Content uploaded by Pernilla Wiebert
Author content
All content in this area was uploaded by Pernilla Wiebert on Sep 10, 2015
Content may be subject to copyright.
ORIGINAL ARTICLE
Testicular germ cell tumours and parental
occupational exposure to pesticides: a register-based
case–control study in the Nordic countries (NORD-
TEST study)
Charlotte Le Cornet,
1,2
Béatrice Fervers,
2,3
Susanne Oksbjerg Dalton,
4
Maria Feychting,
5
Eero Pukkala,
6,7
Tore Tynes,
8,9
Johnni Hansen,
4
Karl-Christian Nordby,
9
Rémi Béranger,
1,2,3
Timo Kauppinen,
10
Sanni Uuksulainen,
10
Pernilla Wiebert,
5
Torill Woldbæk,
9
Niels E Skakkebæk,
11
Ann Olsson,
1,5
Joachim Schüz
1
▸Additional material is
published online only. To view
please visit the journal online
(http://dx.doi.org/10.1136/
oemed-2015-102860).
For numbered affiliations see
end of article.
Correspondence to
Dr Charlotte Le Cornet, Section
of Environment and Radiation,
International Agency for
Research on Cancer (IARC),
150 Cours Albert Thomas,
Lyon 69372, Cedex 08, France;
Lecornetc@students.iarc.fr
Received 29 January 2015
Revised 16 July 2015
Accepted 5 August 2015
To cite: Le Cornet C,
Fervers B, Oksbjerg
Dalton S, et al.Occup
Environ Med Published
Online First: [please include
Day Month Year]
doi:10.1136/oemed-2015-
102860
ABSTRACT
Objectives A potential impact of exposure to
endocrine disruptors, including pesticides, during
intrauterine life, has been hypothesised in testicular
germ cell tumour (TGCT) aetiology, but exposure
assessment is challenging. This large-scale registry-based
case–control study aimed to investigate the association
between parental occupational exposure to pesticides
and TGCT risk in their sons.
Methods Cases born in 1960 or onwards, aged
between 14 and 49 years, and diagnosed between 1978
and 2013 in Denmark, Finland, Norway or Sweden,
were identified from the respective nationwide cancer
registries. Four controls per case were randomly selected
from the general national populations, matched on year
of birth. Information on parental occupation was
collected through censuses or Pension Fund information
and converted into a pesticide exposure index based on
the Finnish National Job-Exposure Matrix.
Results A total of 9569 cases and 32 028 controls
were included. No overall associations were found for
either maternal or paternal exposures and TGCT risk in
their sons, with ORs of 0.83 (95% CI 0.56 to 1.23) and
of 1.03 (0.92 to 1.14), respectively. Country-specific
estimates and stratification by birth cohorts revealed
some heterogeneity. Cryptorchidism, hypospadias and
family history of testicular cancer were risk factors but
adjustment did not change the main results.
Conclusions This is the largest study on prenatal
exposure to pesticides and TGCT risk, overall providing
no evidence of an association. Limitations to assess
individual exposure in registry-based studies might have
contributed to the null result.
INTRODUCTION
Testicular germ cell tumours (TGCT) represent
98% of testicular cancer. Although accounting for
only 1% of all cancers in men, they remain the
most common cancer diagnosed in young men
aged 15–39 years. In 2012, over 23 000 new cases
were estimated in Europe.
1
TGCT affects mainly
Caucasian populations with incidence rates up to
10-fold higher in Europe compared with Asian or
African countries.
2
Incidence rates have increased
steeply over the past decades in industrialised coun-
tries, with the highest incidence rates observed in
Europe and the USA,
3
and the burden is predicted
to rise by 24% by 2025 in Europe.
4
Neighbouring countries show important geo-
graphical variations in incidence rates. An example
of this is the very different incidence rates in
Denmark and Finland: 10.0 and 4.8/100 000 men
in 2010, respectively.
5
Incidence rates also vary
between first and second generations of immi-
grants.
67
Both observations are in support of a pre-
dominant role of environmental factors in TGCT
development, in addition to the role of genetic sus-
ceptibility
8
and family history of TGCT.
9
Despite extensive research, risk factors of TGCT
remain largely unknown. The peak of incidence of
TGCT in young adults, its association with con-
genital malformations such as cryptorchidism and
hypospadias,
10
and the fact that embryonic stem
cell markers have been found in carcinoma in situ
(CIS) cells,
11
the precursor of TGCT, suggest early
determinants of testicular cancer onset.
What this paper adds
▸Prenatal exposure to endocrine disruptors,
including pesticides, has been hypothesised to
increase the risk of testicular germ cell tumours
(TGCT). Only a few rather small studies have
investigated this time window of exposure,
with inconclusive results.
▸This large-scale registry-based case–control
study in the Nordic countries provides no
evidence of an association between parental
occupational exposure to pesticides and TGCT
risk in the offspring.
▸Limitations to assess individual exposure in
registry-based studies might have decreased
our ability to show positive association.
Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860 1
Environment
OEM Online First, published on August 24, 2015 as 10.1136/oemed-2015-102860
Copyright Article author (or their employer) 2015. Produced by BMJ Publishing Group Ltd under licence.
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
According to testicular dysgenesis syndrome (TDS) hypoth-
esis, TGCT, cryptorchidism, hypospadias and several forms of
male infertility
12
share common aetiological factors.
13
The
hypothesis is that exposure to endocrine disruptors during fetal
life disturbs normal gonadal development and particularly the
normal differentiation of primordial germ cells, inducing CIS.
12
The prenatal origin of TGCT is also supported by the observed
associations of TGCT with gestational age, birth weight,
14
sib-ship size and birth order.
15
So far, the investigation of environmental factors associated
with TGCT has focused mainly on exposure during adulthood.
Two recent reviews
16 17
emphasised the need for further
research into exposures during early time windows of exposure.
Maternal exposure to exogenous oestrogens during gestation
has been suggested to induce male reproductive tract abnormal-
ities.
18
Diethylstilbestrol, dichlorodiphenyltrichloroethane, hex-
achlorobenzene (HCB), polychlorinated biphenyls (PCBs) and
some pesticides with endocrine disrupting properties are sus-
pected to disturb the normal genital development by mimicking
oestrogen.
19
Paternal occupational exposures have been seen to
potentially affect cancer susceptibility in the child by transpla-
cental exposures.
20
Despite an observed association with persist-
ent organic pollutants in maternal serum,
21
few epidemiological
studies have investigated the association of prenatal exposures to
pesticides with TGCT risk since 2001, when the TDS hypothesis
was first proposed.
It is methodologically difficult to retrospectively assess pre-
natal exposure. Only a few epidemiological studies
22 23
have
attempted to study parental occupational exposure, in particular
to pesticides, but these have been severely limited by lack of stat-
istical power due to a limited number of cases.
The pooled NORD-TEST study has been designed as a case–
control study, nested within the populations of Denmark,
Finland, Norway and Sweden, utilising the registry data with
the objective to investigate the possible association between par-
ental occupational exposure to pesticides during prenatal
periods and the risk in their sons to develop TGCT in
adulthood.
MATERIALS AND METHODS
Denmark, Finland, Norway and Sweden have population-based
cancer registries since the 1940s to 1960s, and registries record-
ing information on socioeconomic status, family composition
and occupations since the 1960s.
The NORD-TEST study is a registry-based nested case–
control study conducted in Denmark, Finland, Norway and
Sweden. In these countries, all residents have a unique personal
identification number (CPR number) since 1968 in Denmark,
1967 in Finland, 1964 in Norway and 1947 in Sweden, allow-
ing linkage between registries for tracking personal medical
information and occupational status for research purposes (see
online supplementary file 1).
Study population
In order to have information on the parental occupations before
birth and given that the censuses start at the earliest in 1960, all
men diagnosed with TGCT from 1978 onwards were regarded
as eligible. Cases and controls born outside of the included
countries or with a history of cancer (except non-melanoma
skin cancer) prior to the index date of TGCT diagnosis were
not eligible. The last available diagnostic year at the time of data
set creation was in 2003 in Denmark (due to availability of
occupational data, see below), 2012 in Finland, 2010 in
Norway and 2011 in Sweden (see online supplementary file 1).
In addition to date of diagnosis, the study inclusion depended
on the year of the first census available, with Denmark including
men born between 1965 and 1981, Finland including men born
in 1970 or later and Norway and Sweden including men born
in 1960 or later.
Cases were identified in the nationwide population-based
cancer registries using International Classification of Diseases
(ICD) codes as follows: ICD 7:178, ICD 8 and ICD 9:186, ICD
10 and ICD-O-3:C62, Danish revision of ICD 7:178, 278, 378,
478, 578, 678, 878.
Online supplementary file 2 gives the histological codes and
classifications used to categorise TGCT cases into seminoma and
non-seminoma. In Denmark, the classification used was the
extended Danish version of the seventh version of the ICD. We
used the ICD-Oncology (ICD-O-3) in Finland, a Norwegian
version of the ICD-O in Norway and, in Sweden, histology
code from 1958,
24
according to WHO/HS/CANC/24.1,
ICD-O-2 from 1993 and, from 2005 onwards, ICD-O-3.
Four controls per case were randomly selected from the
national population registers of each country matched to case
on year of birth. Overall, 11 111 cases and 38 553 controls
were extracted.
Figure 1 presents the various steps of exclusions for cases and
controls. Men aged between 14 and 49 years at diagnosis were
retained. Only cases with confirmed TGCT subtypes (seminoma
and non-seminoma) were kept. Cases and controls were
excluded when occupations before birth were unavailable for
both parents. The final sample included 9569 cases and 32 028
controls.
Data collection
Exposure assessment
The Finnish National Job-Exposure Matrix (FINJEM) translates
the occupational codes from the Finnish classification into quan-
titative estimates of fungicides, herbicides and insecticides. These
estimations were based on available exposure measurements,
hazard surveys and assessments from occupational hygienists.
25
Fungicide, herbicide and insecticide exposures were calculated as
the product of the proportion of exposed persons (P), and the
mean level (in mg/m
3
) of exposure (L) estimated for each occupa-
tional code.
26
Pesticide exposure was thus then the sum of fungi-
cide, herbicide and insecticide exposures. For parental
occupation, data were available only from one census prior to the
child’s birth, without indication of duration of occupation, and
thus did not allow cumulative exposure assessment. In FINJEM,
six occupational codes are associated with pesticide exposure
(300: farmers, silviculturists, horticulturists, 311: commercial
garden and park workers, 340: forestry workers and lumberjacks,
671: sawyers, 672: plywood and fibreboard workers, 731: cooks
and furnacemen (chemical processes)). FINJEM includes three
dimensions: the quantification of the agents, the occupations and
the calendar periods, taking into account changes of pesticide
exposures over time. In our study, the FINJEM estimates were
used for two distinct periods, 1960–1984 and 1985–1994.
26
Parental job titles (or job codes) for cases and controls were
identified from census data in Finland, Norway and Sweden. Job
codes were collected from six censuses in Sweden (1960, 1970,
1975, 1980, 1985, 1990), from five censuses in Finland (1970,
1975, 1980, 1985, 1990) and from four censuses in Norway
(1960, 1970, 1980, 1990 (1990 being complete only for a subset
of the census subjects)). Missing data from the Norwegian survey
in 1990 were replaced by job titles from the 1980 census when
available, assuming that the impact on the results was negligible
since less than 2% of our study population was born in 1990 or
2 Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
later. In addition, missing job codes between two known censuses
with identical job titles were recoded with the same job codes.
The Finnish occupational classification contains 311 job codes.
27
Swedish and Norwegian job codes were translated to the Finnish
ones.
In Denmark, no census data were collected; we used parental
employment history on a company level prior to the birth of
the sons instead. Since 1964, all employees in Denmark have
been compulsory members of the Supplementary Pension Fund,
which stores historical information on all employments (dates of
start and end, a unique company number and the individual
CPR number of the employee).
28
Afive-digit detailed industry
code for each employee has been applied by Statistics Denmark
based on the International Standard Industrial Classification
(ISIC). A Danish expert ( JH) on occupational exposures,
blinded to case–control status, manually selected the Danish
industry codes with a similar level of exposure to pesticides as
the existing occupations included in FINJEM. Additionally, job-
codes taken from the income tax form were available for
Denmark. Since industry codes include various occupations and
workplaces within a specific industry, a cross-check between job
codes (proxy for white/blue collar) and industry codes allowed
distinguishing between exposed and unexposed job codes within
an industry, and therefore to categorise the pesticide exposure
level to 0 when exposure was unlikely (eg, dentist, care staff,
office assistant).
To target prenatal exposure, the closest maternal and paternal
occupation before the child’s birth was selected from the
Finnish, Norwegian and Swedish censuses. In Denmark, the
longest parental employment held in the last 12 months prior to
the child’s birth was retrieved from the Supplementary Pension
Fund. When the father had no employments in this period, the
longest occupation held within the 2 years prior to the child’s
birth was selected instead.
Maternal and paternal total pesticide, fungicide, herbicide
and insecticide exposures before the child’s birth were cate-
gorised into binary variables (exposed/unexposed) as well as
into four categories of exposure based on quantitative estima-
tions (unexposed, low exposure (<0.01 mg/m
3
), intermediate
exposure (0.01–0.1 mg/m
3
), high exposure (>0.1 mg/m
3
)).
Parental pesticide exposure before birth was the sum of the total
paternal and maternal pesticide exposure before birth.
The parental pesticide exposure was not calculated when one of
the parents had no information on occupation prior to the
child’s birth and the other had an occupation unlikely to be
exposed to pesticides.
Confounding variables
Data on siblings and parents of cases and controls were
retrieved from the central population registries via the CPR
number. Cases, controls, and their parents and siblings were
linked to the respective cancer registries in the four countries in
order to collect testicular cancer family history. The national
cancer registries have operated since 1942 in Denmark, 1953 in
Finland and Norway, and since 1958 in Sweden. Date of diag-
nosis, age at diagnosis, morphology and histology of the cancer
were extracted, for brothers and fathers, from the cancer
registries.
Data on diagnostic and surgical interventions for cryptorchid-
ism and hypospadias were extracted from the birth registry,
which started in 1987 in Finland and 1973 in Sweden, and
from the Hospital Discharge Register, available from 1969 in
Finland and from 1964 in Sweden. In Finland, in addition to
birth registry and hospital discharge registry, the registry of con-
genital malformations from 1980 at the National Institute for
Health and Welfare was also used. In Norway, only information
from the Medical Birth Registry from 1967 onwards was used
to retrieve cryptorchidism and hypospadias. Congenital malfor-
mations were not included in the Danish birth registry until
1996; therefore, they were obtained from the Hospital
Discharge Register available since 1977 for inpatients and from
1995 for outpatients. Online supplementary file 3 shows the
disease codes and operational codes used to identify hypospa-
dias and cryptorchidism. Only the first diagnosis or the first
operation declared for the same congenital malformation was
taken into account.
Statistical analysis
Conditional logistic regression for matched case–control sets
was used to measure the association between parental occupa-
tional exposure to pesticides and TGCT risk, with ORs and
their 95% CI. The impact of well-known risk factors for TGCT
was tested in the model; cryptorchidism (yes/no), hypospadias
Figure 1 Flow chart of inclusion of cases and controls (*Controls not matching the inclusion criteria as well as all controls matched to excluded
cases were excluded).
Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860 3
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
(yes/no), father’s history of testicular cancer ( yes/no) and broth-
ers’history of testicular cancer (yes/no).
These variables were treated as potential confounders in mul-
tivariable models. No confounding was assumed if the expos-
ure–outcome association measure changed less than 15% on
inclusion of the covariate in models.
The analyses were made for all TGCT combined and separ-
ately for seminoma and non-seminoma. The Spearman correl-
ation coefficient was used to estimate the correlation between
paternal and maternal exposure to pesticides. Additional stratifi-
cations were conducted by country, by 5-year birth cohorts, as
well as by year of census, to investigate the potential changes in
exposure over time. A sensitivity analysis was performed
restricted to cases born within 2 years after the parental census
to narrow the time window between the census and the child’s
birth, thereby reducing the risk of misclassification due to job
changes. All analyses were carried out using SAS statistical
package V.9.3 (SAS Institute, Inc, Cary, North Carolina, USA).
NORD-TEST has been approved by the relevant data protec-
tion and ethical committees in Denmark, Finland, Norway and
Sweden, and by the International Agency for Research on
Cancer (IARC) ethics committee.
RESULTS
Overall, 9569 cases and 32 028 controls were included in the
study. Country-specific distributions of cases and controls are
shown in online supplementary file 4. Sweden and Norway
together account for 70% of the total study population. Most
cases (84%) were born before 1980 and 94% were diagnosed in
1990 or later. Eighty-five per cent of the cases were diagnosed
at age 20–40 years with an average age of 32 years for semi-
noma and 26 years for non-seminoma. Seminoma accounts for
44% of the cases.
Information about occupational exposure prior to the child’s
birth was available for 89% of the mothers and 97% of the
fathers (see online supplementary file 4). The median time
between the closest census date prior to child birth and at the
child’s birth was 2 years for fathers and for mothers. Only 0.4%
of mothers were classified as being potentially exposed to pesti-
cides in their occupation prior to the child’s birth, whereas
4.9% of fathers were classified as such.
Cryptorchidism and hypospadias were associated with TGCT
with OR=3.28 (95% CI 2.72 to 3.97) and OR=2.21 (95% CI
1.46 to 3.36), respectively. Associations were also found with
family history of testicular cancer among fathers (OR=3.33
(95% CI 2.45 to 4.54)) and brothers (OR=6.30 (95% CI 4.34
to 9.13)).
Table 1 shows the frequency of estimated occupational pesti-
cide exposure in parents and the association with TGCT in their
sons.
No associations were found for either maternal or paternal
occupational exposure to pesticides and TGCT risk in their
sons, with OR=0.83 (95% CI 0.56 to 1.23) and 1.03 (95% CI
0.92 to 1.14), respectively. Adjustment for cryptorchidism,
hypospadias and family history of testicular cancer did not
materially change the effect estimates: adjusted OR=0.81 (95%
CI 0.55 to 1.21) for maternal and 1.03 (95% CI 0.91 to 1.16)
for paternal occupational exposure. Fungicides represented the
most prevalent pesticide exposure among fathers and mothers
(5.0% and 0.4%, respectively), but the fungicides, insecticides
and herbicides assessed individually, did not show any associ-
ation with TGCT in sons (table 1). Owing to small numbers and
non-significant results, we only show the overall outcomes for
maternal exposure and give further details for paternal exposure
in table 1. When both parents were potentially exposed to pesti-
cides in their occupation before the child’s birth, no association
was observed, even for the highest exposure category. In add-
ition, no correlation was found between maternal and paternal
exposure to occupational pesticides (r=0.13).
The stratification by TGCT subtypes showed results similar to
all types combined (table 1); no association was observed either
for father, mother or for both combined.
Variations were found between the country-specific estimates.
While the OR for paternal exposure to pesticides before the
child’s birth and TGCT risk was increased in Denmark
(OR=2.98 (95% CI 1.61 to 5.52)), the estimate was decreased
for maternal exposure in Sweden (OR=0.49 (95% CI 0.23 to
1.05)), with no other notable associations for any other mater-
nal and/or paternal exposure to pesticides prior to the child’s
birth. Stratification by census did not show any further associa-
tions (data not shown). Investigating the association by 5-year
birth cohorts did not reveal any clear trends by time with most
ORs close to 1, but few significant results were observed,
namely for the birth cohort 1960–1965 with parental exposure
(OR 1.30, CI 1.02 to 1.65; mother only: OR 2.53, CI 0.60 to
10.69) and an inverse association for the birth cohort 1975–
1980 with maternal exposure (OR 0.28, CI 0.09 to 0.92).
The sensitivity analysis using only parental occupational pesti-
cide exposures recorded within 2 years prior to the child’s birth,
based on 20 exposed case mothers and 252 exposed case
fathers, showed no association between exposure to pesticides
and TGCT risk, with an OR=0.97 (0.58 to 1.61) for maternal
exposure and 1.06 (0.91 to 1.23) for paternal exposure.
DISCUSSION
Although the study, to our knowledge, is the largest ever made
on parental occupational exposure associated with TGCT risk,
the overall number of parents with a history of occupational
pesticide exposures was small. Only 0.4% of the mothers and
4.9% of the fathers were classified as possibly exposed to pesti-
cides in their job prior to their child’s birth. Our general find-
ings of no association between parental prenatal occupational
exposure to pesticides and TGCT risk, both for seminoma and
non-seminoma, even among the highest category of parental
exposure prior to the child’s birth, confirmed those of smaller
previous studies.
Our results were not entirely homogeneous across countries
and between birth cohorts. Positive association between paternal
exposure to pesticides before the child’s birth and TGCT was
observed in Denmark, but not in the other countries. In add-
ition, parental pesticide exposure of the 1960–1965 birth
cohort was associated with a significantly increased risk of
TCGT. However, we cannot rule out that these results were
chance findings due to the large number of analyses.
Previous studies from the USA
22
and Denmark
23
did not find
an increased risk of TGCT for sons of parents who had
reported, using self-administered questionnaires, employment in
agriculture during the prenatal period. However, these studies
had limited statistical power because of the small populations
and low prevalence of exposure, resulting in effect estimates
with wide CIs. There may be a slight overlap of data between
our study and the previous Danish case–control study
23
limited
to cases diagnosed between 1986 and 1988, which represents
less than 5% of the Danish cases included in the present study.
A Swedish study investigating testicular cancer occurrence
among offspring of pesticide applicators licensed between 1965
and 1976 found only two cases born between 1958 and 1994,
indicating no increased risk of TGCT among pesticide
4 Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
Table 1 Risk for all types of testicular germ cell tumour (TGCT) combined and for TGCT subtypes in relation to individually and combined maternal and paternal prenatal pesticide exposure
All TGCT Seminoma Non-seminoma
Cases Controls
OR 95% CI
Cases Controls
OR 95% CI
Cases Controls
OR 95% CIN%N % N%N % N%N %
Maternal pesticide exposure before birth
None 8410 99.6 28 616 99.5 3588 99.6 12 124 99.6 4822 99.6 16 492 99.5
Any 33 0.4 136 0.5 0.83 (0.56 to 1.23) 15 0.4 46 0.4 1.03 (0.57 to 1.88) 18 0.4 90 0.5 0.72 (0.43 to 1.21)
Paternal pesticide exposure before birth
None 8793 94.9 29 630 95.1 3811 94.1 12 784 94.5 4982 95.6 16 846 95.5
Any 471 5.1 1543 4.9 1.03 (0.92 to 1.14) 241 5.9 741 5.5 1.09 (0.93 to 1.27) 230 4.4 802 4.5 0.97 (0.83 to 1.13)
None 8793 94.9 29 630 95.1 3811 94.0 12 784 94.5 4982 95.6 16 846 95.4
Low 414 4.5 1348 4.3 1.03 (0.92 to 1.16) 221 5.5 663 4.9 1.13 (0.95 to 1.31) 193 3.7 685 3.9 0.95 (0.81 to 1.12)
Medium 24 0.2 108 0.3 0.76 (0.49 to 1.19) 8 0.2 44 0.3 0.65 (0.30 to 1.39) 16 0.3 64 0.4 0.83 (0.48 to 1.45)
High 33 0.4 87 0.3 1.29 (0.86 to 1.94) 12 0.3 34 0.3 1.21 (0.62 to 2.37) 21 0.4 53 0.3 1.34 (0.81 to 2.25)
Paternal herbicide exposure before birth
None 8850 95.5 29 825 95.7 3831 94.5 12 862 95.1 5019 96.3 16 963 96.1
Any 414 4.5 1348 4.3 1.03 (0.92 to 1.16) 221 5.5 663 4.9 1.11 (0.95 to 1.31) 193 3.7 685 3.9 0.95 (0.81 to 1.12)
Paternal fungicide exposure before birth
None 8793 94.9 29 630 95.1 3811 94.1 12 784 94.5 4982 95.6 16 846 95.5
Any 471 5.1 1543 4.9 1.03 (0.92 to 1.14) 241 5.9 741 5.5 1.09 (0.93 to 1.27) 230 4.4 802 4.5 0.97 (0.83 to 1.13)
Paternal insecticide exposure before birth
None 8938 96.5 30 120 96.6 3883 95.8 13 021 96.3 5055 97.0 17 099 96.9
Any 326 3.5 1053 3.4 1.03 (0.91 to 1.18) 169 4.2 504 3.7 1.11 (0.93 to 1.34) 157 3.0 549 3.1 0.96 (0.80 to 1.15)
Parental pesticide exposure before birth
None 7740 94.0 26 576 94.3 3270 93.0 11 187 93.6 4470 94.8 15 389 94.8
Any 492 6.0 1616 5.7 1.04 (0.93 to 1.17) 248 7.0 765 6.4 1.09 (0.93 to 1.29) 244 5.2 851 5.2 1.00 (0.86 to 1.17)
Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860 5
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
applicators’offspring.
29
Also, in Norway, an investigation into
offspring of parents engaged in agricultural activities did not
show any association between testicular cancer and exposure to
pesticides.
30
A recent study on TGCT used a qualitative job-exposure
matrix ( JEM)
31
to categorise the prenatal paternal and maternal
exposure to endocrine disrupting chemicals (EDCs) as unex-
posed, and possibly or probably exposed; pesticides were
included in the EDCs. No increased risk of TGCTwas observed
for probable paternal and maternal exposure to pesticides prior
to the child’s birth compared with unexposed (OR=1.33 (95%
CI 0.65 to 2.70) and 0.97 (95% CI 0.23 to 4.07), respect-
ively).
32
Components with endocrine disrupting properties such
as HCB, PCBs and chlordanes, have also been studied in mater-
nal serum, and a significantly higher concentration in case
mothers compared with control mothers was reported.
21
However, the serum was sampled at the time of the case’s diag-
nosis, which conveys uncertainties with regard to its representa-
tiveness of past exposures. Also, the recruitment of cases was
not population-based.
The strength of our study is the pooling of cases from four
countries, with full national population coverage providing a
larger sample size than prior studies. The registries provided
multiple variables that have been collected systematically over
time, avoiding recall bias. Even if individual lifestyle factors
cannot be collected in registries, information about the major
known TGCT risk factors were available. Cryptorchidism, hypo-
spadias and family history were associated with an increased
TGCT risk in our population, thereby confirming well-known
associations.
910
Parental occupations were systematically
reported and, if missing values occurred, this was independent
of case–control status.
One limitation of our study was the use of a crude proxy for
the time of the prenatal window of exposure. The occupation
before birth was assumed to best reflect the preconceptional or
pregnancy period. Indeed, the lag time between the child’s birth
and the closest census before birth ranged from 0 to 15 years
and a change of occupation between the census and the child’s
birth may have occurred. Nevertheless, more than 80% of
fathers and mothers had a census recorded within 5 years prior
to the child’s birth and 99% within 10 years. Also, the sensitiv-
ity analysis performed for parents having a census within the
2 years before the child’s birth showed similar results. Some pes-
ticides are chemicals persisting in the body for decades.
Therefore, parental exposures occurring several years before the
child birth may have an impact. However, given that our sensi-
tivity analyses show similar OR for approximately half of the
exposed fathers and mothers, with census within the 2 years
before the child’s birth, compared with the overall OR, an
increased risk for the remaining with larger time lag is unlikely.
Retrospective exposure assessment methods are subject to
exposure misclassification and our objective was to limit mis-
classification as much as possible by choosing the most appropri-
ate JEM. FINJEM was initially developed in 1990 to estimate
occupational exposures in Finland and was later extended to the
neighbouring Nordic countries
27 33
for various exposures but
not pesticides. Thus, variation in pesticide use between countries
and over time might have introduced bias; however, climate and
agricultural practices are expected to differ less between Nordic
countries, and this would more probably have been the case in a
study involving more distant countries.
FINJEM assigns an exposure level to a job title assuming
homogeneity of exposure within jobs, which may lead to expos-
ure misclassification. Variation in exposure levels may be
observed within occupations, leaving non-exposed persons
within the exposed groups. Also, some non-agricultural occupa-
tions have been suggested to also be potentially exposed to pes-
ticides, during some tasks (eg, among carpenters, sawmill
workers, staff and students at research institutions, laboratories
that use pesticides, employees in joint experimental work in
agriculture),
34 35
and were not considered in FINJEM. As well,
individual domestic and environmental level of pesticide expos-
ure was not available in our study, inducing potentially misclassi-
fication bias; hence, the effect may have been diluted. However,
the FINJEM approach allows selection of those occupations
having, a priori, a high level of exposure and a high prevalence
of exposure. If no risk was found with these occupational
codes, it means that the occupational codes with a smaller pro-
portion of exposed parents would not have been associated with
TGCT risk either. Nevertheless, the record linkage approach
precludes differential assessment between cases and controls.
Moreover, assuming a Berkson type error for the misclassifica-
tion induced by the JEM approach would suggest that risk esti-
mates are not affected, but this would also reduce the statistical
power of the study.
36
Given the large size of the study, even if
our power is reduced, the association between parental exposure
to pesticides and TGCT seems to be inexistent or rather weak.
Another point that might have impacted on our results is that
FINJEM attributes a probability of exposure for herbicides, fun-
gicides and insecticides, which are groups of compounds with
different mechanisms, whereas information on probability of
exposure to singular active ingredients is not provided. Only
some pesticides have an endocrine disrupting effect which
might have been diluted by using a JEM that did not consider
exposure heterogeneity within each group of pesticides.
A very small proportion of mothers were potentially occupa-
tionally exposed to pesticides in our study, and this might limit
the interpretation of the association between maternal exposure
to pesticide during prenatal period and TGCT risk. On the
other hand, if there was a modest risk that we missed, the
impact on the overall TGCT incidence would be minimal, given
the rarity of exposure. More fathers were exposed. As hypothe-
sised for childhood brain tumours and leucaemia,
20
paternal
exposure might affect cancer susceptibility in the child either by
transplacental exposures or genetic alterations. Yet, we lack a
strong hypothesis explaining a potential impact of paternal
exposure on TGCT risk. Although paternal exposure to pesti-
cides in Denmark was found associated with TGCT in the off-
spring, the stratified analyses produced a large number of risk
estimates, and the few associations observed are likely to be
effects of chance.
Our study only focused on the prenatal occupational expos-
ure to pesticides. However, despite the fact that no clear risk
factors have been found for TGCT among adulthood exposures,
previous epidemiological studies have reported some associa-
tions among agricultural workers, construction workers,
firemen, policemen and military personnel, as well as among
workers in paper, plastic and metal industries.
16
It has recently
been suggested that a combined effect of prenatal and postnatal
risk factors such as exposure to organochlorine pesticides may
impact on TGCT.
17
Endogenous hormones play an important
role in puberty, and therefore environmental exposures during
this critical period may trigger the development of neoplasm
previously initiated.
CONCLUSION
In this large four-country study on parental exposure to pesti-
cides and TGCT risk in their sons, we found little evidence of
6 Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
any association. Although based on a large study population,
NORD-TEST does not rule out that there may be an effect of
parental pesticide exposure on the TGCT risk in the offspring.
Because of the limitations of NORD-TEST, there is a need for
studies with more accurate assessment of exposure from all
sources (environmental, occupational and domestic).
Furthermore, the hypothesis on a potential combined effect of
prenatal and postnatal risk factors deserves further research.
Author affiliations
1
Section of Environment and Radiation, International Agency for Research on Cancer
(IARC), Lyon, France
2
Unité Cancer et Environnement, Centre Léon Bérard, Lyon, France
3
Université Claude Bernard—Lyon 1, Villeurbanne, France
4
Danish Cancer Society Research Center, Copenhagen, Denmark
5
The Institute of Environmental Medicine, Karolinska Institutet, Stockholm, Sweden
6
Finnish Cancer Registry, Institute for Statistical and Epidemiological Cancer
Research, Helsinki, Finland
7
School of Health Sciences, University of Tampere, Tampere, Finland
8
Kreftregisteret, Cancer Registry of Norway, Oslo, Norway
9
National Institute of Occupational Health, Oslo, Norway
10
Finnish Institute of Occupational Health (FIOH), Helsinki, Finland
11
University Department of Growth and Reproduction, University Hospital of
Copenhagen, Copenhagen, Denmark
Acknowledgements This work was supported by public funding from the Lyric
Grant INCa-DGOS-4664 (Institute of Cancer Research, France), the International
Agency for Research on Cancer (IARC) and the Cancéropôle Lyon Auvergne
Rhône-Alpes (CLARA). The authors would like to acknowledge Marianne
Steding-Jessen and Andrea Meersohn from the Danish Cancer Society Research
Center, Karin Fremling from the Institute of Environmental Medicine, Karolinska
Institutet, and Veronique Luzon from IARC, for their help with the data
management. The Family-Cancer Database was created by linking registers
maintained at Statistics Sweden and the Swedish Cancer Registry. Data from the
Finnish Cancer Registry were extracted by permission for research: Number THL/
1123/5.05.00/2012. Thanks also to the NOCCA team, who approved the research
protocol for the use of FINJEM and thanks for their expertise on occupational
exposure in the Nordic countries.
Contributors CLC, BF, SOD, MF, EP, TT, K-CN, NES, AO and JS developed the
study concept and design. CLC, SOD, MF, EP, TT, K-CN and JH contributed to the
data collection. CLC, JH, TK, SU, PW, TW and AO contributed to the exposure
assessment. CLC, BF, RB, NES, AO and JS conducted the analysis and the
interpretation of the results. CLC, BF and JS drafted the manuscript. All the authors
revised it critically for intellectual content and approved the final version.
Competing interests None declared.
Ethics approval International Agency for Research on Cancer (IARC) ethics
committee, Danish Data protection Board, Finnish National Institute for Health and
Welfare (THL), Regional committees for medical and health research ethics (REK),
Ethical review board of Stockholm.
Provenance and peer review Not commissioned; externally peer reviewed.
REFERENCES
1 Forman D, Bray F, Brewster D, et al.Cancer incidence in five continents, vol. X
(electronic version). Lyon: IARC. 2013. http://ci5.iarc.fr (accessed 15 Sep 2014).
2 Ferlay J, Soerjomataram I, Ervik M, et al.GLOBOCAN 2012 v1.0, cancer incidence
and mortality worldwide: IARC CancerBase No. 11. Lyon, France: International
Agency for Research on Cancer, 2013. http://globocan.iarc.fr (accessed 10 Apr
2014).
3 Chia VM, Quraishi SM, Devesa SS, et al. International trends in the incidence of
testicular cancer, 1973–2002. Cancer Epidemiol Biomarkers Prev 2010;19:1151–9.
4 Le Cornet C, Lortet-Tieulent J, Forman D, et al. Testicular cancer incidence to rise by
25% by 2025 in Europe? Model-based predictions in 40 countries using
population-based registry data. Eur J Cancer 2014;50:831–9.
5 Engholm G, Ferlay J, Christensen N, et al. NORDCAN: cancer incidence, mortality,
prevalence and survival in the Nordic Countries, Association of the Nordic Cancer
Registries. Danish Cancer Society. Version 6.1. Assoc Nord Cancer Regist Danish
Cancer Soc 2014. http://www.ancr.nu (accessed 12 Sep 2014).
6 Schmiedel S, Schüz J, Skakkebaek NE, et al. Testicular germ cell cancer incidence in
an immigration perspective, Denmark, 1978 to 2003. JUrol2010;183:1378–82.
7 Hemminki K, Li X. Cancer risks in Nordic immigrants and their offspring in Sweden.
Eur J Cancer 2002;38:2428–34.
8 Dalgaard MD, Weinhold N, Edsgard D, et al. A genome-wide association study of
men with symptoms of testicular dysgenesis syndrome and its network biology
interpretation. J Med Genet 2012;49:58–65.
9 Greene MH, Kratz CP, Mai PL, et al. Familial testicular germ cell tumors in adults:
2010 summary of genetic risk factors and clinical phenotype. Endocr Relat cancer
2010;17:R109–21.
10 McGlynn KA, Cook MB. Etiologic factors in testicular germ-cell tumors. Future Oncol
2009;5:1389–402.
11 Rajpert-De Meyts E. Developmental model for the pathogenesis of testicular
carcinoma in situ: genetic and environmental aspects. Hum Reprod Update
2006;12:303–23.
12 Skakkebaek NE, Rajpert-De Meyts E, Main KM. Testicular dysgenesis syndrome:
an increasingly common developmental disorder with environmental aspects. Hum
Reprod 2001;16:972–8.
13 Toppari J, Virtanen HE, Main KM, et al. Cryptorchidism and hypospadias as a sign
of testicular dysgenesis syndrome (TDS): environmental connection. Birth defects Res
Part A Clin Mol Teratol 2010;88:910–19.
14 Cook MB, Akre O, Forman D, et al. A systematic review and meta-analysis of
perinatal variables in relation to the risk of testicular cancer-experiences of the son.
Int J Epidemiol 2010;39:1605–18.
15 Cook MB, Akre O, Forman D, et al. A systematic review and meta-analysis of
perinatal variables in relation to the risk of testicular cancer—experiences of the
mother. Int J Epidemiol 2009;38:1532–42.
16 Béranger R, Le Cornet C, Schüz J, et al. Occupational and environmental exposures
associated with testicular germ cell tumours: systematic review of prenatal and
life-long exposures. PLoS ONE 2013;8:e77130.
17 McGlynn KA, Trabert B. Adolescent and adult risk factors for testicular cancer. Nat
Rev Urol 2012;9:339–49.
18 Sharpe RM, Skakkebaek NE. Are oestrogens involved in falling sperm counts and
disorders of the male reproductive tract? Lancet 1993;341:1392–5.
19 Expertise collective INSERM. Cancer et environnement. 2008. http://www.inserm.fr
(accessed 1 Dec 2012).
20 Feychting M, Plato N, Nise G, et al. Paternal occupational exposures and childhood
cancer. Environ Health Perspect 2001;109:193–6.
21 Hardell L, Van BB, Lindstrom G, et al. Increased concentrations of polychlorinated
biphenyls, hexachlorobenzene, and chlordanes in mothers of men with testicular
cancer. Environ Health Perspect 2003;111:930–4.
22 Kardaun JW, Hayes RB, Pottern LM, et al. Testicular cancer in young men and
parental occupational exposure. Am J Ind Med 1991;20:219–27.
23 Møller H, Version U, Articles C. Work in agriculture, childhood residence, nitrate
exposure, and testicular cancer risk: a case-control study in Denmark. Cancer
Epidemiol Biomarkers Prev 1997;6:141–4.
24 WHO (World Health Organization). Statistical code for human tumors. WHO/HS/
CANC, 24.1 and 24.2. Geneva: WHO, 1956.
25 Kauppinen T, Uuksulainen S, Saalo A, et al. Use of the Finnish Information System
on Occupational Exposure (FINJEM) in epidemiologic, surveillance, and other
applications. Ann Occup Hyg 2014;58:380–96.
26 Kauppinen T, Toikkanen J, Pukkala E. From cross-tabulations to multipurpose
exposure information systems: a new job-exposure matrix. Am J Ind Med
1998;33:409–17.
27 Kauppinen T, Heikkilä P, Plato N, et al. Construction of job-exposure
matrices for the Nordic Occupational Cancer Study (NOCCA). Acta Oncol
2009;48:791–800.
28 Hansen J, Lassen CF. The supplementary pension fund register. Scand J Public
Health 2011;39:99–102.
29 Rodvall Y, Dich J, Wiklund K. Cancer risk in offspring of male pesticide applicators
in agriculture in Sweden. Occup Environ Med 2003;60:798–801.
30 Kristensen P, Andersen A, Irgens LM, et al. Cancer in offspring of parents engaged
in agricultural activities in Norway: incidence and risk factors in the farm
environment. Int J Cancer 1996;65:39–50.
31 Van Tongeren M. A job-exposure matrix for potential endocrine-disrupting chemicals
developed for a study into the association between maternal occupational exposure
and hypospadias. Ann Occup Hyg 2002;46:465–77.
32 Nori F, Carbone P, Giordano F, et al. Endocrine-disrupting chemicals and testicular
cancer: a case-control study. Arch Environ Occup Health 2010;61:87–95.
33 Dryver E, Brandt L, Kauppinen T, et al. Occupational exposures and non-Hodgkin’s
lymphoma in Southern Sweden. Int J Occup Environ Health 2004;10:13–21.
34 Provost D, Cantagrel A, Lebailly P, et al. Brain tumours and exposure to
pesticides: a case-control study in southwestern France. Occup Environ Med
2007;64:509–14.
35 Mattilsynet. Statens tilsyn for planter, fisk, dyr og næringsmidler. http://www.
mattilsynet.no/planter_og_dyrking/plantevernmidler/autorisasjonsbevis_for_bruk_av_
plantevernmidler/hvem_kan_faa_autorisasjon_for_bruk_av_plantevernmidler.2612
36 Armstrong BG. Effect of measurement error on epidemiological studies of
environmental and occupational exposures. Occup Environ Med 1998;55:651–6.
Le Cornet C, et al.Occup Environ Med 2015;0:1–7. doi:10.1136/oemed-2015-102860 7
Environment
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from
NORD-TEST study)(Nordic countries control study in the−register-based case occupational exposure to pesticides: a Testicular germ cell tumours and parental
Schüz
Wiebert, Torill Woldbæk, Niels E Skakkebæk, Ann Olsson and Joachim
Nordby, Rémi Béranger, Timo Kauppinen, Sanni Uuksulainen, Pernilla
Feychting, Eero Pukkala, Tore Tynes, Johnni Hansen, Karl-Christian
Charlotte Le Cornet, Béatrice Fervers, Susanne Oksbjerg Dalton, Maria
published online August 24, 2015Occup Environ Med
http://oem.bmj.com/content/early/2015/08/24/oemed-2015-102860
Updated information and services can be found at:
These include:
Material
Supplementary
C1.html
http://oem.bmj.com/content/suppl/2015/08/24/oemed-2015-102860.D
Supplementary material can be found at:
References
#BIBL
http://oem.bmj.com/content/early/2015/08/24/oemed-2015-102860
This article cites 30 articles, 14 of which you can access for free at:
service
Email alerting box at the top right corner of the online article.
Receive free email alerts when new articles cite this article. Sign up in the
Notes
http://group.bmj.com/group/rights-licensing/permissions
To request permissions go to:
http://journals.bmj.com/cgi/reprintform
To order reprints go to:
http://group.bmj.com/subscribe/
To subscribe to BMJ go to:
group.bmj.com on September 10, 2015 - Published by http://oem.bmj.com/Downloaded from