1. No category

Lung Cancer in Women and Type of Dwelling in Relation to
Radon Exposure
Christer Svensson, Göran Pershagen and Julius Klominek
Cancer Res 1989;49:1861-1865.
Updated version
E-mail alerts
Reprints and
Subscriptions
Permissions
Access the most recent version of this article at:
http://cancerres.aacrjournals.org/content/49/7/1861
Sign up to receive free email-alerts related to this article or journal.
To order reprints of this article or to subscribe to the journal, contact the AACR Publications
Department at pubs@aacr.org.
To request permission to re-use all or part of this article, contact the AACR Publications
Department at permissions@aacr.org.
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.
(CANCER RESEARCH 49, 1861-1865, April 1, 1989]
Lung Cancer in Women and Type of Dwelling in Relation to Radon Exposure1
Christer Svensson,2 Goran Pershagen, and Julius Klominek
Department of Cancer Epidemiology, Radiumhemmet, Karolinska Institute, Stockholm, Sweden fC. S.]; Department of Epidemiology, Institute of Environmental
Medicine, Karolinska institute, S-104 01 Stockholm, Sweden [C. S., G. P.], and Department of Lung Medicine, Huddinge University Hospital, Huddinge, Sweden [J.
K.J
ABSTRACT
A case-control study based on interviews with 210 incident female lung
cancer patients, 209 age-matched population controls, and 191 hospital
controls was carried out in Stockholm county, Sweden. Radon measure
ments made in a sample of 303 dwellings, in which the study subjects
had lived, showed that dwellings with ground contact had an average
concentration of approximately 160 Bqm~3, twice the average concentra
tion of other dwellings. A cumulated radon exposure index was calculated
for each subject based on data from the interviews and the measurements.
For the total group of lung cancer a relative risk (RR), adjusted for
smoking, age, and degree of urbanization, of 1.8 (95% confidence interval:
1.2-2.9) and 1.7 (0.9-3.3) associated with "intermediate" and "high"
exposure to radon was found. There was also a significant trend to a
positive dose-response relationship (!',„.,„,
= 0.03). For small cell cancer
the corresponding figures for RRs were 1.9 (0.6-4.5) and 4.7 (1.5-14".2),
respectively
(/',„,„!
= 0.01). There seemed to be a positive interaction
between radon exposure and smoking in relation to lung cancer. The
findings indicate that domestic radon may be of importance for the
induction of lung cancer, particularly for some histológica! types.
INTRODUCTION
Experimental studies on animals show that alpha radiation
from RnDs3 may induce cancer in the respiratory tract (1, 2),
and epidemiológica! studies among miners indicate that such
exposure increases the risk for cancer of the lung and bronchus
(3-7). The radon levels in most homes are much lower than in
the mines where increased lung cancer risks have been found,
but in some dwellings Rn concentrations in the same range can
occur (8,9). There are some epidemiological studies suggesting
increased lung cancer risks associated with domestic Rn expo
sure (10-14). Most of these studies are of limited size, and the
exposure estimates are uncertain.
Human cancer risks associated with exposure to low levels
of ionizing radiation are generally estimated from risks found
at high levels, e.g., among survivors from nuclear bomb explo
sions, patients who have received therapeutic radiological treat
ment, and miners exposed to Rn. For Rn, the risk estimates
derived from studies on miners may not be directly applicable
to the general population. The radiation dose to the respiratory
epithelium per unit exposure varies depending on the particle
content of the atmosphere as well as on the breathing pattern
(15), and interactions with other factors such as smoking may
be of importance. It is therefore important to gain empirical
knowledge of risks encountered in people primarily exposed to
Rn in the home.
To investigate the roles of Rn and some other possible
causative or protective factors in lung carcinogenesis, an epi
demiological case-control study was performed among women
living in Stockholm county, Sweden. This report focuses on
risks associated with domestic exposure to Rn.
MATERIALS AND METHODS
The study consists of 210 incident female cases of lung cancer, 209
age-matched population controls, and 191 hospital controls. The aim
was to interview all patients with a suspected or confirmed primary
pulmonary carcinoma admitted to the three clinical departments of
pulmonary medicine and the only clinical department of thoracic sur
gery in Stockholm county. Those patients, who subsequently were
shown not to have lung cancer constituted the hospital control group.
The population controls were selected at random from the population
register of Stockholm county among women, who were born on the
same day as a case.
The interviews of cases and hospital controls were conducted in the
hospital wards during September 1983 to December 1985. The popu
lation controls were interviewed during a personal visit or on the
telephone following identification of each of the cases. A structured
questionnaire was used, in which detailed information was requested
on smoking habits, exposure to environmental tobacco smoke, and on
all dwellings in Sweden where the subject had lived for at least 2 years
since birth or arrival to Sweden. A history of occupations and infor
mation on consumption of foodstuffs rich in carotenoids, vitamin A,
and vitamin C were obtained as well. An extensive description of the
study group and exposure data obtained is given elsewhere ( 16).
Estimation of Cumulated Rn Exposure. The assessment of exposure
to Rn was based on data from two sources. Firstly, data on types of
dwellings lived in by the study subjects were used as recorded in the
questionnaires. Secondly, Rn measurements were made in a sample of
dwellings in Stockholm county, where the subjects had lived.
In Sweden the ground is the most important source of Rn, particu
larly in dwellings with the highest concentrations (17). The greatest
variation in Rn concentration could thus be expected to occur among
dwellings with ground contact, i.e.. one-family houses or apartments
on the ground floor in multifamily houses without basement. Based on
information provided by the study subjects, the dwellings with ground
contact were pinpointed on geological maps, and the ground, on which
they were built, was classified into one of three risk categories ("high,"
"medium," or "low") by the Swedish Geological Company. The aim
was to make Rn measurements in all dwellings with ground contact
built on ground with high or medium risk of Rn emanation. In addition,
a 10% random sample of dwellings with ground contact on low risk
ground, and of apartments on the ground floor in multifamily houses
with basement were selected for measurements as well as a 2% random
sample of apartments above the ground floor.
According to the information obtained during the interviews, the 610
study subjects had lived in a total of 3,518 dwellings during 2 years or
more. 384 of these (10.9%) were selected for measurements. The house
owners refused measurements for 25 dwellings, and for 17 dwellings
we failed to get in contact with the owner. In 39 dwellings measurements
were not made for various reasons, e.g., because they were no longer
used for residential purposes or because the buildings no longer existed
or could not be located. Measurements were thus made in 303 dwellings,
Received 7/6/88; revised 12/15/88; accepted 12/28/88.
The costs of publication of this article were defrayed in part by the payment
or 78.9% of the original sample.
of page charges. This article must therefore be hereby marked advertisement in
The average Rn gas concentration during a 2-week period was
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' This study has been supported by grants from the Jubilee Fund of the Swedish
measured with a TLD. The instrument is designed by the National
Institute of Radiation Protection in Sweden (18). Dust and RnDs are
National Bank and the National Institute of Radiation Protection.
2To whom requests for reprints should be addressed, at Department of
filtered out, and moisture minimized by diffusion of the air through a
Epidemiology, Institute of Environmental Medicine, Karolinska Institute, P. O.
silica gel. The relative standard deviation between different TLDs is
Box 60208, S-104 01 Stockholm, Sweden.
estimated to be 5-6%. The instruments are calibrated yearly, and the
3The abbreviations used are: RnD, radon daughter; TLD, thermoluminescence
SD of the mean of the calibration constants for different calibration
dosimeter: RR, relative risk; CI, confidence interval.
1861
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.
LUNG CANCER IN WOMEN AND TYPE OF DWELLING
periods is of the same order as between different instruments. For the
ionization chamber used as primary standard the deviation from the
true Rn concentration is in the order of 10% (1 SD).
The measurement program was started in 1985 and concluded in
1987. All measurements were made during the heating season. The
TLDs were deployed by personnel from the National Institute for
Building Research and the Swedish Geological Company. They were
placed in the living room during 1 week and then for another week in
a bedroom. The measurement crew recorded information on building
materials, type of foundation, and type of ventilation. Measurements
of y radiation were also made in order to investigate if aerated concrete
based on alum shale had been used as building material. Such concrete
has a high exhalation of Rn and was used extensively in Sweden from
the 1940s until the mid-1970s. In 1975 approximately 10% of the
residential buildings contained various amounts of this material (8).
Based on the information provided by the study subjects, each
dwelling was classified as having ground contact or not. Each of the
two categories was assigned the arithmetic mean of the Rn concentra
tions in dwellings measured in the category. A "cumulated exposure"
was calculated for each subject for whom we had information on type
of dwelling lived in during at least 75% of the lifetime up to 5 years
prior to the interview (for population controls 5 years prior to the
interview of the corresponding case). Exposure during the last 5 years
was not included in the assessment. An occupancy factor of 0.8 was
assumed. For subjects, with no information on type of dwelling during
some time periods, the average exposure intensity during the years for
which there was information, was used for the missing years.
Statistical Methods. Maximum-likelihood estimates of RRs and CIs
were computed by multiple logistic regression analysis (19), in which
indicator variables represented categories of age, smoking, and Rn
exposure as shown in Tables 3 and 4 as well as degree of urbanization
(city of Stockholm versus other municipalities in Stockholm county).
All P values for trend (/"trend)presented, were estimated from the
regression models using a scoring of 1, 2, and 3 in the three Rn
exposure categories. Analyses were also performed using the estimated
median cumulated Rn exposure in the three categories, but the results
were only marginally affected.
The data presented were based on analyses with both control groups
pooled and unconditional of the matching performed for the population
controls. When the analysis was restricted to the cases and population
controls, similar results were obtained with conditional and uncondi
tional logistic regression analysis.
Table 1 Results from measurements of2"Rn-concentration
Stockholm county
in 303 dwellings in
mean
mean
(Bqm-3)106.7105.1194.798.7107.5153.360.66
(Bqm-3)12-135012-135060-456
(Bqnr3)163.5161.1258.3145.6165.7230.076.283.461.5236.029
Ground
contactType
dwellingOne-family
of
housesGround
withoutbasement*Type
floor
ground"High
of
risk""Medium
risk""Low
risk"Not
contactAbove
ground
floor*Ground
ground
floor with base
ment"Building
materialAlum
shale based con
crete*With
contactWithout
ground
ground con
tactOther
materialsWith
contactWithout
ground
ground con
tactN2452396SO188758391952371525120843Arithmetic
" In multifamily houses.
* Uranium rich building material widely used in Sweden until 1975.
Table 2 RR and CI values for different histological types of lung cancer in
relation to estimated cumulated 2"Rn-exposure among women in Stockholm
Based on information given by the subjects and assuming Rn concentrations
in dwellings with ground contact of 160 Bqm ' and without ground contact of 80
^'J^^ì^All
-4,500° (BqnT3a)
(Bqm-3a)N
P^*80241524171151.8(1.2-2.9)2.7(1.2-5.8)1.9(0.6-4.5)1.3(0.5-1.
RR* (95% CI) N RR* (95% CI)
cancerSquamous
lung
cellSmall
cellAdenocarci-nomaOther
typesControlsN7418152912152>6,000°
RESULTS
Radon Measurements. The distribution of Rn measurement
values was approximately lognormal. In Table 1 the results are
tabulated according to different classification criteria based on
information recorded by the measurement crew. Dwellings with
ground contact had a higher mean Rn concentration than others
(163.5 BqnT3 and 76.2 BqnT3, respectively). Since inspections
and measurements were made in only about 10% of all dwell
ings, where the subjects had lived, calculations of cumulated
exposure were based on classification of types of dwellings as
recorded in the questionnaires. This classification gave quite
similar results:
160.1 Bqm-3; *„
„¿â€žu*,
= 80.0 BqnT
Table 1 shows that dwellings, where aerated concrete based on
alum shale was used, had higher Rn concentrations than others.
However, this information could not be utilized for assessment
of individual exposure experience, since the women themselves
generally did not know whether this material had been used in
the construction of the dwellings where they had lived.
There were no significant differences in Rn concentrations
between dwellings with ground contact on different types of
ground. However, only seven dwellings on low risk ground were
measured, making the estimate in this category uncertain. Ex-
°Arithmetic means (medians) of the three exposure categories 3,728(3,897)
BqnT3a, 5,131(5,046) BqnT3a, and 6,918(6,765) BqnT3a.
* Maximum-likelihood estimates of relative risk adjusted for age, smoking,
and municipality of residency.
elusion of dwellings containing alum shale based concrete did
not substantially affect the comparison between the categories
(none of the seven dwellings on low risk type of ground con
tained alum shale).
Risk Estimates. After restricting the analyses to individuals
for whom we had information regarding types of dwellings lived
in for the required time period, there remained 187 cases (65
adenomatous cell, 48 squamous cell, 41 small cell, 14 undifferentiated, 11 large cell, and eight carcinomas of other histology)
and 337 controls (177 population and 160 hospital controls).
Complete residential histories were available for 170 subjects.
For the remaining 354 study subjects the information in the
questionnaire covered an average of 92.9% of the time period
of interest.
The RRs for different histological types of lung cancer in
relation to estimated cumulated Rn exposure are shown in
Table 2. The RRs for the total group of lung cancer were 1.8
(95% CI, 1.2-2.9) and 1.7 (0.9-3.3) in the intermediate and
high categories of cumulated Rn exposure. There was also a
1862
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.
LUNG CANCER IN WOMEN AND TYPE OF DWELLING
Table 5 RRfor lung cancer in relation to smoking and cumulated â„¢¿Rn-exposure
statistically significant trend (Pmn¿= 0.03). The association
among women in Stockholm
was strongest for small cell carcinoma, which also showed a
Maximum-likelihood estimates adjusted for age and municipality of residency.
statistically significant trend (/Vend = 0.01). The lowest risks
Bqm-3a RR (95% CI)
were found for adenocarcinoma. When squamous cell and small
cell cancers were grouped together, the RRs were 2.3 (1.2-4.3)
Never smoked
and 3.1 (1.3-7.5) in the intermediate and high exposure cate
(0.6-3.5)
(0.3-2.6)
gories, respectively, and the trend was highly significant (/"trend
>0- 10 cig/d
4.8
6.5
= 0.005).
(0.9-5.6)
(2.0-11.4)
(1.6-27.1)
>10 cig/d-4,5001.02.3
6.8
12.3
15.9
Table 3 shows the dose-response relationships within differ
(3.0-15.5)>4,500-6,0001.4
(4.9-31.1)>6,0000.9(4.0-62.9)
ent age strata. The highest risks seem to occur in the oldest age
stratum with RRs of 8.3 (95% CI, 2.4-28.5) and 5.4 (1.4-21.3)
positive interaction between the two exposures. In the highest
in the intermediate and high exposure categories, respectively,
for the total group of lung cancer. The trend was significant as stratum of radon exposure there were relative risks of 6.5 (1.6well (/"trend= 0.03). The differences in mean age between cases 27.1) and 15.9 (4.0-62.9) among smokers of up to 10 cigarettes
and controls within each age stratum was less than one year a day or more, respectively, in relation to nonsmokers in the
low radon exposure stratum. As would be expected the trend in
(not shown in table).
Table 4 shows the dose-response relationships within differ
lung cancer risk was stronger for smoking than for estimated
Rn exposure.
ent smoking strata. The greatest risks are seen for the smokers
of up to 10 cig/d with RRs of 2.2 (95% CI, 1.0-4.6) and 3.1
As mentioned previously, all risk analyses presented were
(1.0-9.4) in the intermediate and high exposure categories,
made with pooled control groups. Calculations with each con
respectively, and the trend was significant (/Vend= 0.03). Most
trol group separately showed similar results, but the risk esti
mates tended to be higher, when population controls were used.
of the exsmokers (those who had stopped smoking more than
This was accounted for by differences in distribution of age and
2 years before the interview) belonged to this smoking category,
and those in particular showed the greatest risks. No clear smoking habits between the control groups. In conjunction with
increases in risk are seen for never smokers. The estimates are the observed differences in RR in different age and smoking
categories, this resulted in different overall risk estimates (com
uncertain, however, because of small numbers.
pare Tables 3 and 4). In an analysis with hospital controls
The interaction between smoking and estimated cumulated
Rn exposure is elucidated in Table 5. There seems to be a simulating cases and population controls in the noncase series,
the "risk" estimates associated with Rn exposure were 1.3 and
1.0 for the intermediate and high exposure categories, respec
tively.
Table 3 Relative risk for lung cancer in different age strata in relation to
estimated cumulated â„¢¿Rn-exposure
among women in Stockholm
Maximum-likelihood
residency.
risk estimates adjusted for smoking and municipality of
_j
a)Age
>4,500-6,000
(Bqm a)
>6,000
(Bqm
DISCUSSION
The main finding of this study was the association between
asso
years)Lung
(60
ciation was strongest for small cell carcinoma. In studies on
cancerControlsAge
miners elevated risks associated with Rn exposure have con
sistently been shown for small cell and squamous cell cancers
years)Lung
(6 1-70
(4-6), and the risk estimates have been highest for small cell
cancerControlsAge
carcinoma. A significantly increased risk for adenocarcinoma
has been noted among U. S. but not among Czechoslovak
older)Lung
(71 years or
uranium miners.
cancerControlsK43942436722N
Several potential confounding factors need to be considered
in the interpretation of the results. Confounding by smoking
and age was controlled in the analysis. Residual positive con
founding by smoking is unlikely since smoking was negatively
Table 4 Relative risk for lung cancer in different smoking strata in relation to
correlated to having lived in dwellings with ground contact, a
estimated cumulated mRn-exposure among women in Stockholm
reflection of the fact that smoking is more common in urbanized
Maximum-likelihood risk estimates adjusted for age and municipality of
areas, where one-family houses are rare and most apartments
residency.
«
s
nn
m
-3
i
>4,500-6,000
>6,000
in multifamily houses do not have ground contact. Neither is
-4,500 (Bqm a)
(Bqm-3a)
(Bqm_3a)
residual confounding by age a likely explanation to the results.
The mean ages for cases and controls differed by less than a
CI)1.6(0.6-4.3)2.2(1.0-4.6)1.8(0.9-3.8)N10561210114RR
(95%
CI)1.1(0.4-3.5)3.1(1.0-9.4)1.9(0.5-7.2)/>„,„<0.870.030.12
(95%
year within each age stratum.
Never
smokedLung
cancerControls>0-10cig/dayLung
Ambient air pollution also seems unlikely as a positive confounder, since dwellings with ground contact are more common
outside the urbanized areas. To some extent this should have
been controlled in our analysis, which included a correlate of
cancerControls>10
degree of urbanization (living in the city of Stockholm).
Whether differences in dietary habits between rural and urban
ized populations may influence the risk for lung cancer is not
cig/dayLung
known, but in general rural populations have a lower incidence
cancerControlsA'95517484849N166131383316RR
of cancer than urbanized even after control of smoking habits
(20, 21). Furthermore, adjustment for carrot consumption did
CI)/»«,,,,2238294929281.7(0.8-3.3)1.1(0.5-2.3)8.3(2.4-28.5)28162815341.0(0.2-5.6)1.4(0.5-3.7)5.4(1.4-21.3)0.290.530.03
RR (95% CI) W RR (95%
lung cancer and estimated cumulated Rn exposure. This
1863
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.
LUNG CANCER IN WOMEN AND TYPE OF DWELLING
not have a noticeable impact on the observed risk associations
between living in dwellings with ground contact and lung cancer
(16).
An adequate selection of study subjects, i.e.. cases and con
trols, is another prerequisite for obtaining unbiased estimates
of the relative risks. There are indications that the cases in our
study constituted a biased sample of all cases diagnosed in
Stockholm county with regard to age and living in the city of
Stockholm (16). A comparison with the regional cancer register
showed that the cases included in the study were on an average
younger and more often living outside the city of Stockholm
than cases not included. Since both age and living in the city of
Stockholm are correlates of our measure of cumulated radon
exposure, it was necessary to control these two factors in the
analysis, particularly in relation to the use of the population
controls.
The differences between RRs observed when using population
and hospital controls separately, were largely due to differences
in distribution over age and smoking strata. Both types of
control groups have potential strengths and weaknesses (22).
The hospital controls were interviewed under conditions very
similar to those of the cases, which intuitively makes it more
likely that the quality of the obtained information may be
similar to that of the cases. On the other hand, there may be
an association, direct or indirect, between the exposure(s) under
study and the reason to their hospitalization. In view of the
diagnoses of the hospital controls (16), this does not seem to
be important taking the group as a whole.
The risk of selection bias among the population controls
would be low since the response rate was comparatively high.
On the other hand, there may be differences in the quality of
information, since the interviews were made under different
circumstances than for the cases. However, empirical data
indicate that the quality of information obtained from popula
tion and hospital controls may be similar (22).
The Rn concentration was monitored with TLDs during a 2week period only. A previous investigation based on 91 of the
303 dwellings measured in this study showed a high correlation
between TLD measurements during 2 weeks and a track detec
tors deployed for 6 months or 1 year (23), indicating that the
TLD measurements can be used for estimation of yearly Rn
exposure. The increased average Rn concentration found in
dwellings with ground contact confirms earlier data (8).
Although the biologically important radiation dose to the
lungs and bronchi is delivered by the RnDs (24), measurements
of indoor Rn gas concentration may be better suited for retro
spectively estimating radiation exposure than direct measure
ments of the daughters. The concentration of airborne RnDs
increases with increasing aerosol content of the atmosphere
(15, 25). It is thereby influenced by life-style factors like smok
ing habits of the dwellers, which may differ markedly between
time periods and dwellers.
The exposure classification for Rn in this study was based on
information regarding ground contact of the dwellings. No
usable data were available on building materials with high
exhalation of Rn, although such materials have been widely
used in Sweden (8). In order to increase the precision of the
exposure estimates, an extended Rn measurement program for
the dwellings of the study subjects is currently under way.
Cumulated Rn exposure would be difficult to estimate on the
individual level, even if good quality measurements of today's
radon concentrations were available for all dwellings of the
study subjects. Previous mesurements indicate that the mean
Rn concentration in Swedish homes has increased over the last
decades (17), but there are probably great individual variations.
In addition, personal life-style is of importance, e.g., ventilation
habits and hours spent in the home.
There were some indications of a positive interaction between
smoking and Rn exposure as estimated in our study. A multi
plicative interaction between smoking and Rn exposure has
been indicated in some studies on miners (26, 27) but not in
another (28). It is important to shed further light on the
interaction between domestic Rn exposure and smoking, both
for the understanding of the mechanisms of cancer induction
and as a basis for optimising preventive measures.
Conclusion
An association was observed between lung cancer and years
lived in dwellings with ground contact, which were shown to
have about twice the Rn concentrations of other dwellings. The
association was particularly strong for small cell carcinoma.
Exposure to Rn seems to be the most plausible explanation to
the findings. More precise exposure data are needed for a
detailed quantitative assessment of dose-response relationships.
ACKNOWLEDGMENTS
We want to thank the Swedish Geological Company and the National
Institute for Building Research for assistance in the classification of
the ground and the measurement program, as well as Gunilla Hedqvist
and Eva-Britt Gustafsson for help in the data collection.
REFERENCES
1. Cross, F. T., Palmer, R. F., Filippy, R. E., Dagle, G. E., and Stuart, B. O.
Carcinogenic effects of radon daughters, uranium ore dust and cigarette
smoke in beagle dogs. Health Phys., 42: 33-52, 1982.
2. Chameaud, .1., Perraud, R., Masse, R., and Lafuma, J. Lung carcinogenesis
during in vivo cigarette smoking and radon daughter exposure in rats. Recent
Results Cancer Res., 82:11-20,
1982.
3. Wagoner, J. K., Archer, V. E., Lundin, F. E., Holaday, D. A., and Lloyd, J.
W. Radiation as the cause of lung cancer among miners. N. Engl. J. Med.,
273:181-188, 1965.
4. Horacek, J., Placek, V., and Sevc, J. Histologie types of bronchogenic cancer
in relation to different conditions of radiation exposure. Cancer (Phila.), 40:
832-835, 1977.
5. Archer, V. E., Saccomanno, G., and Jones, J. H. Frequency of different
histologie types of bronchogenic carcinoma as related to radon exposure.
Cancer (Phila.), 34: 2056-2060, 1974.
6. Run/, E., Sevc, J., Placek, V., and Horacek, J. Lung cancer in man in relation
to different time distribution of radiation exposure. Health Phys., 36: 699706, 1979.
7. Radford, E. P., and St. Clair Renard, K. G. Lung cancer in Swedish iron
miners exposed to low doses of radon daughters. N. Engl. J. Med., 310:
1485-1494, 1984.
8. Swedjemark, G. A., and M ¡ones,L. Exposure of the Swedish population to
radon daughters. In: B. Berglund, T. Lindvall, and J. Sundell (eds.), Proceed
ings of the 3rd International Conference on Indoor Air Quality and Climate,
Vol. 2. Stockholm: Swedish Council for Building Research, 1984.
9. Nero, A. V. Indoor radiation exposures from 2"Rn and its daughters: a view
of the issue. Health Phys., 45: 277-288, 1983.
10. Axelson, O., Edling, C, and Kling, H. Lung cancer and residency—A casereferent study on the possible impact of exposure to radon and its daughters
in dwellings. Scand. J. Work Environ. Health, 5:10-15, 1979.
11. Edling, C., Kling, H., and Axelson, O. Radon in homes—a possible cause of
lung cancer. Scand. J. Work Environ. Health, ¡0:25-34, 1984.
12. Pershagen, G., Damber, L., and Falk, R. Exposure to radon in dwellings and
lung cancer. A pilot study. In: B. Berglund, T. Lindvall, and J. Sundell (eds.),
Proceedings of the 3rd International Conference on Indoor Air Quality and
Climate. Vol. 2. Stockholm: Swedish Council for Building Research. 1984.
13. Svensson, C., Eklund, G., and Pershagen, G. Indoor exposure to radon from
the ground and bronchial cancer in women. Int. Arch. Occup. Environ.
Health, 59:123-131, 1987.
14. Lees, R. E. M., Steele, R., and Roberts, J. H. A case-control study of lung
cancer relative to domestic radon exposure. Int. J. Epidemici., 16: 7-12,
1987.
15. James, A. C. Lung dosimetry. In: W. W. Nazaroff and A. V. Nero (eds.),
Radon and its decay products in indoor air. New York: Wiley-Interscience,
1988.
16. Svensson, C. Lung cancer etiology in women. Doctoral dissertation. The
1864
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.
LUNG CANCER IN WOMEN AND TYPE OF DWELLING
17.
18.
19.
20.
21.
Departments of Oncology and Environmental Hygiene. Stockholm: Karolinska Institute, 1988.
Swedish Radon Commission. Radon in dwellings. Report from the Radon
Commission, Ministry of Agriculture, SOU 1983: 6, p. 21. Stockholm: Liber,
1983.
Burén,
A., Hà kansson, B., Mjönes,L., More, H., Nyblom, L., and Wennberg,
P. An integrating radon monitor. Work document a 82-06. Stockholm: The
National Institute of Radiation Protection, 1982.
Breslow, N. E., and Day, N. E. Statistical methods in cancer research. Vol.
1. The analysis of case-control studies. Lyon: IARC, 1980.
The Cancer Registry. Cancer incidence in Sweden 1983. Stockholm: National
Board of Health and Welfare, 1986.
Cederlöf,R., Doll, R., Fowler, B., Friberg, L., Nelson, N., and Vouk, V. Air
pollution and cancer: risk assessment methodology and epidemiological
evidence. Environ. Health Persp., 22:1-10, 1978.
22. Norell, S. E., and Ahlbom, A. Hospital versus population referents in two
case-referent studies. Scand. J. Work Environ. Health, 13: 62-66, 1987.
23. Svensson, C, Pershagen, G., and Hrubec, Z. A comparative study on different
methods of measuring radon concentrations in homes. Health Phys. (in
press), 1989.
24. Jacobi, W. Activity and potential alpha energy of 2"Radon and "''Radondaughters in different atmospheres. Health Phys., 22:441-450, 1972.
25. Bergman, H., and Axelson, O. Passive smoking and indoor radon concentra
tions. Lancet, 2:1308-1309, 1983.
26. Archer, V. E., Wagoner, J. K.. and Lundin, F. E. Uranium mining and
cigarette smoking effects on man. J. Occup. Med., 15: 204-211, 1973.
27. Damber, L., and Larsson, L. G. Combined effects of mining and smoking in
the causation of lung carcinoma. A case control study in northern Sweden.
Acta Radiol. Oncol., 21: 305-313, 1982.
28. Axelson, O., and Sundell, L. Mining, lung cancer and smoking. Scand. J.
Work Environ. Health, 4:46-52, 1978.
1865
Downloaded from cancerres.aacrjournals.org on June 9, 2014. © 1989 American Association for Cancer Research.