samedi 1 janvier 2000

VOLUME 75

WORLD HEALTH ORGANIZATION
INTERNATIONAL AGENCY FOR RESEARCH ON CANCER

IARC MONOGRAPHS
ON THE EVALUATION
OF CARCINOGENIC
RISKS TO HUMANS

VOLUME 75
IONIZING RADIATION, PART 1:
X- AND GAMMA (γ)-RADIATION,
AND NEUTRONS

2000
I A R C
L Y O N
F R A N C E


This publication represents the views and expert opinions
of an IARC Working Group on the
Evaluation of Carcinogenic Risks to Humans,
which met in Lyon,
26 May–2 June 1999
2000

file mono75.pdf

...
 Table 3. Major radiation accidents (1945–97) and early fatalities in nuclear and non-nuclear industries
Year Place Source Dose (or activity
intake)
No. of persons with
significant exposurea
No. of deaths
1945–46 Los Alamos, USA Criticality ≤ 13 Gy 10 2
1958 Vinèa, Yugoslavia Experimental reactor 2.1–4.4 Gy 8 1
1958 Los Alamos, USA Criticality 0.35–45 Gy 3 1
1960 USSR 137Cs (suicide) ∼15 Gy 1 1
1960 USSR Radium bromide (ingestion) 74 mBq 1 1 (after 4 years)
1961 USSR Submarine accident 10–50 Gy > 30 8
1961 Switzerland 3H 3 Gy 3 1
1961 Idaho Falls, USA Explosion in reactor ≤ 3.5 Gy 7 3
1962 Mexico City, Mexico 60Co capsule 9.9–52 Sv 5 4
1963 China 60Co 0.2–80 Gy 6 2
1964 Federal Republic of
Germany
3H 10 Gy 4 1
1964 Rhode Island, USA Criticality 0.3–46 Gy 4 1
1966 Pennsylvania, USA 198Au Unknown 1 1
1967 USSR X-radiation medical diagnostic
facility
50 Gy 1 1 (after 7 years)
1968 Wisconsin, USA 198Au Unknown 1 1
1968 Chicago, USA 198Au 4–5 Gy (bone marrow) 1 1
1972 Bulgaria 137Cs (suicide) > 200 Gy (local, chest) 1 1
1975 Brescia, Italy 60Co 10 Gy 1 1
1978 Algeria 192Ir ≤ 13 Gy 7 1
1982 Norway 60Co 22 Gy 1 1
1983 Constitu, Argentina Criticality 43 Gy 1 1
1984 Morocco 192Ir Unknown 11 8
1985 China 198Au (mistake in treatment) Unknown, internal 2 1
1985–86 USA Accelerator Unknown 3 2
1986 Chernobyl, USSR Nuclear power plant 1–16 Gy 134 28
1987 Goiânia, Brazil 137Cs ≤ 7 Gy 50b 4
1989 El Salvador 60Co irradiation facility 3–8 Gy 3 1
1990 Israel 60Co irradiation facility > 12 Gy 1 1
1990 Spain Radiotherapy accelerator Unknown 27 ≤ 11
1991 Nesvizh, Belarus 60Co irradiation facility 10 Gy 1 1
1992 China 60Co > 0.25–10 Gy 8 3
1992 USA 192Ir brachytherapy > 1000 Gy (local) 1 1
1994 Tammiku, Estonia 137Cs 1830 Gy (thigh) + 4 Gy
(whole body)
3 1
1996 Costa Rica Radiotherapy Unknown 110 ≤ 40
1997 Kremlev, Sarov, Russian
Federation
Criticality experiment 5–10 Gy 1 1

VOLUME 100

World Health Organization
International Agency for Research on Cancer

IARC Monographs on the Evaluation of
Carcinogenic Risks to Humans

VOLUME 100

A Review of Human Carcinogens

Part D: Radiation

LYON, FRANCE

This publication represents the views and expert opinions
of an IARC Working Group on the
Evaluation of Carcinogenic Risks to Humans,
which met in Lyon,

2–9 June 2009

file mono100D.pdf
...
page 120

(f) Radon
Radon is a noble (chemically inert) gas mostly produced through the radioactive decay of environmental uranium/thorium and their radioactive daughters. All of the isotopes of radon are radioactive: 222Rn is the isotope with the longest radioactive half-life, and its naturally abundant parent is 226Ra, itself a daughter of 238U (see Fig. 1.3), 222Rn is the most prevalent in the environment. 220Rn (also known as thoron) is the only other isotope of radon that is found in any
significant quantity in nature. That isotope and its radioactive daughters typically contribute less than 20% of the total dose from radon, and its contribution is often not included in radon exposure assessments. Henceforth, the term radon
should be taken as referring to Radon-222 unless otherwise indicated.
222Rn is an α-particle emitter with a short half-life of 3.82 days, it decays to polonium-218, which is also an α-emitter, and has in turn further short-lived radioactive daughter products (see Fig. 1.3). The presence of this decay chain greatly increases the overall radiological significance of this isotope. Although 222Rn is a gas, its shortlived progeny are electrically charged particles that can become attached to environmental dust particles in the air, the existence and extent of this ‘attached’ fraction has a considerable impact on dose to the upper airways of the lung.
Like its parent radiosotopes (see Fig. 1.3), 222Rn is omnipresent in nature but levels vary because certain types of rocks and soils (e.g. granite, phosphate rocks, and alum shales) contain more of its parents than others (Appleton, 2007). 222Rn rapidly disperses into the troposphere when it escapes into the free atmosphere, i.e. outside of enclosed spaces. Consequently, concentrations of
222Rn in breathing air in open spaces is relatively low, typically around 10 Bq/m3. 222Rn can also be found in building materials albeit at low concentrations (de Jong et al., 2006). Building materials such as concrete, wallboard, brick and tile usually have concentrations similar to those of major rock types used for their manufacture, and levels also vary according to the type
of rock used for construction (Mustonen, 1984; Ackers et al., 1985). Although building materials generally contribute only a very small percentage
of the indoor air 222Rn concentrations, in a few areas, concrete, blocks, or wallboard incorporating radioactive shale or waste products from
uranium mining can make an important contribution to the indoor 222Rn levels (Man & Yeung, 1998; Åkerblom et al., 2005).

(g) Radium
Radium is a naturally occurring rare earth metal. Ubiquitous in the environment, in small quantities, it is found in soils, uranium/thorium ores (e.g. pitchblende), minerals, ground water, and seawater, because the common radium isotopes are products of the main uranium/thorium decay chains. All the isotopes of radium
are radioactive, 226Ra has the longest half-life, and therefore is the predominant isotope found in nature. 226Ra is an α-particle emitter with a half-life
of 1600 years, and decays to 222Rn, which is also an α-particle emitter.
228Ra is a β and gamma emitter with a half-life of 5.75 years, and decays to actinium-228, which is a β-particle and gamma emitter. 226Ra concentrations in soil vary considerably, typically between 10–50 Bq/kg, with approximately 25 Bq/kg considered to be average (UNSCEAR, 1982), concentration in seawater is
4–5 orders of magnitude lower than this. 223Ra and 224Ra are both α-particle emitters with a half-life of 11.43 days and 3.6 days, respectively. 224Ra can be found in ground water.

...

Page 241

INTERNALIZED α-PARTICLE EMITTING
RADIONUCLIDES
Internalized radionuclides that emit α-particles were considered by a previous IARC
Working Group in 2000 (IARC, 2001). Since that time, new data have become available,
these have been incorporated into the Monograph, and taken into consideration in the
present evaluation.
1. Exposure Data
See Section 1 of the Monograph on X-radiation
and γ-radiation in this volume.
2. Cancer in Humans
2.1 Radon
Radon is a natural radioactive gas produced
by the decay of uranium and thorium, which
are present in all rocks and soils in small quantities.
There are several isotopes of radon, the
most important of which are 222Rn (produced
from 238U) and 220Rn (produced from thorium).
220Rn is also known as thoron because of its
parent radionuclide. In the United Kingdom, it
has been shown that 220Rn delivers much smaller
doses to the public in indoor environments than
222Rn (The Independent Advisory Group on
Ionising Radiation, 2009). Unlike 222Rn, 220Rn is
not formed during the radioactive decay of 238U,
and is hence not present at appreciable levels in
uranium mines.
The epidemiological evidence on the cancer
risks from radon is derived largely from cohort
studies of underground miners that had been
exposed to high levels of radon in the past. More
recently, a series of case–control studies of lower
exposures to residential radon have also been
conducted.
The previous IARC Monograph on radon
IARC (1988) states that radon is a cause of lung
cancer in humans, based on clear excess lung
cancer rates consistently observed in underground
miners, and elevated lung cancer risks
seen in experimental animals exposed to radon.
In a subsequent evaluation by IARC (2001), additional
epidemiological evidence of an increased
lung cancer was also seen in case–control studies
of residential radon. Although results from the
13 case–control studies available at that time
were not conclusive, the Working Group noted
that the risk estimates from a meta-analyses of
eight such studies were consistent with estimates
based on the underground miner data (Lubin &
Boice, 1997).
In a detailed evaluation of the health risks
of radon by the Committee on the Biological
Effects of Ionizing Radiation (BEIR) within the

US National Research Council (BEIR IV, 1988),
it was also reported that radon is a cause of lung
cancer in humans. An important aspect of this
work was the development of risk projection
models for radon-related lung cancer, which
provides estimates of the lung cancer risk associated
with residential radon, depending on age,
time since exposure, and either concentration or
duration of exposure.
In an effort to synthesize the main epidemiological
findings and assist in the evaluation of the
lung cancer risks associated with occupational
and environmental exposure to radon, several
combined analyses of the primary raw data
from studies of radon and lung cancer have been
conducted. Several combined analyses of epidemiological
data from 11 cohorts of underground
miners have been conducted (BEIR IV, 1988;
Lubin et al., 1994; Lubin & Boice, 1997; BEIR
VI, 1999). Howe (2006) conducted a combined
analysis of data from three cohorts of uranium
miners from Canada, and Tomášek et al. (2008)
conducted a combined analysis of Czech and
French uranium miners. Combined analyses
of epidemiological data from seven North
American case–control studies of residential
radon and lung cancer (Krewski et al., 2005), 13
European studies (Darby et al., 2005, 2006), and
two studies from the People’s Republic of China
(Lubin et al., 2004) have also been conducted.
Cancers other than lung cancer, notably
haematopoietic lesions, have been investigated
in some of the cohort studies of miners. Case–
control studies of residential radon and childhood
cancers, including leukaemia, have also
been conducted. Ecological studies of environmental
radon and the risk of lung and other
cancers have been reported, but these are less
informative than the cohort and case–control
studies discussed previously (IARC, 2001).
2.1.1 Occupational studies of underground
miners
(a) Early observations of lung disease in miners
Underground mining was the first occupation
associated with an increased risk of lung
cancer. Metal ores were mined in the Erz mountains
(a range between Bohemia and Saxony), in
Schneeberg from the 1400s and in Joachimsthal
(Jachymov) from the 1500s. As early as the 16th
century, Georg Agricola, in his treatise ‘De re
Metallica’, described exceptionally high mortality
rates from respiratory diseases among miners
in the Erz mountains. The disease in miners
was recognized as cancer in 1879 by Harting &
Hesse (1879). This report provided clinical and
autopsy descriptions of intrathoracic neoplasms
in miners, which were classified as lymphosarcoma.
During the early 20th century, histopathological
review of a series of cases established that
the malignancy prevalent among miners in the
Erz mountains was primary cancer of the lung
(Arnstein, 1913; Rostocki, 1926). Many authors
offered explanations for this excess including
exposures to dusts or metals in the ore (particularly
arsenic). In 1932, Pirchan and Sikl suggested
that radioactivity was the most probable cause
of the cancers observed in Jachymov (Pirchan &
Sikl, 1932).
(b) Cohort studies
The first epidemiological evidence of an
increased lung cancer risk among underground
miners exposed to radon in the Colorado Plateau
was given by Archer et al. (1962). Subsequent analyses
of this cohort were conducted by Wagoner
et al. (1964, 1965) as additional lung cancer cases
accrued; the latter analysis was the first to relate
lung cancer risk to cumulative exposure to radon
progeny in terms of working-level months (WLM).
Stram et al. (1999) conducted detailed analyses
of the effects of uncertainties in radon exposures
within this cohort on radon-related lung cancer
risk estimates. Another early study reported
lung cancer risk in Canadian fluorspar mines in
Newfoundland, where substantial amounts of
water seeping through the mines contain radon
gas (de Villiers, 1966). The first statistical study
on the incidence of lung cancer among uranium
miners from former Czechoslovakia (the Czech
Republic) was published in 1966 by (Rericha
et al., 1966), followed by results on autopsy-verified
lung cancer cases (Horacek, 1968). The first
epidemiological study in uranium miners from
former Czechoslovakia (the Czech Republic)
was initiated in the late 1960s, with first results
reported shortly thereafter (Sevc et al., 1971). In
contrast to other epidemiological studies, there
were hundreds of radon measurements per year
in every mine. As of now, cancer risks in 19
cohorts of underground miners exposed to radon
have been investigated (see Table 2.1 available at
http://monographs.iarc.fr/ENG/Monographs/
vol100D/100D-04-Table2.1.pdf). In each of these
cohorts, occupational exposure to radon decay
products was associated with increased lung
cancer risk.
To increase statistical power, particularly
in quantifying the modifying effect of different
factors related to time or age, attempts were made
to pool individual data from related studies for
the joint estimation of risk and the evaluation of
modifying factors. The first such analyses were
conducted by the BEIR IV committee (BEIR
IV, 1988), and included a combined analysis of
three studies of uranium miners in the Colorado
Plateau, USA, the Eldorado mine in Ontario,
Canada, and Swedish iron miners in Malmberget.
By building on initial work by Lubin et al.
(1994, 1995) and Lubin & Boice (1997), a subsequent
report by the US National Research
Council (Lubin, 2003) extended the combined
analysis to encompass 11 cohorts of underground
miners (see Table 2.1 on-line). An important
aspect of this analysis was the development of
a comprehensive risk model for radon-induced
lung cancer in underground miners taking into
account age, time since exposure, and either
exposure concentration or duration of exposure
(see Table 2.2 available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.2.pdf). The previous risk model developed
by the BEIR IV committee did not consider exposure
concentration or duration. The BEIR VI risk
models indicated that lung cancer risk decreased
with time since exposure and age; for a fixed
cumulative exposure, the risk decreased with
increasing exposure concentration (reflecting an
inverse exposure–rate effect), and increased with
duration of exposure.
Another pooled analysis was conducted in
a joint cohort of Czech and French uranium
miners, including a total of 10100 miners and
574 lung cancers (Tomášek et al., 2008). Cohort
members were subject to relatively low levels of
radon exposure (mostly below 4 working-level
(WL)); exposure measurements were available for
over 96% of the total exposure time experienced
by individuals in this joint cohort. The effect
of the quality of the exposure data in this joint
study was analysed by distinguishing exposures
based on measurements from those that were
estimated or extrapolated. If exposure quality is
not accounted for, the estimated ERR/WLM is
substantially underestimated by a factor of 3.4 in
the French study; however, effect modification by
exposure quality was not observed in this study
with relatively low annual exposures, for which
measurements were almost always available. The
term ERR/WLM quantifies the increased in risk
per exposure in working-level months. More
specifically, WLM is a time-integrated exposure
measure, and it is the product of the time
in working months (170 hours) and workinglevel.
One WL equals any combination of radon
progeny in 1 litre of air that gives the ultimate
emission of 130000 MeV of energy of α-particles.
Consequently, 1 WLM corresponds to 2.08 x 10−5
J/m3 x 170 hours = 3.5 x 10−3 J-hours/m3.
Predictions of lung cancer risk were not
substantially different from those based on the
BEIR VI risk models (Table 2.2 on-line). [The
Working Group noted that a complicating factor
in the interpretation of data on lung cancer
risks among uranium miners from former
Czechoslovakia (the Czech Republic) is the joint
exposure to γ-radiation, which can also increase
lung cancer risk.]
In contrast to the empirical radon risk
projection models developed by the BEIR IV
committee, the BEIR VI committee (BEIR VI,
1999) applied biologically based risk models to
describe the lung cancer risks relation to radon
in the Czech and French cohorts; a discussion
on the interpretation of risk projections derived
from the application of such models to epidemiological
data on radon is provided in Heidenreich
& Paretzke (2004).
Grosche et al. (2006) reported on a new
German cohort of 59000 uranium miners,
with 2388 lung cancer cases. This is the largest
of the miner cohorts investigated to date, and
is comparable in size to the 11 cohorts considered
in the BEIR VI report combined (BEIR VI,
1999). Patterns of risk based on age and exposure
concentration were similar to those found
in the BEIR VI report (BEIR VI, 1999), although
the effect of time since exposure was somewhat
different (Table 2.2 on-line), possibly reflecting
the higher proportion of missing causes of death
in the early years of follow-up. Howe, (2006)
conducted a combined analysis of Canadian
data on uranium miners from the Beaverlodge,
Port Radium, and Port Hope cohorts. The study
included 17660 workers, with 618 cases of lung
cancer. Patterns of lung cancer risk were similar
to those found in the BEIR VI report (Table 2.2
on-line).
(c) Joint effects of radon and smoking on
lung cancer risk
Because tobacco smoking is a powerful risk
factor for lung cancer, the joint effects of radon
and smoking need to be considered. The interactions
between exposure to radon and smoking
in the six studies of miners for which smoking
information was available were investigated
by Lubin et al. (1995). Although some studies
were consistent with additive effects of radon
and tobacco smoke on lung cancer risk, other
interactions between radon and tobacco smoke
in which the joint effects of these two agents
were greater than additive (Table 2.3 available
at http://monographs.iarc.fr/ENG/Monographs/
vol100D/100D-04-Table2.3.pdf). Six of the
above studies were jointly analysed in the BEIR
VI report (BEIR VI, 1999), which suggested a
submultiplicative model. The ratio of ERR/WLM
in non-smokers and smokers was 3.0 (95%CI: 0.3
– 29.2). [The Working Group noted that the confidence
interval of this ratio was relatively wide,
because of the small numbers of lung cancers
(64) among non-smokers.] In these studies, the
radon risk coefficients adjusted for smoking were
not substantially different from those obtained
when smoking was ignored.
(d) Lung cancer risks among haematite
miners
Previous IARC Monographs have implicated
radon as contributing to the excess lung cancer
risk observed in haematite miners (IARC, 1972,
1987, 1988). Volume 43 of the IARC Monographs
(IARC, 1988) states that “underground haematite
mining with exposure to radon is carcinogenic to
humans.” Lawler et al. (1985) noted no increased
mortality in 10403 lung cancer among miners
in Minnesota haematite mines relative to population
rates (SMR, 1.00) with low-grade exposure
to radon daughters and silica dust. Kinlen
& Willows (1988) noted that other iron mines,
like that in Cumbria in the United Kingdom, in
which 864 underground miners were studied,
the SMR for lung cancer among workers was
increased relative to population rates in the
period 1948–67 (SMR, 1.53), but not thereafter
(SMR, 1.13). Radon levels in early periods were
in the range of 0.35–3.2 WL, and decreased to
0.1–0.8 WL, suggesting that radon was the causative
agent. In a study of 5406 haematite miners in
China, a significant excess of lung cancer (SMR,
3.7) was observed, although this was based on
only 29 cases of lung cancer (Chen et al., 1990).
In this study, lung cancer risk increased notably
with increasing radon concentrations and with
increasing dust concentrations; however, the
authors were unable to evaluate the independent
effects of radon and dust, because these two
hazards were positively correlated. Collectively,
these observations provide evidence that radon
increases the risk of lung cancer in haematite
miners.
(e) Leukaemia risks in miners
Health effects of exposure to radon progeny
other than lung cancer, including leukaemia,
have been addressed in several miner studies
(see Table 2.4 available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.4.pdf). A combined analysis of 11 cohorts
of underground miners showed no evidence
of an increased risk of leukaemia (Darby,
1995). However, significant trends in the risk
of leukaemia were found in the Czech study in
relation to duration of exposure (Tomášek et al.,
1993; Tomášek & Zárská, 2004), and to cumulative
joint exposure to radiation from radon gas,
external sources of exposure to γ-radiation, and
long-lived radionuclides (Tomášek & Kubik,
2006). In a separate Czech cohort, the risk of
leukaemia also increased with cumulative radon
exposure (Rericha et al., 2006). Another analysis
of a large German cohort of uranium miners has
shown a significant increase in the incidence of
leukaemia among the highest exposed miners
(Möhner et al., 2006), although leukaemia
mortality was not associated with exposure to
radon progeny (Kreuzer et al., 2008).
(f) Cancers other than lung and leukaemia
Darby et al. (2005) found no evidence of an
increased risk of other cancers in their pooled
analysis of 11 miner cohort studies. In an analysis
of the large German cohort, Kreuzer et al.
(2008) found a statistically significant relationship
between cumulative radon exposure and
mortality from all extra pulmonary cancers
combined; this result persisted after adjustment
for potential confounding by arsenic, dust, longlived
radionuclides and γ-radiation. Increasing
trends in cancer risk were also reported at several
specific sites in this study; however, none of these
trends was significant after adjustment for potential
confounding.
Sevcová (1989) reported that the risk of basal
cell carcinoma among Czech uranium miners
was 2–12 times higher than in the general male
population. The mean equivalent dose in the basal
layer of epidermis was estimated to be 0.6–5.0 Sv,
depending on the duration of exposure (Sevcová
et al., 1978). Based on 27 cases observed during
a 20-year follow-up period, the ERR/Sv was estimated
to be 2.2 (Sevcová, 1989).
2.1.2 Environmental studies of indoor radon
An extensive set of case–control studies of
indoor radon and lung cancer were designed,
and, taken individually, these studies did not
provide conclusive evidence of an association
between indoor radon exposure and lung cancer
risk. Because of the difficulty in identifying the
comparatively small relative risks that would
be anticipated from indoor radon exposure,
combined analyses of these studies were undertaken
in North America, Europe and China
(see Table 2.5 available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.5.pdf). The combined analyses had inclusion
criteria for each study with clear rules for
the selection of persons with lung cancer that
included the following: the selection of controls
so as to be representative of the population from
which the lung cancer cases arose; the availability
of detailed residential histories, compiled
in a similar way both for cases and controls; the
availability of long-term (minimum 2 months)
measurements of radon gas concentrations; and
availability of data on smoking habits for individual
study subjects.
(a) Combined analysis of North American case–
control studies
The combined analysis of seven North
American case–control studies included a total
of 3662 cases and 4966 controls (Krewski et al.,
2005, 2006). All studies used long-term α-particle
track detectors to measure the concentration of
radon progeny in indoor air for 12 months (Field
et al., 2006). Contemporaneous measurements
were made in homes that subjects had occupied
or were currently occupying; these measurements
were used to estimate historical radon
concentrations in those homes. Detectors were
placed in the living areas and bedroom areas of
the home in which subjects had spent the majority
of their time. Conditional likelihood regression
was used to estimate the excess risk of lung
cancer. Table 2.6 (available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.6.pdf) shows the estimated odds ratios for
lung cancer by different concentration levels of
radon, and the excess odds ratios per 100 Bq/m3.
Odds ratios for lung cancer increased with residential
radon concentration. The estimated odds
ratio after exposure to radon at a concentration
of 100 Bq/m3 in the exposure time window of
5–30 years before the index date was 1.11 (95%CI:
1.00–1.28). This estimate is compatible with the
estimate of 1.12 (95%CI: 1.02–1.25) predicted by
downward extrapolation of the miner data.
The examination of potential effect modification
by demographic factors (sex, age,
education level, respondent type) and smoking
variables (smoking status, number of cigarettes
per day, duration of smoking, years since quitting
smoking) showed no evidence of heterogeneity
of radon effects. There was no apparent heterogeneity
in the association by sex, educational level,
type of respondent (proxy or self), or cigarette
smoking, although there was some evidence of
a decreasing radon-associated lung cancer risk
with age (P = 0.23).
Analysis of the effects of radon exposure by
different histological types of lung cancer showed
the largest excess odds ratio (0.23 per 100 Bq/
m3) for small cell carcinoma, although the confidence
limits overlapped with other histological
types of lung cancer. Because of the reduced
number of subjects, all of the confidence limits
for the excess odds ratios for specific histological
types of lung cancer included zero. Analyses
restricted to subsets of the data with presumed
more accurate radon dosimetry (increasing
number of years in the 5–30 exposure time
window and limiting the number of residences
by subjects) resulted in increased estimates of
risk with increasing number of years monitored.
In addition, excess odds ratios were larger when
data were restricted to subjects living in one or
two houses compared with no housing restrictions.
These results provide direct evidence of an
association between residential radon exposure
and lung cancer risk.
(b) Combined analysis of the European case–
control studies
Combined analysis of case–control studies
of indoor radon have also been carried out in
Europe. Darby et al. (2005, 2006) pooled individual
data from all studies and organized them
into a uniform data format to more precisely
estimate the increased risk of lung cancer due to
residential radon exposure, and to determine the
modifying effects of smoking, age, sex, and other
factors.
Data on smoking history and also on radon
exposure history, based on long-term measurements
of radon gas concentrations, were available
for a total of 7148 persons with lung cancer
so as to be representative of the population from
which the lung cancer cases arose; the availability
of detailed residential histories, compiled
in a similar way both for cases and controls; the
availability of long-term (minimum 2 months)
measurements of radon gas concentrations; and
availability of data on smoking habits for individual
study subjects.
(a) Combined analysis of North American case–
control studies
The combined analysis of seven North
American case–control studies included a total
of 3662 cases and 4966 controls (Krewski et al.,
2005, 2006). All studies used long-term α-particle
track detectors to measure the concentration of
radon progeny in indoor air for 12 months (Field
et al., 2006). Contemporaneous measurements
were made in homes that subjects had occupied
or were currently occupying; these measurements
were used to estimate historical radon
concentrations in those homes. Detectors were
placed in the living areas and bedroom areas of
the home in which subjects had spent the majority
of their time. Conditional likelihood regression
was used to estimate the excess risk of lung
cancer. Table 2.6 (available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.6.pdf) shows the estimated odds ratios for
lung cancer by different concentration levels of
radon, and the excess odds ratios per 100 Bq/m3.
Odds ratios for lung cancer increased with residential
radon concentration. The estimated odds
ratio after exposure to radon at a concentration
of 100 Bq/m3 in the exposure time window of
5–30 years before the index date was 1.11 (95%CI:
1.00–1.28). This estimate is compatible with the
estimate of 1.12 (95%CI: 1.02–1.25) predicted by
downward extrapolation of the miner data.
The examination of potential effect modification
by demographic factors (sex, age,
education level, respondent type) and smoking
variables (smoking status, number of cigarettes
per day, duration of smoking, years since quitting
smoking) showed no evidence of heterogeneity
of radon effects. There was no apparent heterogeneity
in the association by sex, educational level,
type of respondent (proxy or self), or cigarette
smoking, although there was some evidence of
a decreasing radon-associated lung cancer risk
with age (P = 0.23).
Analysis of the effects of radon exposure by
different histological types of lung cancer showed
the largest excess odds ratio (0.23 per 100 Bq/
m3) for small cell carcinoma, although the confidence
limits overlapped with other histological
types of lung cancer. Because of the reduced
number of subjects, all of the confidence limits
for the excess odds ratios for specific histological
types of lung cancer included zero. Analyses
restricted to subsets of the data with presumed
more accurate radon dosimetry (increasing
number of years in the 5–30 exposure time
window and limiting the number of residences
by subjects) resulted in increased estimates of
risk with increasing number of years monitored.
In addition, excess odds ratios were larger when
data were restricted to subjects living in one or
two houses compared with no housing restrictions.
These results provide direct evidence of an
association between residential radon exposure
and lung cancer risk.
(b) Combined analysis of the European case–
control studies
Combined analysis of case–control studies
of indoor radon have also been carried out in
Europe. Darby et al. (2005, 2006) pooled individual
data from all studies and organized them
into a uniform data format to more precisely
estimate the increased risk of lung cancer due to
residential radon exposure, and to determine the
modifying effects of smoking, age, sex, and other
factors.
Data on smoking history and also on radon
exposure history, based on long-term measurements
of radon gas concentrations, were available
for a total of 7148 persons with lung cancer (d) Ecological studies of residential radon and
lung cancer
Cohen & Colditz (1995) reported a negative
correlation between radon levels and lung cancer
in over 3000 counties in the USA. Such ecological
studies are subject to several limitations,
including the absence of county-specific data on
smoking, which can confound the association
between ecological indicators of radon exposure
and lung cancer risk. This possibility was
confirmed by (Puskin, 2003), who subsequently
reported that negative correlations were obtained
between county-level radon concentrations and
county-level cancer occurrence rates for cancers
known to be related to tobacco smoking, with no
correlation at the ecological level between radon
and cancers not related to tobacco smoking.
Similarly, Lagarde & Pershagen, (1999) demonstrated
that an increasing trend in lung cancer
risk with increasing exposure to indoor radon
observed in a national Swedish case–control
study became a decreasing trend when information
on radon and lung cancer was aggregated to
the ecological (county) level.
(e) Attributable risk of lung cancer
Darby et al. (2005) estimated the fraction of
the lung cancer burden attributable to indoor
radon in Europe to be about 9%, based on the
relative risk of lung cancer associated with exposure
to indoor radon in the combined analysis
of the 13 European case–control studies, and the
indoor radon concentrations observed in those
studies. In the USA, the BEIR VI committee
(BEIR VI, 1999) used the radon risk projection
models developed on the basis of the miner data,
and data on radon concentrations in US homes
to estimate the attributable fraction to be in the
range of 10–15%, depending on which of the
committee’s two preferred risk models was used.
Brand et al. (2005) used the BEIR VI risk models
and data on radon concentrations in Canadian
homes to obtain an estimate of the attributable
fraction of 8%. Although subject to some uncertainty,
these results suggest that about 8–15%
of the lung cancer deaths in Europe and North
America may be attributed to residential radon
exposure, making radon the second leading
cause of lung cancer death after tobacco smoking
in those regions.
(f) Studies of leukaemia
Lubin et al. (1998) conducted a case–control
study of acute lymphoblastic leukaemia among
children under 15 years of age in the USA in
relation to residential radon exposure (see Table
2.7 available at http://monographs.iarc.fr/ENG/
Monographs/vol100D/100D-04-Table2.7.pdf),
based on 1-year track-etch radon measurements
in all current and previous residences in
which they had lived for at least 6 months. This
study provided no evidence of an association
between indoor radon exposure and childhood
acute lymphoblastic leukaemia. In a subsequent
case–control study of leukaemia and central
nervous system (CNS) tumours (nephroblastoma,
neuroblastoma, and rhabdomyosarcoma),
Kaletsch et al. (1999) found no evidence of an
increased risk of leukaemia of children under
15 years of age in Lower Saxony, Germany.
Steinbuch et al. (1999) reported no increase in
the risk of acute myeloid leukaemia of children
under 18 years of age identified through the
Children’s Cancer Group, which involves over
120 institutions in the USA and Canada. Law et
al. (2000a) did not find evidence of an increased
risk of either acute lymphoblastic leukaemia or
acute myeloid leukaemia in adults 16–69 years
of age in the United Kingdom. Results from
the United Kingdom Childhood Cancer Study,
which included 805 cases of acute lymphoblastic
leukaemia, demonstrated no association between
residential radon and leukaemia (Cartwright
et al., 2002). Raaschou-Nielsen (2008) conducted
a case–control study of 2400 cases of leukaemia,
CNS tumours, and malignant lymphoma in children
under 15 years of age identified through
the Danish Cancer Registry. Cumulative radon
exposure was associated with an increased risk
of acute lymphoblastic leukaemia, with an odds
ratio of 1.63 (95%CI: 1.05–1.23) for children
exposed to more than 890 Bq/m3–years, relative
to children exposed to less than 160 Bq/m3–years.
[The Working Group noted that a strength of this
study was the inclusion of virtually all relevant
cases in Denmark.]
Several ecological studies and surveys
suggested a positive correlation between exposure
to indoor radon and the risk of adult acute
leukaemia (especially myeloid leukaemia) and
childhood leukaemia (Henshaw et al., 1990;
Haque & Kirk, 1992; Kohli et al., 2000; Evrard
et al., 2006). These studies were based on an
ecological design in which radon levels were
regressed against the incidence of several cancer
sites. Average radon concentrations were obtained
from national or county surveys, and recorded as
population-averaged arithmetic means. In some
cases, crude geographic or geological features of
the inhabited areas were used to derive estimates
of levels of radiation emission, and subsequently
used as surrogates for exposure assessment
(Forastiere et al., 1992). [The Working Group
noted that this type of study design has many
limitations, including a lack of measurement of
individual exposure to indoor radiation, a lack
of control population, the difficulty in separating
radon effect from that of indoor γ-radiation, and
the absence of multiple regression analyses of
potential confounders (Eatough & Henshaw,
1994). In addition, ecological studies were often
based on the assumption that national or regional
radon concentrations apply to areas where cancer
registries have been compiled.]
(g) Cancers other than lung and leukaemia
In addition to leukaemia, the case–control
study conducted by (Kaletsch et al., 1999) in
Germany examined the association between
indoor radon and solid tumours. An elevated
odds ratio of 2.61 was reported (95%CI: 0.96
−7.13) for radon exposures above 70 Bq/m3 relative
to lower exposures; and this finding was
based mainly on six CNS tumours, for which
the odds ratio was 3.85 (95%CI: 1.26–11.81).
The United Kingdom Childhood Cancer Study
examined the association between indoor radon
and non-Hodgkin lymphoma, Hodgkin disease,
CNS tumours, and other solid tumours, and
found no association with any of these tumours
(Cartwright et al., 2002). The case–control study
by Raaschou-Nielsen (2008) in Denmark found
no association between indoor radon and either
tumours of the central nervous system or malignant
lymphoma.
Ecological studies have suggested that several
cancers might also be weakly correlated with
indoor radon, especially kidney cancer, prostate
cancer, malignant melanoma, and some childhood
cancers (Butland et al., 1990; Axelson,
1995). However, these studies use ecological indicators
of radon exposure, and do not control for
possible confounders such as indoor γ-radiation
or tobacco smoking.
2.1.3 Synthesis
Cohort studies of underground miners
exposed to high levels of radon (specifically, 222Rn
and its decay products) in the past have consistently
demonstrated an increased risk of lung
cancer, providing sufficient evidence of carcinogenicity
in humans (IARC, 1988). Case–control
studies of residential radon and lung cancer have
added to the weight of epidemiological evidence
linking radon to lung cancer (IARC, 2001).
Since then, combined analyses of data from
seven case–control studies of indoor radon and
lung cancer in North America (Krewski et al.,
2005, 2006), 13 case–control studies in Europe
(Darby et al., 2005, 2006), and two studies in
China (Lubin et al., 2004) have provided clear
evidence of an increased risk of lung cancer due to
radon (specifically, 222Rn and its decay products)
in homes. A large study of uranium miners in
Germany (Grosche et al., 2006) and a joint study
in France and the Czech Republic (Tomášek
et al., 2008) have reaffirmed previous findings
of increased risk of lung cancer in underground
miners exposed to radon.
Cohort studies of underground miners
permit an assessment of cancer risk at multiple
sites; and some evidence of an increased risk of
leukaemia was reported among Czech uranium
miners, although these miners were also exposed
to γ-radiation (a risk factor for leukaemia). Case–
control studies of childhood and adult leukaemia
in relation to indoor radon exposure have mostly
not shown elevated risks, although one study
suggested an increased risk of leukaemia among
children in Denmark. An increased risk of solid
tumours was seen in one case–control study in
Germany; however, this result was based on only
six CNS tumours, and was not confirmed in
other case–control studies.
IARC (1972, 1987, 1988) previously concluded
that haematite miners exposed to radon were at
increased risk of lung cancer. A subsequent study
of haematite miners in China demonstrated
increasing lung cancer risk with increasing
radon concentrations; however, a similar trend
was seen with increasing dust concentrations,
and it was not possible to separate the effects of
radon and dust in this study. The Working Group
reaffirmed the conclusion reached in the earlier
IARC evaluations that radon contributes to the
increased lung cancer risk seen in haematite
miners.
2.2 α-Particle emitters
2.2.1 Radium-224/226/228
The previous IARC Monograph evaluation of
radium-224, radium-226, and radium-228 IARC
(2001) was based on an increased risk of bone
sarcoma associated with all three isotopes, as
well as an increased risk of paranasal sinuses and
mastoid process associated with 226Ra, in cohorts
of radium watch-dial painters who ingested 226Ra
(often in combination with 228Ra), and patients
injected with 224Ra. Few epidemiological analyses
of cancer risk following radium exposure
have been published since then. One of these is
an update to a cohort study of patients injected
with 224Ra in Germany (Wick et al., 2008), while
two recent case–control studies in the USA and
Thailand have considered radium in drinkingwater
(Guse et al., 2002; Hirunwatthanakul et al.,
2006).
(a) Bone
The cohort studies of cancer risk among
radium watch-dial painters in the USA were
initially carried out at the Massachusetts Institute
of Technology (Rowland et al., 1978) and the
Argonne National Laboratory (Stebbings et al.,
1984; Carnes et al., 1997), and later combined
(Rowland et al., 1983; Spiers et al., 1983). Those
studies, in which some of the painters ingested
226Ra (often in combination with 228Ra) by the
practice of ‘pointing’ their paintbrush tips with
their lips, showed consistent increases in the
risk for bone sarcoma related to exposure to
α-particles (Rowland et al., 1978; Stebbings et al.,
1984; see Table 2.8 available at http://monographs.
iarc.fr/ENG/Monographs/vol100D/100D-04-
Table2.8.pdf). Carnes et al. (1997) reported that
both isotopes of radium contributed significantly
and independently to the rate of mortality from
bone sarcomas in multivariate analyses of dose–
response relationships in which the two isotopes
were included as separate variables. The excess
risk for carcinomas of the paranasal sinuses and
mastoid process was associated with internally
deposited 226Ra, but probably not 228Ra (Rowland
et al., 1978). On the other hand, in the studies of
British dial painters who were exposed to lower
doses (none of them engaged in brush pointing),
no bone sarcomas were observed (Baverstock &
Papworth, 1985).
No further updates of bone cancer among
radium watch-dial painters have been published
in recent years, but new analyses using data
from the US studies have appeared. Bijwaard et
al. (2004) developed two-mutation mechanistic
models fitting animal and human data on bone
cancer. They reported that the results using data
for watch-dial painters agree well with those for
studies of radium-exposed beagles. The best fit
for the watch-dial painters had equal cell killing
terms in both mutation rates, but a nearly equally
well-fitting model could be constructed with cell
killing only in the second mutation rate, as in
the analysis of beagle data. In an analysis of data
on bone and sinus cancers for radium watch-dial
painters using a two-mutation model, Leenhouts
& Brugmans (2000) reported that the model
parameters from the best fit were consistent with
cellular radiobiological data. The fitted dose–
response relationships were linear–quadratic
with radium intake and with α-particle radiation
dose, and did not support a model involving a
threshold dose. The risks at low doses were estimated
to be about a factor of 10 lower than those
based on a linear extrapolation from high doses.
Bone sarcomas were the major late effect
among patients with tuberculosis, ankylosing
spondylitis, and other diseases who were treated
with high doses of 224Ra (mean bone surface dose,
30 Gy) in a cohort study in Germany (Nekolla
et al. 2000; see Table 2.8 on-line). Nekolla et al.
(2000) used an improved dosimetry system –
relative to previous analyses in that cohort – with
modified doses to the bone surface, particularly
for exposures at younger ages. Virtually all of
the tumours in the cohort could be attributed
to exposure to radium, reflecting the very high
bone-surface doses received. In contrast to
previous analyses of this cohort, the excess absolute
risk (EAR) decreased with increasing age at
exposure. As before, the EAR for a given total
dose decreased with increasing duration of exposure;
however, there was little evidence of such an
effect at the lower doses received by this cohort,
which was suggested to be in agreement with
microdosimetric considerations and general
radiobiological experience (Nekolla et al., 2000).
Among ankylosing spondylitis patients treated
with lower doses of 224Ra (mean bone-surface
dose, 5 Gy) in another German cohort, there was
an excess of bone cancer relative to population
rates, but based on only four cases (Wick et al.,
1999).
A case–control study of osteosarcoma in
Wisconsin (USA) looked for any correlation
with estimated levels of total α-particle activity
and levels of 226Ra and 228Ra in drinking-water,
by linking measurements to Zone Improvement
Plan (ZIP) codes (Guse et al., 2002; see Table
2.9 available at http://monographs.iarc.fr/ENG/
Monographs/vol100D/100D-04-Table2.9.pdf).
No evidence of an association was found.
However, the study lacked individual exposure
data, other than ZIP code. In addition, the exposures
were much lower than those for the 226Ra
watch-dial painters.
(b) Leukaemia
Wick et al. (2008) and Nekolla et al. (1999,
2000) reported findings for leukaemia in two
separate cohorts of ankylosing spondylitis
patients in Germany (see Table 2.8 on-line).
Exposures from 224Ra in the former cohort were
lower than those in the latter cohort. Wick et
al. (2008) found a significantly raised risk of
leukaemia – particularly myeloid leukaemia –
relative to population rates, which was in line with
experimental findings from mice injected with
varying amounts of this radionuclide. Nekolla et
al. (1999) reported eight leukaemia cases in their
cohort, compared with 3.8 expected from population
rates (P = 0.04). When a 2-year lag was
used, the corresponding P value was 0.08. [The
Working Group noted that although there were
indications of raised leukaemia risks in both of
the 224Ra cohorts, these findings were based on
small numbers of cases and that dose–response
analyses were not performed.]
No excess incidence of leukaemia was observed
among watch-dial painters or among watch-dial
painters with measured body burdens in the USA
(Spiers et al., 1983). However, leukaemia occurred
early in female watch-dial painters and an excess
of leukaemia was observed among male watchdial
painters (Stebbings, 1998).
(c) Other cancers
Little information has appeared since the
previous IARC Monograph (IARC, 2001) on
the association between exposure to radium
and risk of cancers other than bone cancer and
leukaemia. In particular, there are no new findings
for the radium watch-dial painters nor from
the 224Ra medically exposed cohorts. The excess
risk for carcinomas of the paranasal sinuses and
mastoid process seen among US radium watchdial
painters was associated with internally
deposited 226Ra, but probably not 228Ra (Rowland
et al., 1978). In particular, these cancers occurred
mainly among subjects exposed to 226Ra only,
and infrequently among those exposed to both
226Ra and 228Ra (Rowland et al., 1978). High 222Ra
levels were found in the mastoid cavity of subjects
whose body burdens were primarily from 226Ra,
and suggested that radioactive decay of 222Ra
released into this cavity by decay of 226Ra in the
surrounding bone is the cause of these cancers
(Evans, 1966).
In a cohort of USA radium watch-dial
painters, suggestive positive associations were
observed between estimated radium body
burden and lung cancer and multiple myeloma.
These cancers, particularly multiple myeloma,
were more closely associated with duration of
employment than with radium intake (Stebbings
et al., 1984). [The Working Group noted that
duration of employment corresponded to duration
of γ-radiation exposure, and was a surrogate
for cumulative external γ-radiation dose.]
No increased risk of lung cancer was observed in
cohorts of patients injected with 224Ra (Nekolla
et al., 1999; Wick et al., 1999).
Stebbings et al. (1984) also reported an association
between estimated radium burden and
mortality from breast cancer in US radium
watch-dial painters. This association may have
been confounded; in particular, women who had
worked the longest and had had both heavier
exposure to γ-radiation from radium and higher
breast cancer rates tended to have chosen not to
have children. A raised risk of breast cancer was
also observed in a cohort of women in the United
Kingdom who worked with radium paint (one
sided P = 0.077) (Baverstock et al., 1981). Due to
small body burden of radium compared to the US
luminizers, no further analyses were performed
with regard to α-particles. In analyses stratified
for both age at start of luminizing (< 30 versus
≥ 30 years) and γ-radiation dose (< 0.2 versus ≥ 0.2
Gy), the excess risk was seen to be predominant
among the younger age group receiving ≥ 0.2 Gy
of γ-radiation (one-sided P = 0.009). Nekolla et
al. (1999) reported a significantly raised risk of
breast cancer among patients injected with 224Ra.
Such an association was not observed among
patients injected with low-dose 224Ra (Wick et al.,
1999). [The Working Group noted indications of
a raised breast cancer risk in an unexposed group
in the analysis conducted by Nekolla et al. (1999),
suggesting that factors other than radiation may
have contributed to the breast cancer excess seen
in the exposed group].
Statistically significant increases in risk of
soft-tissue sarcomas, kidney cancer, urinary
bladder cancer, liver cancer and thyroid carcinoma
were also reported among patients injected
with high doses of 224Ra (Nekolla et al., 1999),
but not among those who received low doses
(Wick et al., 1999). [The Working Group noted
that although significant increases for the aforementioned
types of cancer were reported by
Nekolla et al. (1999) relative to population rates
(Table 2.8 on-line), the corresponding data for
a control group of unexposed patients were not
presented. For the other 224Ra cohort, Wick et al.
(1999) presented results for both exposed and
unexposed patients (Table 2.8 on-line). However,
in both cohorts, the numbers of cases of specific
cancer types were generally small.]
In Thailand, a case–control study of cancers
of the upper digestive tract reported a statistically
significant association with intakes of radium
in drinking-water, based on small numbers of
cases (Hirunwatthanakul et al., 2006; Table 2.9
on-line). [The Working Group noted that in
contrast to the other study on radium in drinkingwater
(Guse et al., 2002), this study collected
information on individuals’ daily consumption
of drinking-water and on other potential risk
factors, although for the cancer cases (but not the
controls) this information was provided mainly
by relatives.]
(d) Synthesis
As discussed previously by IARC (2001), the
studies of cancer risk among US radium watchdial
painters showed consistent increases in the
risk for bone sarcoma related to exposure to
α-particles, and both 226Ra and 228Ra contributed
significantly and independently to this elevated
risk. The previous Working Group (IARC, 2001)
associated the excess risk for carcinomas of the
paranasal sinuses and mastoid process in this
cohort to internally deposited 226Ra, but probably
not 228Ra. No further data was available to the
Working Group that altered the conclusions in
the previous IARC Monograph.
The most recent analysis of the risk of bone
tumours among patients treated with 224Ra for
tuberculosis or ankylosing spondylitis supports
the strong association observed by the previous
Working Group (IARC, 2001).
There is some evidence of elevated leukaemia
risks in the two cohorts of patients injected with
224Ra cohorts. However, these findings were
based on small numbers of cases, and dose–
response analyses were not performed. No excess
incidence of leukaemia was observed among US
radium watch-dial painters overall. The possibility
that radium isotopes increase leukaemia
risk in humans cannot be ruled out, but the
available evidence did not permit any causal relationship
to be established.
2.2.2 Mixed α-particle emitters
(a) Thorium-232
The previous IARC evaluation of 232Th and its
decay products, administered intravenously as a
colloidal dispersion of 232Th dioxide, was based on
increased risk of primary liver cancer, including
haemangiosarcomas, and leukaemia, excluding
chronic lymphocytic leukaemia (IARC, 2001).
The evidence of cancer risk associated with
Thorotrast (stabilized 232Th dioxide) exposures
came mainly from cohort studies in Denmark,
Germany, Japan, Portugal, and Sweden (IARC,
2001). Thorotrast was used extensively in medical
practice between the 1930s and the 1950s as
a radiographic contrast agent. Owing to its
colloidal nature, Thorotrast is retained mostly in
the reticuloendothelial system (liver, spleen, and
bone marrow) after intravenous injection.
(b) Liver and biliary tract cancers
Cohort studies in Denmark, Germany, Japan,
Portugal, Sweden, and the USA demonstrated
significantly increased risks for liver cancer
(approximately one-third being haemangiosarcomas),
which were significantly correlated with
the volume of Thorotrast injected. The incidence
of and mortality from liver cirrhosis were also
significantly increased in all studies in which
liver cirrhosis was an end-point (Mori et al.,
1999; dos Santos Silva et al., 2003). A combined
analysis of the cohorts of Danish and Swedish
Thorotrast patients (Travis et al., 2003) showed
statistically significant trends with a surrogate
measure for cumulative radiation dose in the
incidence of primary liver cancer and cancer of
the gallbladder. [The Working Group noted that
key strengths of this analysis were the long-term
follow-up, the availability of cancer incidence
data, the large number of cases observed and
the opportunity to conduct a dose–response
analysis, albeit based on a surrogate measure.]
Among patients injected with 20 mL or more
of Thorotrast, the cumulative excess cancer
incidence remained elevated for up to 50 years,
and approached 97%. Analysis of a smaller
cohort of Thorotrast patients in the USA, based
on mortality data, yielded comparable findings
(Travis et al., 2003). An extended mortality
follow-up of Thorotrast patients in Portugal
(dos Santos Silva et al., 2003) showed statistically
significant trends with a surrogate measure
for cumulative radiation dose for all cancers
combined, and for the grouping of liver cancer
and chronic liver diseases. Becker et al. (2008)
described an extended follow-up of mortality
in the German Thorotrast cohort, which is the
largest single study of Thorotrast patients. By the
end of 2004, nearly all of these patients had died.
For all malignant neoplasms and for cancers of
the liver and intrahepatic bile ducts, both the
SMR and the relative risk compared to a control
group increased with increasing time since first
exposure. An earlier analysis of the German
cohort (van Kaick et al., 1999) reported associations
between the amount of Thorotrast injected
and mortality from cancers of the liver, gallbladder
and extrahepatic bile ducts. A Japanese
cohort (Mori et al., 1999) also showed increased
mortality from liver cancer among Thorotrast
patients. In this publication, the risk associated
with increasing time since first exposure was
also reported, but no formal statistical test for
trend was presented (see Table 2.10 available at
http://monographs.iarc.fr/ENG/Monographs/
vol100D/100D-04-Table2.10.pdf).
Results of the continued follow-ups of
Thorotrast exposed patients are summarized in
Table 2.11, available at http://monographs.iarc.fr/
ENG/Monographs/vol100D/100D-04-Table2.11.
pdf.
(c) Haematological malignancies
A significantly increased risk of leukaemia
excluding chronic lymphocytic leukaemia has
been reported in the Thorotrast cohorts in
Denmark, Germany, Japan, Portugal, Sweden,
and the USA. A combined analysis of the cohorts
of Danish and Swedish Thorotrast patients (Travis
et al., 2003) in which the incidence of leukaemias
(excluding chronic lymphocytic leukaemia) was
significantly higher than that among unexposed
patients showed no statistically significant trend
in incidence associated with this dose measure.
Analysis of a smaller cohort of Thorotrast
patients in the USA, based on mortality data,
yielded comparable findings (Travis et al., 2001).
In an extended mortality follow-up of Thorotrast
patients in Portugal (dos Santos Silva et al., 2003),
mortality from benign and malignant haematological
diseases and from leukaemia (excluding
chronic lymphocytic leukaemia) remained high
relative to national rates over the follow-up
period (more than 40 years after administration
of Thorotrast), but did not show a trend with the
surrogate dose measure. In an extended followup
of mortality in the German Thorotrast cohort
(Becker et al. (2008), statistically significantly
elevated risks were seen for malignancies of the
haematopoietic system (particularly myeloid
leukaemia). An earlier analysis of the German
cohort (van Kaick et al., 1999) reported associations
between the amount of Thorotrast injected
and mortality from a grouping of myeloid
leukaemia and myelodysplastic syndrome (see
Table 2.10 on-line).
(d) Other cancers
Increased risks for cancers at other sites were
reported in some studies but not consistently.
A combined analysis of the cohorts of Danish
and Swedish Thorotrast patients (Travis et al.,
2003) showed statistically significant trends with
a surrogate measure for cumulative radiation
dose in the incidence of cancers of the pancreas,
peritoneum and other digestive organs. [The
Working Group noted that the excess risks for
site-specific cancers should be interpreted with
caution because of the potential bias associated
with the selection of cohort participants,
non-comparability of the internal and external
comparison groups, and confounding by indication.]
In an extended follow-up of mortality in the
German Thorotrast cohort (Becker et al., 2008),
statistically significantly elevated risks were
seen for cancer of the pancreas, brain, and prostate.
The earlier analysis of the German cohort
by van Kaick et al. (1999) did not find a raised
risk for cancer of the prostate, but this analysis
(unlike the most recent analysis by Becker et al.,
2008) did not take into account the different
age distributions of the exposed and unexposed
groups (See Table 2.10 on-line).
The Thorotrast studies give mixed results
on lung cancer risk (See Table 2.11 on-line),
although patients given Thorotrast exhale high
concentrations of 220Rn (thoron). [The Working
Group noted that the interpretation of these
findings is hampered by the lack of information
on smoking.] Studies in the USA and China of
workers exposed to thorium by inhalation of
fine particles containing thorium and its decay
products reported a raised risk of lung cancer
relative to national rates and – in an updated
analysis of miners in China (Chen et al., 2003)
– relative to an unexposed control group (see
Table 2.10 on-line). However, this latter study
did not incorporate a dose–response analysis.
Furthermore, the presence in the Chinese mines
of silica dioxide and rare-earth elements raises
concerns about possible confounding. The other
occupational study – of thorium workers in the
USA – did not show an association between lung
cancer and potential for thorium exposure (Liu
et al., 1992; Table 2.10 on-line). Furthermore,
data on smoking were not available for either of
these occupational studies.
(e) Synthesis
Results of the continued follow-up studies of
patients exposed to Thorotrast continue to show
raised risks several decades after first exposure
for all malignant neoplasms combined, with
consistently large relative risks seen for liver
cancer and malignancies of the haematopoietic
system. The risk of liver cancer increased with
increasing values for a surrogate of radiation
dose in analyses of the Danish/Swedish, German,
Portuguese, and US cohorts.
An earlier analysis of the German cohort
reported an association between a measure
of radiation dose and mortality from myeloid
leukaemia and myelodysplastic syndrome. In
contrast, analyses of the grouping of haematopoietic
malignancies in the Danish/Swedish,
Portuguese and US cohorts did not show an
association with a surrogate of dose; however,
these analyses were not conducted specifically
for leukaemia other than chronic lymphocytic
leukaemia in the Portuguese and US cohorts.
Large increased risks of cancers of the extrahepatic
bile ducts and of the gallbladder were
reported in the two largest analyses of Thorotrast
patients (Table 2.11 on-line). In view of their integral
relationship to the liver, in which most of
the injected Thorotrast is deposited, the extrahepatic
bile ducts are likely to receive substantial
α-particle exposure. Although the gallbladder is
a minor site of Thorotrast storage, its location
on the visceral surface of the liver may lead to
continual α-particle exposure.
Pancreatic cancer risk is elevated in the
German and Danish/Swedish Thorotrast cohorts,
although the relative risks tend to be lower
than those for the aforementioned cancer sites
(Table 2.11 on-line). In the latter cohort, there is
also borderline evidence of an association with a
surrogate measure of radiation dose. Doses to the
pancreas are likely to be notably lower than these
to the liver or spleen, although the anatomical
juxtaposition of a portion of the pancreas to the
spleen may have led to higher doses. However,
information on smoking habits is lacking in
these studies. Prostate cancer is also significantly
elevated in the German and Danish/Swedish
Thorotrast cohorts (Table 2.11 on-line), although
there has been no analysis of prostate cancer risk
in relation to the amount of Thorotrast injected
or any other measure of radiation exposure.
Doses to the prostate are likely to be small relative
to those for organs such as the liver.
Raised risks have also been reported among
Thorotrast patients for several other types of
cancer, although the interpretation of these
findings is unclear largely because of possible
confounding by factors that led to the exposure.
This is particularly important for brain cancer
because many of these patients were examined
with Thorotrast for cerebral angiography.
Studies of workers exposed to thorium by
inhalation of fine particles containing thorium
and its decay products, together with some –
but not all –studies of Thorotrast patients have
reported raised risks of lung cancer. However,
possible differences in smoking habits, and –
among miners – possible confounding by other
exposures must also be considered.
Overall, large, statistically significant relative
risks between exposure to Thorotrast and
primary liver cancer, leukaemia (excluding
chronic lymphocytic leukaemia), cancers of the
extrahepatic bile ducts, and cancer of the gallbladder
have been observed in the two largest
analyses and, in several instances, show associations
with measures of exposure. For pancreatic
and prostate cancers, significantly raised risks
were observed in the two largest analyses but
these risks are lower than those for aforementioned
cancer types and confounding cannot
be excluded. No substantive new evidence on
cancer risks following inhalation of 232Th has
appeared since the publication of the previous
IARC Monograph (IARC, 2001).

page 271

3.2.2 Polonium-210
(a) Hamster
By varying the number of repeated intratracheal
instillations that contained 210Po, Little et
al. (1985) were able to show that frank malignant
lung tumour incidence increased with increasing
dose rate. All groups received the same total dose
to the lung (Table 3.1). For these groups, the total
number of instillations was kept constant by
substituting instillation of equal volumes of isotonic
saline when no 210Po was administered. Of
significant note was the decreased tumour incidence
for a group in which a single 210Po instillation
was administered without accompanying
saline instillations. This result reinforced the
conclusions of Shami et al. (1982) who demonstrated
the importance of the saline administrations
given after 210Po in increasing lung tumour
incidence.