?13 log2(a/60)].
There was no evidence that the ERR per gray for external
doses depended on gender (P ? 0.5), attained age (P ?
0.5), or age at hire (P ? 0.38). However, because of the
known difference in baseline lung cancer rates for the two
sexes, we nevertheless provide separate estimates for males
and females; the ERR per gray for females was estimated
to be 1.9 (95% CI: ?0; 11) times that for males. Based on
a model with no modification by gender or attained age,
the P value for the association with external dose was
? 0.001; for males alone, this P value was 0.002.
Although Table 3 and Fig. 1 show different dependences
on gender and attained age for internal and external dose,
in fact they did not differ significantly. In a model in which
the modifying effects of gender and attained age were as-
sumed to be the same for internal and external dose, the
estimated ratio of the coefficients for internal (?-particle)
and external (?-ray) dose was 33 (95% CI: 14; 98). This
can be regarded as an estimate of the relative biological
effectiveness (RBE) of internal (?-particle) dose compared
with external (?-ray) dose.
GILBERT ET AL.
Observed and Expected Deaths from Lung Cancer, and Estimated Excess Deaths Associated with Internal and
Internal ?-particle dose
Estimated as zerob
Estimated as positivec
Could not be estimatedd
Internal ?-particle dose
Estimated as zerob
Estimated as positivec
Could not be estimatedd
Note. Percentages are given in parentheses.
aDeaths that would have occurred in the absence of external or internal radiation exposure.
bPrimarily persons who worked only in reactor or auxiliary plants.
cPrimarily persons who worked in the radiochemical or plutonium plant and were monitored for plutonium.
dWorked in radiochemical or plutonium plant and not monitored for plutonium.
eOf these, 29 lung cancer deaths (27 males, 2 females) were monitored in the last 2 years of follow-up but were considered as unmonitored in
Among workers who were not monitored for plutonium,
statistically significant elevated risks were found in all but
the lowest of the plutonium surrogate categories (P ? 0.005
for category 2; P ? 0.001 for categories 3 and 4), with the
largest risks among females in the highest category. For the
other categories, the ERR did not differ by sex (P ? 0.5
in all cases). The ERR showed a decline with attained age
that was even stronger than that estimated for monitored
workers (Fig. 1). There was no evidence of modification
by age at hire (P ? 0.35).
Table 4 shows the predicted number of deaths associated
with internal and external exposure and also shows the es-
timated number that would have occurred in the absence of
exposure. In males, about 24% of the deaths are attributed
to internal exposure and about 10% to external exposure.
In females, about 57% of the deaths are attributed to inter-
nal exposure and about 8% to external exposure. For both
males and females, the percentage of deaths attributed to
internal exposure is higher among workers in the radio-
chemical and plutonium plants who were monitored for
plutonium than for those who were not. This probably
comes about at least in part because those with the highest
doses were more likely to be monitored. In addition, the
excess among the unmonitored workers is more likely to
be underestimated because the surrogate index is not as
reliable a measure of exposure as lung dose.
Results of Fitting the Excess Absolute Risk (EAR) Model
The third column of Table 3 shows estimated parameters
for the EAR model. For internal lung dose, the ratio of EAR
per 104PY-Gy for females to that for males was 0.43 (95%
CI: 0.24; 0.72). The comparable ratio for external dose was
estimated, with considerable uncertainty, to be 0.18 (95%
CI: ?0; 1.1). For internal dose, both the logarithm of at-
tained age and its square were needed to describe the de-
pendence on attained age, and this resulted in the EAR
increasing to about age 65 and then decreasing as depicted
in Fig. 1. For external dose, the dependence could be de-
scribed with the logarithm of attained age alone. Although
Fig. 1 seems to indicate a markedly different pattern for
internal and external dose at older ages, the pattern for ex-
ternal dose could not be estimated with certainty, and there
was no evidence that the two patterns differed significantly.
There was no evidence of dependence on age at hire for
internal dose (P ? 0.5). There was, however, modest evi-
dence that the EAR for external dose might decrease with
increasing age at hire (P ? 0.045), although we did not
include this dependence in our model.
Among workers whose internal doses could not be esti-
mated, statistically significant elevated EARs were found
in the three highest plutonium surrogate categories (P ?
0.042 for category 2; P ? 0.001 for categories 3 and 4).
Modification by gender was similar to that among those
whose internal doses could be estimated, with the ratio of
EARs for females and males estimated to be 0.44 (95% CI:
0.20; 0.90). The EAR depended on attained age, and in this
case increased until about age 55, and then decreased as
depicted in Fig. 1. There was also evidence of a decrease
in the EAR with age at hire (P ? 0.007), and this was
included in the Table 3 model.
LUNG CANCER IN MAYAK WORKERS
Numbers of Person-Years, Lung Cancer Deaths, and Relative Risks of Lung Cancer Mortality (with 95% CI)
by Categories of Internal Dose to the Lung
risk (95% CI)
risk (95% CI)
1.4 (1.0; 1.8)
2.4 (1.5; 3.6)
10.1 (6.3; 15)
19 (9.5; 35)
33 (14; 67)
0.91 (?0.91; 3.1)
15 (3.0; 38)
7 250(110; 660)
bIncludes only person-years and deaths of those who were monitored for plutonium or worked only in reactor or auxiliary plants. The person-years
and deaths not shown in this table were included in the analysis by estimating parameters for the ERR for each plutonium surrogate category as in
Results of Lung Cancer Mortality Analyses Including Only Mayak Workers Who Were Monitored for
Plutonium or Worked Only in the Reactor or Auxiliary Plantsa
(374 lung cancer
Workers in the main
plants who were hired before 1959 with information
on smoking (278 lung cancer deaths)
Not adjusted for smokingAdjusted for smoking
9.6 (5.7; 17)
Relative risk for smokingb
ERR per gray for internal dose
4.2 (2.8; 6.0)
22 (9.5; 56)
ERR per gray for external dose
Both sexes 0.10 (?0; 0.29)0.065 (?0; 0.25)0.027 (?0; 0.18)
aWorkers in the radiochemical or plutonium plants who were not monitored for plutonium were excluded from these analyses.
bThis is the relative risk of lung cancer for workers who smoked relative to workers who did not smoke.
Additional Results Based on the ERR Model
To further explore the dose response, we estimated the
ERR for each of five categories of internal lung dose. These
are shown in Table 5. For both males and females, risks
increased with increasing lung dose, and risks were signif-
icantly elevated in all but the lowest positive dose category
for females. In addition, we conducted analyses of data in
restricted dose ranges. Evidence of a statistically significant
response was found when analyses were restricted to inter-
nal doses less than 1 Gy (P ? 0.001) and to internal doses
less than 0.5 Gy (P ? 0.044) but not when restricted to
internal doses less than 0.2 Gy, although the risk coefficient
was still positive. The failure to detect elevated risks may
result simply from the reduced statistical power of these
We also investigated the possible modifying effect of
time since exposure for internal dose. In a model with no
modification by attained age, the estimated ERR per gray
(with 95% CI) for internal dose received 5–15, 15–25 and
25? years before the time at risk were respectively ?0.03
(?0; 12), 10.7 (3.8; 16), and 2.2 (0.2; 4.4), and this model
fitted the data somewhat better (P ? 0.044) than a model
with a single coefficient. However, with attained age in the
model, there was no evidence that the ERR per gray dif-
fered by time since exposure (P ? 0.38), although the es-
timates for the three windows showed a similar pattern
(?0.01, 8.2 and 4.4). Furthermore, the addition of attained
age improved the fit of the model even when the ERR per
gray was estimated separately for the three time-since-ex-
posure periods indicated above (P ? 0.048). The estimates
for the three windows suggest that 15-year lagged dose
might fit the data better than 5-year lagged dose; however,
in fact, the deviances for the two models were similar. With
a 15-year lag, the estimated ERR per gray for males at
attained age 60 was 6.0 (95% CI: 4.1; 8.5).
To investigate whether the inclusion of workers in the
radiochemical and plutonium production plants who were
not monitored for plutonium might be distorting results, we
conducted analyses that excluded these workers and person-
years with results shown in the second column of Table 6.
Because of the sparse data on external dose in this restricted
data set, estimates of the ERR per gray for external dose
were based on the two sexes combined. For internal dose,
results are similar to those based on the full cohort (see
GILBERT ET AL.
Estimated Parameters for Lung Cancer Mortality Risks with 95% Confidence Intervals (CI) for Internal Lung
Dose in Mayak Workers and External Lung Dose in the Life Span Study (LSS) Cohort of Japanese Atomic
Bomb Survivors Exposed between the Ages of 15 and 60
ERR per sievertb
at attained age 60c
LSS cohort age 15–60 at exposure
ERR per sievert
at attained age
Total 306,505374 1,797,201 1,130
0.23 (0.16; 0.33)
0.93 (0.46; 1.9)
4.0 (1.9; 8.8)
0.40 (0.032; 0.86)
1.40 (0.76; 2.2)
3.6 (1.2; 11)
By attained age
Excess deaths per
Excess deaths per
104PY-Sv (95% CI)
Under 55 years
1.4 (0.77; 2.3)
5.1 (3.5; 7.1)
5.2 (2.6; 8.6)
1.4 (?0; 12)
1.5 (?0; 3.8)
7.2 (3.4; 11.9)
14.3 (6.6; 23.7)
aIncluding only person-years and lung cancer deaths where internal lung doses could be estimated; however, all workers and person-years were
included in the analyses.
bSieverts were calculated by dividing the dose in grays by a quality factor (QF) of 20.
cBased on a model in which the coefficient of the logarithm of attained age was set equal to ?2.2.
dEAR for females would be a factor of 0.43 smaller than these estimates.
Table 3). However, for external dose, the ERR per gray was
smaller than estimates based on the full data set and did
not differ significantly from zero (P ? 0.12).
To further investigate this, we conducted an analysis in-
cluding all subjects but with separate estimates of the ex-
ternal dose ERR per gray for those with internal doses that
could be estimated and for the remainder. The two estimates
(for both sexes combined) were very similar: 0.16 (95%
CI: 0.03; 0.35) and 0.18 (95% CI: 0.04; 0.37), respectively.
The estimate of 0.16 per gray for those whose internal dos-
es could be estimated is higher than that shown in the sec-
ond column of Table 6. This difference appears to come
about because, with the restricted data used in Table 6,
baseline risks were estimated to be higher than in analyses
based on the full data set.
We also conducted analyses that included data on smok-
ing (ever/never classification), which are currently available
for most main plant workers who were hired before 1959
and whose plutonium doses could be estimated. Among
these workers, 74% of the males were identified as smok-
ers, whereas only 3.4% of the females were identified as
smokers. The third and fourth columns of Table 6 show
results for this subgroup of workers with and without ad-
justment for smoking, which was carried out by including
smoking status as part of the baseline risk. Adjustment for
smoking reduced the ERR per gray for internal dose only
slightly and did not greatly modify the female/male ratio
of the ERR per gray. The estimated coefficient of the log-
arithm of attained age was also unchanged (–4.5 both with
and without smoking adjustment). However, the smoking
adjustment reduced the estimate of the ratio of the baseline
risks for males and females from 11 to 2.0. Also, excluding
auxiliary plant workers and workers hired after 1958 had
little effect on the internal dose estimates as can be seen
by comparing results in columns 1 and 3 of Table 6. Re-
striction of analyses to workers with smoking data reduced
the ERR per gray for external dose and adjustment for
smoking reduced it still further. However, confidence inter-
vals were wide for these restricted analyses.
Results of Parallel Analyses of Data on Mayak Workers
and Japanese Atomic Bomb Survivors
Table 7 shows the distribution of person-years and lung
cancer deaths by gender and attained age for Mayak work-
ers who had internal doses that could be estimated and for
members of the Life Span Study (LSS) cohort of Japanese
atomic bomb survivors who were 15–60 years old at ex-
posure. For the LSS cohort, 68% of the person-years were
in females, whereas the comparable percentage was 24 for
Mayak workers. The LSS cohort is also older with 33% of
the person-years over age 65, compared with 11% for the
Mayak cohort. In comparing risks for the two cohorts, it is
thus important to take account of dependences on gender
and attained age.
When the ERR model was fitted to the LSS data, there
was little evidence of modification by attained age. The
estimated coefficient for the logarithm of attained age
LUNG CANCER IN MAYAK WORKERS
(?0.36) was lower than that for internal dose in Mayak
workers (?3.2), but the two coefficients did not differ sig-
nificantly (P ? 0.12). To compare ERR per sievert for the
LSS and for internal dose in the Mayak cohort, we fitted
models in which the coefficient of the logarithm of attained
age was set equal to ?2.2, obtained by weighting the co-
hort-specific estimates by their inverse variances, and a val-
ue that was compatible with the data from both cohorts.
Results are shown in the top half of Table 7. For Mayak
workers, estimates are expressed per sievert using a quality
factor of 20 as recommended by the ICRP (23) for ?-par-
ticle exposure. Estimates of the ERR per sievert based on
A-bomb survivors are larger than those for Mayak workers,
but there is considerable overlap in the confidence intervals.
The ratios of the coefficients for females and males were
similar for the two cohorts. The assumption of a larger RBE
than 20, as estimated from the Mayak cohort, would make
the Mayak estimates even lower.
When the EAR model was fitted to the LSS data, there
was no evidence of dependence on gender (P ? 0.5). The
ratio of estimates for females and males was 1.3 (95% CI:
0.44; 4.0), higher than the ratio of 0.43 shown in Table 3
for internal dose in Mayak workers (P value for difference
? 0.08). As with Mayak workers, the EAR depended on
attained age, but the dependences in the two cohorts dif-
fered. For this reason, Table 7 shows EARs by categories
of attained age. Results for the LSS cohort are for both
sexes since there was no evidence that the EAR depended
on gender. Results for Mayak are for males, and the EAR
for females would be a factor of 0.43 smaller than the es-
timates shown in Table 7. For Mayak workers, there was
clear evidence of risk from internal dose (P ? 0.001) for
all but the oldest age group. By contrast, for the LSS co-
hort, the strongest evidence of excess risk was for the two
oldest age groups (P ? 0.001), with little evidence of ex-
cess risk before age 55. For attained ages under 65, esti-
mated risks for Mayak workers are significantly higher than
those in the LSS cohort. Although risks for Mayak workers
over age 75 are estimated to be smaller than those in the
LSS, this might be because of the very limited data on
Mayak workers in this age group.
For external dose in Mayak workers, modification by at-
tained age was remarkably similar to that in the LSS cohort.
There was little evidence that the ERR per sievert was mod-
ified by attained age in either cohort, and the EAR in-
creased with attained age with nearly identical estimates of
the coefficients of the logarithm of attained age (5.7 for
Mayak; 5.8 for the LSS). Thus risk coefficients could be
compared directly. For the LSS, the ERR per sievert for
males (with no modification by attained age) was 0.29
(95% CI: 0.02; 0.62), while the EAR for both sexes at
attained age 60 was 2.2 excess deaths per 104PY-Sv (95%
CI: 1.2; 3.6). These estimates are similar to those for ex-
ternal dose in male Mayak workers shown in Table 3 (ERR
per gray ? 0.17; EAR ? 2.4 excess deaths per 104PY-
As in previous analyses based on less extensive data, we
found strong evidence that internal plutonium exposure in-
creases lung cancer risks. Risk increased in a dose-depen-
dent fashion among workers whose internal doses from plu-
tonium exposure could be estimated, and, in addition, risk
increased with the ordered plutonium surrogate variable
among potentially exposed workers whose internal doses
could not be estimated. The dose–response relationship was
adequately described by a linear function, and there was
evidence of risk when analyses were restricted to internal
doses to the lung less than 0.5 Gy. The estimated ERR per
gray for a male at attained age 60 was 4.7 (95% CI: 3.3;
6.7), very similar to the estimate of 4.5 obtained in the most
recent analysis by Kreisheimer et al. (14), which was based
on a subset of the data and did not include consideration
of the modifying effect of attained age.
For internal plutonium exposure, the ERR per gray for
females was estimated to be more than four times that for
males. By contrast, the EAR per 104PY-Gy for females
was less than half that for males. The different patterns for
ERR and EAR models reflect the very strong difference in
baseline risks for the two sexes with baseline risks for
males estimated to be about 11 times those for females;
smoking differences explain much of this difference. Gen-
der differences were also observed for the plutonium sur-
rogate categories, where different mean doses for males and
females in a given category might have contributed.
The ERR per gray and EAR per 104PY-Gy for internal
dose also depended on attained age. The ERR was found
to decrease with attained age, especially among potentially
exposed workers who were not monitored for plutonium.
In this latter group, it was not possible to take account of
the accumulation of dose with age and time since exposure.
Analyses based on EAR models also revealed dependences
on attained age. The EAR per 104PY-Gy among those
whose internal doses could be estimated increased with at-
tained age to about age 65 and then decreased. With both
ERR and EAR models, excess risks had declined to non-
significant levels by the time workers reached age 75; how-
ever, data were very limited for evaluating risks at older
ages (see Table 7). There was little evidence of modification
by age at hire or time since exposure
We also found that external exposure increased lung can-
cer risks. This is in contrast to several earlier analyses of
lung cancer risks (10–14) where there was little evidence
of a dose response for external exposure. However, these
earlier analyses did not include workers in the radiochem-
ical and plutonium plants who were not monitored for plu-
tonium, a restriction that excluded more than half the work-
ers who had either external doses exceeding 1 Gy or ex-
ternal doses in the 0.1–1-Gy range (see Table 1). The use
of the plutonium-potential variable allowed us to include
these workers and thus provided a more powerful assess-
ment of the effects of external dose, and we think this ap-
GILBERT ET AL.
proach is more appropriate for the evaluation of external
dose effects. Kreisheimer et al. (14) estimated the ERR per
gray for external dose in males as 0.06 (95% CI: ?0; 0.20),
not incompatible with the estimate shown in Table 3, and
similar to our results based on restricted data (Table 6). Our
results for external dose are in agreement with analyses by
Shilnikova et al. (7) of lung, liver and bone cancer risks
(as a single category) with most of the deaths due to lung
cancer. These analyses, which included the same workers
as ours, also found a statistically significant association
with external dose and no evidence of modification by gen-
der, attained age, or time since exposure.
There is of course concern that the use of the plutonium-
potential variable may not have provided an adequate ad-
justment for the effects of internal exposure. However, there
is no direct evidence of this since the ERR per gray for
external dose based on those whose plutonium doses could
be estimated was similar to that based on those whose plu-
tonium doses could not be estimated (P ? 0.5). Neverthe-
less, analyses that were restricted to workers whose internal
doses could be estimated (Table 6) yielded lower nonsig-
nificant estimates of the effects of external dose, although
their wide confidence intervals indicated they were com-
patible with estimates based on the entire cohort. Even in
this group, uncertainties in plutonium dosimetry could lead
to inadequate adjustment for internal exposure. External ra-
diation exposure has been found to increase lung cancer
risk in atomic bomb survivors (22), ankylosing spondylitis
patients (24), Hodgkin’s disease patients (25), and peptic
ulcer patients (26). No association was found among tu-
berculosis patients who received protracted exposure from
fluoroscopies, but these patients had underlying lung dis-
ease (27). Although we think that the identified association
between external radiation dose and lung cancer mortality
for Mayak workers is likely to be real, there is potential
for bias in quantifying the dose–response relationship.
Parallel analyses of Mayak workers and the LSS cohort
were also conducted. For external dose, even though one
exposure was fractionated and the other acute, results were
very similar both with respect to the estimated level of the
ERR and EAR and the modifying effects of gender and
For internal dose, both the magnitude of the ERR per
sievert and the sex ratio for the two cohorts were reasonably
comparable. However, the ERR showed a strong decline
with attained age in Mayak workers but not in the LSS
cohort, although it is possible that this difference was due
to chance. Pierce et al. (28) recently conducted analyses of
lung cancer risks in a subset of the LSS that included data
on smoking and found evidence for a strong decline in the
ERR with attained age with the coefficient for the logarithm
of attained age estimated to be ?3.6, similar to the estimate
of ?3.2 for Mayak workers. They also found that the fe-
male/male ratio was reduced when smoking was taken into
Analyses of Mayak workers (internal dose) and the LSS
based on the EAR model revealed differences in both the
gender effect and the pattern of risk for attained age. For
ages under 65, estimated risks for Mayak workers are sig-
nificantly higher than those in the LSS cohort and may
suggest that the relative biological effectiveness (RBE) is
higher than 20.
Estimates of the RBE of dose from plutonium (relative
to external dose) were also higher than the quality factor
(QF) of 20 recommended by the International Commission
on Radiological Protection (ICRP) (23), but the wide con-
fidence intervals include the ICRP value. Estimates of the
RBE are possibly less dependent on particular characteris-
tics (including smoking patterns) of Russian workers than
are the estimated ERR and EAR.
Patterns of baseline risks by gender, attained age, and
birth cohort also differ in the LSS and Mayak cohorts, al-
though we have not explored this in detail. However, we
note that Preston et al. (22) estimate that baseline risks for
LSS males are about 2.4 times those for LSS females,
whereas baseline risks for Mayak males are about 11 times
those for females. In the LSS cohort, the sex ratio for the
ERRs is approximately equal to the sex ratio of the baseline
risks, and thus the EARs for the two sexes are similar. In
the Mayak cohort, the female/male ERR ratio is not as large
as the male/female baseline risk ratio, and thus the EARs
for Mayak females are smaller than those for males. Base-
line rates for Mayak males are much higher than those for
LSS males, a difference that reflects at least in part differ-
ences in smoking habits for the two cohorts. The average
year of birth for Mayak workers was 1932, while that for
LSS members exposed between the ages of 15 and 60 was
1910, a difference that likely contributes to different smok-
It may be more appropriate to compare Mayak risk es-
timates for internal plutonium exposure to estimates based
on underground miners who, like Mayak workers, received
protracted dose from ?-particle emitters. In an analysis of
lung cancer risks in eleven cohorts of male underground
miners (29), the ERR per WLM (working level month) was
found to decrease with increasing dose rate, attained age,
and time since exposure. The ERR per WLM for exposure
received 5–15 years ago at the lowest dose rate (?0.5 WL)
at attained age 55–64 was estimated to be 0.034 per WLM.
If it is assumed that 1 WLM is equal to 5 mGy (30, 31),
this would correspond to 6.8 per gray, slightly higher than
the estimate of 4.7 per gray based on Mayak workers. Es-
timates for higher dose rates in the miners are smaller; for
example, those exposed at 0.5 to 1.0 WL were estimated
to have half the risk of those exposed at the lowest rate.
For both Mayak workers and underground miners, the ERR
decreased with increasing attained age. For the under-
ground miners, the ERR per WLM for attained ages of 55–
64, 65–74 and 75? years relative to those for under age
55 were estimated to be 0.57, 0.29 and 0.09, respectively.
Comparable ratios of ERRs for Mayak workers based on
internal ?-particle dose are similar, 0.67, 0.33 and 0.12,
LUNG CANCER IN MAYAK WORKERS
respectively. By contrast, we did not see the same pattern
of decrease in the ERR with time since exposure that was
observed for miners. Dosimetry limitations, especially the
fact that it was not possible to measure the pattern of in-
ternal lung dose accumulation in individual workers, un-
doubtedly limited our ability to distinguish risks from dose
received in different periods.
Several limitations of this study need to be noted and
could potentially distort models developed in this paper.
First, this a complex data set, and it is not simple to develop
models that adequately describe the dependence of lung
cancer risks on internal and external dose and also on gen-
der, attained age, and age at hire. Our models are undoubt-
edly overly simple, but data do not support more detailed
An important limitation of most results presented in this
paper is their failure to include data on smoking. Results
for internal dose from analyses that were adjusted for smok-
ing using the limited data that are now available did not
differ greatly from analyses that were not so adjusted. Al-
though adjustment for smoking lowered the estimated ERR
per gray for external dose, this may have been due to
chance given the wide confidence limits. The very high
smoking rate in Russian men and the very low rate in Rus-
sian women make it unlikely that smoking is a strong con-
founder. Smoking undoubtedly contributes to the gender
differences that were identified in our analyses, and it might
also contribute to the pattern of risks with attained age.
Efforts are currently under way to use medical records to
improve data on smoking (including amount smoked) for
the full cohort.
Another important limitation of our study relates to dose
estimates. Estimating lung dose from exposure to plutoni-
um and its pattern over time is challenging, and even dose
estimates based on state-of-the-art dosimetry methods are
subject to large uncertainties and potential biases. Smoking
can influence the rate of clearance of plutonium from the
lung, and current plutonium dosimetry models do not take
account of this. Because of differences in smoking habits
for males and females, dosimetry biases might thus affect
estimates of risk coefficient ratios for the two sexes. Ex-
ternal doses are less problematic, but dose estimates for
early years are subject to biases resulting from limitations
in the dosimeters used at that time. For example, film do-
simeters used in the period 1948–1953 had no compensat-
ing filters for high-energy ? particles, which could have led
to overestimation of the external ?-radiation dose for some
The ongoing collaborative Russian and U.S. dosimetry
program continues to improve individual dose estimates for
both external radiation and plutonium. In the near future,
revised plutonium dosimetry will use physiologically based
models of the fate of plutonium in the respiratory tract and
in systemic tissues. These revised models will enable great-
er use of the unique autopsy data assembled at SUBI and
will make it possible to address the potential influence of
the worker’s health status (smoking history, liver disease,
etc.) on the distribution of plutonium in the body. Revised
external dosimetry will take into account the limitations
and sensitivities of the early dosimeters, make dose correc-
tions for photon energy spectra and angular dependence at
specific work locations, and provide new estimates of organ
doses. Doses from the small amount of neutron exposure
will also be estimated for specific work locations and oc-
In spite of the limitations noted above, lung cancer risks
from internal dose appear similar in many ways to such
risks in underground miners who were also exposed to pro-
tracted ?-particle radiation, whereas lung cancer risks from
external dose are similar to such risks in A-bomb survivors.
Clearly it is important to re-evaluate lung cancer risks mak-
ing use of both improved dose estimates and smoking data.
Support for this study has been provided by the U.S. Department of
Energy (DOE) under the auspices of the JCCRER and by the U.S. Na-
tional Cancer Institute through contract N01-CP-11014 to the Radiation
Effects Research Foundation (RERF). The authors thank V. F. Khokry-
akov for the opportunity to use data from his laboratory and E. K. Vas-
ilenko for the use of data on external exposure from the Mayak plant.
This report makes use of data obtained from RERF in Hiroshima, Japan.
RERF is a private, non-profit foundation funded by the Japanese Ministry
of Health, Labour and Welfare (MHLW) and the U.S. DOE, the latter
through the National Academy of Sciences. The conclusions in this report
are those of the authors and do not necessarily reflect the scientific judg-
ment of RERF or its funding agencies.
Received: April 12, 2004; accepted: June 23, 2004
1. G. L. Voelz, J. N. P. Lawrence and E. R. Johnson, Fifty years of
plutonium exposure to the Manhattan Project plutonium workers: An
update. Health Phys. 73, 611–619 (1997).
2. G. S. Wilkinson, G. L. Tietjen, L. D. Wiggs, W. A. Galke, J. F.
Acquavella, M. Reyes, G. L. Voelz and R. J. Waxweiler, Mortality
among plutonium and other radiation workers at a plutonium weap-
ons facility. Am. J. Epidemiol. 125, 231–250 (1987).
3. R. Z. Omar, J. A. Barber and P. J. Smith, Cancer mortality and mor-
bidity among plutonium workers at the Sellafield plant of British
Nuclear Fuels. Br. J. Cancer 79, 1288–1301 (1998).
4. S. Wing, D. Richardson, S. Wolf and G. Mihlan, Plutonium-related
work and cause-specific mortality at the United Sates Department of
Energy Hanford Site. Am. J. Ind. Med. 45, 153–164 (2004).
5. National Research Council, Committee on the Biological Effects of
Ionizing Radiation, Health Effects of Radon and Other Internally
Deposited Alpha Emitters (BEIR IV). National Academy Press, Wash-
ington, DC, 1988.
6. NCRP, Scientific Basis for Evaluating the Risks to Populations From
Space Applications of Plutonium. Report No. 131, National Council
of Radiation Protection and Measurements, Bethesda, MD, 2001.
7. N. S. Shilnikova, D. L. Preston, E. Ron, E. S. Gilbert, E. K. Vassi-
lenko, S. A. Romanov, I. S. Kuznetsova, M. E. Sokolnikov, P. V.
Okatenko and N. A. Koshurnikova, Cancer mortality risk among
workers at the Mayak Nuclear. Radiat. Res. 159, 787–798 (2003).
8. N. A. Koshurnikova, E. S. Gilbert, M. Sokolnikov, V. F. Khokhry-
akov, S. Miller, D. L. Preston, S. A. Romanov, N. S. Shilnikova,
K. G. Suslova and V. V. Vostrotin, Bone cancers in Mayak workers.
Radiat. Res. 154, 237–245 (2000).
516 Download full-text
GILBERT ET AL.
9. E. S. Gilbert, N. A. Koshurnikova, M. Sokolnikov, V. F. Khokhry-
akov, S. Miller, D. L. Preston, S. A. Romanov, N. S. Shilnikova,
K. G. Suslova and V. V. Vostrotin, Liver cancers in Mayak workers.
Radiat. Res. 154, 246–252 (2000).
10. Z. B. Tokarskaya, N. D. Okladnikova, Z. D. Belyaeva and E. G.
Drozhko, Multifactorial analysis of radiation and nonradiation factors
on the lung cancer incidences among workers of the Mayak Nuclear
Enterprise. Health Phys. 69, 356–366 (1997).
11. Z. B. Tokarskaya, B. R. Scott, G. V. Zhutova, N. D. Okladnikova,
Z. D. Belyaeva, B. F. Khokhryakov, H. Schollnberger and E. K. Vas-
ilenko, Interaction of radiation and smoking in lung cancer induction
among workers at the Mayak Nuclear Enterprise. Health Phys. 83,
12. N. A. Koshurnikova, M. G. Bolotnikova, L. A. Ilyin, I. B. Keirim-
Markus, Z. S. Menshikh, P. V. Okatenko, S. A. Romanov, V. I. Tsvet-
kov and N. S. Shilnikova, Lung cancer risk due to exposure to in-
corporated plutonium. Radiat. Res. 149, 366–371 (1998).
13. M. Kreisheimer, N. A. Koshurnikova, E. Nekolla, V. F. Khokhryakov,
S. A. Romanov, M. E. Sokolnikov, N. S. Sokolnikov, P. V. Okatenko
and A. M. Kellerer, Lung cancer mortality among male nuclear work-
ers of the Mayak facilities in the former Soviet Union. Radiat. Res.
154, 3–11 (2000).
14. M. Kreisheimer, M. E. Sokolnikov, N. A. Koshurnikova, V. F. Khokh-
ryakov, S. A. Romanov, N. S. Shilnikova, P. V. Okatenko, E. A.
Nekolla and A. M. Kellerer, Lung cancer mortality among nuclear
workers of the Mayak facilities in the former Soviet-Union. An up-
dated analysis considering smoking as the main confounding factor.
Radiat. Environ. Biophys. 42, 129–135 (2003).
15. N. A. Koshurnikova, N. S. Shilnikova, P. V. Okatenko, V. V. Kreslov,
M. G. Bolotnikova, M. E. Sokolnikov, V. F. Khokhryakov, K. G.
Suslova, E. K. Vassilenko and S. A. Romanov, Characteristics of the
cohort of workers at the Mayak nuclear complex. Radiat. Res. 152,
16. V. Khokhryakov, K. Suslova, E. Aladova, E. Vasilenko, S. C. Miller,
D. M. Slaughter and M. P. Krahenbuhl, Development of an improved
dosimetry system for the workers at the Mayak Production Associ-
ation. Health Phys. 79, 72–76 (2000).
17. V. F. Khokhryakov, Z. S. Menshikh, K. G. Suslova, T. I. Kudryav-
tseva, Z. B. Tokarskaya and S. A. Romanov, Plutonium excretion
model for the healthy man. Radiat. Prot. Dosim. 53, 235–239 (1994).
18. V. F. Khokhryakov, K. G. Suslova, V. V. Vostrotin, S. A. Romanov,
Z. S. Menshikh, T. I. Kudryavtseva, R. E. Filipy, S. C. Miller and
M. P. Krahenbuhl, The development of the plutonium lung clearance
model for exposure estimation of the Mayak Production Association,
nuclear plant workers. Health Phys. 82, 425–431 (2002).
19. K. G. Suslova, V. F. Khokhryakov, Z. B. Tokarskaya, Z. B. Nifatov,
M. P. Krahenbuhl and S. C. Miller, Extrapulmonary organ distribution
of plutonium in healthy workers exposed by chronic inhalation at
The ‘‘Mayak’’ Production Association. Health Phys. 82, 432–444
20. M. P. Krahenbuhl, D. M. Slaughter, J. L. Wilde, J. D. Bess, S. C.
Miller, V. F. Khokhryakov, K. G. Suslova, V. V. Vostrotin, S. A.
Romanov and T. I. Kudryavtseva, The historical and current appli-
cation of the FIB-1 model to assess organ dose from plutonium in-
takes in Mayak workers. Health Phys. 82, 445–454 (2002).
21. D. L. Preston, J. H. Lubin and D. A. Pierce, EPICURE User’s Guide.
HiroSoft International Corporation, Seattle, WA, 1993.
22. D. L. Preston, Y. Shimizu, D. A. Pierce, A. Suyama and K. Mabuchi,
Studies of mortality of atomic bomb survivors Report 13: Solid can-
cer and noncancer disease mortality: 1950–1997. Radiat. Res. 160,
23. ICRP, 1990 Recommendations of the International Commission on
Radiological Protection. Publication 60, Annals of the ICRP, Vol.
21, No. 1–3, Pergamon Press, Oxford, 1991.
24. H. A. Weiss, S. C. Darby and R. Doll, Cancer mortality following
x-ray treatment for ankylosing spondylitis. Int. J. Cancer 59, 327–
25. L. B. Travis, M. Gospodarowicz, R. E. Curtis, E. A. Clarke, M.
Andersson, B. Glimelius, T. Joensuu, C. F. Lynch, F. E. van Leeuwen
and E. S. Gilbert, Lung cancer following chemotherapy and radio-
therapy for Hodgkin’s disease. J. Natl. Cancer Inst. 94, 182–192
26. Z. A. Carr, R. A. Kleinerman, M. Stovall, R. M. Weinstock, M. L.
Griem and C. E. Land, Malignant neoplasms after radiation therapy
for peptic ulcer. Radiat. Res. 157, 668–677 (2002).
27. G. R. Howe, Lung cancer mortality between 1950 and 1987 after
exposure to fractionated moderate-dose-rate ionizing radiation in the
Canadian fluoroscopy cohort study and a comparison with lung can-
cer mortality in the atomic bomb survivors study. Radiat. Res. 142,
28. D. A. Pierce, G. B. Sharp and K. Mabuchi, Joint effects of radiation
and smoking for lung cancer risk among atomic bomb survivors.
Radiat. Res. 159, 511–520 (2003).
29. National Research Council, Committee on the Biological Effects of
Exposure to Radon, Health Effects of Exposure to Radon (BEIR VI).
National Academy Press, Washington, DC, 1999.
30. NCRP, Evaluation of Occupational and Environmental Exposures to
Radon and Radon Daughters in the United States. Report No. 78,
National Council on Radiation Protection and Measurements, Be-
thesda, MD, 1984.
31. National Academy of Sciences/National Research Council, Compar-
ative Dosimetry of Radon in Minas and Homes. National Academy
Press, Washington, DC, 1991.